Compare commits

..

63 commits

Author SHA1 Message Date
Simon Willison
32845d88d5 Deploy canned queries to latest.datasette.io again with workaround for #2916
Should fix some broken links reported in #2912
2026-09-15 12:12:22 -07:00
Aurora
d60d88488f
docs: fix dead Starlette, Uvicorn and Docker Hub links (#2912)
www.starlette.io and www.uvicorn.org no longer resolve (NXDOMAIN); both
projects now publish their documentation at starlette.dev and uvicorn.dev
respectively. The Docker Hub link was missing the /r/ path segment, which
returns a 404.

- docs/plugin_hooks.rst: www.starlette.io/middleware/ -> starlette.dev/middleware/
- docs/changelog.rst (x3): www.uvicorn.org/ -> uvicorn.dev/
- docs/changelog.rst: hub.docker.com/datasetteproject/... -> hub.docker.com/r/datasetteproject/...

All four replacement URLs verified to return HTTP 200.
2026-09-15 11:57:39 -07:00
Alex Garcia
374b194ff5
Add /-/tasks introspection endpoint for supervised background tasks (#2892)
Co-authored-by: Simon Willison <swillison@gmail.com>
2026-09-15 11:56:53 -07:00
Alex Garcia
784695aea6 Ensure startup() runs before any plugin ASGI middleware (#2891) 2026-09-15 11:22:45 -07:00
Alex Garcia
cca08d2886 Add shutdown() plugin hook with ordered graceful teardown (#2890) 2026-09-15 11:09:57 -07:00
Alex Garcia
0c31778580 Add datasette.add_background_task() with supervised launch after startup (#2889) 2026-09-15 10:55:54 -07:00
Simon Willison
b0407867cc Tweak display of count all button, refs #2914 2026-09-15 10:16:16 -07:00
Simon Willison
1538832830 /db/table/-/count? endpoint, fixed 'count all' button
Closes #2914
2026-09-15 10:03:17 -07:00
Alex Garcia
063eeae83d
Slim down CI runs per commit (#2911)
- Only run push-triggered workflows on main; PRs are covered by
  pull_request, so branch commits no longer run everything twice
- Test SQLite versions: single job, Python 3.13 + SQLite 3.25
- Fold coverage into the Python 3.14 test job (including serial
  tests) and remove test-coverage.yml; bump codecov-action to v5
- Cancel superseded in-progress runs on pull requests
- Prettier now also runs on pull requests

Claude-Session: https://claude.ai/code/session_012G7fa8HqJy9AkSjkocetAz
2026-09-15 08:51:24 -07:00
Simon Willison
b338c6f5f6
Migrate from httpx to httpx2, closes #2879
https://claude.ai/code/session_01Xdqoneq8ddvruVZETo6rFf
2026-09-10 19:44:49 -07:00
Simon Willison
61400fba1a Fix Docker release builds on main with Bookworm base image
Apply the Dockerfile fix from 0.65.x to avoid expired Bullseye security repository metadata.

Original fix: 91fa786de9
2026-09-10 18:52:19 -07:00
Simon Willison
186be52863 Publish stable and latest documentation databases to S3 2026-09-10 18:29:54 -07:00
Simon Willison
36acd1ea92 Publish packages when releases are published 2026-09-10 16:53:13 -07:00
Simon Willison
5e7cdaabbd Release 1.0a39 2026-09-10 16:52:52 -07:00
Simon Willison
92c7d4b608 Limit derived-table permissions to one source hop
Simplify the solution to 5de0c1724e - avoid contextvar.
2026-09-10 16:52:52 -07:00
Simon Willison
f70edbfa60 Filter incoming foreign-key relationships by view permission 2026-09-10 16:52:52 -07:00
Simon Willison
b97bb5f016 Reconcile write-timeout regression with the per-call limit 2026-09-10 16:52:26 -07:00
Simon Willison
e036907fc3 Reject structured row writes to virtual and shadow tables 2026-09-10 16:52:25 -07:00
Simon Willison
3f8d8417f6 Inherit source permissions for FTS vocabulary tables 2026-09-10 16:52:25 -07:00
Simon Willison
d334539a1e Deny SQLite statistics table access through a default hook 2026-09-10 16:52:25 -07:00
Simon Willison
628cec8f0c Block framing of stored-query mutation forms 2026-09-09 08:39:03 -07:00
Simon Willison
506c4bb522 Match table permission identities using SQLite case semantics 2026-09-09 08:39:03 -07:00
Simon Willison
e429bd2efa Reuse trusted magic parameter bindings for CSV exports 2026-09-09 08:39:03 -07:00
Alex Garcia
c6ba7b3298 Refuse API token creation from restricted actors 2026-09-09 08:38:40 -07:00
Simon Willison
ac2a9a43a5 Add a default execution time limit to writes 2026-09-08 21:17:03 -07:00
Simon Willison
9d3d741620 Require view permission before using row labels in flash messages 2026-09-08 21:16:36 -07:00
Simon Willison
ceef351622 Protect personalized dynamic responses from shared caching 2026-09-08 21:16:36 -07:00
Simon Willison
7e6039b8df Normalize URL column schemes consistently 2026-09-08 21:16:35 -07:00
Simon Willison
d43a04eb54 Authorize row resources before resolving primary keys 2026-09-08 21:16:35 -07:00
Simon Willison
8b10f58e1b Reject untrusted table-valued PRAGMA reads 2026-09-08 21:16:35 -07:00
Simon Willison
4b8f3b484d Keep private row and table responses out of shared caches 2026-09-08 21:16:35 -07:00
Simon Willison
1be4df77ac Reject invalid token expiry input 2026-09-08 21:16:35 -07:00
Simon Willison
6aa58bf4e5 Authorize configured full-text search targets 2026-09-08 21:16:35 -07:00
Simon Willison
22c601b3d0 Redact configuration keys case-insensitively 2026-09-08 21:16:35 -07:00
Simon Willison
e949ae46de Use unshadowable table classification 2026-09-08 21:16:35 -07:00
Simon Willison
a365903d56 Require view permission before returning written rows 2026-09-08 21:16:35 -07:00
Simon Willison
4c56ce2103 Escape identifiers in upsert row readback 2026-09-08 21:16:35 -07:00
Simon Willison
158c88f259 Escape primary-key cell values in row pages 2026-09-08 21:16:34 -07:00
Simon Willison
35232b5c37 Escape primary-key identifiers in row queries 2026-09-08 21:16:34 -07:00
Simon Willison
c01e95f3bd Filter foreign-key helper targets by view permission 2026-09-08 21:14:10 -07:00
Simon Willison
bf348a22fc Escape LIKE metacharacters in FTS detection 2026-09-08 21:12:23 -07:00
Simon Willison
59618371e9 Validate URL before rendering column links 2026-09-08 21:11:44 -07:00
Simon Willison
5de0c1724e Viewing derived table requires permission for both table and its source
Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 18:09:23 -07:00
Simon Willison
d06737b6f4 Fix CREATE VIEW analysis on Python 3.10 2026-09-08 10:49:52 -07:00
Simon Willison
6473a7ecb0 Clearly document relationship between execute-sql and facets
Refs GHSA-5fff-xcm9-q6vh

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:49:52 -07:00
Simon Willison
f6d0f9bd38 detect_fts() now uses parameterized SQL
Refs GHSA-jcvx-2fh3-pjfp

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:20 -07:00
Simon Willison
c899beaebe escape_sqlite() against column names
Refs GHSA-jcvx-2fh3-pjfp

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:20 -07:00
Simon Willison
3ae092896d Only allow /db/name/-/schema against tables and views
Refs GHSA-926p-cw2f-643h

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:20 -07:00
Simon Willison
5d9a74f370 Check view-table as part of /-/schema and /db/-/schema
Refs GHSA-926p-cw2f-643h

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
01bf476d51 Require view-instance permission for /-/allowed
Refs GHSA-hp2x-vx2r-6vxg

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
4904249025 Use .textContent for column names in column chooser
Refs GHSA-hp2x-vx2r-6vxg

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
435e55ff0a Remove JSON syntax highlighting
Refs GHSA-hp2x-vx2r-6vxg

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
f8e8e65af7 actor cookie respects expire_after
Refs GHSA-53fc-rhfg-h7qp

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
577aeb73f0 Disallow ?_through= if user lacks view-table permission
Refs GHSA-53fc-rhfg-h7qp

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
c280c47424 POST /db/-/create checks table-level insert/update/alter permissions
Refs GHSA-53fc-rhfg-h7qp

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
4d0a2f2e84 execute-write: Check view-table for every table in a CREATE VIEW
Refs GHSA-53fc-rhfg-h7qp

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Simon Willison
c7944fc454 Skip deploy if environment variables are missing 2026-09-08 10:48:19 -07:00
Simon Willison
bdaa8cc76c Disable extension loading once --load-extension extensions are loaded
Refs GHSA-2mvv-ffvc-q5p6

Co-authored-by: Alex Garcia <15178711+asg017@users.noreply.github.com>
2026-09-08 10:48:19 -07:00
Zain Dana Harper
7403ae68bb Give each non-blocking write a distinct task id, refs #2860, #2859
execute_write_fn(fn, block=False) is documented to return "a UUID representing
the queued task". Two things stopped that being true.

_send_to_write_thread() derived the id from uuid.uuid5(NAMESPACE_DNS,
"datasette.io"), which is deterministic, so every non-blocking write in every
database in every process returned 3f143baa-4e3d-5842-a36f-4fa2f683b72f. A
constant cannot identify a particular task. Now uuid4(). Refs #2860.

With num_sql_threads=0 there is no write thread, so execute_write_fn took the
synchronous branch and `result` was the write function's return value, normally
None. The block=False path then unpacked it unconditionally and raised
TypeError: cannot unpack non-iterable NoneType object. The non-threaded branch
now returns the same (task_id, reply_future) shape, with the future already
resolved because the write has finished, so both modes share one code path.
Refs #2859.

test_execute_write_fn_block_false only asserted isinstance(task_id, uuid.UUID),
which a constant satisfies. The new test is parametrized over threaded and
non-threaded and asserts two calls return different ids, so either regression
fails it.
2026-09-08 10:03:52 -07:00
Simon Willison
bdc9731740
check-latest: true, add 3.15 to test matrix, to test RCs (#2895)
See https://simonwillison.net/2026/Sep/1/python-315-rc-2/
2026-09-01 13:37:15 -07:00
Alex Garcia
3e018bb1b5
Run startup via ASGI lifespan instead of waiting for the first request (#2887)
* Run startup via ASGI lifespan instead of waiting for the first request
* Ensure immutable table counts still precompute when startup ran first
2026-09-01 09:39:25 -07:00
Alex Garcia
e78b8a2e6a
Run datasette serve startup and uvicorn on a single event loop (#2886)
* Run datasette serve startup and uvicorn on a single event loop
* Move the serve-subprocess test plumbing into a conftest fixture
* Fix datasette-litestream URL and trim marker-task test comments
* Explain why serve_with_plugins needs a subprocess and plugin files
* Apply ruff 0.16 and black fixes
* Tweaked some comments
2026-09-01 09:32:37 -07:00
Simon Willison
0337fba234
disable_fts() before dropping table
Closes #2874
2026-08-10 15:03:29 -07:00
91 changed files with 5562 additions and 564 deletions

View file

@ -14,24 +14,46 @@ jobs:
deploy: deploy:
runs-on: ubuntu-latest runs-on: ubuntu-latest
steps: steps:
- name: Check deployment prerequisites
id: deployment-prerequisites
env:
GCP_SA_KEY: ${{ secrets.GCP_SA_KEY }}
LATEST_DATASETTE_SECRET: ${{ secrets.LATEST_DATASETTE_SECRET }}
run: |
missing=()
for variable in GCP_SA_KEY LATEST_DATASETTE_SECRET; do
if [[ -z "${!variable:-}" ]]; then
missing+=("$variable")
fi
done
if (( ${#missing[@]} )); then
echo "::notice::Skipping deployment because required environment variables are missing: ${missing[*]}"
echo "available=false" >> "$GITHUB_OUTPUT"
else
echo "available=true" >> "$GITHUB_OUTPUT"
fi
- name: Check out datasette - name: Check out datasette
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
uses: actions/checkout@v7 uses: actions/checkout@v7
- name: Set up Python - name: Set up Python
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
uses: actions/setup-python@v6 uses: actions/setup-python@v6
with: with:
python-version: "3.13" python-version: "3.13"
cache: pip cache: pip
- name: Install Python dependencies - name: Install Python dependencies
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
run: | run: |
python -m pip install --upgrade pip python -m pip install --upgrade pip
python -m pip install . --group dev python -m pip install . --group dev
python -m pip install sphinx-to-sqlite==0.1a1 python -m pip install sphinx-to-sqlite==0.1a1 "s3-credentials>=0.17"
- name: Run tests - name: Run tests
if: ${{ github.ref == 'refs/heads/main' }} if: ${{ steps.deployment-prerequisites.outputs.available == 'true' && github.ref == 'refs/heads/main' }}
run: | run: |
pytest -n auto -m "not serial" pytest -n auto -m "not serial"
pytest -m "serial" pytest -m "serial"
- name: Build fixtures.db and other files needed to deploy the demo - name: Build fixtures.db and other files needed to deploy the demo
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
run: |- run: |-
python tests/fixtures.py \ python tests/fixtures.py \
fixtures.db \ fixtures.db \
@ -39,14 +61,18 @@ jobs:
fixtures-metadata.json \ fixtures-metadata.json \
plugins \ plugins \
--extra-db-filename extra_database.db --extra-db-filename extra_database.db
# Package the config with the plugins, excluding test-only plugin secrets
# that reference temporary files outside the deployed container.
jq 'del(.plugins)' fixtures-config.json > plugins/fixtures-config.json
- name: Build docs.db - name: Build docs.db
if: ${{ github.ref == 'refs/heads/main' }} if: ${{ steps.deployment-prerequisites.outputs.available == 'true' && github.ref == 'refs/heads/main' }}
run: |- run: |-
cd docs cd docs
DISABLE_SPHINX_INLINE_TABS=1 sphinx-build -b xml . _build DISABLE_SPHINX_INLINE_TABS=1 sphinx-build -b xml . _build
sphinx-to-sqlite ../docs.db _build sphinx-to-sqlite ../docs.db _build
cd .. cd ..
- name: Set up the alternate-route demo - name: Set up the alternate-route demo
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
run: | run: |
echo ' echo '
from datasette import hookimpl from datasette import hookimpl
@ -58,6 +84,7 @@ jobs:
' > plugins/alternative_route.py ' > plugins/alternative_route.py
cp fixtures.db fixtures2.db cp fixtures.db fixtures2.db
- name: And the counters writable stored query demo - name: And the counters writable stored query demo
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
run: | run: |
cat > plugins/counters.py <<EOF cat > plugins/counters.py <<EOF
from datasette import hookimpl from datasette import hookimpl
@ -97,12 +124,15 @@ jobs:
# cat metadata.json # cat metadata.json
- id: auth - id: auth
name: Authenticate to Google Cloud name: Authenticate to Google Cloud
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
uses: google-github-actions/auth@v3 uses: google-github-actions/auth@v3
with: with:
credentials_json: ${{ secrets.GCP_SA_KEY }} credentials_json: ${{ secrets.GCP_SA_KEY }}
- name: Set up Cloud SDK - name: Set up Cloud SDK
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
uses: google-github-actions/setup-gcloud@v3 uses: google-github-actions/setup-gcloud@v3
- name: Deploy to Cloud Run - name: Deploy to Cloud Run
if: ${{ steps.deployment-prerequisites.outputs.available == 'true' }}
env: env:
LATEST_DATASETTE_SECRET: ${{ secrets.LATEST_DATASETTE_SECRET }} LATEST_DATASETTE_SECRET: ${{ secrets.LATEST_DATASETTE_SECRET }}
run: |- run: |-
@ -117,16 +147,16 @@ jobs:
--plugins-dir=plugins \ --plugins-dir=plugins \
--branch=$GITHUB_SHA \ --branch=$GITHUB_SHA \
--version-note=$GITHUB_SHA \ --version-note=$GITHUB_SHA \
--extra-options="--setting template_debug 1 --setting trace_debug 1 --crossdb --root" \ --extra-options="--config plugins/fixtures-config.json --setting template_debug 1 --setting trace_debug 1 --crossdb --root" \
--install 'datasette-ephemeral-tables>=0.2.2' \ --install 'datasette-ephemeral-tables>=0.2.2' \
--service "datasette-latest$SUFFIX" \ --service "datasette-latest$SUFFIX" \
--secret $LATEST_DATASETTE_SECRET --secret $LATEST_DATASETTE_SECRET
- name: Deploy to docs as well (only for main) - name: Upload latest documentation database to S3 (only for main)
if: ${{ github.ref == 'refs/heads/main' }} if: ${{ steps.deployment-prerequisites.outputs.available == 'true' && github.ref == 'refs/heads/main' }}
env:
AWS_ACCESS_KEY_ID: ${{ secrets.S3_DATASETTE_DOCS_ACCESS_KEY }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.S3_DATASETTE_DOCS_SECRET_KEY }}
run: |- run: |-
# Deploy docs.db to a different service # Keep development documentation separate from the stable release database.
datasette publish cloudrun docs.db \ s3-credentials put-object datasette-docs latest/docs.db docs.db \
--branch=$GITHUB_SHA \ --content-type application/octet-stream
--version-note=$GITHUB_SHA \
--extra-options="--setting template_debug 1" \
--service=datasette-docs-latest

View file

@ -2,9 +2,15 @@ name: Playwright
on: on:
push: push:
branches:
- main
pull_request: pull_request:
workflow_dispatch: workflow_dispatch:
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
permissions: permissions:
contents: read contents: read

View file

@ -1,6 +1,15 @@
name: Check JavaScript for conformance with Prettier name: Check JavaScript for conformance with Prettier
on: [push] on:
push:
branches:
- main
pull_request:
workflow_dispatch:
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
permissions: permissions:
contents: read contents: read

View file

@ -2,7 +2,7 @@ name: Publish Python Package
on: on:
release: release:
types: [created] types: [published]
permissions: permissions:
contents: read contents: read
@ -51,6 +51,8 @@ jobs:
- name: Publish - name: Publish
uses: pypa/gh-action-pypi-publish@release/v1 uses: pypa/gh-action-pypi-publish@release/v1
# After the first non-prerelease 1.0 release, disable this job on 0.65.x,
# even for later 0.65 releases, so they cannot overwrite the 1.0 stable docs.
deploy_static_docs: deploy_static_docs:
runs-on: ubuntu-latest runs-on: ubuntu-latest
needs: [deploy] needs: [deploy]
@ -66,26 +68,20 @@ jobs:
- name: Install dependencies - name: Install dependencies
run: | run: |
python -m pip install . --group dev python -m pip install . --group dev
python -m pip install sphinx-to-sqlite==0.1a1 python -m pip install sphinx-to-sqlite==0.1a1 "s3-credentials>=0.17"
- name: Build docs.db - name: Build docs.db
run: |- run: |-
cd docs cd docs
DISABLE_SPHINX_INLINE_TABS=1 sphinx-build -b xml . _build DISABLE_SPHINX_INLINE_TABS=1 sphinx-build -b xml . _build
sphinx-to-sqlite ../docs.db _build sphinx-to-sqlite ../docs.db _build
cd .. cd ..
- id: auth - name: Upload stable documentation database to S3
name: Authenticate to Google Cloud env:
uses: google-github-actions/auth@v2 AWS_ACCESS_KEY_ID: ${{ secrets.S3_DATASETTE_DOCS_ACCESS_KEY }}
with: AWS_SECRET_ACCESS_KEY: ${{ secrets.S3_DATASETTE_DOCS_SECRET_KEY }}
credentials_json: ${{ secrets.GCP_SA_KEY }}
- name: Set up Cloud SDK
uses: google-github-actions/setup-gcloud@v3
- name: Deploy stable-docs.datasette.io to Cloud Run
run: |- run: |-
gcloud config set run/region us-central1 s3-credentials put-object datasette-docs docs.db docs.db \
gcloud config set project datasette-222320 --content-type application/octet-stream
datasette publish cloudrun docs.db \
--service=datasette-docs-stable
deploy_docker: deploy_docker:
runs-on: ubuntu-latest runs-on: ubuntu-latest

View file

@ -1,6 +1,15 @@
name: Check spelling in documentation name: Check spelling in documentation
on: [push, pull_request] on:
push:
branches:
- main
pull_request:
workflow_dispatch:
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
permissions: permissions:
contents: read contents: read

View file

@ -1,40 +0,0 @@
name: Calculate test coverage
on:
push:
branches:
- main
pull_request:
branches:
- main
permissions:
contents: read
jobs:
test:
runs-on: ubuntu-latest
steps:
- name: Check out datasette
uses: actions/checkout@v7
- name: Set up Python
uses: actions/setup-python@v6
with:
python-version: '3.12'
cache: 'pip'
cache-dependency-path: '**/pyproject.toml'
- name: Install Python dependencies
run: |
python -m pip install --upgrade pip
python -m pip install . --group dev
python -m pip install pytest-cov
- name: Run tests
run: |-
ls -lah
cat .coveragerc
pytest -m "not serial" --cov=datasette --cov-config=.coveragerc --cov-report xml:coverage.xml --cov-report term -x
ls -lah
- name: Upload coverage report
uses: codecov/codecov-action@v1
with:
token: ${{ secrets.CODECOV_TOKEN }}
file: coverage.xml

View file

@ -2,9 +2,15 @@ name: Test in Pyodide with shot-scraper
on: on:
push: push:
branches:
- main
pull_request: pull_request:
workflow_dispatch: workflow_dispatch:
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
permissions: permissions:
contents: read contents: read

View file

@ -1,6 +1,15 @@
name: Test SQLite versions name: Test SQLite versions
on: [push, pull_request] on:
push:
branches:
- main
pull_request:
workflow_dispatch:
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
permissions: permissions:
contents: read contents: read
@ -12,10 +21,10 @@ jobs:
strategy: strategy:
matrix: matrix:
platform: [ubuntu-latest] platform: [ubuntu-latest]
python-version: ["3.10", "3.11", "3.12", "3.13", "3.14"] python-version: ["3.13"]
sqlite-version: [ sqlite-version: [
#"3", # latest version #"3", # latest version
"3.46", #"3.46",
#"3.45", #"3.45",
#"3.27", #"3.27",
#"3.26", #"3.26",

View file

@ -1,6 +1,15 @@
name: Test name: Test
on: [push, pull_request] on:
push:
branches:
- main
pull_request:
workflow_dispatch:
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
permissions: permissions:
contents: read contents: read
@ -11,16 +20,20 @@ jobs:
strategy: strategy:
fail-fast: false fail-fast: false
matrix: matrix:
python-version: ["3.10", "3.11", "3.12", "3.13", "3.14"] python-version: ["3.10", "3.11", "3.12", "3.13", "3.14", "3.15"]
include:
- python-version: "3.14"
coverage: true
steps: steps:
- uses: actions/checkout@v7 - uses: actions/checkout@v7
- name: Set up Python ${{ matrix.python-version }} - name: Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v6 uses: actions/setup-python@v7
with: with:
python-version: ${{ matrix.python-version }} python-version: ${{ matrix.python-version }}
allow-prereleases: true allow-prereleases: true
cache: pip cache: pip
cache-dependency-path: pyproject.toml cache-dependency-path: pyproject.toml
check-latest: true
- name: Build extension for --load-extension test - name: Build extension for --load-extension test
run: |- run: |-
(cd tests && gcc ext.c -fPIC -shared -o ext.so) (cd tests && gcc ext.c -fPIC -shared -o ext.so)
@ -28,12 +41,27 @@ jobs:
run: | run: |
pip install . --group dev pip install . --group dev
pip freeze pip freeze
- name: Install pytest-cov
if: ${{ matrix.coverage }}
run: pip install pytest-cov
- name: Run tests - name: Run tests
run: | run: |
pytest -n auto -m "not serial" if [ "${{ matrix.coverage }}" = "true" ]; then
pytest -m "serial" COV="--cov=datasette --cov-config=.coveragerc"
pytest -n auto -m "not serial" $COV --cov-report=
pytest -m "serial" $COV --cov-append --cov-report xml:coverage.xml --cov-report term
else
pytest -n auto -m "not serial"
pytest -m "serial"
fi
# And the test that exceeds a localhost HTTPS server # And the test that exceeds a localhost HTTPS server
tests/test_datasette_https_server.sh tests/test_datasette_https_server.sh
- name: Upload coverage report
if: ${{ matrix.coverage }}
uses: codecov/codecov-action@v5
with:
token: ${{ secrets.CODECOV_TOKEN }}
files: coverage.xml
- name: Black - name: Black
run: | run: |
black --version black --version

View file

@ -1,4 +1,4 @@
FROM python:3.11.0-slim-bullseye as build FROM python:3.11-slim-bookworm AS build
# Version of Datasette to install, e.g. 0.55 # Version of Datasette to install, e.g. 0.55
# docker build . -t datasette --build-arg VERSION=0.55 # docker build . -t datasette --build-arg VERSION=0.55

View file

@ -1,6 +1,7 @@
from datasette.permissions import Permission # noqa from datasette.permissions import Permission # noqa
from datasette.version import __version_info__, __version__ # noqa from datasette.version import __version_info__, __version__ # noqa
from datasette.events import Event # noqa from datasette.events import Event # noqa
from datasette.background_tasks import BackgroundTask, BackgroundTaskSupervisor # noqa
from datasette.tokens import TokenHandler, TokenInvalid, TokenRestrictions # noqa from datasette.tokens import TokenHandler, TokenInvalid, TokenRestrictions # noqa
from datasette.utils.asgi import ( # noqa from datasette.utils.asgi import ( # noqa
Forbidden, Forbidden,

View file

@ -28,7 +28,7 @@ import urllib.parse
from concurrent import futures from concurrent import futures
from pathlib import Path from pathlib import Path
import httpx import httpx2
from itsdangerous import BadSignature, URLSafeSerializer from itsdangerous import BadSignature, URLSafeSerializer
from jinja2 import ( from jinja2 import (
ChoiceLoader, ChoiceLoader,
@ -42,6 +42,7 @@ from jinja2.exceptions import TemplateNotFound
from markupsafe import Markup, escape from markupsafe import Markup, escape
from . import stored_queries, write_sql from . import stored_queries, write_sql
from .background_tasks import BackgroundTask, BackgroundTaskSupervisor
from .column_types import SQLiteType from .column_types import SQLiteType
from .csrf import CrossOriginProtectionMiddleware from .csrf import CrossOriginProtectionMiddleware
from .database import Database, QueryInterrupted from .database import Database, QueryInterrupted
@ -145,6 +146,7 @@ from .views.stored_queries import (
) )
from .views.table import ( from .views.table import (
TableAutocompleteView, TableAutocompleteView,
TableCountView,
TableDropView, TableDropView,
TableFragmentView, TableFragmentView,
TableInsertView, TableInsertView,
@ -315,7 +317,7 @@ def _permission_cache_key(actor, action, parent, child):
actor_key = ( actor_key = (
json.dumps(actor, sort_keys=True, default=repr) if actor is not None else None json.dumps(actor, sort_keys=True, default=repr) if actor is not None else None
) )
return (actor_key, action, parent, child) return (actor_key, action.name, parent, action.normalize_child(child))
async def favicon(request, send): async def favicon(request, send):
@ -422,6 +424,7 @@ class Datasette:
default_deny=False, default_deny=False,
): ):
self._startup_invoked = False self._startup_invoked = False
self._shutdown_invoked = False
self._closed = False self._closed = False
assert config_dir is None or isinstance( assert config_dir is None or isinstance(
config_dir, Path config_dir, Path
@ -453,8 +456,11 @@ class Datasette:
self.databases = collections.OrderedDict() self.databases = collections.OrderedDict()
self.actions = {} # .invoke_startup() will populate this self.actions = {} # .invoke_startup() will populate this
self._column_types = {} # .invoke_startup() will populate this self._column_types = {} # .invoke_startup() will populate this
self._setup_db_done = False
self._suppress_background_tasks = False
try: try:
self._refresh_schemas_lock = asyncio.Lock() self._refresh_schemas_lock = asyncio.Lock()
self._startup_lock = asyncio.Lock()
except RuntimeError as rex: except RuntimeError as rex:
# Workaround for intermittent test failure, see: # Workaround for intermittent test failure, see:
# https://github.com/simonw/datasette/issues/1802 # https://github.com/simonw/datasette/issues/1802
@ -462,8 +468,10 @@ class Datasette:
loop = asyncio.new_event_loop() loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop) asyncio.set_event_loop(loop)
self._refresh_schemas_lock = asyncio.Lock() self._refresh_schemas_lock = asyncio.Lock()
self._startup_lock = asyncio.Lock()
else: else:
raise raise
self._background_tasks = BackgroundTaskSupervisor(self)
self.crossdb = crossdb self.crossdb = crossdb
self.nolock = nolock self.nolock = nolock
if memory or crossdb or not self.files: if memory or crossdb or not self.files:
@ -1529,15 +1537,28 @@ class Datasette:
conn.row_factory = sqlite3.Row conn.row_factory = sqlite3.Row
conn.text_factory = lambda x: str(x, "utf-8", "replace") conn.text_factory = lambda x: str(x, "utf-8", "replace")
if self.sqlite_extensions and database != INTERNAL_DB_NAME: if self.sqlite_extensions and database != INTERNAL_DB_NAME:
# Extension loading is only enabled for as long as it takes to
# load the configured extensions. Leaving it enabled would let
# anyone who can execute SQL call load_extension() themselves.
conn.enable_load_extension(True) conn.enable_load_extension(True)
for extension in self.sqlite_extensions: try:
# "extension" is either a string path to the extension for extension in self.sqlite_extensions:
# or a 2-item tuple that specifies which entrypoint to load. # "extension" is either a string path to the extension
if isinstance(extension, tuple): # or a 2-item tuple that specifies which entrypoint to load.
path, entrypoint = extension if isinstance(extension, tuple):
conn.execute("SELECT load_extension(?, ?)", [path, entrypoint]) path, entrypoint = extension
else: if sys.version_info >= (3, 12):
conn.execute("SELECT load_extension(?)", [extension]) conn.load_extension(path, entrypoint=entrypoint)
else:
# Connection.load_extension() only gained the
# entrypoint argument in Python 3.12
conn.execute(
"SELECT load_extension(?, ?)", [path, entrypoint]
)
else:
conn.load_extension(extension)
finally:
conn.enable_load_extension(False)
if self.setting("cache_size_kb"): if self.setting("cache_size_kb"):
conn.execute(f"PRAGMA cache_size=-{self.setting('cache_size_kb')}") conn.execute(f"PRAGMA cache_size=-{self.setting('cache_size_kb')}")
# pylint: disable=no-member # pylint: disable=no-member
@ -1730,8 +1751,145 @@ class Datasette:
sql, params = await build_allowed_resources_sql( sql, params = await build_allowed_resources_sql(
self, actor, action, parent=parent, include_is_private=include_is_private self, actor, action, parent=parent, include_is_private=include_is_private
) )
if action == "view-table":
sql, params = await self._apply_derived_table_permissions_to_sql(
sql,
params,
actor=actor,
parent=parent,
include_is_private=include_is_private,
)
return ResourcesSQL(sql, params) return ResourcesSQL(sql, params)
async def _allowed_derived_table_source(
self, database, source, *, actor, dependencies
):
"""Check an immediate source, denying sources that are themselves derived."""
if any(
TableResource.normalize_child(table)
== TableResource.normalize_child(source)
for table in dependencies
):
return False
# The source has no dependency in this map. Evaluate its own permission
# and prerequisites without starting another dependency check.
verdicts = await self._allowed_many(
actions=["view-table"],
resource=TableResource(database, source),
actor=actor,
check_derived=False,
)
return verdicts["view-table"]
async def _apply_derived_table_permissions_to_sql(
self,
sql,
params,
*,
actor,
parent,
include_is_private,
):
databases = (
[(parent, self.databases[parent])]
if parent in self.databases
else ([] if parent is not None else list(self.databases.items()))
)
dependency_maps = dict(
zip(
(name for name, _ in databases),
await asyncio.gather(
*(db.derived_table_dependencies() for _, db in databases)
),
)
)
dependencies = [
(database_name, child, source)
for database_name, dependency_map in dependency_maps.items()
for child, source in dependency_map.items()
]
if not dependencies:
return sql, params
sources = sorted(
{(database_name, source) for database_name, _, source in dependencies}
)
actor_verdicts = await asyncio.gather(
*(
self._allowed_derived_table_source(
database_name,
source,
actor=actor,
dependencies=dependency_maps[database_name],
)
for database_name, source in sources
)
)
actor_allowed = dict(zip(sources, actor_verdicts))
anonymous_allowed = {}
if include_is_private:
anonymous_verdicts = await asyncio.gather(
*(
self._allowed_derived_table_source(
database_name,
source,
actor=None,
dependencies=dependency_maps[database_name],
)
for database_name, source in sources
)
)
anonymous_allowed = dict(zip(sources, anonymous_verdicts))
wrapped_params = dict(params)
derived_rows = [
[
database_name,
child,
int(actor_allowed[(database_name, source)]),
*(
[int(anonymous_allowed[(database_name, source)])]
if include_is_private
else []
),
]
for database_name, child, source in dependencies
]
derived_param = "_datasette_derived_permissions"
while derived_param in wrapped_params:
derived_param += "_"
wrapped_params[derived_param] = json.dumps(derived_rows)
derived_columns = "parent, child, source_allowed"
select_columns = "allowed.parent, allowed.child, allowed.reason"
if include_is_private:
derived_columns += ", source_anonymous_allowed"
select_columns += (
", CASE WHEN derived.source_anonymous_allowed = 0 "
"THEN 1 ELSE allowed.is_private END AS is_private"
)
wrapped_sql = f"""
WITH derived_permissions({derived_columns}) AS (
SELECT
json_extract(value, '$[0]'),
json_extract(value, '$[1]'),
json_extract(value, '$[2]')
{", json_extract(value, '$[3]')" if include_is_private else ""}
FROM json_each(:{derived_param})
),
allowed AS (
{sql}
)
SELECT {select_columns}
FROM allowed
LEFT JOIN derived_permissions AS derived
ON allowed.parent = derived.parent AND allowed.child = derived.child COLLATE NOCASE
WHERE COALESCE(derived.source_allowed, 1) = 1
ORDER BY allowed.parent, allowed.child
""".strip()
return wrapped_sql, wrapped_params
async def allowed_resources( async def allowed_resources(
self, self,
action: str, action: str,
@ -1934,6 +2092,12 @@ class Datasette:
) )
# {"edit-schema": True, "drop-table": True, "insert-row": False} # {"edit-schema": True, "drop-table": True, "insert-row": False}
""" """
return await self._allowed_many(
actions=actions, resource=resource, actor=actor, check_derived=True
)
async def _allowed_many(self, *, actions, resource, actor, check_derived):
"""Evaluate permissions, optionally applying the one-hop source policy."""
from datasette.permissions import ( from datasette.permissions import (
_permission_check_cache, _permission_check_cache,
_skip_permission_checks, _skip_permission_checks,
@ -1971,7 +2135,7 @@ class Datasette:
to_check = [] to_check = []
for name in expanded: for name in expanded:
if cache is not None: if cache is not None:
key = _permission_cache_key(actor, name, parent, child) key = _permission_cache_key(actor, self.actions[name], parent, child)
if key in cache: if key in cache:
final[name] = cache[key] final[name] = cache[key]
continue continue
@ -1987,6 +2151,28 @@ class Datasette:
child=child, child=child,
) )
if (
check_derived
and "view-table" in to_check
and raw.get("view-table")
and isinstance(resource, TableResource)
and parent in self.databases
):
dependencies = await self.databases[parent].derived_table_dependencies()
source = next(
(
source
for table, source in dependencies.items()
if TableResource.normalize_child(table)
== TableResource.normalize_child(child)
),
None,
)
if source is not None:
raw["view-table"] = await self._allowed_derived_table_source(
parent, source, actor=actor, dependencies=dependencies
)
def resolve(name): def resolve(name):
# final verdict = own rules AND verdict of also_requires chain # final verdict = own rules AND verdict of also_requires chain
if name in final: if name in final:
@ -2004,7 +2190,9 @@ class Datasette:
# Cache the freshly computed checks # Cache the freshly computed checks
if cache is not None: if cache is not None:
for name in to_check: for name in to_check:
cache[_permission_cache_key(actor, name, parent, child)] = final[name] cache[
_permission_cache_key(actor, self.actions[name], parent, child)
] = final[name]
# Log every check (including cache hits) for the debug page, # Log every check (including cache hits) for the debug page,
# dependencies before the actions that required them # dependencies before the actions that required them
@ -2278,6 +2466,21 @@ class Datasette:
) )
return d return d
def _tasks(self):
return {
"tasks": [
{
"name": t.name,
"state": t.state,
"function": t.function,
"started_at": t.started_at,
"exception": repr(t.exception) if t.exception else None,
}
for t in self._background_tasks.tasks()
],
"launched": self._background_tasks.launched,
}
def _actor(self, request): def _actor(self, request):
return {"actor": request.actor} return {"actor": request.actor}
@ -2446,7 +2649,7 @@ class Datasette:
): ):
data = {"a": actor} data = {"a": actor}
if expire_after: if expire_after:
expires_at = int(time.time()) + (24 * 60 * 60) expires_at = int(time.time()) + expire_after
data["e"] = baseconv.base62.encode(expires_at) data["e"] = baseconv.base62.encode(expires_at)
response.set_cookie("ds_actor", self.sign(data, "actor")) response.set_cookie("ds_actor", self.sign(data, "actor"))
@ -2566,6 +2769,12 @@ class Datasette:
), ),
r"/-/threads(\.(?P<format>json))?$", r"/-/threads(\.(?P<format>json))?$",
) )
add_route(
JsonDataView.as_view(
self, "tasks.json", self._tasks, permission="permissions-debug"
),
r"/-/tasks(\.(?P<format>json))?$",
)
add_route( add_route(
JsonDataView.as_view( JsonDataView.as_view(
self, self,
@ -2740,6 +2949,10 @@ class Datasette:
TableSetColumnTypeView.as_view(self), TableSetColumnTypeView.as_view(self),
r"/(?P<database>[^\/\.]+)/(?P<table>[^\/\.]+)/-/set-column-type$", r"/(?P<database>[^\/\.]+)/(?P<table>[^\/\.]+)/-/set-column-type$",
) )
add_route(
TableCountView.as_view(self),
r"/(?P<database>[^\/\.]+)/(?P<table>[^\/\.]+)/-/count$",
)
add_route( add_route(
TableFragmentView.as_view(self), TableFragmentView.as_view(self),
r"/(?P<database>[^\/\.]+)/(?P<table>[^\/\.]+)/-/fragment$", r"/(?P<database>[^\/\.]+)/(?P<table>[^\/\.]+)/-/fragment$",
@ -2803,26 +3016,127 @@ class Datasette:
raise RowNotFound(db.name, table_name, pk_values) raise RowNotFound(db.name, table_name, pk_values)
return ResolvedRow(db, table_name, sql, params, pks, pk_values, results.first()) return ResolvedRow(db, table_name, sql, params, pks, pk_values, results.first())
async def _startup_sequence(self):
"""Idempotently run the full startup sequence: table counts for
immutable databases, then invoke_startup(). Safe to call more than
once and safe to call concurrently - callers block until whichever
call got there first has finished.
This is the single entry point used by both AsgiLifespan (so
real deployments finish startup before accepting requests) and
AsgiRunOnFirstRequest (the fallback for hosts that never send
lifespan events, e.g. DatasetteClient's httpx2.ASGITransport), and
`datasette serve` (cli.py) calls it too. The fast path below checks
both `_startup_invoked` and `_setup_db_done` - not just the former -
so that a bare `await ds.invoke_startup()` made by a caller ahead of
`_startup_sequence()` (which only sets `_startup_invoked`) can't
make this method skip the immutable-database table-count precompute.
"""
if self._startup_invoked and self._setup_db_done:
return
async with self._startup_lock:
if self._startup_invoked and self._setup_db_done:
return
if not self._setup_db_done:
# First time server starts up, calculate table counts for
# immutable databases
for database in self.databases.values():
if not database.is_mutable:
await database.table_counts(limit=60 * 60 * 1000)
self._setup_db_done = True
await self.invoke_startup()
def add_background_task(self, func, name=None) -> BackgroundTask:
"""Register a piece of supervised background work, typically from
a plugin's ``startup`` hook.
``func`` must be a coroutine function taking one positional
argument, the ``Datasette`` instance - core calls ``func(self)``.
Callable any time after ``__init__``: if background tasks haven't
launched yet (the common case - most callers are ``startup`` hooks,
which run before launch), this buffers the registration until they
do; if they've already launched (e.g. called from a request
handler after the server is up), the task starts immediately.
Returns a :class:`~datasette.background_tasks.BackgroundTask`
handle (``.name``, ``.state``, ``.task``, ``.exception``,
``.started_at``, ``.function``, ``.cancel()``).
``name`` defaults to ``func.__qualname__``; on a name collision a
``-2``, ``-3``, ... suffix is appended, since names are how
``/-/tasks`` and log messages identify work.
"""
return self._background_tasks.add(func, name=name)
async def start_background_tasks(self):
"""Run startup (if it hasn't run yet) and launch every registered
background task.
Public entry point for tests, embedders, and headless CLIs (the
``datasette-rss``-style ``fetch --due`` shape) that want supervised
background tasks without running a server - equivalent to what
happens automatically via ASGI lifespan / the first-request
fallback in a served deployment.
"""
await self.invoke_startup()
await self._background_tasks.launch_all()
async def _launch_background_tasks(self):
"""Idempotently launch every registered background task. Private:
this is the entry point wired into the lifecycle trigger lists
(the second entry in both ``AsgiLifespan`` and
``AsgiRunOnFirstRequest``'s ``on_startup``, after
``_startup_sequence``) - not something plugins or embedders should
call directly; use ``add_background_task`` /
``start_background_tasks`` instead.
Positioned after ``_startup_sequence`` in both trigger lists so
launch always happens once every plugin's ``startup`` hook has had
a chance to register work - the ordering guarantee that makes
``add_background_task`` useful. No-ops when
``_suppress_background_tasks`` is set (the ``--get`` CLI path: its
one-shot TestClient request flows through the full ASGI stack,
including the first-request fallback, but must never launch
long-lived background work).
"""
if self._suppress_background_tasks:
return
await self._background_tasks.launch_all()
async def invoke_shutdown(self):
"""Run the graceful teardown sequence: plugin ``shutdown`` hooks,
then cancel and drain supervised background tasks, then close
every database.
"""
if self._shutdown_invoked:
return
self._shutdown_invoked = True
for hook in pm.hook.shutdown(datasette=self):
try:
await await_me_maybe(hook)
except Exception:
logging.getLogger("datasette").exception("shutdown hook failed")
await self._background_tasks.cancel_all(grace=5.0)
self.close()
def app(self): def app(self):
"""Returns an ASGI app function that serves the whole of Datasette""" """Returns an ASGI app function that serves the whole of Datasette"""
routes = self._routes() routes = self._routes()
async def setup_db():
# First time server starts up, calculate table counts for immutable databases
for database in self.databases.values():
if not database.is_mutable:
await database.table_counts(limit=60 * 60 * 1000)
async def _close_on_shutdown():
self.close()
asgi = CrossOriginProtectionMiddleware(DatasetteRouter(self, routes), self) asgi = CrossOriginProtectionMiddleware(DatasetteRouter(self, routes), self)
if self.setting("trace_debug"): if self.setting("trace_debug"):
asgi = AsgiTracer(asgi) asgi = AsgiTracer(asgi)
asgi = AsgiLifespan(asgi, on_shutdown=[_close_on_shutdown]) asgi = AsgiLifespan(
asgi = AsgiRunOnFirstRequest(asgi, on_startup=[setup_db, self.invoke_startup]) asgi,
on_startup=[self._startup_sequence, self._launch_background_tasks],
on_shutdown=[self.invoke_shutdown],
)
for wrapper in pm.hook.asgi_wrapper(datasette=self): for wrapper in pm.hook.asgi_wrapper(datasette=self):
asgi = wrapper(asgi) asgi = wrapper(asgi)
asgi = AsgiRunOnFirstRequest(
asgi,
on_startup=[self._startup_sequence, self._launch_background_tasks],
)
return asgi return asgi
@ -2860,6 +3174,50 @@ class DatasetteRouter:
receive, receive,
max_post_body_bytes=self.ds.setting("max_post_body_bytes"), max_post_body_bytes=self.ds.setting("max_post_body_bytes"),
) )
match, view = resolve_routes(self.routes, path)
is_static = view is favicon or getattr(view, "_datasette_static", False)
original_send = send
async def send(message):
if message["type"] == "http.response.start" and not (
is_static and message["status"] in (200, 304)
):
# Decide privacy after rendering, including for streaming responses
# and error handlers. A public primary resource can still include
# private labels, actor navigation, or cookie-dependent content.
headers = list(message.get("headers", []))
personalized = (
request.actor is not None
or "cookie" in request.headers
or "authorization" in request.headers
or any(key.lower() == b"set-cookie" for key, _ in headers)
)
if personalized:
headers = [
(key, value)
for key, value in headers
if key.lower() != b"cache-control"
]
headers.append((b"cache-control", b"private, no-store"))
# Anonymous responses must not be reused for credentialed requests.
# Preserve any additional variation specified by views or plugins.
vary = [
part.strip()
for key, value in headers
if key.lower() == b"vary"
for part in value.split(b",")
if part.strip()
]
if b"*" not in vary:
for name in (b"Cookie", b"Authorization"):
if name.lower() not in {part.lower() for part in vary}:
vary.append(name)
headers = [(k, v) for k, v in headers if k.lower() != b"vary"]
headers.append((b"vary", b", ".join(vary)))
message = dict(message, headers=headers)
await original_send(message)
# Populate request_messages if ds_messages cookie is present # Populate request_messages if ds_messages cookie is present
try: try:
request._messages = self.ds.unsign( request._messages = self.ds.unsign(
@ -2899,8 +3257,7 @@ class DatasetteRouter:
return await self.handle_401(request, send, token_error) return await self.handle_401(request, send, token_error)
scope_modifications["actor"] = actor or default_actor scope_modifications["actor"] = actor or default_actor
scope = dict(scope, **scope_modifications) scope = dict(scope, **scope_modifications)
request.scope = scope
match, view = resolve_routes(self.routes, path)
if match is None: if match is None:
return await self.handle_404(request, send) return await self.handle_404(request, send)
@ -3215,14 +3572,14 @@ class DatasetteClient:
with _DatasetteClientContext(): with _DatasetteClientContext():
if skip_permission_checks: if skip_permission_checks:
with SkipPermissions(): with SkipPermissions():
async with httpx.AsyncClient( async with httpx2.AsyncClient(
transport=httpx.ASGITransport(app=self.app), transport=httpx2.ASGITransport(app=self.app),
cookies=kwargs.pop("cookies", None), cookies=kwargs.pop("cookies", None),
) as client: ) as client:
return await getattr(client, method)(self._fix(path), **kwargs) return await getattr(client, method)(self._fix(path), **kwargs)
else: else:
async with httpx.AsyncClient( async with httpx2.AsyncClient(
transport=httpx.ASGITransport(app=self.app), transport=httpx2.ASGITransport(app=self.app),
cookies=kwargs.pop("cookies", None), cookies=kwargs.pop("cookies", None),
) as client: ) as client:
return await getattr(client, method)(self._fix(path), **kwargs) return await getattr(client, method)(self._fix(path), **kwargs)
@ -3269,10 +3626,10 @@ class DatasetteClient:
method: HTTP method (e.g., "GET", "POST", "PUT") method: HTTP method (e.g., "GET", "POST", "PUT")
path: The path to request path: The path to request
skip_permission_checks: If True, bypass all permission checks for this request skip_permission_checks: If True, bypass all permission checks for this request
**kwargs: Additional arguments to pass to httpx **kwargs: Additional arguments to pass to httpx2
Returns: Returns:
httpx.Response: The response from the request httpx2.Response: The response from the request
""" """
from datasette.permissions import SkipPermissions from datasette.permissions import SkipPermissions
@ -3281,16 +3638,16 @@ class DatasetteClient:
with _DatasetteClientContext(): with _DatasetteClientContext():
if skip_permission_checks: if skip_permission_checks:
with SkipPermissions(): with SkipPermissions():
async with httpx.AsyncClient( async with httpx2.AsyncClient(
transport=httpx.ASGITransport(app=self.app), transport=httpx2.ASGITransport(app=self.app),
cookies=kwargs.pop("cookies", None), cookies=kwargs.pop("cookies", None),
) as client: ) as client:
return await client.request( return await client.request(
method, self._fix(path, avoid_path_rewrites), **kwargs method, self._fix(path, avoid_path_rewrites), **kwargs
) )
else: else:
async with httpx.AsyncClient( async with httpx2.AsyncClient(
transport=httpx.ASGITransport(app=self.app), transport=httpx2.ASGITransport(app=self.app),
cookies=kwargs.pop("cookies", None), cookies=kwargs.pop("cookies", None),
) as client: ) as client:
return await client.request( return await client.request(

View file

@ -0,0 +1,227 @@
"""
Supervised background-task registration for Datasette core.
Plugins that need long-lived background work (a polling loop, a queue
consumer, a scheduled job runner) register it with
``datasette.add_background_task(func, name=None)`` - typically from a
``startup`` plugin hook - instead of fire-and-forgetting their own
``asyncio.create_task()``. Core owns:
- **references**: every launched ``asyncio.Task`` is kept alive on a
:class:`BackgroundTaskSupervisor`, so it can never be silently garbage
collected the way an unreferenced ``create_task()`` call can be;
- **launch timing**: registered work is buffered until
:meth:`BackgroundTaskSupervisor.launch_all` runs, which core arranges to
happen only after *every* plugin's ``startup`` hook has finished - so
a task that depends on another plugin having registered something first
doesn't need ``tryfirst=True`` ordering tricks;
- **crash surfacing**: an unhandled exception in a background task is
logged with its full traceback to the ``datasette.background_tasks``
logger and recorded on the handle, instead of becoming an "Task
exception was never retrieved" warning nobody sees;
- **cancellation**: :meth:`BackgroundTaskSupervisor.cancel_all` cancels
every task still running and waits (with a grace period) for them to
actually stop.
"""
from __future__ import annotations
import asyncio
import datetime
import functools
import logging
from collections.abc import Awaitable, Callable
logger = logging.getLogger("datasette.background_tasks")
def _utcnow_iso() -> str:
return datetime.datetime.now(datetime.timezone.utc).isoformat()
def _function_path(func: Callable) -> str:
"""Describe the callable without guessing which plugin registered it."""
while isinstance(func, functools.partial):
func = func.func
if not hasattr(func, "__qualname__"):
func = type(func).__call__
return f"{func.__module__}.{func.__qualname__}"
class BackgroundTask:
"""A handle to a single piece of supervised background work.
States: ``registered`` (added but not yet launched) -> ``running`` ->
one of ``completed`` (returned cleanly), ``crashed`` (raised an
exception other than ``CancelledError`` - see ``.exception``), or
``cancelled`` (``.cancel()`` was called, or it was still running at
shutdown).
"""
def __init__(
self,
name: str,
func: Callable[[object], Awaitable[None]],
):
self.name = name
self.state = "registered"
self.task: asyncio.Task | None = None
self.exception: BaseException | None = None
self.started_at: str | None = None
self.function = _function_path(func)
self._func = func
self._supervisor: BackgroundTaskSupervisor | None = None
def cancel(self) -> None:
"""Cancel this task.
If it has already been launched, cancels the underlying
``asyncio.Task`` - its state becomes ``cancelled`` once the
cancellation is observed (asynchronously, via the task's done
callback). If it has not been launched yet, this is a no-op as
far as asyncio is concerned (there's no task to cancel) but it
deregisters the handle from its supervisor so it never runs.
"""
if self.task is not None:
self.task.cancel()
elif self._supervisor is not None:
self._supervisor._deregister(self)
def __repr__(self) -> str:
return f"<BackgroundTask name={self.name!r} state={self.state!r}>"
class BackgroundTaskSupervisor:
"""Owns registration and launch of every :class:`BackgroundTask` for a
single ``Datasette`` instance.
Registration (:meth:`add`) is separate from launch
(:meth:`launch_all`): plugins register work whenever convenient
(typically from a ``startup`` hook, but request handlers can register
dynamic per-job work too), and it either sits buffered until
:meth:`launch_all` runs, or - if :meth:`launch_all` has already run -
starts immediately.
Strong references to every :class:`BackgroundTask` (and its
``asyncio.Task``) are kept for the life of the instance, by design -
that's what makes the enrichments-style "fire-and-forget task gets
garbage collected mid-flight" bug impossible here. There is currently
no pruning of completed/crashed/cancelled tasks, so a plugin that
dynamically registers many short-lived tasks over a long process
lifetime (a per-job registration pattern, e.g. one task per queued
job) will grow this list without bound. That's an accepted v1
trade-off in favour of full introspection (``/-/tasks``); revisit
with a pruning or capping policy if unbounded growth is reported in
practice.
"""
def __init__(self, datasette):
self._datasette = datasette
self._tasks: list[BackgroundTask] = []
self._names = set()
self._launched = False
self._lock = asyncio.Lock()
def add(self, func, name=None) -> BackgroundTask:
base_name = name or getattr(func, "__qualname__", None) or repr(func)
actual_name = self._unique_name(base_name)
handle = BackgroundTask(actual_name, func)
handle._supervisor = self
self._tasks.append(handle)
self._names.add(actual_name)
if self._launched:
self._launch_one(handle)
return handle
def _unique_name(self, base_name: str) -> str:
if base_name not in self._names:
return base_name
n = 2
while f"{base_name}-{n}" in self._names:
n += 1
return f"{base_name}-{n}"
def _deregister(self, handle: BackgroundTask) -> None:
try:
self._tasks.remove(handle)
except ValueError:
pass
self._names.discard(handle.name)
def _launch_one(self, handle: BackgroundTask) -> None:
handle.state = "running"
handle.started_at = _utcnow_iso()
handle.task = asyncio.create_task(
handle._func(self._datasette), name=handle.name
)
handle.task.add_done_callback(functools.partial(_on_task_done, handle))
async def launch_all(self) -> None:
"""Launch every currently-registered task that hasn't launched
yet. Idempotent and safe to call concurrently: subsequent (or
racing) calls are no-ops once the first has set ``self._launched``.
"""
if self._launched:
return
async with self._lock:
if self._launched:
return
self._launched = True
for handle in list(self._tasks):
if handle.task is None:
self._launch_one(handle)
async def cancel_all(self, grace: float = 5.0) -> None:
"""Cancel every task that isn't already done, then wait up to
``grace`` seconds for them to actually finish. Stragglers still
running after that are logged by name (but left to finish or not
on their own - this does not forcibly kill them, asyncio has no
mechanism for that).
"""
handles_by_task = {
handle.task: handle for handle in self._tasks if handle.task is not None
}
pending = [task for task in handles_by_task if not task.done()]
for task in pending:
task.cancel()
if not pending:
return
_done, not_done = await asyncio.wait(pending, timeout=grace)
if not_done:
names = sorted(handles_by_task[task].name for task in not_done)
logger.warning(
"%d background task(s) did not finish within the %.1fs grace "
"period after cancellation: %s",
len(names),
grace,
", ".join(names),
)
def tasks(self) -> list[BackgroundTask]:
"""Return every registered :class:`BackgroundTask`, launched or
not, in registration order. Used by the ``/-/tasks`` debug
endpoint.
"""
return list(self._tasks)
@property
def launched(self) -> bool:
"""Whether :meth:`launch_all` has run yet - lets ``/-/tasks``
distinguish "no tasks registered" from "tasks registered but
nothing has armed the launch yet" without reaching for the
private ``_launched`` attribute.
"""
return self._launched
def _on_task_done(handle: BackgroundTask, task: asyncio.Task) -> None:
if task.cancelled():
handle.state = "cancelled"
return
exc = task.exception()
if exc is not None:
handle.state = "crashed"
handle.exception = exc
logger.error("Background task %r crashed", handle.name, exc_info=exc)
return
handle.state = "completed"

View file

@ -663,16 +663,6 @@ def serve(
# Private utility mechanism for writing unit tests # Private utility mechanism for writing unit tests
return ds return ds
# Run async soundness checks before startup hooks, since invoke_startup
# now populates internal tables which requires querying each database
run_sync(lambda: check_databases(ds))
# Run the "startup" plugin hooks
try:
run_sync(ds.invoke_startup)
except StartupError as e:
raise click.ClickException(e.args[0])
if headers and not get: if headers and not get:
raise click.ClickException("--headers can only be used with --get") raise click.ClickException("--headers can only be used with --get")
@ -680,6 +670,19 @@ def serve(
raise click.ClickException("--token can only be used with --get") raise click.ClickException("--token can only be used with --get")
if get: if get:
# --get means we don't run Uvicorn at all
run_sync(lambda: check_databases(ds))
try:
run_sync(ds.invoke_startup)
except StartupError as e:
raise click.ClickException(e.args[0])
# --get never launches background tasks: TestClient's request below
# flows through the full ASGI stack, including the
# AsgiRunOnFirstRequest fallback, which would otherwise launch them.
ds._suppress_background_tasks = True
client = TestClient(ds) client = TestClient(ds)
request_headers = {} request_headers = {}
if token: if token:
@ -704,34 +707,54 @@ def serve(
sys.exit(exit_code) sys.exit(exit_code)
return return
# Start the server # check_databases, invoke_startup() and the uvicorn server all run on a
url = None # single event loop, so that anything a plugin's "startup" hook schedules
if root: # on the loop (asyncio.create_task, Lock/Queue/Event objects, ...) is
ds.root_enabled = True # still alive when the server starts handling requests.
url = "http://{}:{}{}?token={}".format( async def _serve_async():
host, port, ds.urls.path("-/auth-token"), ds._root_token # Populate internal catalog tables before invoke_startup
) await check_databases(ds)
click.echo(url)
if open_browser: # Run the full startup sequence (immutable-database table-count
if url is None: # precompute + the "startup" plugin hooks) via the same entry point
# Figure out most convenient URL - to table, database or homepage # AsgiLifespan/AsgiRunOnFirstRequest use, so it's not skipped when
path = run_sync(lambda: initial_path_for_datasette(ds)) # uvicorn's lifespan.startup fires moments later.
url = f"http://{host}:{port}{path}" try:
webbrowser.open(url) await ds._startup_sequence()
uvicorn_kwargs = { except StartupError as e:
"host": host, raise click.ClickException(e.args[0])
"port": port,
"log_level": "info", # Start the server
"lifespan": "on", url = None
"workers": 1, if root:
} ds.root_enabled = True
if uds: url = "http://{}:{}{}?token={}".format(
uvicorn_kwargs["uds"] = uds host, port, ds.urls.path("-/auth-token"), ds._root_token
if ssl_keyfile: )
uvicorn_kwargs["ssl_keyfile"] = ssl_keyfile click.echo(url)
if ssl_certfile: if open_browser:
uvicorn_kwargs["ssl_certfile"] = ssl_certfile if url is None:
uvicorn.run(ds.app(), **uvicorn_kwargs) # Figure out most convenient URL - to table, database or homepage
path = await initial_path_for_datasette(ds)
url = f"http://{host}:{port}{path}"
webbrowser.open(url)
uvicorn_kwargs = {
"host": host,
"port": port,
"log_level": "info",
"lifespan": "on",
"workers": 1,
}
if uds:
uvicorn_kwargs["uds"] = uds
if ssl_keyfile:
uvicorn_kwargs["ssl_keyfile"] = ssl_keyfile
if ssl_certfile:
uvicorn_kwargs["ssl_certfile"] = ssl_certfile
server = uvicorn.Server(uvicorn.Config(ds.app(), **uvicorn_kwargs))
await server.serve()
asyncio.run(_serve_async())
@cli.command() @cli.command()

View file

@ -29,7 +29,7 @@ from .utils import (
table_columns, table_columns,
) )
from .utils.sql_analysis import SQLAnalysis, analyze_sql_tables from .utils.sql_analysis import SQLAnalysis, analyze_sql_tables
from .utils.sqlite import sqlite_hidden_table_names from .utils.sqlite import sqlite_derived_table_dependencies, sqlite_hidden_table_names
connections = threading.local() connections = threading.local()
@ -85,6 +85,7 @@ class Database:
self.cached_hash = None self.cached_hash = None
self.cached_size = None self.cached_size = None
self._cached_table_counts = None self._cached_table_counts = None
self._cached_derived_table_dependencies = None
self._write_thread = None self._write_thread = None
self._write_queue = None self._write_queue = None
self._closed = False self._closed = False
@ -246,17 +247,29 @@ class Database:
return_all=False, return_all=False,
returning_limit=EXECUTE_WRITE_RETURNING_LIMIT, returning_limit=EXECUTE_WRITE_RETURNING_LIMIT,
transaction=True, transaction=True,
time_limit_ms=2000,
): ):
self._check_not_closed() self._check_not_closed()
if returning_limit < 0: if returning_limit < 0:
raise ValueError("returning_limit must be >= 0") raise ValueError("returning_limit must be >= 0")
def _inner(conn): def execute_sql(conn):
cursor = conn.execute(sql, params or []) cursor = conn.execute(sql, params or [])
return ExecuteWriteResult.from_cursor( return ExecuteWriteResult.from_cursor(
cursor, return_all=return_all, returning_limit=returning_limit cursor, return_all=return_all, returning_limit=returning_limit
) )
def _inner(conn):
try:
if time_limit_ms is None:
return execute_sql(conn)
with sqlite_timelimit(conn, time_limit_ms):
return execute_sql(conn)
except (sqlite3.OperationalError, sqlite3.DatabaseError) as e:
if e.args == ("interrupted",):
raise QueryInterrupted(e, sql, params)
raise
with trace("sql", database=self.name, sql=sql.strip(), params=params): with trace("sql", database=self.name, sql=sql.strip(), params=params):
results = await self.execute_write_fn( results = await self.execute_write_fn(
_inner, block=block, request=request, transaction=transaction _inner, block=block, request=request, transaction=transaction
@ -354,6 +367,15 @@ class Database:
result = fn(self._write_connection) result = fn(self._write_connection)
else: else:
result = fn(self._write_connection) result = fn(self._write_connection)
if not block:
# There is no write thread here, so the write has already
# finished. Hand back the same (task_id, reply_future) shape
# _send_to_write_thread() returns, with the future already
# resolved, so the block=False path below is identical in
# both modes.
reply_future = asyncio.get_running_loop().create_future()
reply_future.set_result(result)
result = (uuid.uuid4(), reply_future)
else: else:
result = await self._send_to_write_thread( result = await self._send_to_write_thread(
fn, block=block, transaction=transaction fn, block=block, transaction=transaction
@ -425,7 +447,7 @@ class Database:
) )
self._write_thread.name = f"_execute_writes for database {self.name}" self._write_thread.name = f"_execute_writes for database {self.name}"
self._write_thread.start() self._write_thread.start()
task_id = uuid.uuid5(uuid.NAMESPACE_DNS, "datasette.io") task_id = uuid.uuid4()
loop = asyncio.get_running_loop() loop = asyncio.get_running_loop()
reply_future = loop.create_future() reply_future = loop.create_future()
self._write_queue.put( self._write_queue.put(
@ -759,6 +781,17 @@ class Database:
return hidden_tables return hidden_tables
async def derived_table_dependencies(self):
"""Return implementation tables and the tables they derive from."""
schema_version = (await self.execute("PRAGMA schema_version")).first()[0]
if (
self._cached_derived_table_dependencies is None
or self._cached_derived_table_dependencies[0] != schema_version
):
dependencies = await self.execute_fn(sqlite_derived_table_dependencies)
self._cached_derived_table_dependencies = (schema_version, dependencies)
return self._cached_derived_table_dependencies[1]
async def view_names(self): async def view_names(self):
results = await self.execute("select name from sqlite_master where type='view'") results = await self.execute("select name from sqlite_master where type='view'")
return [r[0] for r in results.rows] return [r[0] for r in results.rows]

View file

@ -6,6 +6,17 @@ import markupsafe
from datasette import hookimpl from datasette import hookimpl
from datasette.column_types import ColumnType, SQLiteType from datasette.column_types import ColumnType, SQLiteType
_HTTP_URL_RE = re.compile(r"https?://\S+", re.IGNORECASE)
def _normalize_http_url(value):
if not isinstance(value, str):
return None
normalized = value.strip()
if not _HTTP_URL_RE.fullmatch(normalized):
return None
return normalized
class UrlColumnType(ColumnType): class UrlColumnType(ColumnType):
name = "url" name = "url"
@ -15,7 +26,10 @@ class UrlColumnType(ColumnType):
async def render_cell(self, value, column, table, database, datasette, request): async def render_cell(self, value, column, table, database, datasette, request):
if not value or not isinstance(value, str): if not value or not isinstance(value, str):
return None return None
escaped = markupsafe.escape(value.strip()) normalized = _normalize_http_url(value)
if normalized is None:
return markupsafe.escape(value.strip())
escaped = markupsafe.escape(normalized)
return markupsafe.Markup(f'<a href="{escaped}">{escaped}</a>') return markupsafe.Markup(f'<a href="{escaped}">{escaped}</a>')
async def validate(self, value, datasette): async def validate(self, value, datasette):
@ -23,7 +37,7 @@ class UrlColumnType(ColumnType):
return None return None
if not isinstance(value, str): if not isinstance(value, str):
return "URL must be a string" return "URL must be a string"
if not re.match(r"^https?://\S+$", value.strip()): if _normalize_http_url(value) is None:
return "Invalid URL" return "Invalid URL"
return None return None

View file

@ -92,6 +92,13 @@ class ConfigPermissionProcessor:
# Tables implicitly reference their parent databases # Tables implicitly reference their parent databases
self.restricted_databases.update(db for db, _ in self.restricted_tables) self.restricted_databases.update(db for db, _ in self.restricted_tables)
# Resolve identity keys once per action, rather than scanning the
# restriction allowlist for every configured table's allow block.
self.restricted_table_keys = {
(db, self.action_obj.normalize_child(table) if self.action_obj else table)
for db, table in self.restricted_tables
}
def evaluate_allow_block(self, allow_block: Any) -> bool | None: def evaluate_allow_block(self, allow_block: Any) -> bool | None:
"""Evaluate an allow block against the current actor.""" """Evaluate an allow block against the current actor."""
if allow_block is None: if allow_block is None:
@ -125,8 +132,10 @@ class ConfigPermissionProcessor:
if parent: if parent:
table_restrictions = (self.restrictions.get("r", {}) or {}).get(parent, {}) table_restrictions = (self.restrictions.get("r", {}) or {}).get(parent, {})
if child: if child:
table_actions = table_restrictions.get(child, []) child_key = (
if self.action_checks.intersection(table_actions): self.action_obj.normalize_child(child) if self.action_obj else child
)
if (parent, child_key) in self.restricted_table_keys:
return True return True
else: else:
# Parent query should proceed if any child in this database is allowlisted # Parent query should proceed if any child in this database is allowlisted

View file

@ -185,11 +185,15 @@ def restrictions_allow_action(
# Check table/resource level # Check table/resource level
if resource is not None and not isinstance(resource, str) and len(resource) == 2: if resource is not None and not isinstance(resource, str) and len(resource) == 2:
database, table = resource database, table = resource
table_allowed = restrictions.get("r", {}).get(database, {}).get(table) action_obj = datasette.actions.get(action)
if table_allowed is not None: normalize = action_obj.normalize_child if action_obj else lambda name: name
assert isinstance(table_allowed, list) for table_name, table_allowed in (
if to_check.intersection(table_allowed): restrictions.get("r", {}).get(database, {}).items()
return True ):
if normalize(table_name) == normalize(table):
assert isinstance(table_allowed, list)
if to_check.intersection(table_allowed):
return True
# This action is not explicitly allowed, so reject it # This action is not explicitly allowed, so reject it
return False return False

View file

@ -0,0 +1,25 @@
"""Default table-access policy for SQLite optimizer statistics."""
import json
from datasette import hookimpl
from datasette.permissions import PermissionSQL
@hookimpl
def permission_resources_sql(action):
if action != "view-table":
return None
return PermissionSQL(
sql="""
SELECT database_name AS parent, value AS child, 0 AS allow,
'SQLite statistics tables are denied by default' AS reason
FROM catalog_databases
CROSS JOIN json_each(:sqlite_statistics_names)
""",
params={
"sqlite_statistics_names": json.dumps(
["sqlite_stat1", "sqlite_stat2", "sqlite_stat3", "sqlite_stat4"]
)
},
)

View file

@ -2,7 +2,7 @@ import json
from typing import ClassVar from typing import ClassVar
from datasette import hookimpl from datasette import hookimpl
from datasette.resources import DatabaseResource from datasette.resources import DatabaseResource, TableResource
from datasette.utils.asgi import BadRequest from datasette.utils.asgi import BadRequest
from datasette.views.base import DatasetteError from datasette.views.base import DatasetteError
@ -51,13 +51,20 @@ def search_filters(request, database, table, datasette):
human_descriptions = [] human_descriptions = []
extra_context = {} extra_context = {}
# Figure out which fts_table to use # Figure out which trusted fts_table to use. Query string parameters can
# repeat this mapping (for backwards compatibility), but must not select
# a different table or primary key.
table_metadata = await datasette.table_config(database, table) table_metadata = await datasette.table_config(database, table)
db = datasette.get_database(database) db = datasette.get_database(database)
fts_table = request.args.get("_fts_table") fts_table = table_metadata.get("fts_table")
fts_table = fts_table or table_metadata.get("fts_table")
fts_table = fts_table or await db.fts_table(table) fts_table = fts_table or await db.fts_table(table)
fts_pk = request.args.get("_fts_pk", table_metadata.get("fts_pk", "rowid")) fts_pk = table_metadata.get("fts_pk", "rowid")
requested_fts_table = request.args.get("_fts_table")
requested_fts_pk = request.args.get("_fts_pk")
if (requested_fts_table and requested_fts_table != fts_table) or (
requested_fts_pk and requested_fts_pk != fts_pk
):
raise BadRequest("Invalid _fts_table or _fts_pk")
search_args = { search_args = {
key: request.args[key] key: request.args[key]
for key in request.args for key in request.args
@ -75,6 +82,11 @@ def search_filters(request, database, table, datasette):
extra_context["supports_search"] = bool(fts_table) extra_context["supports_search"] = bool(fts_table)
if fts_table and search_args: if fts_table and search_args:
await datasette.ensure_permission(
action="view-table",
resource=TableResource(database=database, table=fts_table),
actor=request.actor,
)
if "_search" in search_args: if "_search" in search_args:
# Simple ?_search=xxx # Simple ?_search=xxx
search = search_args["_search"] search = search_args["_search"]
@ -135,6 +147,11 @@ def through_filters(request, database, table, datasette):
through_table = through_data["table"] through_table = through_data["table"]
other_column = through_data["column"] other_column = through_data["column"]
value = through_data["value"] value = through_data["value"]
await datasette.ensure_permission(
action="view-table",
resource=TableResource(database=database, table=through_table),
actor=request.actor,
)
db = datasette.get_database(database) db = datasette.get_database(database)
outgoing_foreign_keys = await db.foreign_keys_for_table(through_table) outgoing_foreign_keys = await db.foreign_keys_for_table(through_table)
fk_to_us = next( fk_to_us = next(

View file

@ -9,6 +9,11 @@ def startup(datasette):
"""Fires directly after Datasette first starts running""" """Fires directly after Datasette first starts running"""
@hookspec
def shutdown(datasette):
"""Called once when the Datasette server is shutting down"""
@hookspec @hookspec
def asgi_wrapper(datasette): def asgi_wrapper(datasette):
"""Returns an ASGI middleware callable to wrap our ASGI application with""" """Returns an ASGI middleware callable to wrap our ASGI application with"""

View file

@ -3,6 +3,10 @@ from abc import ABC, abstractmethod
from dataclasses import dataclass from dataclasses import dataclass
from typing import Any, NamedTuple from typing import Any, NamedTuple
_SQLITE_IDENTIFIER_CASE = str.maketrans(
"ABCDEFGHIJKLMNOPQRSTUVWXYZ", "abcdefghijklmnopqrstuvwxyz"
)
# Context variable to track when permission checks should be skipped # Context variable to track when permission checks should be skipped
_skip_permission_checks = contextvars.ContextVar( _skip_permission_checks = contextvars.ContextVar(
"skip_permission_checks", default=False "skip_permission_checks", default=False
@ -49,6 +53,15 @@ class Resource(ABC):
# Class-level metadata (subclasses must define these) # Class-level metadata (subclasses must define these)
name: str = None # e.g., "table", "database", "model" name: str = None # e.g., "table", "database", "model"
parent_class: type["Resource"] | None = None # e.g., DatabaseResource for tables parent_class: type["Resource"] | None = None # e.g., DatabaseResource for tables
case_insensitive_child: bool = False
@classmethod
def normalize_child(cls, child: str | None) -> str | None:
"""Return a comparison key without changing the resource's display name."""
if cls.case_insensitive_child and child is not None:
# Match SQLite NOCASE: fold ASCII only, not Unicode lower/casefold.
return child.translate(_SQLITE_IDENTIFIER_CASE)
return child
# Instance-level optional extra attributes # Instance-level optional extra attributes
reasons: list[str] | None = None reasons: list[str] | None = None
@ -146,6 +159,11 @@ class Action:
resource_class: type[Resource] | None = None resource_class: type[Resource] | None = None
also_requires: str | None = None # Optional action name that must also be allowed also_requires: str | None = None # Optional action name that must also be allowed
def normalize_child(self, child: str | None) -> str | None:
if self.resource_class is None:
return child
return self.resource_class.normalize_child(child)
@property @property
def takes_parent(self) -> bool: def takes_parent(self) -> bool:
""" """

View file

@ -18,6 +18,7 @@ DEFAULT_PLUGINS = (
"datasette.actor_auth_cookie", "datasette.actor_auth_cookie",
"datasette.default_permissions", "datasette.default_permissions",
"datasette.default_permissions.tokens", "datasette.default_permissions.tokens",
"datasette.default_permissions.sqlite_statistics",
"datasette.default_actions", "datasette.default_actions",
"datasette.default_column_types", "datasette.default_column_types",
"datasette.default_magic_parameters", "datasette.default_magic_parameters",

View file

@ -25,6 +25,7 @@ class TableResource(Resource):
name = "table" name = "table"
parent_class = DatabaseResource parent_class = DatabaseResource
case_insensitive_child = True
def __init__(self, database: str, table: str): def __init__(self, database: str, table: str):
super().__init__(parent=database, child=table) super().__init__(parent=database, child=table)

View file

@ -216,6 +216,49 @@ a:active {
text-decoration: underline; text-decoration: underline;
} }
.table-summary .count-all ~ .table-summary-description {
margin-left: 0.5rem;
}
.table-summary .count-error:not(:empty) {
display: block;
margin-top: 0.25rem;
font-size: 0.875rem;
font-weight: 400;
line-height: 1.5;
}
button.count-all {
background: none;
border: none;
padding: 3px 0;
margin-left: 0.25rem;
color: #276890;
font-family: inherit;
font-size: 0.8125rem;
font-weight: 400;
line-height: 1.5;
cursor: pointer;
}
button.count-all:hover,
button.count-all:focus-visible {
text-decoration: underline;
}
button.count-all:disabled {
color: #596478;
cursor: wait;
}
@media (pointer: coarse) {
button.count-all {
min-height: 44px;
padding-left: 7px;
padding-right: 7px;
}
}
button.button-as-link { button.button-as-link {
background: none; background: none;
border: none; border: none;

View file

@ -472,11 +472,13 @@ class ColumnChooser extends HTMLElement {
<span class="drag-item-check"> <span class="drag-item-check">
<input type="checkbox" ${this._checked.has(col) ? "checked" : ""}> <input type="checkbox" ${this._checked.has(col) ? "checked" : ""}>
</span> </span>
<span class="drag-item-label">${col}</span> <span class="drag-item-label"></span>
</label> </label>
<div class="drop-indicator"></div> <div class="drop-indicator"></div>
`; `;
li.querySelector(".drag-item-label").textContent = col;
li.querySelector("input").addEventListener("change", (e) => { li.querySelector("input").addEventListener("change", (e) => {
e.target.checked ? this._checked.add(col) : this._checked.delete(col); e.target.checked ? this._checked.add(col) : this._checked.delete(col);
this._updateCounts(); this._updateCounts();

View file

@ -1,56 +0,0 @@
/*
https://github.com/luyilin/json-format-highlight
From https://unpkg.com/json-format-highlight@1.0.1/dist/json-format-highlight.js
MIT Licensed
*/
(function (global, factory) {
typeof exports === "object" && typeof module !== "undefined"
? (module.exports = factory())
: typeof define === "function" && define.amd
? define(factory)
: (global.jsonFormatHighlight = factory());
})(this, function () {
"use strict";
var defaultColors = {
keyColor: "dimgray",
numberColor: "lightskyblue",
stringColor: "lightcoral",
trueColor: "lightseagreen",
falseColor: "#f66578",
nullColor: "cornflowerblue",
};
function index(json, colorOptions) {
if (colorOptions === void 0) colorOptions = {};
if (!json) {
return;
}
if (typeof json !== "string") {
json = JSON.stringify(json, null, 2);
}
var colors = Object.assign({}, defaultColors, colorOptions);
json = json.replace(/&/g, "&").replace(/</g, "<").replace(/>/g, ">");
return json.replace(
/("(\\u[a-zA-Z0-9]{4}|\\[^u]|[^\\"])*"(\s*:)?|\b(true|false|null)\b|-?\d+(?:\.\d*)?(?:[eE][+]?\d+)?)/g,
function (match) {
var color = colors.numberColor;
if (/^"/.test(match)) {
color = /:$/.test(match) ? colors.keyColor : colors.stringColor;
} else {
color = /true/.test(match)
? colors.trueColor
: /false/.test(match)
? colors.falseColor
: /null/.test(match)
? colors.nullColor
: color;
}
return '<span style="color: ' + color + '">' + match + "</span>";
},
);
}
return index;
});

View file

@ -860,10 +860,45 @@ function openColumnChooser() {
}); });
} }
function initCountAll() {
var button = document.querySelector(".count-all");
if (!button) {
return;
}
button.addEventListener("click", async function () {
var count = document.querySelector(".table-count");
var error = document.querySelector(".count-error");
button.disabled = true;
button.textContent = "Counting…";
error.textContent = "";
try {
var response = await fetch(button.dataset.countUrl + location.search, {
method: "POST",
headers: {
Accept: "application/json",
},
});
var data = await response.json();
if (!response.ok || !data.ok) {
throw new Error((data.errors || ["Count failed"]).join(" "));
}
count.textContent =
data.count.toLocaleString("en-US") +
(data.count === 1 ? " row" : " rows");
button.remove();
} catch (ex) {
error.textContent = ex.message || "Count failed";
button.disabled = false;
button.textContent = "count all";
}
});
}
// Ensures Table UI is initialized only after the Manager is ready. // Ensures Table UI is initialized only after the Manager is ready.
document.addEventListener("datasette_init", function (evt) { document.addEventListener("datasette_init", function (evt) {
const { detail: manager } = evt; const { detail: manager } = evt;
initCountAll();
initializeColumnActions(manager); initializeColumnActions(manager);
// Main table // Main table

View file

@ -3,7 +3,6 @@
{% block title %}API Explorer{% endblock %} {% block title %}API Explorer{% endblock %}
{% block extra_head %} {% block extra_head %}
<script src="{{ static('json-format-highlight-1.0.1.js') }}"></script>
{% endblock %} {% endblock %}
{% block content %} {% block content %}
@ -126,7 +125,7 @@ getForm.addEventListener("submit", (ev) => {
document.getElementById('response-status').textContent = response.status; document.getElementById('response-status').textContent = response.status;
return response.json(); return response.json();
}).then((data) => { }).then((data) => {
output.querySelector('pre').innerHTML = jsonFormatHighlight(data); output.querySelector('pre').textContent = JSON.stringify(data, null, 2);
errorList.style.display = 'none'; errorList.style.display = 'none';
}).catch((error) => { }).catch((error) => {
alert(error); alert(error);
@ -174,7 +173,7 @@ postForm.addEventListener("submit", (ev) => {
} else { } else {
errorList.style.display = 'none'; errorList.style.display = 'none';
} }
output.querySelector('pre').innerHTML = jsonFormatHighlight(data); output.querySelector('pre').textContent = JSON.stringify(data, null, 2);
output.style.display = 'block'; output.style.display = 'block';
}).catch(err => { }).catch(err => {
alert("Error: " + err); alert("Error: " + err);

View file

@ -3,7 +3,6 @@
{% block title %}Allowed Resources{% endblock %} {% block title %}Allowed Resources{% endblock %}
{% block extra_head %} {% block extra_head %}
<script src="{{ static('json-format-highlight-1.0.1.js') }}"></script>
{% include "_permission_ui_styles.html" %} {% include "_permission_ui_styles.html" %}
{% include "_debug_common_functions.html" %} {% include "_debug_common_functions.html" %}
{% endblock %} {% endblock %}
@ -198,7 +197,7 @@ function displayResults(data) {
} }
// Update raw JSON // Update raw JSON
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data); document.getElementById('raw-json').textContent = JSON.stringify(data, null, 2);
} }
function displayError(data) { function displayError(data) {
@ -208,7 +207,7 @@ function displayError(data) {
resultsContent.innerHTML = `<div class="error-message">Error: ${escapeHtml(data.error || 'Unknown error')}</div>`; resultsContent.innerHTML = `<div class="error-message">Error: ${escapeHtml(data.error || 'Unknown error')}</div>`;
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data); document.getElementById('raw-json').textContent = JSON.stringify(data, null, 2);
} }
// Disable child input if parent is empty // Disable child input if parent is empty

View file

@ -3,7 +3,6 @@
{% block title %}Explain a permission decision{% endblock %} {% block title %}Explain a permission decision{% endblock %}
{% block extra_head %} {% block extra_head %}
<script src="{{ static('json-format-highlight-1.0.1.js') }}"></script>
{% include "_permission_ui_styles.html" %} {% include "_permission_ui_styles.html" %}
{% include "_debug_common_functions.html" %} {% include "_debug_common_functions.html" %}
<style> <style>
@ -238,7 +237,7 @@ function displayResult(data) {
displayRules(data.explanation); displayRules(data.explanation);
displayRestrictions(data.explanation.restrictions); displayRestrictions(data.explanation.restrictions);
displayRequirements(data.explanation.required_actions); displayRequirements(data.explanation.required_actions);
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data); document.getElementById('raw-json').textContent = JSON.stringify(data, null, 2);
} }
function displayRules(explanation) { function displayRules(explanation) {
@ -298,7 +297,7 @@ function displayError(data) {
document.getElementById('matching-rules').innerHTML = ''; document.getElementById('matching-rules').innerHTML = '';
document.getElementById('restrictions-section').style.display = 'none'; document.getElementById('restrictions-section').style.display = 'none';
document.getElementById('requirements-section').style.display = 'none'; document.getElementById('requirements-section').style.display = 'none';
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data); document.getElementById('raw-json').textContent = JSON.stringify(data, null, 2);
} }
form.addEventListener('submit', event => { form.addEventListener('submit', event => {

View file

@ -3,7 +3,6 @@
{% block title %}Permission Rules{% endblock %} {% block title %}Permission Rules{% endblock %}
{% block extra_head %} {% block extra_head %}
<script src="{{ static('json-format-highlight-1.0.1.js') }}"></script>
{% include "_permission_ui_styles.html" %} {% include "_permission_ui_styles.html" %}
{% include "_debug_common_functions.html" %} {% include "_debug_common_functions.html" %}
{% endblock %} {% endblock %}
@ -185,7 +184,7 @@ function displayResults(data) {
} }
// Update raw JSON // Update raw JSON
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data); document.getElementById('raw-json').textContent = JSON.stringify(data, null, 2);
} }
function displayError(data) { function displayError(data) {
@ -195,7 +194,7 @@ function displayError(data) {
resultsContent.innerHTML = `<div class="error-message">Error: ${escapeHtml(data.error || 'Unknown error')}</div>`; resultsContent.innerHTML = `<div class="error-message">Error: ${escapeHtml(data.error || 'Unknown error')}</div>`;
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data); document.getElementById('raw-json').textContent = JSON.stringify(data, null, 2);
} }
</script> </script>

View file

@ -1,6 +1,6 @@
{% extends "base.html" %} {% extends "base.html" %}
{% block title %}{{ database }}: {{ table }}: {% if count_truncated %}&gt;{{ "{:,}".format(count - 1) }} rows{% elif count or count == 0 %}{{ "{:,}".format(count) }} row{% if count == 1 %}{% else %}s{% endif %}{% endif %}{% if human_description_en %} {{ human_description_en }}{% endif %}{% endblock %} {% block title %}{{ database }}: {{ table }}: {% if count_truncated %}{{ "{:,}".format(count - 1) }}+ rows{% elif count or count == 0 %}{{ "{:,}".format(count) }} row{% if count == 1 %}{% else %}s{% endif %}{% endif %}{% if human_description_en %} {{ human_description_en }}{% endif %}{% endblock %}
{% block extra_head %} {% block extra_head %}
{{- super() -}} {{- super() -}}
@ -47,11 +47,12 @@
{% endif %} {% endif %}
{% if count or human_description_en %} {% if count or human_description_en %}
<h3> <h3 class="table-summary">
{% if count_truncated %}&gt;{{ "{:,}".format(count - 1) }} rows {% if count_truncated %}<span class="table-count" aria-live="polite">{{ "{:,}".format(count - 1) }}+ rows</span>
{% if allow_execute_sql and query.sql %} <a class="count-sql" style="font-size: 0.8em;" href="{{ urls.database_query(database, count_sql) }}">count all</a>{% endif %} <button type="button" class="count-all" data-count-url="{{ urls.table(database, table) }}/-/count">count all</button>
<span class="count-error" role="alert"></span>
{% elif count or count == 0 %}{{ "{:,}".format(count) }} row{% if count == 1 %}{% else %}s{% endif %}{% endif %} {% elif count or count == 0 %}{{ "{:,}".format(count) }} row{% if count == 1 %}{% else %}s{% endif %}{% endif %}
{% if human_description_en %}{{ human_description_en }}{% endif %} {% if human_description_en %}<span class="table-summary-description">{{ human_description_en }}</span>{% endif %}
</h3> </h3>
{% endif %} {% endif %}

View file

@ -820,7 +820,8 @@ def detect_spatialite(conn):
def detect_fts(conn, table): def detect_fts(conn, table):
"""Detect if table has a corresponding FTS virtual table and return it""" """Detect if table has a corresponding FTS virtual table and return it"""
rows = conn.execute(detect_fts_sql(table)).fetchall() sql, params = detect_fts_sql(table)
rows = conn.execute(sql, params).fetchall()
if len(rows) == 0: if len(rows) == 0:
return None return None
else: else:
@ -828,18 +829,26 @@ def detect_fts(conn, table):
def detect_fts_sql(table): def detect_fts_sql(table):
return r""" escaped_table = table.replace("\\", "\\\\").replace("%", "\\%").replace("_", "\\_")
select name from sqlite_master return (
where rootpage = 0 r"""
and ( select name from sqlite_master
sql like '%VIRTUAL TABLE%USING FTS%content="{table}"%' where rootpage = 0
or sql like '%VIRTUAL TABLE%USING FTS%content=[{table}]%' and (
or ( sql like :fts_double_quoted escape char(92)
tbl_name = "{table}" or sql like :fts_bracket_quoted escape char(92)
and sql like '%VIRTUAL TABLE%USING FTS%' or (
tbl_name = :table
and sql like '%VIRTUAL TABLE%USING FTS%'
)
) )
) """,
""".format(table=table.replace("'", "''")) {
"fts_double_quoted": f'%VIRTUAL TABLE%USING FTS%content="{escaped_table}"%',
"fts_bracket_quoted": f"%VIRTUAL TABLE%USING FTS%content=[{escaped_table}]%",
"table": table,
},
)
def detect_json1(conn=None): def detect_json1(conn=None):
@ -1557,7 +1566,13 @@ async def row_sql_params_pks(db, table, pk_values):
if use_rowid: if use_rowid:
select = "rowid, *" select = "rowid, *"
pks = ["rowid"] pks = ["rowid"]
wheres = [f'"{pk}"=:p{i}' for i, pk in enumerate(pks)] wheres = []
for i, pk in enumerate(pks):
escaped_pk = escape_sqlite(pk)
# Preserve the historic always-quoted SQL exposed by _extra=query
if escaped_pk == pk:
escaped_pk = f'"{pk}"'
wheres.append(f"{escaped_pk}=:p{i}")
sql = f"select {select} from {escape_sqlite(table)} where {' AND '.join(wheres)}" sql = f"select {select} from {escape_sqlite(table)} where {' AND '.join(wheres)}"
params = {} params = {}
for i, pk_value in enumerate(pk_values): for i, pk_value in enumerate(pk_values):
@ -1729,7 +1744,7 @@ def redact_keys(original: dict, key_patterns: Iterable) -> dict:
return { return {
k: ( k: (
redact(v) redact(v)
if not any(pattern in k for pattern in key_patterns) if not any(pattern in k.casefold() for pattern in key_patterns)
else "***" else "***"
) )
for k, v in data.items() for k, v in data.items()

View file

@ -29,6 +29,15 @@ from datasette.utils.permissions import gather_permission_sql_from_hooks
if TYPE_CHECKING: if TYPE_CHECKING:
from datasette.app import Datasette from datasette.app import Datasette
from datasette.permissions import Action
def _child_collation(action: "Action") -> str:
"""Match resource identity without changing the spelling returned by SQL."""
resource_class = action.resource_class
if resource_class is not None and resource_class.case_insensitive_child:
return "NOCASE"
return "BINARY"
async def build_allowed_resources_sql( async def build_allowed_resources_sql(
@ -149,6 +158,7 @@ async def _build_single_action_sql(
raise ValueError(f"Unknown action: {action}") raise ValueError(f"Unknown action: {action}")
# Get base resources SQL from the resource class # Get base resources SQL from the resource class
child_collation = _child_collation(action_obj)
base_resources_sql = await action_obj.resource_class.resources_sql( base_resources_sql = await action_obj.resource_class.resources_sql(
datasette, actor=actor datasette, actor=actor
) )
@ -185,7 +195,7 @@ async def _build_single_action_sql(
if permission_sql.sql is None: if permission_sql.sql is None:
continue continue
rule_sqls.append(f""" rule_sqls.append(f"""
SELECT parent, child, allow, reason, '{permission_sql.source}' AS source_plugin FROM ( SELECT parent, child COLLATE {child_collation} AS child, allow, reason, '{permission_sql.source}' AS source_plugin FROM (
{permission_sql.sql} {permission_sql.sql}
) )
""".strip()) """.strip())
@ -299,9 +309,9 @@ async def _build_single_action_sql(
query_parts.extend( query_parts.extend(
["anon_child_agg AS ("] ["anon_child_agg AS ("]
+ _anon_agg( + _anon_agg(
"parent, child,", f"parent, child COLLATE {child_collation} AS child,",
"parent IS NOT NULL AND child IS NOT NULL", "parent IS NOT NULL AND child IS NOT NULL",
"parent, child", f"parent, child COLLATE {child_collation}",
) )
+ ["),", "anon_parent_agg AS ("] + ["),", "anon_parent_agg AS ("]
+ _anon_agg("parent,", "parent IS NOT NULL AND child IS NULL", "parent") + _anon_agg("parent,", "parent IS NOT NULL AND child IS NULL", "parent")
@ -382,7 +392,8 @@ async def _build_single_action_sql(
# Wrap each restriction_sql in a subquery to avoid operator precedence issues # Wrap each restriction_sql in a subquery to avoid operator precedence issues
# with UNION ALL inside the restriction SQL statements # with UNION ALL inside the restriction SQL statements
restriction_intersect = "\nINTERSECT\n".join( restriction_intersect = "\nINTERSECT\n".join(
f"SELECT * FROM ({sql})" for sql in restriction_sqls f"SELECT parent, child COLLATE {child_collation} AS child FROM ({sql})"
for sql in restriction_sqls
) )
# Decompose by NULL-pattern so the final filter can use pure-equality # Decompose by NULL-pattern so the final filter can use pure-equality
# EXISTS lookups (satisfiable via automatic indexes) instead of a # EXISTS lookups (satisfiable via automatic indexes) instead of a
@ -480,6 +491,7 @@ async def build_permission_rules_sql(
union_parts = [] union_parts = []
all_params = {} all_params = {}
restriction_sqls = [] restriction_sqls = []
child_collation = _child_collation(action_obj)
for permission_sql in permission_sqls: for permission_sql in permission_sqls:
all_params.update(permission_sql.params or {}) all_params.update(permission_sql.params or {})
@ -493,7 +505,7 @@ async def build_permission_rules_sql(
continue continue
union_parts.append(f""" union_parts.append(f"""
SELECT parent, child, allow, reason, '{permission_sql.source}' AS source_plugin FROM ( SELECT parent, child COLLATE {child_collation} AS child, allow, reason, '{permission_sql.source}' AS source_plugin FROM (
{permission_sql.sql} {permission_sql.sql}
) )
""".strip()) """.strip())
@ -564,6 +576,7 @@ async def check_permissions_for_actions(
verdicts = {} verdicts = {}
for i, (action, permission_sqls) in enumerate(zip(unique_actions, gathered)): for i, (action, permission_sqls) in enumerate(zip(unique_actions, gathered)):
child_collation = _child_collation(datasette.actions[action])
prefix = f"a{i}_" prefix = f"a{i}_"
rule_parts = [] rule_parts = []
restriction_parts = [] restriction_parts = []
@ -589,7 +602,7 @@ async def check_permissions_for_actions(
if sql is None: if sql is None:
continue continue
rule_parts.append( rule_parts.append(
f"SELECT parent, child, allow, reason, '{permission_sql.source}' AS source_plugin FROM (\n{sql}\n)" f"SELECT parent, child COLLATE {child_collation} AS child, allow, reason, '{permission_sql.source}' AS source_plugin FROM (\n{sql}\n)"
) )
if not rule_parts: if not rule_parts:
@ -623,7 +636,8 @@ async def check_permissions_for_actions(
if restriction_parts: if restriction_parts:
# Database-level restrictions (parent, NULL) match all children # Database-level restrictions (parent, NULL) match all children
restriction_intersect = "\nINTERSECT\n".join( restriction_intersect = "\nINTERSECT\n".join(
f"SELECT * FROM ({sql})" for sql in restriction_parts f"SELECT parent, child COLLATE {child_collation} AS child FROM ({sql})"
for sql in restriction_parts
) )
ctes.append(f"a{i}_restriction AS (\n{restriction_intersect}\n)") ctes.append(f"a{i}_restriction AS (\n{restriction_intersect}\n)")
verdict_sql = f"""({verdict_sql}) AND EXISTS ( verdict_sql = f"""({verdict_sql}) AND EXISTS (
@ -770,6 +784,7 @@ async def _explain_single_action(
db = datasette.get_internal_database() db = datasette.get_internal_database()
matched_rules = [] matched_rules = []
restrictions = [] restrictions = []
child_collation = _child_collation(datasette.actions[action])
for permission_sql in permission_sqls: for permission_sql in permission_sqls:
params = dict(permission_sql.params or {}) params = dict(permission_sql.params or {})
@ -784,7 +799,7 @@ async def _explain_single_action(
SELECT parent, child, allow, reason SELECT parent, child, allow, reason
FROM ({permission_sql.sql}) AS permission_rules FROM ({permission_sql.sql}) AS permission_rules
WHERE (parent IS NULL OR parent = :{parent_param}) WHERE (parent IS NULL OR parent = :{parent_param})
AND (child IS NULL OR child = :{child_param}) AND (child IS NULL OR child COLLATE {child_collation} = :{child_param})
""", """,
params, params,
) )
@ -811,7 +826,7 @@ async def _explain_single_action(
SELECT EXISTS( SELECT EXISTS(
SELECT 1 FROM ({permission_sql.restriction_sql}) AS restriction_rules SELECT 1 FROM ({permission_sql.restriction_sql}) AS restriction_rules
WHERE (parent IS NULL OR parent = :{parent_param}) WHERE (parent IS NULL OR parent = :{parent_param})
AND (child IS NULL OR child = :{child_param}) AND (child IS NULL OR child COLLATE {child_collation} = :{child_param})
) AS resource_is_in_allowlist ) AS resource_is_in_allowlist
""", """,
params, params,

View file

@ -1,3 +1,4 @@
import asyncio
import json import json
import re import re
from http.cookies import Morsel, SimpleCookie from http.cookies import Morsel, SimpleCookie
@ -300,12 +301,24 @@ class AsgiLifespan:
while True: while True:
message = await receive() message = await receive()
if message["type"] == "lifespan.startup": if message["type"] == "lifespan.startup":
for fn in self.on_startup: try:
await fn() for fn in self.on_startup:
await fn()
except Exception as e: # noqa: BLE001
await send(
{"type": "lifespan.startup.failed", "message": str(e)}
)
return
await send({"type": "lifespan.startup.complete"}) await send({"type": "lifespan.startup.complete"})
elif message["type"] == "lifespan.shutdown": elif message["type"] == "lifespan.shutdown":
for fn in self.on_shutdown: try:
await fn() for fn in self.on_shutdown:
await fn()
except Exception as e: # noqa: BLE001
await send(
{"type": "lifespan.shutdown.failed", "message": str(e)}
)
return
await send({"type": "lifespan.shutdown.complete"}) await send({"type": "lifespan.shutdown.complete"})
return return
else: else:
@ -485,6 +498,8 @@ def asgi_static(root_path, chunk_size=4096, headers=None, content_type=None):
await asgi_send_html(send, "404: File not found", 404) await asgi_send_html(send, "404: File not found", 404)
return return
# Only the actual static-file handler can bypass dynamic response privacy.
inner_static._datasette_static = True
return inner_static return inner_static
@ -624,10 +639,23 @@ class AsgiRunOnFirstRequest:
self.asgi = asgi self.asgi = asgi
self.on_startup = on_startup self.on_startup = on_startup
self._started = False self._started = False
# Guards against concurrent early requests interleaving with startup:
# without this, several requests could all observe `_started is
# False` and proceed before any of them finish running the hooks.
self._lock = asyncio.Lock()
async def __call__(self, scope, receive, send): async def __call__(self, scope, receive, send):
if not self._started: # Leave "lifespan" scope events alone - this shim only exists as a
self._started = True # fallback for hosts that never send them. It wraps AsgiLifespan, so
for hook in self.on_startup: # if it ran on_startup here too, a startup exception would escape
await hook() # before AsgiLifespan's own try/except got a chance to turn it into
# a lifespan.startup.failed message.
if scope["type"] != "lifespan" and not self._started:
async with self._lock:
# Re-check: another request may have finished startup while
# we were waiting for the lock.
if not self._started:
for hook in self.on_startup:
await hook()
self._started = True
return await self.asgi(scope, receive, send) return await self.asgi(scope, receive, send)

View file

@ -1,6 +1,8 @@
import sys
from dataclasses import dataclass from dataclasses import dataclass
from typing import Literal from typing import Literal
from datasette.utils import escape_sqlite
from datasette.utils.sqlite import SQLiteTableType, sqlite3, sqlite_table_type from datasette.utils.sqlite import SQLiteTableType, sqlite3, sqlite_table_type
SQLOperation = Literal[ SQLOperation = Literal[
@ -195,6 +197,16 @@ def _allow_authorizer_action(*args):
return sqlite3.SQLITE_OK return sqlite3.SQLITE_OK
def _disable_authorizer(conn):
# Python 3.11 added support for unregistering an authorizer using None.
# On Python 3.10, None is installed as the callback instead, and the next
# statement fails with "not authorized" when sqlite3 tries to call it.
if sys.version_info >= (3, 11):
conn.set_authorizer(None)
else:
conn.set_authorizer(_allow_authorizer_action)
def analyze_sql_tables( def analyze_sql_tables(
conn, conn,
sql: str, sql: str,
@ -208,7 +220,9 @@ def analyze_sql_tables(
This function is synchronous and connection-based. It temporarily installs a This function is synchronous and connection-based. It temporarily installs a
SQLite authorizer, prepares ``EXPLAIN <sql>``, and returns the operation SQLite authorizer, prepares ``EXPLAIN <sql>``, and returns the operation
callbacks observed while SQLite compiles the statement. callbacks observed while SQLite compiles the statement. ``CREATE VIEW`` is
additionally executed inside a rolled-back savepoint so its source-table reads
can be discovered by analyzing a query against the temporary view.
""" """
operations: dict[OperationKey, set[str]] = {} operations: dict[OperationKey, set[str]] = {}
@ -481,7 +495,7 @@ def analyze_sql_tables(
conn, key.table, schema=key.sqlite_schema conn, key.table, schema=key.sqlite_schema
) )
finally: finally:
conn.set_authorizer(None) _disable_authorizer(conn)
has_schema_operation = any( has_schema_operation = any(
key.target_type in {"table", "index", "view", "trigger", "virtual-table"} key.target_type in {"table", "index", "view", "trigger", "virtual-table"}
@ -532,7 +546,7 @@ def analyze_sql_tables(
return None return None
return table_kind_cache[(key.sqlite_schema, key.table)] return table_kind_cache[(key.sqlite_schema, key.table)]
return SQLAnalysis( analysis = SQLAnalysis(
operations=tuple( operations=tuple(
Operation( Operation(
operation=key.operation, operation=key.operation,
@ -549,3 +563,58 @@ def analyze_sql_tables(
for key, columns in operations.items() for key, columns in operations.items()
) )
) )
# SQLite does not resolve the SELECT body of a view when preparing CREATE
# VIEW, so its authorizer does not report reads from the view's source
# tables. Temporarily create the view, analyze a query against it (which
# does resolve the body), then roll the schema change back. Database-level
# callers use an isolated writable connection for this analysis.
create_view_operations = tuple(
operation
for operation in analysis.operations
if operation.operation == "create" and operation.target_type == "view"
)
if not create_view_operations:
return analysis
savepoint = "datasette_analyze_create_view"
conn.execute(f"SAVEPOINT {savepoint}")
try:
conn.execute(sql, params if params is not None else {})
dependency_reads = []
for view_operation in create_view_operations:
if view_operation.sqlite_schema is None or view_operation.table is None:
raise sqlite3.OperationalError(
"Could not determine the created view name"
)
quoted_schema = escape_sqlite(view_operation.sqlite_schema)
quoted_view = escape_sqlite(view_operation.table)
qualified_view = f"{quoted_schema}.{quoted_view}"
view_analysis = analyze_sql_tables(
conn,
f"SELECT * FROM {qualified_view}",
database_name=database_name,
schema_to_database=schema_to_database,
)
dependency_reads.extend(
operation
for operation in view_analysis.operations
if operation.operation == "read"
and not (
operation.sqlite_schema == view_operation.sqlite_schema
and operation.table == view_operation.table
)
)
finally:
conn.execute(f"ROLLBACK TO {savepoint}")
conn.execute(f"RELEASE {savepoint}")
existing_operations = set(analysis.operations)
return SQLAnalysis(
operations=analysis.operations
+ tuple(
operation
for operation in dependency_reads
if operation not in existing_operations
)
)

View file

@ -15,8 +15,17 @@ if hasattr(sqlite3, "enable_callback_tracebacks"):
_cached_sqlite_version = None _cached_sqlite_version = None
_cached_supports_returning = None _cached_supports_returning = None
SQLiteTableType = Literal["table", "view", "virtual", "shadow"] SQLiteTableType = Literal["table", "view", "virtual", "shadow"]
_SQLITE_IDENTIFIER_RE = (
r"""(?:"(?:[^"]|"")*"|'(?:[^']|'')*'|`(?:[^`]|``)*`|\[[^\]]*\]|[^\s.()'"`\[\]]+)"""
)
_VIRTUAL_TABLE_MODULE_RE = re.compile( _VIRTUAL_TABLE_MODULE_RE = re.compile(
r"\bCREATE\s+VIRTUAL\s+TABLE\b.*?\bUSING\s+([^\s(]+)", r"^\s*CREATE\s+VIRTUAL\s+TABLE\b\s*(?:IF\s+NOT\s+EXISTS\s+)?"
+ _SQLITE_IDENTIFIER_RE
+ r"(?:\s*\.\s*"
+ _SQLITE_IDENTIFIER_RE
+ r")?\s*\bUSING\b\s*("
+ _SQLITE_IDENTIFIER_RE
+ r")",
re.IGNORECASE | re.DOTALL, re.IGNORECASE | re.DOTALL,
) )
_VIRTUAL_TABLE_SHADOW_SUFFIXES = { _VIRTUAL_TABLE_SHADOW_SUFFIXES = {
@ -83,19 +92,53 @@ def sqlite_table_type(
) -> SQLiteTableType | None: ) -> SQLiteTableType | None:
if supports_table_list(): if supports_table_list():
try: try:
query = "select type from pragma_table_list where name = ?" # Use the "PRAGMA table_list" statement form rather than the
params: tuple[str, ...] = (table,) # pragma_table_list(...) table-valued function. The
# table-valued function is resolved like an ordinary relation
# name, so an attacker-created table or view literally named
# "pragma_table_list" can shadow it and spoof the reported
# type (e.g. claiming a virtual table is an ordinary table).
# The PRAGMA statement form is a distinct piece of SQL syntax
# that always invokes SQLite's built-in pragma, so it cannot
# be shadowed by a user-created relation.
if schema is not None: if schema is not None:
query += " and schema = ?" query = f"PRAGMA {_quote_identifier(schema)}.table_list"
params = (table, schema) else:
row = conn.execute(query, params).fetchone() query = "PRAGMA table_list"
if row is not None and row[0] in {"table", "view", "virtual", "shadow"}: cursor = conn.execute(query)
return row[0] columns = [description[0] for description in cursor.description]
for row in cursor.fetchall():
record = dict(zip(columns, row))
if record.get("name") != table:
continue
if schema is not None and record.get("schema") != schema:
continue
row_type = record.get("type")
if row_type in {"table", "view", "virtual", "shadow"}:
return row_type
except sqlite3.DatabaseError: except sqlite3.DatabaseError:
pass pass
return _sqlite_table_type_from_schema(conn, table, schema=schema) return _sqlite_table_type_from_schema(conn, table, schema=schema)
def check_structured_write_table(conn, table: str, *, allow_missing=False):
"""Validate a row-write target on the connection that will perform the write."""
# SQLite resolves identifiers case-insensitively. The create API must not
# treat a differently cased existing name as a missing table.
row = conn.execute(
"select name from main.sqlite_master where name = ? collate nocase "
"and type in ('table', 'view')",
(table,),
).fetchone()
if row is None and allow_missing:
return
if row is not None and sqlite_table_type(conn, row[0]) == "table":
return
# Virtual table modules can interpret row writes as administrative operations.
# Their shadow tables are internal storage, not independently writable data.
raise ValueError("Structured writes require an ordinary table")
def sqlite_hidden_table_names(conn, *, schema: str | None = "main") -> list[str]: def sqlite_hidden_table_names(conn, *, schema: str | None = "main") -> list[str]:
schema_table = _sqlite_schema_table(schema) schema_table = _sqlite_schema_table(schema)
try: try:
@ -118,6 +161,63 @@ def sqlite_hidden_table_names(conn, *, schema: str | None = "main") -> list[str]
return sorted(hidden_tables) + content_fts_tables return sorted(hidden_tables) + content_fts_tables
def sqlite_derived_table_dependencies(
conn, *, schema: str | None = "main"
) -> dict[str, str]:
"""Return implementation table -> logical/content table dependencies.
``PRAGMA table_list`` safely identifies virtual and shadow tables, but
does not report which virtual table owns a shadow table or which table is
named by an FTS ``content=`` option. Derive those relationships from
``sqlite_master`` DDL and the documented shadow-table suffixes.
Database errors propagate: failed discovery must not be mistaken for an
empty dependency map and cached as permission to skip inheritance.
"""
schema_table = _sqlite_schema_table(schema)
rows = conn.execute(
f"select name, sql from {schema_table} where type = 'table'"
).fetchall()
table_names = {row[0] for row in rows}
# SQLite identifiers fold ASCII letters only.
identifier_case = str.maketrans(
"ABCDEFGHIJKLMNOPQRSTUVWXYZ", "abcdefghijklmnopqrstuvwxyz"
)
canonical_names = {name.translate(identifier_case): name for name in table_names}
dependencies = {}
for virtual_table, sql in rows:
module = _virtual_table_module(sql)
if module is None:
continue
# SQLite's documented shadow tables are implementation details of
# their logical virtual table.
for suffix in _VIRTUAL_TABLE_SHADOW_SUFFIXES.get(module, ()):
shadow_table = virtual_table + suffix
if shadow_table in table_names:
dependencies[shadow_table] = virtual_table
# An external-content FTS table can expose values fetched from its
# content table, so it must also depend on that table's permission.
if module in {"fts3", "fts4", "fts5"}:
content_table = _fts_external_content_table(sql)
if content_table:
dependencies[virtual_table] = content_table
if module in {"fts5vocab", "fts4aux"}:
source = _fts_vocabulary_source(sql, module, schema or "main")
source = (
canonical_names.get(source.translate(identifier_case))
if source
else None
)
# An unresolved source is itself derived, so the one-hop policy denies it.
dependencies[virtual_table] = source or virtual_table
return dependencies
def _sqlite_table_type_from_schema( def _sqlite_table_type_from_schema(
conn, conn,
table: str, table: str,
@ -184,10 +284,151 @@ def _quote_identifier(value: str) -> str:
def _virtual_table_module(sql: str | None) -> str | None: def _virtual_table_module(sql: str | None) -> str | None:
if not sql: if not sql:
return None return None
match = _VIRTUAL_TABLE_MODULE_RE.search(_strip_sql_comments(sql))
if match is None:
return None
return _unquote_sql_value(match.group(1)).lower()
def _fts_external_content_table(sql: str | None) -> str | None:
"""Extract the external ``content=`` table from an FTS declaration."""
if not sql:
return None
sql = _strip_sql_comments(sql)
match = _VIRTUAL_TABLE_MODULE_RE.search(sql) match = _VIRTUAL_TABLE_MODULE_RE.search(sql)
if match is None: if match is None:
return None return None
return match.group(1).strip("\"'[]`").lower() open_paren = sql.find("(", match.end())
if open_paren == -1:
return None
close_paren = sql.rfind(")")
if close_paren <= open_paren:
return None
for argument in _split_sql_arguments(sql[open_paren + 1 : close_paren]):
key, separator, value = argument.partition("=")
if not separator or key.strip().lower() != "content":
continue
return _unquote_sql_value(value.strip())
return None
def _fts_vocabulary_source(sql: str, module: str, schema: str) -> str | None:
"""Resolve a vocabulary source within the current SQLite schema.
Cross-schema sources cannot be represented by the dependency map and
are conservatively left unresolved.
"""
sql = _strip_sql_comments(sql)
match = _VIRTUAL_TABLE_MODULE_RE.search(sql)
if match is None:
return None
start = sql.find("(", match.end())
end = sql.rfind(")")
if start < 0 or end <= start:
return None
arguments = [
_unquote_sql_value(arg.strip())
for arg in _split_sql_arguments(sql[start + 1 : end])
]
expected = 2 if module == "fts5vocab" else 1
if len(arguments) == expected:
return arguments[0]
if len(arguments) == expected + 1 and arguments[0].lower() == schema.lower():
return arguments[1]
return None
def _split_sql_arguments(arguments: str) -> list[str]:
"""Split comma-separated SQLite arguments without splitting quoted text."""
parts = []
start = 0
quote = None
closing_quote = None
index = 0
while index < len(arguments):
char = arguments[index]
if quote is None:
if char in {"'", '"', "`", "["}:
quote = char
closing_quote = "]" if char == "[" else char
elif char == ",":
parts.append(arguments[start:index])
start = index + 1
elif char == closing_quote:
# Single/double/backtick quoting escapes the delimiter by
# doubling it. Square-bracket identifiers do not.
if (
quote != "["
and index + 1 < len(arguments)
and arguments[index + 1] == closing_quote
):
index += 1
else:
quote = None
closing_quote = None
index += 1
parts.append(arguments[start:])
return parts
def _strip_sql_comments(sql: str) -> str:
"""Remove SQLite comments while preserving quoted strings/identifiers."""
output = []
quote = None
closing_quote = None
index = 0
while index < len(sql):
char = sql[index]
next_char = sql[index + 1] if index + 1 < len(sql) else ""
if quote is None:
if char in {"'", '"', "`", "["}:
quote = char
closing_quote = "]" if char == "[" else char
output.append(char)
elif char == "-" and next_char == "-":
index += 2
while index < len(sql) and sql[index] not in "\r\n":
index += 1
output.append(" ")
continue
elif char == "/" and next_char == "*":
index += 2
while index + 1 < len(sql) and sql[index : index + 2] != "*/":
index += 1
index = min(index + 2, len(sql))
output.append(" ")
continue
else:
output.append(char)
else:
output.append(char)
if char == closing_quote:
if (
quote != "["
and index + 1 < len(sql)
and sql[index + 1] == closing_quote
):
output.append(sql[index + 1])
index += 1
else:
quote = None
closing_quote = None
index += 1
return "".join(output)
def _unquote_sql_value(value: str) -> str:
if len(value) < 2:
return value
pairs = {"'": "'", '"': '"', "`": "`", "[": "]"}
closing = pairs.get(value[0])
if closing is None or value[-1] != closing:
return value
unquoted = value[1:-1]
if value[0] != "[":
unquoted = unquoted.replace(closing * 2, closing)
return unquoted
def _is_fts_content_virtual_table(sql: str | None) -> bool: def _is_fts_content_virtual_table(sql: str | None) -> bool:

View file

@ -4,7 +4,7 @@ from urllib.parse import urlencode
from asgiref.sync import async_to_sync from asgiref.sync import async_to_sync
# These wrapper classes pre-date the introduction of # These wrapper classes pre-date the introduction of
# datasette.client and httpx to Datasette. They could # datasette.client and httpx2 to Datasette. They could
# be removed if the Datasette tests are modified to # be removed if the Datasette tests are modified to
# call datasette.client directly. # call datasette.client directly.

View file

@ -1,2 +1,2 @@
__version__ = "1.0a38" __version__ = "1.0a39"
__version_info__ = tuple(__version__.split(".")) __version_info__ = tuple(__version__.split("."))

View file

@ -40,7 +40,11 @@ from datasette.write_sql import QueryWriteRejected
from . import Context from . import Context
from .base import DatasetteError, View, stream_csv from .base import DatasetteError, View, stream_csv
from .query_helpers import _ensure_stored_query_execution_permissions, _table_columns from .query_helpers import (
_block_framing,
_ensure_stored_query_execution_permissions,
_table_columns,
)
from .table_create_alter import _create_table_ui_context from .table_create_alter import _create_table_ui_context
from .table_extras import ( from .table_extras import (
QueryExtraContext, QueryExtraContext,
@ -857,7 +861,8 @@ class QueryView(View):
raise DatasetteError("?sql= is required", status=400) raise DatasetteError("?sql= is required", status=400)
async def fetch_data_for_csv(request, _next=None): async def fetch_data_for_csv(request, _next=None):
results = await db.execute(sql, params, truncate=True) # Reuse the trusted magic parameter values prepared above.
results = await db.execute(sql, params_for_query, truncate=True)
data = {"rows": results.rows, "columns": results.columns} data = {"rows": results.rows, "columns": results.columns}
return data, None, None return data, None, None
@ -1140,6 +1145,8 @@ class QueryView(View):
assert False, f"Invalid format: {format_}" assert False, f"Invalid format: {format_}"
if datasette.cors: if datasette.cors:
add_cors_headers(r.headers) add_cors_headers(r.headers)
if stored_query_write and format_ == "html":
_block_framing(r)
return r return r

View file

@ -1,6 +1,7 @@
import re import re
from urllib.parse import urlencode from urllib.parse import urlencode
from datasette.database import QueryInterrupted
from datasette.resources import DatabaseResource from datasette.resources import DatabaseResource
from datasette.utils import UNSTABLE_API_MESSAGE, sqlite3 from datasette.utils import UNSTABLE_API_MESSAGE, sqlite3
from datasette.utils.asgi import Response from datasette.utils.asgi import Response
@ -384,7 +385,7 @@ class ExecuteWriteView(BaseView):
try: try:
execute_write_kwargs = {"request": request} execute_write_kwargs = {"request": request}
cursor = await db.execute_write(sql, params, **execute_write_kwargs) cursor = await db.execute_write(sql, params, **execute_write_kwargs)
except sqlite3.DatabaseError as ex: except (QueryInterrupted, sqlite3.DatabaseError) as ex:
message = str(ex) message = str(ex)
if wants_json: if wants_json:
return _block_framing(Response.error([message], 400)) return _block_framing(Response.error([message], 400))

View file

@ -28,9 +28,11 @@ from datasette.utils import (
path_with_format, path_with_format,
path_with_removed_args, path_with_removed_args,
sqlite3, sqlite3,
tilde_decode,
to_css_class, to_css_class,
) )
from datasette.utils.asgi import Forbidden, NotFound, PayloadTooLarge, Response from datasette.utils.asgi import Forbidden, NotFound, PayloadTooLarge, Response
from datasette.utils.sqlite import check_structured_write_table
from . import Context, from_extra from . import Context, from_extra
from .base import BaseView, DatasetteError, stream_csv from .base import BaseView, DatasetteError, stream_csv
@ -137,6 +139,12 @@ class RowContext(Context):
) )
async def _database_and_table_resource_from_request(datasette, request):
db = await datasette.resolve_database(request)
table = tilde_decode(request.url_vars["table"])
return db, table, TableResource(database=db.name, table=table)
class RowView(BaseView): class RowView(BaseView):
name = "row" name = "row"
@ -263,7 +271,7 @@ class RowView(BaseView):
if ttl is None or not ttl.isdigit(): if ttl is None or not ttl.isdigit():
ttl = self.ds.setting("default_cache_ttl") ttl = self.ds.setting("default_cache_ttl")
return self.set_response_headers(response, ttl) return self.set_response_headers(response, ttl, request)
async def html(self, request, data, extra_template_data, templates): async def html(self, request, data, extra_template_data, templates):
extras = {} extras = {}
@ -376,36 +384,50 @@ class RowView(BaseView):
}, },
) )
def set_response_headers(self, response, ttl): def set_response_headers(self, response, ttl, request=None):
private = getattr(request, "_datasette_private_response", False)
# Set far-future cache expiry # Set far-future cache expiry
if self.ds.cache_headers and response.status == 200: if self.ds.cache_headers and response.status == 200:
ttl = int(ttl) if private:
if ttl == 0: # This response is only visible to the current actor (denied
ttl_header = "no-cache" # to anonymous requests), so it must never be stored by a
# shared cache/CDN - and ?_ttl= must not override that.
response.headers["Cache-Control"] = "private, no-store"
response.headers["Vary"] = "Cookie"
else: else:
ttl_header = f"max-age={ttl}" ttl = int(ttl)
response.headers["Cache-Control"] = ttl_header if ttl == 0:
ttl_header = "no-cache"
else:
ttl_header = f"max-age={ttl}"
response.headers["Cache-Control"] = ttl_header
response.headers["Referrer-Policy"] = "no-referrer" response.headers["Referrer-Policy"] = "no-referrer"
if self.ds.cors: if self.ds.cors:
add_cors_headers(response.headers) add_cors_headers(response.headers)
return response return response
async def data(self, request, default_labels=False): async def data(self, request, default_labels=False):
resolved = await self.ds.resolve_row(request) db, table, resource = await _database_and_table_resource_from_request(
db = resolved.db self.ds, request
)
database = db.name database = db.name
table = resolved.table
pk_values = resolved.pk_values
# Ensure user has permission to view this row # Check the URL resource before resolving the row, so a denied request
# cannot distinguish an existing primary key from a missing one.
visible, private = await self.ds.check_visibility( visible, private = await self.ds.check_visibility(
request.actor, request.actor,
action="view-table", action="view-table",
resource=TableResource(database=database, table=table), resource=resource,
) )
if not visible: if not visible:
raise Forbidden("You do not have permission to view this table") raise Forbidden("You do not have permission to view this table")
# Record whether this response is private (visible to this actor
# only) so set_response_headers() can set appropriate Cache-Control
# headers, regardless of which output format ends up being rendered.
request._datasette_private_response = private
resolved = await self.ds.resolve_row(request)
pk_values = resolved.pk_values
results = await resolved.db.execute( results = await resolved.db.execute(
resolved.sql, resolved.params, truncate=True resolved.sql, resolved.params, truncate=True
) )
@ -482,8 +504,8 @@ class RowView(BaseView):
for row in display_rows: for row in display_rows:
for cell in row: for cell in row:
if cell["column"] in pk_set: if cell["column"] in pk_set:
cell["value"] = markupsafe.Markup( cell["value"] = markupsafe.Markup("<strong>{}</strong>").format(
"<strong>{}</strong>".format(cell["value"]) cell["value"]
) )
label_column = await db.label_column_for_table(table) if is_table else None label_column = await db.label_column_for_table(table) if is_table else None
@ -556,7 +578,7 @@ class RowView(BaseView):
"private": private, "private": private,
"columns": reordered_columns, "columns": reordered_columns,
"foreign_key_tables": await self.foreign_key_tables( "foreign_key_tables": await self.foreign_key_tables(
database, table, pk_values database, table, pk_values, actor=request.actor
), ),
"database_color": db.color, "database_color": db.color,
"display_columns": display_columns, "display_columns": display_columns,
@ -633,12 +655,23 @@ class RowView(BaseView):
), ),
) )
async def foreign_key_tables(self, database, table, pk_values): async def foreign_key_tables(self, database, table, pk_values, *, actor):
if len(pk_values) != 1: if len(pk_values) != 1:
return [] return []
db = self.ds.databases[database] db = self.ds.databases[database]
all_foreign_keys = await db.get_all_foreign_keys() all_foreign_keys = await db.get_all_foreign_keys()
foreign_keys = all_foreign_keys[table]["incoming"] foreign_keys = []
table_permissions = {}
for fk in all_foreign_keys[table]["incoming"]:
other_table = fk["other_table"]
if other_table not in table_permissions:
table_permissions[other_table] = await self.ds.allowed(
action="view-table",
resource=TableResource(database=database, table=other_table),
actor=actor,
)
if table_permissions[other_table]:
foreign_keys.append(fk)
if len(foreign_keys) == 0: if len(foreign_keys) == 0:
return [] return []
@ -695,9 +728,24 @@ def _truncated_row_flash_label(label):
return label[: ROW_FLASH_LABEL_MAX_LENGTH - 1] + "\u2026" return label[: ROW_FLASH_LABEL_MAX_LENGTH - 1] + "\u2026"
async def _row_flash_message(db, action, resolved, row=None): async def _row_flash_message(
datasette, request, action, resolved, row=None, *, refresh_row=False
):
pk_label = ", ".join(resolved.pk_values) pk_label = ", ".join(resolved.pk_values)
label_column = await db.label_column_for_table(resolved.table) # Mutation permission does not grant access to stored row labels.
if not await datasette.allowed(
action="view-table",
resource=TableResource(database=resolved.db.name, table=resolved.table),
actor=request.actor,
):
return f"{action} row {pk_label}"
if refresh_row and row is None:
results = await resolved.db.execute(
resolved.sql, resolved.params, truncate=True
)
row = results.first()
label_column = await resolved.db.label_column_for_table(resolved.table)
label = row_label_from_label_column(row or resolved.row, label_column) label = row_label_from_label_column(row or resolved.row, label_column)
if label: if label:
label = _truncated_row_flash_label(label) label = _truncated_row_flash_label(label)
@ -710,22 +758,28 @@ async def _resolve_row_and_check_permission(datasette, request, permission):
from datasette.app import DatabaseNotFound, RowNotFound, TableNotFound from datasette.app import DatabaseNotFound, RowNotFound, TableNotFound
try: try:
resolved = await datasette.resolve_row(request) _, _, resource = await _database_and_table_resource_from_request(
datasette, request
)
except DatabaseNotFound as e: except DatabaseNotFound as e:
return False, Response.error([f"Database not found: {e.database_name}"], 404) return False, Response.error([f"Database not found: {e.database_name}"], 404)
# Check the URL resource before resolving the row, so a denied request
# cannot distinguish an existing primary key from a missing one.
if not await datasette.allowed(
action=permission,
resource=resource,
actor=request.actor,
):
return False, Response.error(["Permission denied"], 403)
try:
resolved = await datasette.resolve_row(request)
except TableNotFound as e: except TableNotFound as e:
return False, Response.error([f"Table not found: {e.table}"], 404) return False, Response.error([f"Table not found: {e.table}"], 404)
except RowNotFound as e: except RowNotFound as e:
return False, Response.error([f"Record not found: {e.pk_values}"], 404) return False, Response.error([f"Record not found: {e.pk_values}"], 404)
# Ensure user has permission to delete this row
if not await datasette.allowed(
action=permission,
resource=TableResource(database=resolved.db.name, table=resolved.table),
actor=request.actor,
):
return False, Response.error(["Permission denied"], 403)
return True, resolved return True, resolved
@ -744,6 +798,7 @@ class RowDeleteView(BaseView):
# Delete table # Delete table
def delete_row(conn): def delete_row(conn):
check_structured_write_table(conn, resolved.table)
sqlite_utils.Database(conn)[resolved.table].delete(resolved.pk_values) sqlite_utils.Database(conn)[resolved.table].delete(resolved.pk_values)
try: try:
@ -765,7 +820,7 @@ class RowDeleteView(BaseView):
table_url = self.ds.urls.table(resolved.db.name, resolved.table) table_url = self.ds.urls.table(resolved.db.name, resolved.table)
self.ds.add_message( self.ds.add_message(
request, request,
await _row_flash_message(resolved.db, "Deleted", resolved), await _row_flash_message(self.ds, request, "Deleted", resolved),
self.ds.INFO, self.ds.INFO,
) )
return Response.json({"ok": True, "redirect": str(table_url)}, status=200) return Response.json({"ok": True, "redirect": str(table_url)}, status=200)
@ -826,6 +881,7 @@ class RowUpdateView(BaseView):
return Response.error(["Permission denied for alter-table"], 403) return Response.error(["Permission denied for alter-table"], 403)
def update_row(conn): def update_row(conn):
check_structured_write_table(conn, resolved.table)
sqlite_utils.Database(conn)[resolved.table].update( sqlite_utils.Database(conn)[resolved.table].update(
resolved.pk_values, update, alter=alter resolved.pk_values, update, alter=alter
) )
@ -838,7 +894,14 @@ class RowUpdateView(BaseView):
result = {"ok": True} result = {"ok": True}
returned_row = None returned_row = None
if data.get("return"): # Only read back and disclose the stored row if the actor is also
# allowed to view this table - update-row alone must not be usable
# to read data the actor cannot otherwise see.
if data.get("return") and await self.ds.allowed(
action="view-table",
resource=TableResource(database=resolved.db.name, table=resolved.table),
actor=request.actor,
):
results = await resolved.db.execute( results = await resolved.db.execute(
resolved.sql, resolved.params, truncate=True resolved.sql, resolved.params, truncate=True
) )
@ -855,16 +918,15 @@ class RowUpdateView(BaseView):
) )
if request.args.get("_message"): if request.args.get("_message"):
message_row = returned_row
if message_row is None:
results = await resolved.db.execute(
resolved.sql, resolved.params, truncate=True
)
message_row = results.first()
self.ds.add_message( self.ds.add_message(
request, request,
await _row_flash_message( await _row_flash_message(
resolved.db, "Updated", resolved, row=message_row self.ds,
request,
"Updated",
resolved,
row=returned_row,
refresh_row=True,
), ),
self.ds.INFO, self.ds.INFO,
) )

View file

@ -311,6 +311,7 @@ class AllowedResourcesView(BaseView):
has_json_alternate = False has_json_alternate = False
async def get(self, request): async def get(self, request):
await self.ds.ensure_permission(action="view-instance", actor=request.actor)
await self.ds.refresh_schemas() await self.ds.refresh_schemas()
# Check if user has permissions-debug (to show sensitive fields) # Check if user has permissions-debug (to show sensitive fields)
@ -796,6 +797,8 @@ class CreateTokenView(BaseView):
raise Forbidden( raise Forbidden(
"Token authentication cannot be used to create additional tokens" "Token authentication cannot be used to create additional tokens"
) )
if "_r" in request.actor:
raise Forbidden("Restricted actors cannot create API tokens")
async def shared(self, request): async def shared(self, request):
self.check_permission(request) self.check_permission(request)
@ -873,6 +876,11 @@ class CreateTokenView(BaseView):
else: else:
errors.append("Invalid expire duration unit") errors.append("Invalid expire duration unit")
if errors:
context = await self.shared(request)
context["errors"] = errors
return await self.render(["create_token.html"], request, context)
# Are there any restrictions? # Are there any restrictions?
from datasette.tokens import TokenRestrictions from datasette.tokens import TokenRestrictions
@ -1261,14 +1269,21 @@ class SchemaBaseView(BaseView):
has_json_alternate = False has_json_alternate = False
async def get_database_schema(self, database_name): async def get_database_schema(self, database_name, actor):
"""Get schema SQL for a database.""" """Get schema SQL for a database."""
db = self.ds.databases[database_name] db = self.ds.databases[database_name]
result = await db.execute( allowed_tables_page = await self.ds.allowed_resources(
"select group_concat(sql, ';' || CHAR(10)) as schema from sqlite_master where sql is not null" "view-table", actor, parent=database_name
)
allowed_table_names = {
resource.child async for resource in allowed_tables_page.all()
}
result = await db.execute(
"select tbl_name, sql from sqlite_master where sql is not null"
)
return ";\n".join(
row["sql"] for row in result.rows if row["tbl_name"] in allowed_table_names
) )
row = result.first()
return row["schema"] if row and row["schema"] else ""
def format_json_response(self, data): def format_json_response(self, data):
"""Format data as JSON response with CORS headers if needed.""" """Format data as JSON response with CORS headers if needed."""
@ -1330,7 +1345,7 @@ class InstanceSchemaView(SchemaBaseView):
# Get schema for each database # Get schema for each database
schemas = [] schemas = []
for database_name in allowed_databases: for database_name in allowed_databases:
schema = await self.get_database_schema(database_name) schema = await self.get_database_schema(database_name, request.actor)
schemas.append({"database": database_name, "schema": schema}) schemas.append({"database": database_name, "schema": schema})
if format_ == "json": if format_ == "json":
@ -1371,7 +1386,7 @@ class DatabaseSchemaView(SchemaBaseView):
if database_name not in self.ds.databases: if database_name not in self.ds.databases:
return self.format_error_response("Database not found", format_) return self.format_error_response("Database not found", format_)
schema = await self.get_database_schema(database_name) schema = await self.get_database_schema(database_name, request.actor)
if format_ == "json": if format_ == "json":
return self.format_json_response( return self.format_json_response(
@ -1410,7 +1425,8 @@ class TableSchemaView(SchemaBaseView):
# Get schema for the table # Get schema for the table
db = self.ds.databases[database_name] db = self.ds.databases[database_name]
result = await db.execute( result = await db.execute(
"select sql from sqlite_master where name = ? and sql is not null", "select sql from sqlite_master where name = ? "
"and type in ('table', 'view') and sql is not null",
[table_name], [table_name],
) )
row = result.first() row = result.first()

View file

@ -279,7 +279,7 @@ class QueryCreateView(BaseView):
), ),
) )
response.status = status response.status = status
return response return _block_framing(response)
async def get(self, request): async def get(self, request):
db = await self.ds.resolve_database(request) db = await self.ds.resolve_database(request)
@ -527,7 +527,7 @@ class QueryEditView(BaseView):
), ),
) )
response.status = status response.status = status
return response return _block_framing(response)
async def get(self, request): async def get(self, request):
db, query_name, existing = await self._load(request) db, query_name, existing = await self._load(request)
@ -639,15 +639,17 @@ class QueryDeleteView(BaseView):
return Response.error( return Response.error(
["Trusted queries cannot be deleted using the API"], 403 ["Trusted queries cannot be deleted using the API"], 403
) )
return await self.render( return _block_framing(
["query_delete.html"], await self.render(
request, ["query_delete.html"],
{ request,
"database": db.name, {
"database_color": db.color, "database": db.name,
"query": stored_query_to_dict(existing), "database_color": db.color,
"query_url": self.ds.urls.table(db.name, query_name), "query": stored_query_to_dict(existing),
}, "query_url": self.ds.urls.table(db.name, query_name),
},
)
) )
async def post(self, request): async def post(self, request):

View file

@ -57,6 +57,7 @@ from datasette.utils.asgi import (
Request, Request,
Response, Response,
) )
from datasette.utils.sqlite import check_structured_write_table
from . import Context, from_extra from . import Context, from_extra
from .base import BaseView, DatasetteError, stream_csv from .base import BaseView, DatasetteError, stream_csv
@ -1126,6 +1127,7 @@ class TableInsertView(BaseView):
row_pk_values_for_later = [tuple(row[pk] for pk in pks) for row in rows] row_pk_values_for_later = [tuple(row[pk] for pk in pks) for row in rows]
def insert_or_upsert_rows(conn): def insert_or_upsert_rows(conn):
check_structured_write_table(conn, table_name)
table = sqlite_utils.Database(conn)[table_name] table = sqlite_utils.Database(conn)[table_name]
kwargs = {} kwargs = {}
if upsert: if upsert:
@ -1157,17 +1159,32 @@ class TableInsertView(BaseView):
# TODO: narrow to expected write errors so Datasette bugs surface as 500s # TODO: narrow to expected write errors so Datasette bugs surface as 500s
return Response.error([str(e)]) return Response.error([str(e)])
result = {"ok": True} result = {"ok": True}
# Only read back and disclose stored rows if the actor is also
# allowed to view this table - insert-row/update-row alone must
# not be usable to read data the actor cannot otherwise see.
if should_return and not await self.ds.allowed(
action="view-table",
resource=TableResource(database=database_name, table=table_name),
actor=request.actor,
):
should_return = False
if should_return: if should_return:
if upsert: if upsert:
# Fetch based on initial input IDs # Fetch based on initial input IDs
where_clause = " OR ".join( where_clause = " OR ".join(
["({})".format(" AND ".join(f"{pk} = ?" for pk in pks))] [
"({})".format(
" AND ".join(f"{escape_sqlite(pk)} = ?" for pk in pks)
)
]
* len(row_pk_values_for_later) * len(row_pk_values_for_later)
) )
args = list(itertools.chain.from_iterable(row_pk_values_for_later)) args = list(itertools.chain.from_iterable(row_pk_values_for_later))
fetched_rows = await db.execute( fetched_rows = await db.execute(
"select {}* from [{}] where {}".format( "select {}* from {} where {}".format(
"rowid, " if pks == ["rowid"] else "", table_name, where_clause "rowid, " if pks == ["rowid"] else "",
escape_sqlite(table_name),
where_clause,
), ),
args, args,
) )
@ -1382,7 +1399,9 @@ class TableDropView(BaseView):
"database": database_name, "database": database_name,
"table": table_name, "table": table_name,
"row_count": ( "row_count": (
await db.execute(f"select count(*) from [{table_name}]") await db.execute(
f"select count(*) from {escape_sqlite(table_name)}"
)
).single_value(), ).single_value(),
"message": 'Pass "confirm": true to confirm', "message": 'Pass "confirm": true to confirm',
}, },
@ -1391,7 +1410,9 @@ class TableDropView(BaseView):
# Drop table # Drop table
def drop_table(conn): def drop_table(conn):
sqlite_utils.Database(conn)[table_name].drop() table = sqlite_utils.Database(conn)[table_name]
table.disable_fts()
table.drop()
await db.execute_write_fn(drop_table, request=request) await db.execute_write_fn(drop_table, request=request)
await self.ds.track_event( await self.ds.track_event(
@ -1407,6 +1428,42 @@ class TableDropView(BaseView):
return Response.json({"ok": True}, status=200) return Response.json({"ok": True}, status=200)
class TableCountView(BaseView):
name = "table-count"
async def post(self, request):
try:
return await self.count(request)
except (NotFound, Forbidden, BadRequest, DatasetteError) as ex:
return Response.error(str(ex), status=ex.status)
async def count(self, request):
resolved = await self.ds.resolve_table(request)
visible, _private = await self.ds.check_visibility(
request.actor,
action="view-table",
resource=TableResource(database=resolved.db.name, table=resolved.table),
)
if not visible:
raise Forbidden("You do not have permission to view this table")
_, where_clauses, params, _, _ = await _table_filters(
self.ds, request, resolved.db.name, resolved.table
)
sql = f"select count(*) from {escape_sqlite(resolved.table)}"
if where_clauses:
sql += " where " + " and ".join(where_clauses)
try:
results = await resolved.db.execute(sql, params)
except QueryInterrupted:
return Response.error("Count query timed out", status=400)
except (sqlite3.OperationalError, InvalidSql) as ex:
return Response.error(str(ex), status=400)
return Response.json(
{"ok": True, "count": results.single_value()},
headers={"Cache-Control": "no-store"},
)
class TableFragmentView(BaseView): class TableFragmentView(BaseView):
name = "table-fragment" name = "table-fragment"
@ -1693,13 +1750,22 @@ async def table_view(datasette, request):
if ttl is None or not ttl.isdigit(): if ttl is None or not ttl.isdigit():
ttl = datasette.setting("default_cache_ttl") ttl = datasette.setting("default_cache_ttl")
private = getattr(request, "_datasette_private_response", False)
if datasette.cache_headers and response.status == 200: if datasette.cache_headers and response.status == 200:
ttl = int(ttl) if private:
if ttl == 0: # This response is only visible to the current actor (denied to
ttl_header = "no-cache" # anonymous requests), so it must never be stored by a shared
# cache/CDN - and ?_ttl= must not be able to override that.
response.headers["Cache-Control"] = "private, no-store"
response.headers["Vary"] = "Cookie"
else: else:
ttl_header = f"max-age={ttl}" ttl = int(ttl)
response.headers["Cache-Control"] = ttl_header if ttl == 0:
ttl_header = "no-cache"
else:
ttl_header = f"max-age={ttl}"
response.headers["Cache-Control"] = ttl_header
# Referrer policy # Referrer policy
response.headers["Referrer-Policy"] = "no-referrer" response.headers["Referrer-Policy"] = "no-referrer"
@ -1923,6 +1989,47 @@ async def table_view_traced(datasette, request):
return r return r
async def _table_filters(datasette, request, database_name, table_name):
# Arguments that start with _ and don't contain a __ are
# special - things like ?_search= - and should not be
# treated as filters.
filter_args = []
for key in request.args:
if not (key.startswith("_") and "__" not in key):
for v in request.args.getlist(key):
filter_args.append((key, v))
# Build where clauses from query string arguments
filters = Filters(sorted(filter_args))
where_clauses, params = filters.build_where_clauses(table_name)
# Execute filters_from_request plugin hooks - including the default
# ones that live in datasette/filters.py
extra_context_from_filters = {}
extra_human_descriptions = []
for hook in pm.hook.filters_from_request(
request=request,
table=table_name,
database=database_name,
datasette=datasette,
):
filter_arguments = await await_me_maybe(hook)
if filter_arguments:
where_clauses.extend(filter_arguments.where_clauses)
params.update(filter_arguments.params)
extra_human_descriptions.extend(filter_arguments.human_descriptions)
extra_context_from_filters.update(filter_arguments.extra_context)
return (
filters,
where_clauses,
params,
extra_human_descriptions,
extra_context_from_filters,
)
async def table_view_data( async def table_view_data(
datasette, datasette,
request, request,
@ -1947,6 +2054,10 @@ async def table_view_data(
) )
if not visible: if not visible:
raise Forbidden("You do not have permission to view this table") raise Forbidden("You do not have permission to view this table")
# Record whether this response is private (visible to this actor only)
# so the outer table_view() can set appropriate Cache-Control headers,
# regardless of which output format ends up being rendered.
request._datasette_private_response = private
# Redirect based on request.args, if necessary # Redirect based on request.args, if necessary
redirect_response = await _redirect_if_needed(datasette, request, resolved) redirect_response = await _redirect_if_needed(datasette, request, resolved)
@ -1997,36 +2108,13 @@ async def table_view_data(
table_metadata = await datasette.table_config(database_name, table_name) table_metadata = await datasette.table_config(database_name, table_name)
# Arguments that start with _ and don't contain a __ are (
# special - things like ?_search= - and should not be filters,
# treated as filters. where_clauses,
filter_args = [] params,
for key in request.args: extra_human_descriptions,
if not (key.startswith("_") and "__" not in key): extra_context_from_filters,
for v in request.args.getlist(key): ) = await _table_filters(datasette, request, database_name, table_name)
filter_args.append((key, v))
# Build where clauses from query string arguments
filters = Filters(sorted(filter_args))
where_clauses, params = filters.build_where_clauses(table_name)
# Execute filters_from_request plugin hooks - including the default
# ones that live in datasette/filters.py
extra_context_from_filters = {}
extra_human_descriptions = []
for hook in pm.hook.filters_from_request(
request=request,
table=table_name,
database=database_name,
datasette=datasette,
):
filter_arguments = await await_me_maybe(hook)
if filter_arguments:
where_clauses.extend(filter_arguments.where_clauses)
params.update(filter_arguments.params)
extra_human_descriptions.extend(filter_arguments.human_descriptions)
extra_context_from_filters.update(filter_arguments.extra_context)
# Deal with custom sort orders # Deal with custom sort orders
sortable_columns = await _sortable_columns_for_table( sortable_columns = await _sortable_columns_for_table(
@ -2430,9 +2518,12 @@ async def _next_value_and_url(
except IndexError: except IndexError:
# sort/sort_desc column missing from SELECT - look up value by PK instead # sort/sort_desc column missing from SELECT - look up value by PK instead
prefix_where_clause = " and ".join( prefix_where_clause = " and ".join(
f"[{pk}] = :pk{i}" for i, pk in enumerate(pks) f"{escape_sqlite(pk)} = :pk{i}" for i, pk in enumerate(pks)
)
prefix_lookup_sql = (
f"select {escape_sqlite(sort or sort_desc)} "
f"from {escape_sqlite(table_name)} where {prefix_where_clause}"
) )
prefix_lookup_sql = f"select [{sort or sort_desc}] from [{table_name}] where {prefix_where_clause}"
prefix = ( prefix = (
await db.execute( await db.execute(
prefix_lookup_sql, prefix_lookup_sql,

View file

@ -27,7 +27,15 @@ from datasette.utils import (
table_column_details, table_column_details,
) )
from datasette.utils.asgi import NotFound, PayloadTooLarge, Response from datasette.utils.asgi import NotFound, PayloadTooLarge, Response
from datasette.utils.sqlite import sqlite_hidden_table_names from datasette.utils.permissions import (
SKIP_PERMISSION_CHECKS,
gather_permission_sql_from_hooks,
resolve_permissions_with_candidates,
)
from datasette.utils.sqlite import (
check_structured_write_table,
sqlite_hidden_table_names,
)
from .base import BaseView from .base import BaseView
@ -122,6 +130,30 @@ def _public_foreign_key_target(target):
} }
async def _filter_visible_foreign_key_targets(datasette, actor, database_name, targets):
if not targets:
return []
permission_sqls = await gather_permission_sql_from_hooks(
datasette=datasette,
actor=actor,
action="view-table",
)
if permission_sqls is SKIP_PERMISSION_CHECKS:
return targets
candidate_tables = list(dict.fromkeys(target["fk_table"] for target in targets))
permission_rows = await resolve_permissions_with_candidates(
datasette.get_internal_database(),
actor,
permission_sqls,
[(database_name, table_name) for table_name in candidate_tables],
"view-table",
)
visible_tables = {row["child"] for row in permission_rows if bool(row["allow"])}
return [target for target in targets if target["fk_table"] in visible_tables]
def _singular(name): def _singular(name):
if name.endswith("ies") and len(name) > 3: if name.endswith("ies") and len(name) > 3:
return name[:-3] + "y" return name[:-3] + "y"
@ -821,16 +853,18 @@ class TableCreateView(BaseView):
ignore = create_request.ignore ignore = create_request.ignore
replace = create_request.replace replace = create_request.replace
table_name = create_request.table
table_exists = await db.table_exists(table_name)
table_resource = TableResource(database=database_name, table=table_name)
# Replacing rows requires update-row permission # Replacing rows requires update-row permission
if replace and not await self.ds.allowed( if replace and not await self.ds.allowed(
action="update-row", action="update-row",
resource=DatabaseResource(database=database_name), resource=table_resource,
actor=request.actor, actor=request.actor,
): ):
return Response.error(["Permission denied: need update-row"], 403) return Response.error(["Permission denied: need update-row"], 403)
table_name = create_request.table
table_exists = await db.table_exists(table_name)
columns = create_request.columns columns = create_request.columns
rows = create_request.rows_list rows = create_request.rows_list
@ -838,7 +872,7 @@ class TableCreateView(BaseView):
# Must have insert-row permission # Must have insert-row permission
if not await self.ds.allowed( if not await self.ds.allowed(
action="insert-row", action="insert-row",
resource=DatabaseResource(database=database_name), resource=table_resource,
actor=request.actor, actor=request.actor,
): ):
return Response.error(["Permission denied: need insert-row"], 403) return Response.error(["Permission denied: need insert-row"], 403)
@ -857,7 +891,7 @@ class TableCreateView(BaseView):
if create_request.alter: if create_request.alter:
if not await self.ds.allowed( if not await self.ds.allowed(
action="alter-table", action="alter-table",
resource=DatabaseResource(database=database_name), resource=table_resource,
actor=request.actor, actor=request.actor,
): ):
return Response.error( return Response.error(
@ -893,6 +927,7 @@ class TableCreateView(BaseView):
) )
def create_table(conn): def create_table(conn):
check_structured_write_table(conn, table_name, allow_missing=True)
db_for_write = sqlite_utils.Database(conn) db_for_write = sqlite_utils.Database(conn)
table = db_for_write[table_name] table = db_for_write[table_name]
if rows: if rows:
@ -1012,6 +1047,9 @@ class DatabaseForeignKeyTargetsView(BaseView):
for target in (await db.execute(FOREIGN_KEY_TARGETS_SQL)).dicts() for target in (await db.execute(FOREIGN_KEY_TARGETS_SQL)).dicts()
if target["fk_table"] not in hidden_tables if target["fk_table"] not in hidden_tables
] ]
targets = await _filter_visible_foreign_key_targets(
self.ds, request.actor, database_name, targets
)
return Response.json( return Response.json(
{ {
"ok": True, "ok": True,
@ -1050,6 +1088,15 @@ class TableForeignKeySuggestionsView(BaseView):
source_columns, targets, current_by_column = await db.execute_fn( source_columns, targets, current_by_column = await db.execute_fn(
lambda conn: _foreign_key_suggestion_metadata(conn, table_name) lambda conn: _foreign_key_suggestion_metadata(conn, table_name)
) )
targets = await _filter_visible_foreign_key_targets(
self.ds, request.actor, database_name, targets
)
visible_target_tables = {target["fk_table"] for target in targets}
current_by_column = {
column: current
for column, current in current_by_column.items()
if current["fk_table"] in visible_target_tables
}
columns = [] columns = []
options_by_column = {} options_by_column = {}

View file

@ -1206,7 +1206,10 @@ class ForeignKeyTablesExtra(Extra):
async def resolve(self, context): async def resolve(self, context):
return await context.foreign_key_tables( return await context.foreign_key_tables(
context.database_name, context.table_name, context.pk_values context.database_name,
context.table_name,
context.pk_values,
actor=context.request.actor,
) )

View file

@ -83,6 +83,22 @@ def decision_for_write_sql_operation(
) )
if operation.operation == "function": if operation.operation == "function":
return IgnoreWriteSqlOperation("SQL function") return IgnoreWriteSqlOperation("SQL function")
if (
operation.operation == "read"
and operation.target_type == "table"
and operation.table is not None
and operation.table_kind is None
and operation.table.lower().startswith("pragma_")
):
# Eponymous table-valued PRAGMA functions (e.g. pragma_table_info("secret"))
# report a read of the synthetic "pragma_table_info" table, not of the
# table passed as an argument. That means a view-table denial on the real
# table is never consulted, so these could otherwise be used to read
# schema metadata (column names, table lists, ...) for tables the actor
# is not allowed to view. Reject them outright in untrusted write SQL,
# including inside CREATE VIEW bodies (whose reads are discovered here
# via the rolled-back dependency-read analysis above).
return UnsupportedWriteSqlOperation(unsupported_message)
if ( if (
operation.operation == "read" operation.operation == "read"
and operation.target_type == "table" and operation.target_type == "table"

View file

@ -158,6 +158,15 @@ Datasette resolves matching rules from most specific to least specific:
This means a resource-level allow can provide an exception to a parent-level deny. It also means that two plugins which disagree at the same level resolve to deny. This means a resource-level allow can provide an exception to a parent-level deny. It also means that two plugins which disagree at the same level resolve to deny.
For table and view permissions, resource names use SQLite's case-insensitive
identifier matching: ``Secret``, ``secret`` and ``SECRET`` identify the same
table. This applies to configuration rules, plugin rules and token restrictions.
Only ASCII letters are case-insensitive; non-ASCII characters remain distinct.
Conflicting rules for different spellings of the same name follow the usual
deny-wins rule at the same scope. Names retain their original spelling in
resource listings and permission explanations. Database names, stored query
names and other resource types remain case-sensitive.
.. list-table:: Permission rule examples .. list-table:: Permission rule examples
:header-rows: 1 :header-rows: 1
@ -182,6 +191,18 @@ This means a resource-level allow can provide an exception to a parent-level den
The built-in public defaults are global allow rules for actions such as ``view-instance``, ``view-database`` and ``view-table``. They follow the same precedence rules as configuration and plugin rules. The ``--default-deny`` option prevents Datasette from contributing those default allow rules. The built-in public defaults are global allow rules for actions such as ``view-instance``, ``view-database`` and ``view-table``. They follow the same precedence rules as configuration and plugin rules. The ``--default-deny`` option prevents Datasette from contributing those default allow rules.
The built-in ``datasette.default_permissions.sqlite_statistics`` plugin denies
``view-table`` for ``sqlite_stat1``, ``sqlite_stat2``, ``sqlite_stat3`` and
``sqlite_stat4``. These table-level denials also apply to root users and take
precedence over configuration or plugin allow rules at the same scope.
This controls table access and listings, without changing ``execute-sql`` or
SQLite's internal use of statistics.
A plugin can replace this policy by unregistering
``datasette.default_permissions.sqlite_statistics`` through ``datasette.pm``
and registering its own permission hook. Plugin registration is process-wide:
replacing this policy affects every Datasette instance in that process.
Datasette performs checks using :ref:`datasette_allowed`, which accepts keyword arguments for ``action``, ``resource`` and an optional ``actor``. Datasette performs checks using :ref:`datasette_allowed`, which accepts keyword arguments for ``action``, ``resource`` and an optional ``actor``.
``resource`` should be an instance of the appropriate ``Resource`` subclass from :mod:`datasette.resources`—for example ``InstanceResource()``, ``DatabaseResource(database="...``)`` or ``TableResource(database="...", table="...")``. This defaults to ``InstanceResource()`` if not specified. ``resource`` should be an instance of the appropriate ``Resource`` subclass from :mod:`datasette.resources`—for example ``InstanceResource()``, ``DatabaseResource(database="...``)`` or ``TableResource(database="...", table="...")``. This defaults to ``InstanceResource()`` if not specified.
@ -771,6 +792,8 @@ Datasette defaults to allowing any site visitor to execute their own custom SQL
Access to this ability is controlled by the :ref:`actions_execute_sql` permission. Access to this ability is controlled by the :ref:`actions_execute_sql` permission.
This permission does not apply to structured table-browsing operations where Datasette constructs the SQL, such as sorting, column filters and :ref:`facets`. Faceting is controlled separately by the :ref:`setting_allow_facet` setting.
The easiest way to disable arbitrary SQL queries is using the :ref:`default_allow_sql setting <setting_default_allow_sql>` when you first start Datasette running. The easiest way to disable arbitrary SQL queries is using the :ref:`default_allow_sql setting <setting_default_allow_sql>` when you first start Datasette running.
You can alternatively use an ``"allow_sql"`` block to control who is allowed to execute arbitrary SQL queries. You can alternatively use an ``"allow_sql"`` block to control who is allowed to execute arbitrary SQL queries.
@ -1359,6 +1382,12 @@ view-table
Actor is allowed to view a table (or view) page, e.g. https://latest.datasette.io/fixtures/complex_foreign_keys Actor is allowed to view a table (or view) page, e.g. https://latest.datasette.io/fixtures/complex_foreign_keys
Derived implementation tables require access to their immediate source: FTS and RTree shadow tables require access to their virtual table, external-content FTS tables require access to their content table, and FTS vocabulary tables (``fts5vocab`` and ``fts4aux``) require access to their FTS table. The derived table's own permission rules also apply.
Access is always denied if the source table is itself derived, or if a vocabulary table's source cannot be identified.
The same rules apply to individual permission checks and table listings, including whether they are private. If a database error prevents dependency discovery, the check or listing fails with an error instead of ignoring the dependencies. Failed discovery results are not cached, so later checks can retry.
``resource`` - ``datasette.resources.TableResource(database, table)`` ``resource`` - ``datasette.resources.TableResource(database, table)``
``database`` is the name of the database (string) ``database`` is the name of the database (string)
@ -1521,6 +1550,8 @@ execute-sql
Actor is allowed to run arbitrary read-only SQL queries against a specific database using the :ref:`custom SQL query page <pages_custom_sql_queries>`, e.g. https://latest.datasette.io/fixtures/-/query?sql=select+100 Actor is allowed to run arbitrary read-only SQL queries against a specific database using the :ref:`custom SQL query page <pages_custom_sql_queries>`, e.g. https://latest.datasette.io/fixtures/-/query?sql=select+100
This action also controls raw SQL supplied using ``?_where=``. It does not control structured table-browsing features such as :ref:`facets`, which use SQL generated by Datasette and are controlled by :ref:`setting_allow_facet`.
``resource`` - ``datasette.resources.DatabaseResource(database)`` ``resource`` - ``datasette.resources.DatabaseResource(database)``
``database`` is the name of the database (string) ``database`` is the name of the database (string)

View file

@ -4,6 +4,67 @@
Changelog Changelog
========= =========
.. _unreleased:
Unreleased
----------
- Fixed incorrect counts when clicking **count all** on filtered table pages. The button now uses a new :ref:`POST count endpoint <TableCountView>`. (:issue:`2914`)
- Datasette now uses `httpx2 <https://httpx2.pydantic.dev/>`__, the Pydantic-maintained continuation of `httpx <https://www.python-httpx.org/>`__, in place of ``httpx``. The public API is the same, but responses returned by :ref:`internals_datasette_client` are now ``httpx2.Response`` objects rather than ``httpx.Response``. Plugins that use ``isinstance()`` checks against ``httpx.Response`` should be updated to use ``httpx2``. **Plugins that use httpx without explicitly depending on it** will need to add an explicit dependency or switch to `httpx2`.
Background tasks
~~~~~~~~~~~~~~~~
Datasette plugins can now use **background tasks** to run code independent of the Datasette request/response cycle.
- New :ref:`datasette_add_background_task` API: plugins register supervised, long-lived background work - typically from a ``startup`` hook - and these will be launched after every ``startup`` hook has run. Tasks are cancelled (with a five-second grace period) on shutdown.
- New ``/-/tasks`` JSON debug endpoint lists every supervised background task and its state, in the style of ``/-/threads``. See :ref:`JsonDataView_tasks`. It requires the ``permissions-debug`` permission.
- New :ref:`plugin_hook_shutdown` plugin hook, called during graceful shutdown (Ctrl-C, ``SIGTERM``) before background tasks are cancelled and before database connections are closed. It is not called on a hard kill (``SIGKILL``).
- Plugin ``asgi_wrapper`` middleware now always runs *after* startup has completed.
- If your plugin uses ``asgi_wrapper`` to start background tasks on the first incoming request, you should migrate to ``datasette.add_background_task()`` instead. `datasette-cron <https://datasette.io/plugins/datasette-cron>`__ and `datasette-enrichments <https://datasette.io/plugins/datasette-enrichments>`__ are being migrated to this pattern.
.. _v1_0_a39:
1.0a39 (2026-09-10)
-------------------
This alpha release includes security fixes for permissions, SQL construction, HTML rendering, authentication and caching, plus improvements to application startup and write execution.
See `0.65.4 <https://docs.datasette.io/en/stable/changelog.html#v0-65-4>`__ for fixes that have been backported to the stable 0.65.x branch.
The Datasette blog `has more details on these releases <https://datasette.io/blog/2026/september-security-releases/>`__.
Some of the security fixes include:
- Table and view permission checks now take SQLite's case-insensitive names into account. See :ref:`authentication_permissions_explained`.
- Viewing a full-text search index table now checks you have permission to view the table from which it draws its content.
- Viewing SQLite statistics tables (``sqlite_stat1`` through ``sqlite_stat4``) is now denied by a default.
- Table schema display now obeys the ``view-table`` permission.
- Table filters using ``?_through=`` require permission to view the intermediate table.
- Foreign-key target and suggestion APIs, incoming foreign-key relationships and their row counts now respect ``view-table`` permission.
- Row endpoints check permissions before resolving primary keys, to avoid revealing the existence of an otherwise invisible primary key.
- Improved permission checks for the create-table API. See :ref:`json_api_write`.
- The write SQL interface now checks ``view-table`` permission for tables referenced by ``CREATE VIEW`` statements.
- Fixed SQL identifier escaping for column names from untrusted database schemas.
- Fixed HTML escaping for column names from untrusted database schemas.
- URL columns now render links only for validated HTTP or HTTPS URLs.
- Private and personalized dynamic responses now use ``Cache-Control: private, no-store``. Anonymous dynamic responses vary by ``Cookie`` and ``Authorization``.
- Actor cookies now respect ``expire_after``.
- Restricted actors can no longer create API tokens.
- Stored-query create, edit and delete forms now block framing to prevent clickjacking.
- Configuration secret redaction now matches key names case-insensitively.
- SQLite extension loading is disabled after extensions supplied using ``--load-extension`` have been loaded.
Other improvements and fixes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- :ref:`db.execute_write() <database_execute_write>` now has a default execution time limit of 2,000ms. Plugins can override this using ``time_limit_ms=`` or disable it using ``time_limit_ms=None``. This limit is independent of the ``sql_time_limit_ms`` setting for read queries.
- Application startup now runs through ASGI lifespan events before requests are accepted, with a first-request fallback for hosts without lifespan support. Thanks, `Alex Garcia <https://github.com/asg017>`__. (:pr:`2887`)
- ``datasette serve`` now runs startup hooks and Uvicorn on the same event loop, preserving background tasks started by plugins. The minimum Uvicorn version is now 0.29. Thanks, `Alex Garcia <https://github.com/asg017>`__. (:pr:`2886`)
- Non-blocking writes using ``execute_write_fn(..., block=False)`` now return a distinct task UUID for every call and work correctly with ``num_sql_threads=0``. Thanks, `Zain Dana Harper <https://github.com/HarperZ9>`__. (:issue:`2860`, :issue:`2859`)
- Dropping a table now disables its full-text search index first. (:issue:`2874`)
- Fixed ``CREATE VIEW`` SQL analysis on Python 3.10.
.. _v1_0_a38: .. _v1_0_a38:
1.0a38 (2026-08-06) 1.0a38 (2026-08-06)
@ -1092,7 +1153,7 @@ Features
- New ``--nolock`` option for ignoring file locks when opening read-only databases. (:issue:`1744`) - New ``--nolock`` option for ignoring file locks when opening read-only databases. (:issue:`1744`)
- Spaces in the database names in URLs are now encoded as ``+`` rather than ``~20``. (:issue:`1701`) - Spaces in the database names in URLs are now encoded as ``+`` rather than ``~20``. (:issue:`1701`)
- ``<Binary: 2427344 bytes>`` is now displayed as ``<Binary: 2,427,344 bytes>`` and is accompanied by tooltip showing "2.3MB". (:issue:`1712`) - ``<Binary: 2427344 bytes>`` is now displayed as ``<Binary: 2,427,344 bytes>`` and is accompanied by tooltip showing "2.3MB". (:issue:`1712`)
- The base Docker image used by ``datasette publish cloudrun``, ``datasette package`` and the `official Datasette image <https://hub.docker.com/datasetteproject/datasette>`__ has been upgraded to ``3.10.6-slim-bullseye``. (:issue:`1768`) - The base Docker image used by ``datasette publish cloudrun``, ``datasette package`` and the `official Datasette image <https://hub.docker.com/r/datasetteproject/datasette>`__ has been upgraded to ``3.10.6-slim-bullseye``. (:issue:`1768`)
- Canned writable queries against immutable databases now show a warning message. (:issue:`1728`) - Canned writable queries against immutable databases now show a warning message. (:issue:`1728`)
- ``datasette publish cloudrun`` has a new ``--timeout`` option which can be used to increase the time limit applied by the Google Cloud build environment. Thanks, Tim Sherratt. (:pr:`1717`) - ``datasette publish cloudrun`` has a new ``--timeout`` option which can be used to increase the time limit applied by the Google Cloud build environment. Thanks, Tim Sherratt. (:pr:`1717`)
- ``datasette publish cloudrun`` has new ``--min-instances`` and ``--max-instances`` options. (:issue:`1779`) - ``datasette publish cloudrun`` has new ``--min-instances`` and ``--max-instances`` options. (:issue:`1779`)
@ -2121,7 +2182,7 @@ If you are still running Python 3.5 you should stick with ``0.30.2``, which you
- Removed obsolete ``?_group_count=col`` feature (:issue:`504`) - Removed obsolete ``?_group_count=col`` feature (:issue:`504`)
- Improved user interface and documentation for ``datasette publish cloudrun`` (:issue:`608`) - Improved user interface and documentation for ``datasette publish cloudrun`` (:issue:`608`)
- Tables with indexes now show the ``CREATE INDEX`` statements on the table page (:issue:`618`) - Tables with indexes now show the ``CREATE INDEX`` statements on the table page (:issue:`618`)
- Current version of `uvicorn <https://www.uvicorn.org/>`__ is now shown on ``/-/versions`` - Current version of `uvicorn <https://uvicorn.dev/>`__ is now shown on ``/-/versions``
- Python 3.8 is now supported! (:issue:`622`) - Python 3.8 is now supported! (:issue:`622`)
- Python 3.5 is no longer supported. - Python 3.5 is no longer supported.
@ -2172,7 +2233,7 @@ If you are still running Python 3.5 you should stick with ``0.30.2``, which you
0.29.2 (2019-07-13) 0.29.2 (2019-07-13)
------------------- -------------------
- Bumped `Uvicorn <https://www.uvicorn.org/>`__ to 0.8.4, fixing a bug where the query string was not included in the server logs. (:issue:`559`) - Bumped `Uvicorn <https://uvicorn.dev/>`__ to 0.8.4, fixing a bug where the query string was not included in the server logs. (:issue:`559`)
- Fixed bug where the navigation breadcrumbs were not displayed correctly on the page for a custom query. (:issue:`558`) - Fixed bug where the navigation breadcrumbs were not displayed correctly on the page for a custom query. (:issue:`558`)
- Fixed bug where custom query names containing unicode characters caused errors. - Fixed bug where custom query names containing unicode characters caused errors.
@ -2194,7 +2255,7 @@ ASGI, new plugin hooks, facet by date and much, much more...
ASGI ASGI
~~~~ ~~~~
`ASGI <https://asgi.readthedocs.io/>`__ is the Asynchronous Server Gateway Interface standard. I've been wanting to convert Datasette into an ASGI application for over a year - `Port Datasette to ASGI #272 <https://github.com/simonw/datasette/issues/272>`__ tracks thirteen months of intermittent development - but with Datasette 0.29 the change is finally released. This also means Datasette now runs on top of `Uvicorn <https://www.uvicorn.org/>`__ and no longer depends on `Sanic <https://github.com/huge-success/sanic>`__. `ASGI <https://asgi.readthedocs.io/>`__ is the Asynchronous Server Gateway Interface standard. I've been wanting to convert Datasette into an ASGI application for over a year - `Port Datasette to ASGI #272 <https://github.com/simonw/datasette/issues/272>`__ tracks thirteen months of intermittent development - but with Datasette 0.29 the change is finally released. This also means Datasette now runs on top of `Uvicorn <https://uvicorn.dev/>`__ and no longer depends on `Sanic <https://github.com/huge-success/sanic>`__.
I wrote about the significance of this change in `Porting Datasette to ASGI, and Turtles all the way down <https://simonwillison.net/2019/Jun/23/datasette-asgi/>`__. I wrote about the significance of this change in `Porting Datasette to ASGI, and Turtles all the way down <https://simonwillison.net/2019/Jun/23/datasette-asgi/>`__.

View file

@ -14,6 +14,8 @@ Here's `an example <https://congress-legislators.datasettes.com/legislators/legi
Facets can be specified in two ways: using query string parameters, or in ``metadata.json`` configuration for the table. Facets can be specified in two ways: using query string parameters, or in ``metadata.json`` configuration for the table.
Facet queries are generated by Datasette and summarize rows the actor already has permission to view. They do not require the :ref:`actions_execute_sql` permission. Use the :ref:`setting_allow_facet` setting to control whether users can request facets using query string parameters.
Facets in query strings Facets in query strings
----------------------- -----------------------

View file

@ -52,11 +52,11 @@ Configuring full-text search for a table or view
If a table has a corresponding FTS table set up using the ``content=`` argument to ``CREATE VIRTUAL TABLE`` shown below, Datasette will detect it automatically and add a search interface to the table page for that table. If a table has a corresponding FTS table set up using the ``content=`` argument to ``CREATE VIRTUAL TABLE`` shown below, Datasette will detect it automatically and add a search interface to the table page for that table.
You can also manually configure which table should be used for full-text search using query string parameters or table configuration in ``datasette.yaml`` (see :ref:`table_configuration_fts`). You can set the associated FTS table for a specific table and you can also set one for a view - if you do that, the page for that SQL view will offer a search option. You can also manually configure which table should be used for full-text search using table configuration in ``datasette.yaml`` (see :ref:`table_configuration_fts`). You can set the associated FTS table for a specific table and you can also set one for a view - if you do that, the page for that SQL view will offer a search option.
Use ``?_fts_table=x`` to over-ride the FTS table for a specific page. If the primary key was something other than ``rowid`` you can use ``?_fts_pk=col`` to set that as well. This is particularly useful for views, for example: The legacy ``?_fts_table=x`` and ``?_fts_pk=col`` query string parameters are accepted only if they exactly match the configured or automatically detected FTS mapping. They cannot be used to select a different FTS table or primary key. This prevents a public table from being used to probe the contents of a private FTS table.
https://latest.datasette.io/fixtures/searchable_view?_fts_table=searchable_fts&_fts_pk=pk Searching also requires the current actor to have ``view-table`` permission for the FTS table itself, in addition to permission to view the table or view being searched.
The ``fts_table`` metadata property can be used to specify an associated FTS table. If the primary key column in your table which was used to populate the FTS table is something other than ``rowid``, you can specify the column to use with the ``fts_pk`` property. The ``fts_table`` metadata property can be used to specify an associated FTS table. If the primary key column in your table which was used to populate the FTS table is something other than ``rowid``, you can specify the column to use with the ``fts_pk`` property.

View file

@ -1403,7 +1403,139 @@ Release all resources held by this ``Datasette`` instance. This calls :ref:`data
If a call to ``Database.close()`` on one of the attached databases raises an exception, ``Datasette.close()`` will continue trying to close the remaining databases and will re-raise the first exception after every database has been processed. If a call to ``Database.close()`` on one of the attached databases raises an exception, ``Datasette.close()`` will continue trying to close the remaining databases and will re-raise the first exception after every database has been processed.
When Datasette is being served over ASGI the ``close()`` method is wired up to the lifespan shutdown event, so resources are released cleanly on ``SIGTERM`` / ``SIGINT``. When Datasette is being served over ASGI the ``close()`` method is wired up to the lifespan shutdown event, so resources are released cleanly on ``SIGTERM`` / ``SIGINT``. See :ref:`datasette_lifecycle` for where ``close()`` fits into the full startup-to-shutdown sequence.
.. _datasette_lifecycle:
Application lifecycle
---------------------
Datasette guarantees a fixed sequence of events between the moment a ``Datasette`` instance is constructed and the moment its resources are released:
1. ``Datasette(...)`` — the constructor runs synchronously and does not run plugin hooks.
2. **Startup**``await datasette.invoke_startup()`` runs once: it populates the internal database's catalog of table schemas (:ref:`internals_internal`), loads canned queries and column type configuration, then calls every registered :ref:`plugin_hook_startup` hook, in plugin registration order. When Datasette is being served, table-count precomputation for immutable databases runs immediately before this, as part of the same startup sequence.
3. **Background-task launch** — once *every* ``startup`` hook has finished (not before), every task registered with :ref:`datasette_add_background_task` — by any plugin — is launched. A task registered by one plugin's ``startup`` hook can safely depend on state set up by another plugin's ``startup`` hook, because launch only happens after the whole round of hooks completes.
4. **Serving** — the instance handles requests (or, for headless or CLI use, does whatever the embedding program does with it).
5. **Shutdown** — triggered by the ASGI ``lifespan.shutdown`` event (Ctrl-C, ``SIGTERM``) or the end of a ``datasette serve`` process: every :ref:`plugin_hook_shutdown` hook runs first, while background tasks are still alive, so a plugin can tell its own task to wind down gracefully; every still-running background task is then cancelled and given a five-second grace period to actually stop; finally every database connection is released via :ref:`datasette_close`.
.. admonition:: Startup hooks run on the event loop that serves requests
In every trigger path below, ``startup`` hooks run on the same ``asyncio`` event loop that goes on to accept connections. It is safe to create loop-bound primitives — ``asyncio.Lock``, ``asyncio.Queue``, ``asyncio.Event``, a raw ``asyncio.create_task()`` call — inside a ``startup`` hook, and to register long-lived background work with :ref:`datasette_add_background_task` there. This was not always true: older Datasette versions ran startup on a temporary event loop in the CLI that was closed before the server's own loop was created, which could silently kill anything scheduled on it.
Three trigger paths
~~~~~~~~~~~~~~~~~~~
- **``datasette serve`` (CLI)** — startup and ``uvicorn.Server.serve()`` both run inside a single ``asyncio.run()`` call, so there is exactly one event loop for the whole life of the process.
- **ASGI lifespan**``Datasette.app()`` wires startup and background-task launch into the ``on_startup`` list, and shutdown into the ``on_shutdown`` list, of an internal ``AsgiLifespan`` wrapper. A spec-compliant ASGI server (uvicorn, hypercorn, and others) sends the ``lifespan.startup`` message and waits for ``lifespan.startup.complete`` before delivering any ``http`` or ``websocket`` scope, so startup — including every plugin's own internal-database migrations — is guaranteed to have finished before any request reaches Datasette, including requests seen by plugin :ref:`asgi_wrapper <plugin_asgi_wrapper>` middleware. If a ``startup`` hook raises, ``AsgiLifespan`` sends ``lifespan.startup.failed`` with the exception message instead of hanging or crashing ambiguously, so the host can abort the boot cleanly.
- **First-request fallback** — an internal ``AsgiRunOnFirstRequest`` wrapper runs the same startup work as a safety net for hosts that never send ASGI lifespan events at all: some ASGI mounts, a bare ``app()`` embedded inside another framework, and :ref:`datasette.client <internals_datasette_client>` / test clients, which drive requests directly over ``httpx2.ASGITransport`` without ever emitting ``lifespan.startup``. It runs startup exactly once, the first time any non-lifespan scope arrives, guarded by a lock so that concurrent early requests can't run it twice.
All three paths call the same idempotent internal methods, so it is safe for more than one of them to fire — lifespan startup completing and then a first request arriving afterwards is a no-op the second time. A host that never sends lifespan events and never goes through the CLI degrades to first-request timing: startup runs on the first request instead of before it, exactly as Datasette always worked prior to this lifecycle guarantee. This is a deliberate fallback rather than a regression — see :ref:`datasette_add_background_task` for how to opt out of launching background tasks (the ``--get`` CLI path) or drive startup and launch explicitly (tests, headless embedders).
.. _datasette_add_background_task:
.add_background_task(func, name=None)
-------------------------------------
``func`` - async callable
A coroutine function taking one positional argument, the ``Datasette`` instance. Core calls ``await func(datasette)``.
``name`` - string, optional
A name for the task, used to identify it in the ``/-/tasks`` introspection endpoint (:ref:`JsonDataView_tasks`) and in log messages. Defaults to ``func.__qualname__``. If the resulting name collides with an already-registered task, a ``-2``, ``-3``, ... suffix is appended.
Registers a piece of supervised, long-lived background work — typically called from a :ref:`plugin_hook_startup` hook, though it can be called at any point after the instance exists, including from a request handler. Returns a :ref:`BackgroundTask <BackgroundTask>` handle.
Registration is separate from launch. Calling this from a ``startup`` hook — the common case — buffers the task; core launches every registered task once *all* ``startup`` hooks have completed, as described in :ref:`datasette_lifecycle`. Calling it after launch has already happened — for example from a request handler, to start a per-job task dynamically — starts the task immediately instead.
.. code-block:: python
import asyncio
from datasette import hookimpl
async def poll_for_updates(datasette):
while True:
await do_one_poll(datasette)
await asyncio.sleep(60)
@hookimpl
def startup(datasette):
datasette.add_background_task(
poll_for_updates, name="my-plugin-poller"
)
Core owns the task for the rest of the process's life:
- **A strong reference is kept forever**, so the task can never be silently garbage collected the way an unreferenced ``asyncio.create_task()`` call can be.
- **A crash is logged, not swallowed.** If ``func`` raises anything other than ``asyncio.CancelledError``, the exception (with its traceback) is logged to the ``datasette.background_tasks`` logger and recorded on the handle's ``.exception``, and the task's ``.state`` becomes ``crashed``.
- **Cancellation is coordinated.** On shutdown, every task that is still running is cancelled and given a grace period to stop — see :ref:`datasette_lifecycle`.
Launch matrix
~~~~~~~~~~~~~
Whether registered tasks actually launch depends on how the instance is being run:
.. list-table::
:header-rows: 1
* - Trigger
- Launches registered tasks?
* - ASGI lifespan (real server deployments)
- Yes, after ``lifespan.startup`` completes
* - First-request fallback (lifespan-less hosts)
- Yes, on the first request — parity with the lifespan case
* - ``datasette serve --get``
- Never
* - Tests / headless embedders
- Only if you call :ref:`datasette_start_background_tasks` explicitly
``datasette --get`` never launches background tasks, even though its one-shot request flows through the same first-request fallback as everything else: it sets an internal flag before making that request specifically to suppress the launch, since a one-shot CLI invocation has no server loop left running afterwards to keep any launched tasks alive.
.. _BackgroundTask:
BackgroundTask objects
~~~~~~~~~~~~~~~~~~~~~~
``add_background_task()`` returns a ``BackgroundTask`` handle with the following attributes:
``.name`` - string
The task's (unique) name.
``.state`` - string
One of ``registered`` (added but not yet launched), ``running``, ``completed`` (returned cleanly), ``crashed`` (raised an exception) or ``cancelled``.
``.task`` - ``asyncio.Task`` or ``None``
The underlying ``asyncio.Task``, once launched. ``None`` while still ``registered``.
``.exception`` - ``BaseException`` or ``None``
The exception that crashed the task, if ``.state`` is ``crashed``.
``.started_at`` - string or ``None``
ISO 8601 UTC timestamp of when the task was launched.
``.function`` - string
The callable's dotted module and qualified name, for example ``my_plugin.jobs.poll_for_updates``.
``.cancel()``
Cancel the task. If it has already launched, this cancels the underlying ``asyncio.Task````.state`` becomes ``cancelled`` once the cancellation is observed. If it has not launched yet, it is removed from the queue so it never runs.
This is also the shape of each entry returned by the ``/-/tasks`` JSON introspection endpoint — see :ref:`JsonDataView_tasks`.
.. _datasette_start_background_tasks:
await .start_background_tasks()
-------------------------------
Runs startup (if it has not already run) and launches every task registered with :ref:`datasette_add_background_task`. This is the explicit equivalent of what happens automatically via ASGI lifespan or the first-request fallback in a served deployment — the entry point for tests and headless embedders (a cron-style CLI command that wants supervised background work without running a server) that need background tasks without going through either of those paths.
.. code-block:: python
datasette = Datasette(memory=True)
await datasette.start_background_tasks()
.. note::
``start_background_tasks()`` calls ``invoke_startup()`` internally, **not** the fuller startup sequence a served instance uses — so calling it directly, without a prior request through ``datasette.client``, skips the immutable-database table-count precompute that a real server performs as part of startup. This only matters if your code inspects table counts before any request has been made; if you also exercise the instance via ``datasette.client`` (which arms the first-request fallback, and therefore the full startup sequence including table counts), or don't care about table counts up front, there is nothing to worry about.
.. _datasette_track_event: .. _datasette_track_event:
@ -1594,32 +1726,32 @@ datasette.client
Plugins can make internal simulated HTTP requests to the Datasette instance within which they are running. This ensures that all of Datasette's external JSON APIs are also available to plugins, while avoiding the overhead of making an external HTTP call to access those APIs. Plugins can make internal simulated HTTP requests to the Datasette instance within which they are running. This ensures that all of Datasette's external JSON APIs are also available to plugins, while avoiding the overhead of making an external HTTP call to access those APIs.
The ``datasette.client`` object is a wrapper around the `HTTPX Python library <https://www.python-httpx.org/>`__, providing an async-friendly API that is similar to the widely used `Requests library <https://requests.readthedocs.io/>`__. The ``datasette.client`` object is a wrapper around the `HTTPX2 Python library <https://httpx2.pydantic.dev/>`__, providing an async-friendly API that is similar to the widely used `Requests library <https://requests.readthedocs.io/>`__.
It offers the following methods: It offers the following methods:
``await datasette.client.get(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.get(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal GET request against that path. Execute an internal GET request against that path.
``await datasette.client.post(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.post(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal POST request. Use ``data={"name": "value"}`` to pass form parameters. Execute an internal POST request. Use ``data={"name": "value"}`` to pass form parameters.
``await datasette.client.options(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.options(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal OPTIONS request. Execute an internal OPTIONS request.
``await datasette.client.head(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.head(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal HEAD request. Execute an internal HEAD request.
``await datasette.client.put(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.put(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal PUT request. Execute an internal PUT request.
``await datasette.client.patch(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.patch(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal PATCH request. Execute an internal PATCH request.
``await datasette.client.delete(path, **kwargs)`` - returns HTTPX Response ``await datasette.client.delete(path, **kwargs)`` - returns HTTPX2 Response
Execute an internal DELETE request. Execute an internal DELETE request.
``await datasette.client.request(method, path, **kwargs)`` - returns HTTPX Response ``await datasette.client.request(method, path, **kwargs)`` - returns HTTPX2 Response
Execute an internal request with the given HTTP method against that path. Execute an internal request with the given HTTP method against that path.
These methods can be used with :ref:`internals_datasette_urls` - for example: These methods can be used with :ref:`internals_datasette_urls` - for example:
@ -1636,7 +1768,7 @@ These methods can be used with :ref:`internals_datasette_urls` - for example:
``datasette.client`` methods automatically take the current :ref:`setting_base_url` setting into account, whether or not you use the ``datasette.urls`` family of methods to construct the path. ``datasette.client`` methods automatically take the current :ref:`setting_base_url` setting into account, whether or not you use the ``datasette.urls`` family of methods to construct the path.
For documentation on available ``**kwargs`` options and the shape of the HTTPX Response object refer to the `HTTPX Async documentation <https://www.python-httpx.org/async/>`__. For documentation on available ``**kwargs`` options and the shape of the HTTPX2 Response object refer to the `HTTPX2 Async documentation <https://httpx2.pydantic.dev/async/>`__.
.. _internals_datasette_client_actor: .. _internals_datasette_client_actor:
@ -2023,8 +2155,8 @@ Example usage:
.. _database_execute_write: .. _database_execute_write:
await db.execute_write(sql, params=None, block=True, request=None, return_all=False, returning_limit=10, transaction=True) await db.execute_write(sql, params=None, block=True, request=None, return_all=False, returning_limit=10, transaction=True, time_limit_ms=2000)
-------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------
SQLite only allows one database connection to write at a time. Datasette handles this for you by maintaining a queue of writes to be executed against a given database. Plugins can submit write operations to this queue and they will be executed in the order in which they are received. SQLite only allows one database connection to write at a time. Datasette handles this for you by maintaining a queue of writes to be executed against a given database. Plugins can submit write operations to this queue and they will be executed in the order in which they are received.
@ -2063,6 +2195,13 @@ Each call to ``execute_write()`` will be executed inside a transaction. Pass
``transaction=False`` for statements such as ``VACUUM`` that cannot run inside ``transaction=False`` for statements such as ``VACUUM`` that cannot run inside
a transaction. a transaction.
Write statements have a default time limit of 2,000ms. Pass a different value
using ``time_limit_ms=`` or use ``time_limit_ms=None`` to allow the statement to
run without a time limit.
This write limit is independent of the ``sql_time_limit_ms`` setting used for
read queries. Changing that setting does not change the default write limit.
.. _database_execute_write_script: .. _database_execute_write_script:
await db.execute_write_script(sql, block=True) await db.execute_write_script(sql, block=True)
@ -2623,12 +2762,12 @@ This example uses trace to record the start, end and duration of any HTTP GET re
.. code-block:: python .. code-block:: python
from datasette.tracer import trace from datasette.tracer import trace
import httpx import httpx2
async def fetch_url(url): async def fetch_url(url):
with trace("fetch-url", url=url): with trace("fetch-url", url=url):
async with httpx.AsyncClient() as client: async with httpx2.AsyncClient() as client:
return await client.get(url) return await client.get(url)
.. _internals_tracer_trace_child_tasks: .. _internals_tracer_trace_child_tasks:

View file

@ -9,7 +9,7 @@ Each of these pages can be viewed in your browser. Add ``.json`` to the URL to g
JSON responses that return an object include an ``"ok": true`` key, consistent with the rest of the :ref:`JSON API <json_api>`. JSON responses that return an object include an ``"ok": true`` key, consistent with the rest of the :ref:`JSON API <json_api>`.
The introspection endpoints documented on this page are covered by the :ref:`JSON API stability promise <json_api_stability>`, with the exception of the debug endpoints ``/-/threads`` and ``/-/actions``, whose shapes may change in future releases. The introspection endpoints documented on this page are covered by the :ref:`JSON API stability promise <json_api_stability>`, with the exception of the debug endpoints ``/-/threads``, ``/-/tasks`` and ``/-/actions``, whose shapes may change in future releases.
.. _JsonDataView_metadata: .. _JsonDataView_metadata:
@ -278,6 +278,42 @@ Shows details of threads and ``asyncio`` tasks. This endpoint requires the ``per
] ]
} }
.. _JsonDataView_tasks:
/-/tasks
--------
Shows the state of every supervised background task registered with :ref:`datasette.add_background_task() <datasette_add_background_task>`; see also :ref:`BackgroundTask <BackgroundTask>` for what each field below means, and :ref:`datasette_lifecycle` for when tasks are launched. This endpoint requires the ``permissions-debug`` permission, since a crashed task's ``exception`` field can reveal internals such as file paths or query text:
.. code-block:: json
{
"ok": true,
"tasks": [
{
"name": "my_plugin.poll_for_updates",
"state": "running",
"function": "my_plugin.poll_for_updates",
"started_at": "2026-07-30T12:00:00+00:00",
"exception": null
},
{
"name": "my_plugin.broken_task",
"state": "crashed",
"function": "my_plugin.broken_task",
"started_at": "2026-07-30T12:00:00+00:00",
"exception": "ValueError('something went wrong')"
}
],
"launched": true
}
Each entry's ``function`` identifies the callable by its dotted module and qualified name.
Each entry's ``state`` is one of ``registered`` (added but not yet launched), ``running``, ``completed``, ``crashed`` or ``cancelled``. ``exception`` is a one-line ``repr()`` of the exception for a ``crashed`` task, or ``null`` otherwise - the full traceback is written to the ``datasette.background_tasks`` logger instead, to keep this payload skimmable.
The top-level ``launched`` flag reports whether the instance has run its one-time background task launch (after ``startup`` hooks finish, or via lifespan/first-request/:ref:`start_background_tasks() <datasette_start_background_tasks>`). It distinguishes "no tasks have been registered" (``tasks`` is empty either way) from "tasks are registered but nothing has armed the launch yet" (``launched`` is ``false`` and every task's ``state`` is still ``registered``) - useful when debugging a host that never triggers Datasette's lifespan events.
.. _JsonDataView_actor: .. _JsonDataView_actor:
/-/actor /-/actor

View file

@ -48,6 +48,7 @@ Some JSON endpoints are **exempt** from this promise:
debug playground. debug playground.
- Debug and support endpoints are documented so you can use them, but their - Debug and support endpoints are documented so you can use them, but their
JSON shapes are not frozen: :ref:`/-/threads <JsonDataView_threads>`, JSON shapes are not frozen: :ref:`/-/threads <JsonDataView_threads>`,
:ref:`/-/tasks <JsonDataView_tasks>`,
:ref:`/-/actions <JsonDataView_actions>`, :ref:`/-/actions <JsonDataView_actions>`,
the :ref:`permission debug endpoints <PermissionsDebugView>` the :ref:`permission debug endpoints <PermissionsDebugView>`
(``/-/allowed``, ``/-/rules``, ``/-/check``) and the (``/-/allowed``, ``/-/rules``, ``/-/check``) and the
@ -1326,6 +1327,23 @@ The following extras are available for arbitrary SQL query responses and stored,
.. [[[end]]] .. [[[end]]]
.. _TableCountView:
Counting all matching rows
--------------------------
``POST /<database>/<table>/-/count`` returns an exact count of the rows matching the table's query string filters::
POST /fixtures/facetable/-/count?state=CA
{"ok": true, "count": 10}
The endpoint supports the same column, search and plugin filters as the table page. Pagination and display options such as ``_next``, ``_size`` and ``_sort`` do not affect the count.
This requires ``view-table`` permission. ``execute-sql`` permission is only needed if using ``_where`` filters.
Unlike the ``count`` extra, this count is not capped by the row count limit. The usual SQL time limit still applies; a timed-out count returns a 400 JSON error.
.. _TableAutocompleteView: .. _TableAutocompleteView:
Table autocomplete Table autocomplete
@ -1661,6 +1679,8 @@ The request body is always parsed as JSON, regardless of the request's ``Content
The row-based write APIs can write :ref:`binary values in JSON <binary_json_format>` using Datasette's Base64 representation for BLOB data. The row-based write APIs can write :ref:`binary values in JSON <binary_json_format>` using Datasette's Base64 representation for BLOB data.
Structured inserts, upserts, updates and deletes only support ordinary SQLite tables. Virtual tables and their internal shadow tables are rejected, including when adding rows to an existing table through the create-table API. Writes to ordinary content tables can still update full-text search indexes through configured triggers.
.. _ExecuteWriteView: .. _ExecuteWriteView:
Executing write SQL Executing write SQL

View file

@ -1107,7 +1107,7 @@ Return an `ASGI <https://asgi.readthedocs.io/>`__ middleware wrapper function th
This is a very powerful hook. You can use it to manipulate the entire Datasette response, or even to configure new URL routes that will be handled by your own custom code. This is a very powerful hook. You can use it to manipulate the entire Datasette response, or even to configure new URL routes that will be handled by your own custom code.
You can write your ASGI code directly against the low-level specification, or you can use the middleware utilities provided by an ASGI framework such as `Starlette <https://www.starlette.io/middleware/>`__. You can write your ASGI code directly against the low-level specification, or you can use the middleware utilities provided by an ASGI framework such as `Starlette <https://starlette.dev/middleware/>`__.
This example plugin adds a ``x-databases`` HTTP header listing the currently attached databases: This example plugin adds a ``x-databases`` HTTP header listing the currently attached databases:
@ -1157,7 +1157,7 @@ Examples: `datasette-cors <https://datasette.io/plugins/datasette-cors>`__, `dat
startup(datasette) startup(datasette)
------------------ ------------------
This hook fires when the Datasette application server first starts up. This hook fires when the Datasette application server first starts up. It runs on the same event loop that goes on to serve requests, so it is safe to create loop-bound primitives and register background work here — see :ref:`datasette_lifecycle` for the full guarantee and the three ways startup can be triggered.
Here is an example that validates required plugin configuration. The server will fail to start and show an error if the validation check fails: Here is an example that validates required plugin configuration. The server will fail to start and show an error if the validation check fails:
@ -1195,6 +1195,7 @@ Potential use-cases:
* Create database tables that a plugin needs on startup * Create database tables that a plugin needs on startup
* Validate the configuration for a plugin on startup, and raise an error if it is invalid * Validate the configuration for a plugin on startup, and raise an error if it is invalid
* Raise a ``datasette.utils.StartupError("message")`` exception to prevent Datasette from starting and display that message to the user. * Raise a ``datasette.utils.StartupError("message")`` exception to prevent Datasette from starting and display that message to the user.
* Register supervised long-lived background work using :ref:`datasette_add_background_task`, which core launches once every plugin's ``startup()`` hook has finished.
.. note:: .. note::
@ -1211,6 +1212,31 @@ Potential use-cases:
Examples: `datasette-saved-queries <https://datasette.io/plugins/datasette-saved-queries>`__, `datasette-init <https://datasette.io/plugins/datasette-init>`__ Examples: `datasette-saved-queries <https://datasette.io/plugins/datasette-saved-queries>`__, `datasette-init <https://datasette.io/plugins/datasette-init>`__
.. _plugin_hook_shutdown:
shutdown(datasette)
-------------------
This hook fires once, when the Datasette application server is shutting down gracefully - triggered by the ASGI ``lifespan.shutdown`` event, which includes pressing Ctrl-C or sending ``SIGTERM`` to a ``datasette serve`` process. It is not called on a hard kill (``SIGKILL``), since there is no opportunity to run any code in that case.
Like ``startup()``, this can be a regular function or it can return an async function to be awaited.
It runs before Datasette cancels any background tasks it is supervising (see :ref:`datasette_add_background_task`) and before it closes its database connections, so you can use it to tell your plugin's own background work to stop gracefully while a database connection is still available to write out any final state. See :ref:`datasette_lifecycle` for exactly where this fits into the full startup-to-shutdown sequence:
.. code-block:: python
@hookimpl
def shutdown(datasette):
async def inner():
db = datasette.get_database()
await db.execute_write(
"insert into shutdown_log (at) values (datetime('now'))"
)
return inner
If your ``shutdown()`` hook raises an exception it will be logged but not re-raised, so one plugin's broken shutdown code cannot prevent other plugins - or Datasette itself - from finishing their own teardown.
.. _plugin_hook_actor_from_request: .. _plugin_hook_actor_from_request:
actor_from_request(datasette, request) actor_from_request(datasette, request)

View file

@ -261,6 +261,15 @@ If you run ``datasette plugins --all`` it will include default plugins that ship
"permission_resources_sql" "permission_resources_sql"
] ]
}, },
{
"name": "datasette.default_permissions.sqlite_statistics",
"static": false,
"templates": false,
"version": null,
"hooks": [
"permission_resources_sql"
]
},
{ {
"name": "datasette.default_permissions.tokens", "name": "datasette.default_permissions.tokens",
"static": false, "static": false,

View file

@ -71,6 +71,8 @@ Should users be able to execute arbitrary SQL queries by default?
Setting this to ``off`` causes permission checks for :ref:`actions_execute_sql` to fail by default. Setting this to ``off`` causes permission checks for :ref:`actions_execute_sql` to fail by default.
This setting controls the ability to submit arbitrary SQL. It does not disable structured table-browsing features that use SQL generated by Datasette, such as sorting, column filters and :ref:`facets`. Use :ref:`setting_allow_facet` to control whether users can request facets.
:: ::
datasette mydatabase.db --setting default_allow_sql off datasette mydatabase.db --setting default_allow_sql off
@ -254,6 +256,8 @@ Default HTTP caching max-age header in seconds, used for ``Cache-Control: max-ag
datasette mydatabase.db --setting default_cache_ttl 60 datasette mydatabase.db --setting default_cache_ttl 60
Dynamic responses for authenticated actors, requests with cookies or an ``Authorization`` header, and responses that set cookies use ``Cache-Control: private, no-store``. This takes precedence over ``default_cache_ttl`` and ``?_ttl=``, even when cache headers are otherwise disabled. Anonymous dynamic responses vary by ``Cookie`` and ``Authorization``. Static assets retain their own cache policy.
.. _setting_cache_size_kb: .. _setting_cache_size_kb:
cache_size_kb cache_size_kb

View file

@ -25,7 +25,7 @@ If you use the template described in :ref:`writing_plugins_cookiecutter` your pl
) )
This test uses the :ref:`internals_datasette_client` object to exercise a test instance of Datasette. ``datasette.client`` is a wrapper around the `HTTPX <https://www.python-httpx.org/>`__ Python library which can imitate HTTP requests using ASGI. This is the recommended way to write tests against a Datasette instance. This test uses the :ref:`internals_datasette_client` object to exercise a test instance of Datasette. ``datasette.client`` is a wrapper around the `HTTPX2 <https://httpx2.pydantic.dev/>`__ Python library which can imitate HTTP requests using ASGI. This is the recommended way to write tests against a Datasette instance.
This test also uses the `pytest-asyncio <https://pypi.org/project/pytest-asyncio/>`__ package to add support for ``async def`` test functions running under pytest. This test also uses the `pytest-asyncio <https://pypi.org/project/pytest-asyncio/>`__ package to add support for ``async def`` test functions running under pytest.
@ -78,9 +78,19 @@ Creating a ``Datasette()`` instance like this as useful shortcut in tests, but t
datasette = Datasette(memory=True) datasette = Datasette(memory=True)
await datasette.invoke_startup() await datasette.invoke_startup()
This method registers any :ref:`plugin_hook_startup` or :ref:`plugin_hook_prepare_jinja2_environment` plugins that might themselves need to make async calls. This method registers any :ref:`plugin_hook_startup` or :ref:`plugin_hook_prepare_jinja2_environment` plugins that might themselves need to make async calls. It runs on the same event loop that runs your test, matching the guarantee described in :ref:`datasette_lifecycle`.
If you are using ``await datasette.client.get()`` and similar methods then you don't need to worry about this - Datasette automatically calls ``invoke_startup()`` the first time it handles a request. If you are using ``await datasette.client.get()`` and similar methods then you don't need to worry about this - Datasette automatically calls ``invoke_startup()`` the first time it handles a request, via the first-request fallback described in :ref:`datasette_lifecycle`.
If your plugin also registers work with :ref:`datasette_add_background_task` (typically from a ``startup`` hook) and your test needs that work to actually run, call ``await datasette.start_background_tasks()`` as well - ``invoke_startup()`` alone only runs ``startup`` hooks, it does not launch anything they registered:
.. code-block:: python
datasette = Datasette(memory=True)
await datasette.start_background_tasks()
# Any tasks registered by a startup() hook are now running
A request made through ``datasette.client`` arms both startup and background-task launch automatically, since they're both part of the same first-request fallback - ``start_background_tasks()`` is for tests that need tasks running without making an HTTP request first.
.. _testing_plugins_datasette_fixtures_database: .. _testing_plugins_datasette_fixtures_database:
@ -154,7 +164,7 @@ If you need to opt out of this behavior, add the following to your ``pytest.ini`
Using datasette.client in tests Using datasette.client in tests
------------------------------- -------------------------------
The :ref:`internals_datasette_client` mechanism is designed for use in tests. It provides access to a pre-configured `HTTPX async client <https://www.python-httpx.org/async/>`__ instance that can make GET, POST and other HTTP requests against a Datasette instance from inside a test. The :ref:`internals_datasette_client` mechanism is designed for use in tests. It provides access to a pre-configured `HTTPX2 async client <https://httpx2.pydantic.dev/async/>`__ instance that can make GET, POST and other HTTP requests against a Datasette instance from inside a test.
A simple test looks like this: A simple test looks like this:
@ -273,22 +283,22 @@ If you want to create that test database repeatedly for every individual test fu
.. _testing_plugins_pytest_httpx: .. _testing_plugins_pytest_httpx:
Testing outbound HTTP calls with pytest-httpx Testing outbound HTTP calls with pytest-httpx2
--------------------------------------------- ----------------------------------------------
If your plugin makes outbound HTTP calls - for example datasette-auth-github or datasette-import-table - you may need to mock those HTTP requests in your tests. If your plugin makes outbound HTTP calls - for example datasette-auth-github or datasette-import-table - you may need to mock those HTTP requests in your tests.
The `pytest-httpx <https://pypi.org/project/pytest-httpx/>`__ package is a useful library for mocking calls. It can be tricky to use with Datasette though since it mocks all HTTPX requests, and Datasette's own testing mechanism uses HTTPX internally. The `pytest-httpx2 <https://pypi.org/project/pytest-httpx2/>`__ package provides a ``httpx2_mock`` fixture, built on `respx <https://lundberg.github.io/respx/>`__, for mocking outbound calls made using HTTPX2.
To avoid breaking your tests, you can return ``["localhost"]`` from the ``non_mocked_hosts()`` fixture. Datasette's own ``datasette.client`` mechanism uses HTTPX2 internally too, but those requests are passed directly to the ASGI application rather than being sent over the network, so they are not affected by the mock.
As an example, here's a very simple plugin which executes an HTTP response and returns the resulting content: As an example, here's a very simple plugin which executes an HTTP request and returns the resulting content:
.. code-block:: python .. code-block:: python
from datasette import hookimpl from datasette import hookimpl
from datasette.utils.asgi import Response from datasette.utils.asgi import Response
import httpx import httpx2
@hookimpl @hookimpl
@ -306,27 +316,18 @@ As an example, here's a very simple plugin which executes an HTTP response and r
</form>""") </form>""")
vars = await request.post_vars() vars = await request.post_vars()
url = vars["url"] url = vars["url"]
return Response.text(httpx.get(url).text) return Response.text(httpx2.get(url).text)
Here's a test for that plugin that mocks the HTTPX outbound request: Here's a test for that plugin that mocks the HTTPX2 outbound request:
.. code-block:: python .. code-block:: python
from datasette.app import Datasette from datasette.app import Datasette
import pytest
@pytest.fixture async def test_outbound_http_call(httpx2_mock):
def non_mocked_hosts(): httpx2_mock.get("https://www.example.com/").respond(
# This ensures httpx-mock will not affect Datasette's own text="Hello world"
# httpx calls made in the tests by datasette.client:
return ["localhost"]
async def test_outbound_http_call(httpx_mock):
httpx_mock.add_response(
url="https://www.example.com/",
text="Hello world",
) )
datasette = Datasette([], memory=True) datasette = Datasette([], memory=True)
response = await datasette.client.post( response = await datasette.client.post(
@ -335,11 +336,13 @@ Here's a test for that plugin that mocks the HTTPX outbound request:
) )
assert response.text == "Hello world" assert response.text == "Hello world"
outbound_request = httpx_mock.get_request() outbound_request = httpx2_mock.calls.last.request
assert ( assert (
outbound_request.url == "https://www.example.com/" outbound_request.url == "https://www.example.com/"
) )
If your plugin still makes its outbound calls using the original ``httpx`` library you can continue to mock those using `pytest-httpx <https://pypi.org/project/pytest-httpx/>`__.
.. _testing_plugins_register_in_test: .. _testing_plugins_register_in_test:
Registering a plugin for the duration of a test Registering a plugin for the duration of a test

View file

@ -28,9 +28,9 @@ dependencies = [
"click-default-group>=1.2.3", "click-default-group>=1.2.3",
"Jinja2>=2.10.3", "Jinja2>=2.10.3",
"hupper>=1.9", "hupper>=1.9",
"httpx>=0.20,<1.0", "httpx2>=2.0",
"pluggy>=1.0", "pluggy>=1.0",
"uvicorn>=0.11", "uvicorn>=0.29",
"aiofiles>=0.4", "aiofiles>=0.4",
"PyYAML>=5.3", "PyYAML>=5.3",
"mergedeep>=1.1.1", "mergedeep>=1.1.1",

View file

@ -2,13 +2,14 @@ import importlib.metadata
import os import os
import pathlib import pathlib
import re import re
import socket
import subprocess import subprocess
import sys import sys
import tempfile import tempfile
import time import time
from dataclasses import dataclass from dataclasses import dataclass
import httpx import httpx2
import pytest import pytest
import pytest_asyncio import pytest_asyncio
@ -32,17 +33,31 @@ UNDOCUMENTED_PERMISSIONS = {
} }
def wait_until_responds(url, timeout=5.0, client=httpx, **kwargs): def wait_until_responds(url, timeout=5.0, client=httpx2, process=None, **kwargs):
start = time.time() start = time.time()
while time.time() - start < timeout: while time.time() - start < timeout:
# If the server died there is no point waiting out the timeout - fail
# now, with its output, instead of after `timeout` seconds of silence
if process is not None and process.poll() is not None:
raise AssertionError(
"Server exited early with returncode {}\n{}".format(
process.returncode, process.stdout.read().decode("utf-8")
)
)
try: try:
client.get(url, **kwargs) client.get(url, **kwargs)
return return
except httpx.ConnectError: except httpx2.TransportError:
time.sleep(0.1) time.sleep(0.1)
raise AssertionError(f"Timed out waiting for {url} to respond") raise AssertionError(f"Timed out waiting for {url} to respond")
def find_free_port():
with socket.socket() as sock:
sock.bind(("127.0.0.1", 0))
return sock.getsockname()[1]
@pytest.fixture @pytest.fixture
def bare_ds(): def bare_ds():
""" """
@ -277,8 +292,8 @@ def ds_unix_domain_socket_server(tmp_path_factory):
cwd=tempfile.gettempdir(), cwd=tempfile.gettempdir(),
) )
# Poll until available # Poll until available
transport = httpx.HTTPTransport(uds=uds) transport = httpx2.HTTPTransport(uds=uds)
client = httpx.Client(transport=transport) client = httpx2.Client(transport=transport)
try: try:
wait_until_responds( wait_until_responds(
"http://localhost/_memory.json", timeout=30.0, client=client "http://localhost/_memory.json", timeout=30.0, client=client
@ -301,6 +316,71 @@ def ds_unix_domain_socket_server(tmp_path_factory):
pass pass
@pytest.fixture
def serve_with_plugins(tmp_path):
"""Factory fixture for starting ``datasette serve`` in a subprocess with
plugins written to a temporary ``--plugins-dir``.
For tests that need the real serve path: event-loop wiring, exit codes,
signals. The usual in-process ``pm.register`` plugin pattern can't reach
a subprocess, so plugin source is written out as importable files instead.
Unlike ``ds_localhost_http_server`` this is function-scoped and takes a
fresh port each time, because each test needs its own plugins. Call it as::
proc, port = serve_with_plugins({"my_plugin": PLUGIN_SOURCE})
``plugins`` maps module name to Python source. Pass
``wait_for_startup=False`` when the server is expected to fail during
startup rather than begin serving. Extra CLI arguments are passed through.
Every process started is terminated when the test ends.
"""
processes = []
def start(plugins, *extra_args, wait_for_startup=True):
plugins_dir = tmp_path / "plugins"
plugins_dir.mkdir(exist_ok=True)
for module_name, source in plugins.items():
(plugins_dir / f"{module_name}.py").write_text(source, "utf-8")
port = find_free_port()
proc = subprocess.Popen(
[
sys.executable,
"-m",
"datasette",
"--memory",
"--plugins-dir",
str(plugins_dir),
"-h",
"127.0.0.1",
"-p",
str(port),
*extra_args,
],
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
# Avoid FileNotFoundError: [Errno 2] No such file or directory:
cwd=tempfile.gettempdir(),
)
processes.append(proc)
if wait_for_startup:
wait_until_responds(
f"http://127.0.0.1:{port}/-/versions.json", process=proc
)
return proc, port
yield start
for proc in processes:
if proc.poll() is None:
proc.terminate()
try:
proc.wait(timeout=5)
except subprocess.TimeoutExpired:
proc.kill()
proc.wait()
# Import fixtures from fixtures.py to make them available # Import fixtures from fixtures.py to make them available
from .fixtures import ( # noqa: F401 from .fixtures import ( # noqa: F401
TEMP_PLUGIN_SECRET_FILE, TEMP_PLUGIN_SECRET_FILE,

View file

@ -5,6 +5,7 @@ import pytest
from datasette.app import Datasette from datasette.app import Datasette
from datasette.plugins import DEFAULT_PLUGINS from datasette.plugins import DEFAULT_PLUGINS
from datasette.resources import DatabaseResource, TableResource
from datasette.utils import UNSTABLE_API_MESSAGE, escape_sqlite, tilde_encode from datasette.utils import UNSTABLE_API_MESSAGE, escape_sqlite, tilde_encode
from datasette.utils.sqlite import sqlite_version from datasette.utils.sqlite import sqlite_version
from datasette.version import __version__ from datasette.version import __version__
@ -101,14 +102,11 @@ async def test_database_page(ds_client):
"tags", "tags",
} }
# Expected hidden tables # The external-content index is visible, but its shadow tables need a
# second dependency hop and are excluded by the one-hop permission policy.
expected_hidden_tables = { expected_hidden_tables = {
"no_primary_key", "no_primary_key",
"searchable_fts", "searchable_fts",
"searchable_fts_config",
"searchable_fts_data",
"searchable_fts_docsize",
"searchable_fts_idx",
} }
# Verify all expected tables exist # Verify all expected tables exist
@ -458,6 +456,67 @@ async def test_row_foreign_key_tables(ds_client):
] ]
@pytest.mark.asyncio
async def test_row_foreign_key_tables_omit_denied_tables(request):
actor = {"id": "reader"}
ds = Datasette(
memory=True,
default_deny=True,
config={
"databases": {
"data": {
"tables": {
"parents": {"permissions": {"view-table": True}},
"private_children": {"permissions": {"view-table": False}},
}
}
}
},
)
request.addfinalizer(ds.close)
db = ds.add_memory_database("fk_count_leak", name="data")
await db.execute_write("create table parents (id integer primary key, name text)")
await db.execute_write("""
create table private_children (
id integer primary key,
parent_id integer references parents(id)
)
""")
await db.execute_write("insert into parents values (1, 'Public parent')")
await db.execute_write("""
insert into private_children (id, parent_id) values
(1, 1),
(2, 1),
(3, 1)
""")
await ds.invoke_startup()
parent = TableResource(database="data", table="parents")
private_children = TableResource(database="data", table="private_children")
assert await ds.allowed(action="view-table", resource=parent, actor=actor)
assert not await ds.allowed(
action="view-table", resource=private_children, actor=actor
)
assert not await ds.allowed(
action="execute-sql",
resource=DatabaseResource(database="data"),
actor=actor,
)
direct_child = await ds.client.get("/data/private_children.json", actor=actor)
assert direct_child.status_code == 403
parent_response = await ds.client.get(
"/data/parents/1.json?_extra=foreign_key_tables", actor=actor
)
assert parent_response.status_code == 200
foreign_key_tables = parent_response.json().get("foreign_key_tables", [])
assert foreign_key_tables == [], (
"denied child table name, foreign-key column, and row count disclosed: "
f"{foreign_key_tables}"
)
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_row_extras(ds_client): async def test_row_extras(ds_client):
response = await ds_client.get( response = await ds_client.get(
@ -894,10 +953,7 @@ async def test_hidden_sqlite_stat1_table():
await db.execute_write("analyze") await db.execute_write("analyze")
data = (await ds.client.get("/db.json?_show_hidden=1")).json() data = (await ds.client.get("/db.json?_show_hidden=1")).json()
tables = [(t["name"], t["hidden"]) for t in data["tables"]] tables = [(t["name"], t["hidden"]) for t in data["tables"]]
assert tables in ( assert tables == [("normal", False)]
[("normal", False), ("sqlite_stat1", True)],
[("normal", False), ("sqlite_stat1", True), ("sqlite_stat4", True)],
)
@pytest.mark.asyncio @pytest.mark.asyncio

View file

@ -1,6 +1,7 @@
import time import time
import pytest import pytest
import sqlite_utils
from datasette.app import Datasette from datasette.app import Datasette
from datasette.events import RenameTableEvent from datasette.events import RenameTableEvent
@ -67,6 +68,82 @@ BASE64_WRITE_API_VALUE = {"$base64": True, "encoded": "AAEC/f7/"}
BASE64_WRITE_API_LITERAL = '{"$base64": true, "encoded": "AAEC/f7/"}' BASE64_WRITE_API_LITERAL = '{"$base64": true, "encoded": "AAEC/f7/"}'
@pytest.mark.asyncio
@pytest.mark.parametrize("use_fallback", (False, True))
@pytest.mark.parametrize(
"operation", ("insert", "upsert", "update", "delete", "create", "create_uppercase")
)
@pytest.mark.parametrize(
"module,definition,values,shadow_suffix",
(
("fts5", "body", "'original'", "_content"),
("fts4", "body", "'original'", "_content"),
("rtree", "id, minx, maxx", "1, 0, 1", "_rowid"),
),
)
@pytest.mark.parametrize("shadow", (False, True))
async def test_structured_writes_require_ordinary_tables(
ds_write,
monkeypatch,
use_fallback,
operation,
module,
definition,
values,
shadow_suffix,
shadow,
):
if use_fallback:
monkeypatch.setattr("datasette.utils.sqlite.supports_table_list", lambda: False)
db = ds_write.get_database("data")
await db.execute_write(f"create virtual table indexed using {module}({definition})")
await db.execute_write(f"insert into indexed values ({values})")
table = "indexed" + (shadow_suffix if shadow else "")
row = (await db.execute(f"select rowid, * from {escape_sqlite(table)}")).dicts()[0]
pks = await db.primary_keys(table)
pk_value = row[pks[0] if pks else "rowid"]
before = await db.execute_fn(lambda conn: list(conn.iterdump()))
if operation in ("create", "create_uppercase"):
path = "/data/-/create"
body = {
"table": table.upper() if operation == "create_uppercase" else table,
"rows": [row],
}
elif operation in ("update", "delete"):
path = f"/data/{table}/{pk_value}/-/{operation}"
body = {"update": row} if operation == "update" else {}
else:
path = f"/data/{table}/-/{operation}"
body = {"rows": [row]}
response = await ds_write.client.post(
path, json=body, headers=_headers(write_token(ds_write))
)
assert response.status_code == 400, response.text
assert response.json()["errors"] == ["Structured writes require an ordinary table"]
assert await db.execute_fn(lambda conn: list(conn.iterdump())) == before
@pytest.mark.asyncio
async def test_structured_writes_to_content_table_maintain_fts(ds_write):
db = ds_write.get_database("data")
await db.execute_write_fn(
lambda conn: sqlite_utils.Database(conn)["docs"].enable_fts(
["title"], create_triggers=True
)
)
response = await ds_write.client.post(
"/data/docs/-/insert",
json={"row": {"id": 1, "title": "ordinary content"}},
headers=_headers(write_token(ds_write)),
)
assert response.status_code == 201, response.text
matches = await db.execute(
"select rowid from docs_fts where docs_fts match ?", ["ordinary"]
)
assert [row[0] for row in matches.rows] == [1]
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_base64_write_api_create_table_infers_blob_and_raw_escapes(ds_write): async def test_base64_write_api_create_table_infers_blob_and_raw_escapes(ds_write):
token = write_token(ds_write) token = write_token(ds_write)
@ -1295,7 +1372,7 @@ async def test_alter_table_foreign_key_without_fk_column_requires_single_pk(ds_w
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_foreign_key_suggestions(ds_write): async def test_foreign_key_suggestions(ds_write):
token = write_token(ds_write, permissions=["at"]) token = write_token(ds_write, permissions=["alter-table", "view-table"])
db = ds_write.get_database("data") db = ds_write.get_database("data")
await db.execute_write("create table owners (id integer primary key)") await db.execute_write("create table owners (id integer primary key)")
await db.execute_write("insert into owners (id) values (1), (2), (3)") await db.execute_write("insert into owners (id) values (1), (2), (3)")
@ -1361,7 +1438,7 @@ async def test_foreign_key_suggestions_permission_denied(ds_write):
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_foreign_key_suggestions_fail_open(ds_write, monkeypatch): async def test_foreign_key_suggestions_fail_open(ds_write, monkeypatch):
token = write_token(ds_write, permissions=["at"]) token = write_token(ds_write, permissions=["alter-table", "view-table"])
db = ds_write.get_database("data") db = ds_write.get_database("data")
await db.execute_write("create table owners (id integer primary key)") await db.execute_write("create table owners (id integer primary key)")
@ -1392,7 +1469,7 @@ async def test_foreign_key_suggestions_fail_open(ds_write, monkeypatch):
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_foreign_key_targets(ds_write): async def test_foreign_key_targets(ds_write):
token = write_token(ds_write, permissions=["ct"]) token = write_token(ds_write, permissions=["create-table", "view-table"])
db = ds_write.get_database("data") db = ds_write.get_database("data")
await db.execute_write("create table owners (id integer primary key)") await db.execute_write("create table owners (id integer primary key)")
await db.execute_write("create table categories (slug varchar(30) primary key)") await db.execute_write("create table categories (slug varchar(30) primary key)")
@ -1725,6 +1802,42 @@ async def test_drop_table(ds_write, scenario):
assert (await ds_write.client.get("/data/docs")).status_code == 404 assert (await ds_write.client.get("/data/docs")).status_code == 404
@pytest.mark.asyncio
async def test_drop_table_cleans_up_fts(ds_write):
db = ds_write.get_database("data")
def enable_fts(conn):
sqlite_utils.Database(conn)["docs"].enable_fts(["title"], create_triggers=True)
await db.execute_write_fn(enable_fts)
assert {
row[0]
for row in await db.execute(
"select name from sqlite_master where type = 'table' and name like 'docs_fts%'"
)
} == {
"docs_fts",
"docs_fts_config",
"docs_fts_data",
"docs_fts_docsize",
"docs_fts_idx",
}
response = await ds_write.client.post(
"/data/docs/-/drop",
json={"confirm": True},
headers=_headers(write_token(ds_write)),
)
assert response.json() == {"ok": True}
assert [
row[0]
for row in await db.execute(
"select name from sqlite_master where type = 'table' and name like 'docs_fts%'"
)
] == []
@pytest.mark.asyncio @pytest.mark.asyncio
@pytest.mark.parametrize( @pytest.mark.parametrize(
"input,expected_status,expected_response,expected_events", "input,expected_status,expected_response,expected_events",
@ -2708,3 +2821,119 @@ async def test_create_using_alter_against_existing_table(
insert_rows_event = ds_write._tracked_events[1] insert_rows_event = ds_write._tracked_events[1]
assert insert_rows_event.name == "insert-rows" assert insert_rows_event.name == "insert-rows"
assert insert_rows_event.num_rows == 1 assert insert_rows_event.num_rows == 1
@pytest.mark.asyncio
@pytest.mark.parametrize(
("denied_action", "request_body"),
(
(
"insert-row",
{
"table": "salaries",
"rows": [{"id": 9, "note": "INJ-VIA-CREATE"}],
},
),
(
"update-row",
{
"table": "salaries",
"rows": [{"id": 1, "note": "REPLACED"}],
"pk": "id",
"replace": True,
},
),
(
"alter-table",
{
"table": "salaries",
"rows": [{"id": 9, "note": "INSERTED", "extra": "NEW"}],
"alter": True,
},
),
),
)
async def test_create_table_existing_table_respects_table_level_denial(
denied_action, request_body
):
# GHSA-53fc-rhfg-h7qp issue 2: POST /db/-/create against an existing table
# inserts rows into it, so insert-row (and update-row / alter-table) must be
# checked against the TableResource, not just the DatabaseResource.
ds = Datasette(
memory=True,
config={
"databases": {
# id=editor user has each permission at the database level, but
# the selected action is explicitly denied on the salaries table
"data": {
"permissions": {
"create-table": {"id": "editor"},
"insert-row": {"id": "editor"},
"update-row": {"id": "editor"},
"alter-table": {"id": "editor"},
},
"tables": {
"salaries": {"permissions": {denied_action: False}},
},
}
}
},
)
db = ds.add_memory_database(
f"create_table_existing_table_denied_{denied_action}", name="data"
)
await db.execute_write("create table salaries (id integer primary key, note text)")
await db.execute_write("insert into salaries values (1, 'TOPSECRET-A')")
await ds.invoke_startup()
if denied_action == "insert-row":
# Sanity: direct insert into salaries is denied for this actor
direct = await ds.client.post(
"/data/salaries/-/insert",
actor={"id": "editor"},
json={"row": {"id": 9, "note": "INJ-DIRECT"}},
)
assert direct.status_code == 403
response = await ds.client.post(
"/data/-/create",
actor={"id": "editor"},
json=request_body,
)
assert response.status_code == 403, response.json()
assert response.json()["errors"] == [f"Permission denied: need {denied_action}"]
rows = (await db.execute("select id, note from salaries order by id")).rows
assert [tuple(r) for r in rows] == [(1, "TOPSECRET-A")]
assert await db.table_columns("salaries") == ["id", "note"]
@pytest.mark.asyncio
async def test_create_table_respects_predeclared_table_level_denial():
ds = Datasette(
memory=True,
config={
"databases": {
"data": {
"permissions": {
"create-table": {"id": "editor"},
"insert-row": {"id": "editor"},
},
"tables": {
"planned_table": {"permissions": {"insert-row": False}},
},
}
}
},
)
db = ds.add_memory_database("create_table_predeclared_denial", name="data")
await ds.invoke_startup()
response = await ds.client.post(
"/data/-/create",
actor={"id": "editor"},
json={"table": "planned_table", "rows": [{"id": 1}]},
)
assert response.status_code == 403, response.json()
assert response.json()["errors"] == ["Permission denied: need insert-row"]
assert not await db.table_exists("planned_table")

View file

@ -1,4 +1,5 @@
import time import time
from unittest.mock import AsyncMock
import pytest import pytest
from bs4 import BeautifulSoup as Soup from bs4 import BeautifulSoup as Soup
@ -237,6 +238,35 @@ def test_auth_create_token(
assert response3.json["actor"]["id"] == "test" assert response3.json["actor"]["id"] == "test"
@pytest.mark.asyncio
@pytest.mark.parametrize("method", ["GET", "POST"])
@pytest.mark.parametrize(
"restrictions",
[
{},
{"a": ["vi"]},
{"d": {"db": ["vd"]}},
{"r": {"db": {"t1": ["vt"]}}},
],
ids=["empty", "instance", "database", "table"],
)
async def test_auth_create_token_not_allowed_for_restricted_actors(
bare_ds, monkeypatch, method, restrictions
):
create_token = AsyncMock()
monkeypatch.setattr(bare_ds, "create_token", create_token)
response = await bare_ds.client.request(
method,
"/-/create-token",
actor={"id": "test", "_r": restrictions},
)
assert response.status_code == 403
assert "Restricted actors cannot create API tokens" in response.text
create_token.assert_not_called()
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_auth_create_token_not_allowed_for_tokens(ds_client): async def test_auth_create_token_not_allowed_for_tokens(ds_client):
ds_tok = ds_client.ds.sign( ds_tok = ds_client.ds.sign(
@ -524,3 +554,25 @@ async def test_root_without_root_enabled_no_special_permissions(ds_client):
) )
is not True is not True
), "Root without root_enabled should not automatically get set-column-type" ), "Root without root_enabled should not automatically get set-column-type"
@pytest.mark.parametrize("expire_after", (1, 300, 3600, 30 * 24 * 60 * 60))
def test_set_actor_cookie_honours_expire_after(expire_after):
# GHSA-53fc-rhfg-h7qp issue 4: expire_after is documented as a number of
# seconds, but every value was being replaced with 24 hours.
from datasette.app import Datasette
from datasette.utils.asgi import Response
ds = Datasette(memory=True)
response = Response.text("")
before = int(time.time())
ds.set_actor_cookie(response, {"id": "test"}, expire_after=expire_after)
after = int(time.time())
(header,) = response._set_cookie_headers
assert header.startswith("ds_actor=")
value = header[len("ds_actor=") :].split(";", 1)[0]
data = ds.unsign(value, "actor")
assert data["a"] == {"id": "test"}
expires_at = baseconv.base62.decode(data["e"])
assert before + expire_after <= expires_at <= after + expire_after

View file

@ -0,0 +1,381 @@
"""
Tests for datasette.add_background_task() / start_background_tasks() and the
BackgroundTask / BackgroundTaskSupervisor machinery in
datasette/background_tasks.py.
"""
import asyncio
import contextlib
import logging
import httpx2
import pytest
from datasette import hookimpl
from datasette.app import Datasette
from datasette.plugins import pm
async def _drive_lifespan_startup(app):
"""Send a single lifespan.startup message into app's ASGI lifespan loop
and return the list of messages sent back, without ever sending
lifespan.shutdown. Copied from tests/test_lifespan.py's helper of the
same name - mirrors what a real server does: after startup completes
it parks waiting for the next event, and we cancel that wait once
we've observed the startup response.
"""
messages_sent = []
startup_responded = asyncio.Event()
delivered = False
async def receive():
nonlocal delivered
if not delivered:
delivered = True
return {"type": "lifespan.startup"}
await asyncio.Event().wait()
async def send(message):
messages_sent.append(message)
startup_responded.set()
task = asyncio.create_task(app({"type": "lifespan"}, receive, send))
try:
await asyncio.wait_for(startup_responded.wait(), timeout=5)
finally:
task.cancel()
with contextlib.suppress(asyncio.CancelledError):
await task
return messages_sent
@pytest.mark.asyncio
async def test_tasks_registered_in_startup_hook_run_after_lifespan_startup():
# Two tasks registered by one plugin's startup hook - order preserved,
# both running after lifespan startup completes, and no HTTP request
# of any kind is issued anywhere in this test.
events = []
async def task_one(datasette):
events.append("task_one")
# Wait indefinitely to simulate long-lived background work, keeping
# the task "running" for the assertions below until cleanup cancels it.
await asyncio.Event().wait()
async def task_two(datasette):
events.append("task_two")
await asyncio.Event().wait()
class TwoTaskPlugin:
__name__ = "TwoTaskPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
datasette.add_background_task(task_one, name="task-one")
datasette.add_background_task(task_two, name="task-two")
return inner
ds = Datasette(memory=True)
pm.register(TwoTaskPlugin(), name="two_task_plugin")
try:
app = ds.app()
messages = await _drive_lifespan_startup(app)
assert {"type": "lifespan.startup.complete"} in messages
handles = ds._background_tasks.tasks()
assert [h.name for h in handles] == ["task-one", "task-two"]
# Let both tasks run their first line of code.
await asyncio.sleep(0)
assert handles[0].state == "running"
assert handles[1].state == "running"
assert events == ["task_one", "task_two"]
finally:
pm.unregister(name="two_task_plugin")
await ds._background_tasks.cancel_all(grace=1.0)
@pytest.mark.asyncio
async def test_launch_waits_for_every_startup_hook_before_running_any_task():
# PluginA registers a task from its startup hook; PluginB does the
# same from ITS startup hook, which runs after PluginA's (forced with
# tryfirst=True on A). Even though A's registration happens first,
# A's task body must not actually execute until every startup hook -
# including B's - has finished, since launch only happens after
# invoke_startup() completes. This is the ordering guarantee that
# dissolves datasette-cron's tryfirst=True launch hack.
hook_call_order = []
seen_names_when_a_ran = {}
async def task_a(datasette):
seen_names_when_a_ran["names"] = [
h.name for h in datasette._background_tasks.tasks()
]
async def task_b(datasette):
pass
class PluginA:
__name__ = "PluginA"
@hookimpl(tryfirst=True)
def startup(self, datasette):
async def inner():
hook_call_order.append("A")
datasette.add_background_task(task_a, name="task-a")
return inner
class PluginB:
__name__ = "PluginB"
@hookimpl
def startup(self, datasette):
async def inner():
hook_call_order.append("B")
datasette.add_background_task(task_b, name="task-b")
return inner
ds = Datasette(memory=True)
pm.register(PluginA(), name="plugin_a")
pm.register(PluginB(), name="plugin_b")
try:
await ds.start_background_tasks()
# Confirm A's startup hook really did run (and register task-a)
# strictly before B's startup hook ran.
assert hook_call_order == ["A", "B"]
handles = ds._background_tasks.tasks()
await asyncio.wait_for(asyncio.gather(*[h.task for h in handles]), timeout=5)
# Yet by the time task-a's own body executed (after launch, which
# only happens once every startup hook - including B's - has
# finished), task-b was already registered.
assert "task-b" in seen_names_when_a_ran["names"]
finally:
pm.unregister(name="plugin_a")
pm.unregister(name="plugin_b")
@pytest.mark.asyncio
async def test_concurrent_first_requests_launch_background_tasks_exactly_once():
launch_count = {"n": 0}
async def counting_task(datasette):
launch_count["n"] += 1
class CountingTaskPlugin:
__name__ = "CountingTaskPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
datasette.add_background_task(counting_task, name="counting-task")
return inner
ds = Datasette(memory=True)
pm.register(CountingTaskPlugin(), name="counting_task_plugin")
try:
app = ds.app()
transport = httpx2.ASGITransport(app=app)
async with httpx2.AsyncClient(
transport=transport, base_url="http://localhost"
) as client:
responses = await asyncio.gather(
*[client.get("/-/versions.json") for _ in range(10)]
)
assert all(response.status_code == 200 for response in responses)
handles = ds._background_tasks.tasks()
assert len(handles) == 1
await asyncio.wait_for(handles[0].task, timeout=5)
assert launch_count["n"] == 1
finally:
pm.unregister(name="counting_task_plugin")
@pytest.mark.asyncio
async def test_post_launch_registration_starts_immediately_and_cancel_works():
ds = Datasette(memory=True)
await ds.start_background_tasks() # nothing registered yet, but launched
started = asyncio.Event()
async def long_running(datasette):
started.set()
await asyncio.Event().wait()
handle = ds.add_background_task(long_running, name="dynamic-task")
# Registered after launch: starts immediately rather than sitting in
# "registered" limbo.
assert handle.state == "running"
assert handle.task is not None
await asyncio.wait_for(started.wait(), timeout=5)
assert handle.state == "running"
handle.cancel()
with pytest.raises(asyncio.CancelledError):
await handle.task
await asyncio.sleep(0)
assert handle.state == "cancelled"
@pytest.mark.asyncio
async def test_pre_launch_registration_starts_as_registered():
ds = Datasette(memory=True)
async def task(datasette):
pass
handle = ds.add_background_task(task, name="buffered-task")
assert handle.state == "registered"
assert handle.task is None
handle.cancel() # not yet launched: deregisters instead of cancelling
assert handle not in ds._background_tasks.tasks()
@pytest.mark.asyncio
async def test_crashing_task_logs_traceback_and_state_is_crashed(caplog):
ds = Datasette(memory=True)
await ds.start_background_tasks()
survivor_ran = asyncio.Event()
async def crashing_task(datasette):
raise RuntimeError("kaboom")
async def survivor(datasette):
survivor_ran.set()
with caplog.at_level(logging.ERROR, logger="datasette.background_tasks"):
crash_handle = ds.add_background_task(crashing_task, name="crashing_task")
survivor_handle = ds.add_background_task(survivor, name="survivor")
await asyncio.wait_for(
asyncio.gather(
crash_handle.task, survivor_handle.task, return_exceptions=True
),
timeout=5,
)
assert crash_handle.state == "crashed"
assert isinstance(crash_handle.exception, RuntimeError)
assert str(crash_handle.exception) == "kaboom"
# The crash must not affect any other task.
assert survivor_ran.is_set()
assert survivor_handle.state == "completed"
assert "crashing_task" in caplog.text
assert "kaboom" in caplog.text
assert "Traceback" in caplog.text
assert "RuntimeError" in caplog.text
def test_name_collisions_get_suffixed_and_explicit_names_are_respected():
ds = Datasette(memory=True)
async def noop(datasette):
pass
async def another_noop(datasette):
pass
h1 = ds.add_background_task(noop, name="dup")
h2 = ds.add_background_task(another_noop, name="dup")
h3 = ds.add_background_task(noop, name="dup")
assert [h1.name, h2.name, h3.name] == ["dup", "dup-2", "dup-3"]
h_explicit = ds.add_background_task(noop, name="explicit-name")
assert h_explicit.name == "explicit-name"
h_default = ds.add_background_task(noop)
assert h_default.name == noop.__qualname__
@pytest.mark.asyncio
async def test_start_background_tasks_on_bare_datasette():
# The headless-CLI path (datasette-rss's `fetch --due` shape): no
# server, no lifespan, no first HTTP request - just an explicit call.
ran = asyncio.Event()
async def task(datasette):
ran.set()
ds = Datasette([])
assert ds._startup_invoked is False
handle = ds.add_background_task(task, name="headless-task")
assert handle.state == "registered"
await ds.start_background_tasks()
assert ds._startup_invoked is True
await asyncio.wait_for(ran.wait(), timeout=5)
await asyncio.wait_for(handle.task, timeout=5)
# handle.task being done only guarantees the coroutine has returned,
# not that our done-callback (which updates handle.state) has run yet -
# asyncio schedules done-callbacks via call_soon, and awaiting an
# already-done future/task returns immediately without giving the loop
# a chance to drain its ready queue. Yield once to let it run.
await asyncio.sleep(0)
assert handle.state == "completed"
@pytest.mark.asyncio
async def test_cancel_all_cancels_running_tasks_and_leaves_completed_alone():
ds = Datasette(memory=True)
await ds.start_background_tasks()
async def forever(datasette):
await asyncio.Event().wait()
async def quick(datasette):
return "done"
forever_handle = ds.add_background_task(forever, name="forever")
quick_handle = ds.add_background_task(quick, name="quick")
await asyncio.wait_for(quick_handle.task, timeout=5)
assert quick_handle.state == "completed"
await ds._background_tasks.cancel_all(grace=1.0)
assert forever_handle.state == "cancelled"
assert quick_handle.state == "completed"
@pytest.mark.asyncio
async def test_cancel_all_logs_stragglers_that_outlive_the_grace_period(caplog):
ds = Datasette(memory=True)
await ds.start_background_tasks()
async def stubborn(datasette):
with contextlib.suppress(asyncio.CancelledError):
await asyncio.sleep(10)
# Swallowing CancelledError above and returning normally simulates
# a task that ignores cancellation for longer than the grace period.
await asyncio.sleep(10)
handle = ds.add_background_task(stubborn, name="stubborn-task")
# Let the task actually start running and reach its first sleep (inside
# the CancelledError-suppressing block) before cancelling it - a task
# cancelled before it has ever run its first step never enters that
# block at all (the throw happens before the coroutine body starts),
# so it would finish cancelling immediately instead of behaving like a
# straggler.
await asyncio.sleep(0)
with caplog.at_level(logging.WARNING, logger="datasette.background_tasks"):
await ds._background_tasks.cancel_all(grace=0.1)
assert "stubborn-task" in caplog.text
# Clean up: actually cancel it now that the test has made its
# assertion, so it doesn't leak past the end of the test.
handle.task.cancel()
with contextlib.suppress(asyncio.CancelledError):
await handle.task

View file

@ -52,6 +52,56 @@ def test_serve_with_get(tmp_path_factory):
pm.unregister(to_unregister) pm.unregister(to_unregister)
def test_serve_with_get_does_not_launch_background_tasks(tmp_path_factory):
# --get must never launch background tasks, even though its TestClient
# request
# flows through the full ASGI stack (including the AsgiRunOnFirstRequest
# fallback that would otherwise launch them). The plugin's startup hook
# itself still runs (registration happens) - only the launch is
# suppressed, so the sentinel file the background task would write must
# never appear.
plugins_dir = tmp_path_factory.mktemp("plugins_for_get_background_tasks")
sentinel = plugins_dir / "sentinel.txt"
(plugins_dir / "bg_task_for_get.py").write_text(
textwrap.dedent(
f"""
from datasette import hookimpl
@hookimpl
def startup(datasette):
async def inner():
async def task(datasette):
with open("{sentinel!s}", "w") as fp:
fp.write("ran")
datasette.add_background_task(task, name="get-sentinel-task")
return inner
""",
),
"utf-8",
)
runner = CliRunner()
result = runner.invoke(
cli,
[
"serve",
"--memory",
"--plugins-dir",
str(plugins_dir),
"--get",
"/_memory/-/query.json?sql=select+1",
],
)
assert result.exit_code == 0, result.output
assert not sentinel.exists()
to_unregister = next(
p for p in pm.get_plugins() if p.__name__ == "bg_task_for_get.py"
)
pm.unregister(to_unregister)
def test_serve_with_get_headers(): def test_serve_with_get_headers():
runner = CliRunner() runner = CliRunner()
result = runner.invoke( result = runner.invoke(

View file

@ -1,12 +1,15 @@
import signal
import socket import socket
import subprocess
import time
import httpx import httpx2
import pytest import pytest
@pytest.mark.serial @pytest.mark.serial
def test_serve_localhost_http(ds_localhost_http_server): def test_serve_localhost_http(ds_localhost_http_server):
response = httpx.get("http://localhost:8041/_memory.json") response = httpx2.get("http://localhost:8041/_memory.json")
assert { assert {
"database": "_memory", "database": "_memory",
"path": "/_memory", "path": "/_memory",
@ -20,11 +23,205 @@ def test_serve_localhost_http(ds_localhost_http_server):
) )
def test_serve_unix_domain_socket(ds_unix_domain_socket_server): def test_serve_unix_domain_socket(ds_unix_domain_socket_server):
_, uds = ds_unix_domain_socket_server _, uds = ds_unix_domain_socket_server
transport = httpx.HTTPTransport(uds=uds) transport = httpx2.HTTPTransport(uds=uds)
client = httpx.Client(transport=transport) client = httpx2.Client(transport=transport)
response = client.get("http://localhost/_memory.json") response = client.get("http://localhost/_memory.json")
assert { assert {
"database": "_memory", "database": "_memory",
"path": "/_memory", "path": "/_memory",
"tables": [], "tables": [],
}.items() <= response.json().items() }.items() <= response.json().items()
# Shaped after datasette-litestream's startup hook, which schedules a
# background task with asyncio.get_running_loop().create_task(...):
# https://github.com/datasette/datasette-litestream
MARKER_TASK_PLUGIN = """
import asyncio
from datasette import hookimpl
from datasette.utils.asgi import Response
@hookimpl
def startup(datasette):
datasette._startup_calls = getattr(datasette, "_startup_calls", 0) + 1
async def _mark():
# Must await before setting the flag: a task with no internal
# await point could finish on the throwaway loop before it
# closed, masking the regression this test guards against.
await asyncio.sleep(0.2)
datasette._marker_task_ran = True
asyncio.get_running_loop().create_task(_mark())
@hookimpl
def register_routes():
async def marker_status(datasette):
return Response.json(
{
"marker_task_ran": getattr(datasette, "_marker_task_ran", False),
"startup_calls": getattr(datasette, "_startup_calls", 0),
}
)
return [(r"^/-/marker-task-ran$", marker_status)]
"""
STARTUP_ERROR_PLUGIN = """
from datasette import hookimpl
from datasette.utils import StartupError
@hookimpl
def startup(datasette):
raise StartupError("boom from plugin")
"""
@pytest.mark.serial
def test_startup_hook_background_task_runs_on_serving_loop(serve_with_plugins):
"""
Litestream-shaped regression test: a startup hook that does
asyncio.get_running_loop().create_task(...) must have that task
actually execute before/while the server is handling requests. This
only holds if invoke_startup() and uvicorn.Server.serve() share one
event loop. This test fails against unmodified main, where
invoke_startup() runs on a throwaway loop that is closed before
uvicorn opens its own loop to serve.
"""
_, port = serve_with_plugins({"marker_task_plugin": MARKER_TASK_PLUGIN})
# The fixture has already waited for the server to answer requests. The
# marker task deliberately awaits before setting its flag, so poll for a
# moment rather than assuming it landed before the first request arrived.
deadline = time.time() + 3.0
payload = {}
while time.time() < deadline:
payload = httpx2.get(
f"http://127.0.0.1:{port}/-/marker-task-ran", timeout=1.0
).json()
if payload["marker_task_ran"]:
break
time.sleep(0.05)
assert payload.get("marker_task_ran"), (
"The startup hook's asyncio.create_task(...) never ran - "
"invoke_startup() and the server are not sharing an event loop"
)
# Polling above means this test would also pass if the startup hook were
# re-run on the serving loop by the first-request fallback - which would
# hide exactly the bug being tested. invoke_startup() is idempotent today
# so that cannot happen; assert it explicitly so that if the idempotency
# guard is ever removed this test fails loudly instead of silently
# becoming a no-op.
assert payload["startup_calls"] == 1, (
"startup hook ran {} times - the marker may have been set by a "
"re-run on the serving loop rather than by the original task".format(
payload["startup_calls"]
)
)
@pytest.mark.serial
def test_startup_error_fails_fast_before_port_binds(serve_with_plugins):
"""
A "startup" plugin hook that raises StartupError must fail fast: print
the message, exit non-zero, and never accept a connection on the port -
the failure must happen before uvicorn.Server binds the socket.
"""
proc, port = serve_with_plugins(
{"startup_error_plugin": STARTUP_ERROR_PLUGIN}, wait_for_startup=False
)
stdout, _ = proc.communicate(timeout=15)
output = stdout.decode("utf-8")
assert proc.returncode not in (0, None), output
assert "boom from plugin" in output, output
# Nothing is listening on the port now the process has exited. This
# confirms the socket was not left bound; on its own it cannot prove the
# failure preceded the bind, since a port nothing ever touched also
# refuses connections.
with (
pytest.raises(OSError),
socket.create_connection(("127.0.0.1", port), timeout=0.2),
):
pass
# Verify that SIGTERM and SIGINT sent to `datasette serve` trigger uvicorn's
# lifespan.shutdown event and run the plugin shutdown hooks. The plugin below
# writes a sentinel file from its shutdown hook so the tests can check that
# cleanup ran after the server subprocess exits.
SHUTDOWN_SENTINEL_PLUGIN_TEMPLATE = """
import pathlib
from datasette import hookimpl
SENTINEL_PATH = {sentinel_path!r}
@hookimpl
def shutdown(datasette):
pathlib.Path(SENTINEL_PATH).write_text("shutdown ran", "utf-8")
"""
def _start_serve_with_shutdown_sentinel(serve_with_plugins, tmp_path):
sentinel_path = tmp_path / "shutdown-sentinel.txt"
proc, _ = serve_with_plugins(
{
"shutdown_sentinel_plugin": SHUTDOWN_SENTINEL_PLUGIN_TEMPLATE.format(
sentinel_path=str(sentinel_path)
)
}
)
return proc, sentinel_path
@pytest.mark.serial
def test_sigterm_runs_shutdown_hooks(serve_with_plugins, tmp_path):
ds_proc, sentinel_path = _start_serve_with_shutdown_sentinel(
serve_with_plugins, tmp_path
)
assert not sentinel_path.exists()
ds_proc.send_signal(signal.SIGTERM)
try:
ds_proc.wait(timeout=10)
except subprocess.TimeoutExpired:
ds_proc.kill()
ds_proc.wait()
raise AssertionError(
"datasette serve did not exit within 10s of SIGTERM\n"
+ ds_proc.stdout.read().decode("utf-8")
)
output = ds_proc.stdout.read().decode("utf-8")
assert sentinel_path.exists(), (
"shutdown hook never wrote its sentinel file after SIGTERM\n" + output
)
assert sentinel_path.read_text("utf-8") == "shutdown ran"
@pytest.mark.serial
@pytest.mark.skipif(
not hasattr(signal, "SIGINT"), reason="Requires signal.SIGINT support"
)
def test_sigint_runs_shutdown_hooks(serve_with_plugins, tmp_path):
ds_proc, sentinel_path = _start_serve_with_shutdown_sentinel(
serve_with_plugins, tmp_path
)
assert not sentinel_path.exists()
ds_proc.send_signal(signal.SIGINT)
try:
ds_proc.wait(timeout=10)
except subprocess.TimeoutExpired:
ds_proc.kill()
ds_proc.wait()
raise AssertionError(
"datasette serve did not exit within 10s of SIGINT\n"
+ ds_proc.stdout.read().decode("utf-8")
)
output = ds_proc.stdout.read().decode("utf-8")
assert sentinel_path.exists(), (
"shutdown hook never wrote its sentinel file after SIGINT\n" + output
)
assert sentinel_path.read_text("utf-8") == "shutdown ran"

View file

@ -0,0 +1,412 @@
import pytest
from datasette import hookimpl
from datasette.app import Datasette
from datasette.permissions import Action, PermissionSQL, _permission_check_cache
from datasette.resources import DatabaseResource, TableResource
from datasette.utils.sqlite import sqlite3, sqlite_derived_table_dependencies
@pytest.mark.asyncio
@pytest.mark.parametrize("fts_module", ["fts4", "fts5"])
@pytest.mark.parametrize("actor", [None, {"id": "root"}], ids=["anonymous", "root"])
async def test_derived_permissions_allow_one_hop_but_deny_nested_sources(
fts_module, actor
):
class InspectPlugin:
@hookimpl
def register_actions(self):
return [
Action(
name="inspect-derived",
description="Inspect a table",
resource_class=TableResource,
also_requires="view-table",
)
]
@hookimpl
def permission_resources_sql(self, action):
if action == "inspect-derived":
return PermissionSQL(
sql="SELECT NULL AS parent, NULL AS child, 1 AS allow, 'inspect allowed' AS reason"
)
ds = Datasette(memory=True)
ds.pm.register(InspectPlugin(), name="inspect-derived-test")
db = ds.add_memory_database(
f"derived_one_hop_{fts_module}_{actor is not None}", name="data"
)
await db.execute_write("create table Documents (body text)")
await db.execute_write(
f"create virtual table Search using {fts_module}(body, content='Documents')"
)
await db.execute_write(
f"create virtual table Nested using {fts_module}(body, content='sEaRcH')"
)
await ds.invoke_startup()
token = _permission_check_cache.set({})
try:
# Both direct permissions are allowed, but a derived source makes its
# dependent unavailable even to an actor who can view the whole chain.
# Check and cache Search first so its cached grant cannot grant Nested.
for table, expected in (
("Documents", True),
("Search", True),
("Nested", False),
("Search_docsize", False),
):
for spelling in (table, table.upper(), table.lower()):
assert await ds.allowed_many(
actions=["view-table", "inspect-derived"],
resource=TableResource("data", spelling),
actor=actor,
) == {"view-table": expected, "inspect-derived": expected}
page = await ds.allowed_resources(
"view-table", actor, parent="data", include_is_private=True, limit=1000
)
allowed = {resource.child for resource in page.resources}
assert {"Documents", "Search"}.issubset(allowed)
assert "Nested" not in allowed
assert "Search_docsize" not in allowed
finally:
_permission_check_cache.reset(token)
ds.pm.unregister(name="inspect-derived-test")
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("listing", [False, True], ids=["individual", "listing"])
async def test_derived_permission_discovery_error_is_retried(monkeypatch, listing):
ds = Datasette(memory=True)
db = ds.add_memory_database(f"derived_discovery_error_{listing}", name="data")
await db.execute_write("create table documents (id integer primary key)")
await ds.invoke_startup()
class UnavailableSchema:
def execute(self, sql):
raise sqlite3.DatabaseError("schema temporarily unavailable")
async def check():
if listing:
return await ds.allowed_resources("view-table", parent="data")
return await ds.allowed(
action="view-table", resource=TableResource("data", "documents")
)
token = _permission_check_cache.set({})
try:
with monkeypatch.context() as patch:
patch.setattr(
"datasette.database.sqlite_derived_table_dependencies",
lambda conn: sqlite_derived_table_dependencies(UnavailableSchema()),
)
with pytest.raises(sqlite3.DatabaseError, match="schema temporarily"):
await check()
# Failed discovery must not cache an empty map or a permission grant.
assert db._cached_derived_table_dependencies is None
assert not _permission_check_cache.get()
result = await check()
if listing:
assert [resource.child for resource in result.resources] == ["documents"]
else:
assert result is True
assert db._cached_derived_table_dependencies is not None
finally:
_permission_check_cache.reset(token)
@pytest.mark.asyncio
@pytest.mark.parametrize("fts_module", ("fts4", "fts5"))
async def test_external_content_fts_inherits_content_table_view_permission(fts_module):
actor = {"id": "reader"}
secret_marker = "ISSUE_17_EXTERNAL_CONTENT_FTS_SECRET"
ds = Datasette(
memory=True,
config={
"permissions": {
"view-instance": {"id": "reader"},
"view-database": {"id": "reader"},
"view-table": {"id": "reader"},
"execute-sql": {"id": "nobody"},
},
"databases": {
"data": {
"tables": {
"secret": {"permissions": {"view-table": False}},
}
}
},
},
)
db = ds.add_memory_database(f"issue_17_{fts_module}_permissions", name="data")
await db.execute_write("create table secret (id integer primary key, body text)")
await db.execute_write(
"insert into secret (body) values (?)",
[secret_marker],
)
fts_options = "body, content='secret'"
if fts_module == "fts5":
fts_options += ", content_rowid='id'"
await db.execute_write(
f"create virtual table secret_fts using {fts_module}({fts_options})"
)
await db.execute_write("insert into secret_fts(secret_fts) values ('rebuild')")
await ds.invoke_startup()
try:
assert "secret_fts" in await db.hidden_table_names()
assert (
await ds.allowed(
action="execute-sql",
resource=DatabaseResource("data"),
actor=actor,
)
is False
)
direct = await ds.client.get("/data/secret.json", actor=actor)
assert direct.status_code == 403
companion = await ds.client.get(
"/data/secret_fts.json?_shape=array",
actor=actor,
)
assert companion.status_code in (403, 404), (
"An automatically hidden external-content FTS table must inherit "
"the content table's view denial or be unavailable: "
f"{companion.text}"
)
assert secret_marker not in companion.text
finally:
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("fts_module", ("fts4", "fts5"))
@pytest.mark.parametrize("contentless", (False, True), ids=("internal", "contentless"))
async def test_fts_shadow_tables_inherit_logical_table_view_permission(
fts_module, contentless
):
table_config = {
"secret_fts": {"permissions": {"view-table": False}},
# An explicit allow on one implementation table must not override
# the logical FTS table's denial.
"secret_fts_docsize": {"permissions": {"view-table": True}},
}
ds = Datasette(
memory=True,
config={
"permissions": {
"view-instance": True,
"view-database": True,
"view-table": True,
"execute-sql": False,
},
"databases": {"data": {"tables": table_config}},
},
)
db = ds.add_memory_database(
f"issue_17_{fts_module}_{'contentless' if contentless else 'internal'}",
name="data",
)
options = "body, content=''" if contentless else "body"
await db.execute_write(
f"create virtual table secret_fts using {fts_module}({options})"
)
await db.execute_write(
"insert into secret_fts(rowid, body) values (1, 'ISSUE_17_SHADOW_SECRET')"
)
await ds.invoke_startup()
try:
dependencies = await db.derived_table_dependencies()
shadow_tables = sorted(
table for table, source in dependencies.items() if source == "secret_fts"
)
assert shadow_tables
assert "secret_fts_docsize" in shadow_tables
for shadow_table in shadow_tables:
assert (
await ds.allowed(
action="view-table",
resource=TableResource("data", shadow_table),
)
is False
)
response = await ds.client.get(f"/data/{shadow_table}.json?_shape=array")
assert response.status_code == 403
assert "ISSUE_17_SHADOW_SECRET" not in response.text
allowed = await ds.allowed_resources("view-table", parent="data", limit=1000)
allowed_names = {resource.child for resource in allowed.resources}
assert not set(shadow_tables).intersection(allowed_names)
database_json = await ds.client.get("/data.json")
assert database_json.status_code == 200
for shadow_table in shadow_tables:
assert shadow_table not in database_json.text
schema_json = await ds.client.get("/data/-/schema.json")
assert schema_json.status_code == 200
for shadow_table in shadow_tables:
assert shadow_table not in schema_json.text
finally:
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize(
"content_allowed,companion_allowed,expected",
(
(False, True, False),
(True, False, False),
(True, True, True),
),
)
async def test_external_content_and_companion_permissions_are_both_required(
content_allowed, companion_allowed, expected
):
ds = Datasette(
memory=True,
default_deny=True,
config={
"permissions": {
"view-instance": True,
"view-database": True,
},
"databases": {
"data": {
"tables": {
"secret": {"permissions": {"view-table": content_allowed}},
"secret_fts": {
"permissions": {"view-table": companion_allowed}
},
}
}
},
},
)
db = ds.add_memory_database(
f"issue_17_explicit_{int(content_allowed)}_{int(companion_allowed)}",
name="data",
)
await db.execute_write("create table secret(id integer primary key, body text)")
await db.execute_write("insert into secret(body) values ('ISSUE_17_MATRIX_SECRET')")
await db.execute_write(
"create virtual table secret_fts using fts5("
"body, content='secret', content_rowid='id')"
)
await db.execute_write("insert into secret_fts(secret_fts) values ('rebuild')")
await ds.invoke_startup()
try:
assert (
await ds.allowed(
action="view-table",
resource=TableResource("data", "secret_fts"),
)
is expected
)
response = await ds.client.get("/data/secret_fts.json?_shape=array")
assert response.status_code == (200 if expected else 403)
if not expected:
assert "ISSUE_17_MATRIX_SECRET" not in response.text
finally:
ds.close()
@pytest.mark.asyncio
async def test_derived_tables_propagate_private_flag_and_route_permissions():
actor = {"id": "reader"}
ds = Datasette(
memory=True,
config={
"permissions": {
"view-instance": True,
"view-database": True,
"view-table": True,
},
"databases": {
"data": {
"tables": {
"secret": {"permissions": {"view-table": {"id": "reader"}}}
}
}
},
},
)
db = ds.add_memory_database("issue_17_private_flag", name="data")
await db.execute_write("create table secret(id integer primary key, body text)")
await db.execute_write("insert into secret(body) values ('PRIVATE')")
await db.execute_write(
"create virtual table secret_fts using fts5("
"body, content='secret', content_rowid='id')"
)
await db.execute_write("insert into secret_fts(secret_fts) values ('rebuild')")
await ds.invoke_startup()
try:
actor_page = await ds.allowed_resources(
"view-table", actor, parent="data", include_is_private=True, limit=1000
)
actor_resources = {
resource.child: resource for resource in actor_page.resources
}
derived_names = set(await db.derived_table_dependencies())
assert "secret_fts" in actor_resources
assert actor_resources["secret_fts"].private
# Shadow tables depend on the already-derived external-content FTS
# table, so they remain unavailable even to the permitted reader.
assert not (derived_names - {"secret_fts"}).intersection(actor_resources)
anonymous_page = await ds.allowed_resources(
"view-table", parent="data", limit=1000
)
anonymous_names = {resource.child for resource in anonymous_page.resources}
assert not derived_names.intersection(anonymous_names)
for path in (
"/data/secret_fts.json?_facet=body",
"/data/secret_fts.csv",
"/data/secret_fts/-/autocomplete?q=PRIVATE",
"/data/secret_fts/-/schema.json",
):
denied = await ds.client.get(path)
assert denied.status_code == 403
allowed = await ds.client.get(path, actor=actor)
assert allowed.status_code == 200
finally:
ds.close()
@pytest.mark.asyncio
async def test_cyclic_derived_table_dependencies_fail_closed():
ds = Datasette(memory=True)
db = ds.add_memory_database("issue_17_cycle", name="data")
await db.execute_write(
"create virtual table first_fts using fts5(body, content='second_fts')"
)
await db.execute_write(
"create virtual table second_fts using fts5(body, content='first_fts')"
)
await ds.invoke_startup()
try:
for table in ("first_fts", "second_fts"):
assert (
await ds.allowed(
action="view-table", resource=TableResource("data", table)
)
is False
)
page = await ds.allowed_resources("view-table", parent="data", limit=1000)
allowed_names = {resource.child for resource in page.resources}
assert "first_fts" not in allowed_names
assert "second_fts" not in allowed_names
finally:
ds.close()

View file

@ -36,8 +36,10 @@ def test_homepage(app_client_two_attached_databases):
h2 = soup.select("h2")[0] h2 = soup.select("h2")[0]
assert "extra database" == h2.text.strip() assert "extra database" == h2.text.strip()
counts_p, links_p = h2.find_all_next("p")[:2] counts_p, links_p = h2.find_all_next("p")[:2]
# Shadow tables of the external-content index are denied, so they do not
# contribute to the table or row totals.
assert ( assert (
"2 rows in 1 table, 5 rows in 4 hidden tables, 1 view" == counts_p.text.strip() "2 rows in 1 table, 2 rows in 1 hidden table, 1 view" == counts_p.text.strip()
) )
# We should only show visible, not hidden tables here: # We should only show visible, not hidden tables here:
table_links = [ table_links = [

View file

@ -15,11 +15,16 @@ from datasette.database import (
DatasetteClosedError, DatasetteClosedError,
ExecuteWriteResult, ExecuteWriteResult,
MultipleValues, MultipleValues,
QueryInterrupted,
Results, Results,
_deliver_write_result, _deliver_write_result,
) )
from datasette.utils import Column from datasette.utils import Column
from datasette.utils.sqlite import sqlite3, supports_returning from datasette.utils.sqlite import (
sqlite3,
sqlite_derived_table_dependencies,
supports_returning,
)
requires_sqlite_returning = pytest.mark.skipif( requires_sqlite_returning = pytest.mark.skipif(
not supports_returning(), reason="SQLite does not support RETURNING" not supports_returning(), reason="SQLite does not support RETURNING"
@ -38,6 +43,31 @@ async def test_execute(db):
assert 15 == len(results) assert 15 == len(results)
@pytest.mark.asyncio
async def test_derived_dependency_cache_survives_failed_refresh(monkeypatch):
ds = Datasette(memory=True)
db = ds.add_memory_database(uuid.uuid4().hex, name="data")
await db.derived_table_dependencies()
previous_cache = db._cached_derived_table_dependencies
await db.execute_write("create table dependency_cache_refresh (id integer)")
class UnavailableSchema:
def execute(self, sql):
raise sqlite3.DatabaseError("schema temporarily unavailable")
with monkeypatch.context() as patch:
patch.setattr(
"datasette.database.sqlite_derived_table_dependencies",
lambda conn: sqlite_derived_table_dependencies(UnavailableSchema()),
)
with pytest.raises(sqlite3.DatabaseError, match="schema temporarily"):
await db.derived_table_dependencies()
assert db._cached_derived_table_dependencies == previous_cache
await db.derived_table_dependencies()
assert db._cached_derived_table_dependencies[0] != previous_cache[0]
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_results_first(db): async def test_results_first(db):
assert None is (await db.execute("select * from facetable where pk > 100")).first() assert None is (await db.execute("select * from facetable where pk > 100")).first()
@ -478,6 +508,31 @@ async def test_view_names(db):
] ]
@pytest.mark.asyncio
async def test_execute_write_custom_time_limit():
ds = Datasette(settings={"sql_time_limit_ms": 1})
db = ds.add_memory_database(uuid.uuid4().hex, name="write_limits")
await ds.invoke_startup()
# Bounded work from PR #51; even without a limit this finishes on its own.
sql = (
"with recursive c(x) as "
"(select 1 union all select x+1 from c where x < 800000) "
"select x from c where x < 0"
)
try:
await db.execute_write("create table items(value integer)")
with pytest.raises(QueryInterrupted):
await db.execute(sql)
# Writes take their own explicit limit, independent of the read setting.
with pytest.raises(QueryInterrupted):
await db.execute_write(f"insert into items(value) {sql}", time_limit_ms=1)
# Interruption must leave the connection available for subsequent writes.
await db.execute_write("insert into items(value) values (1)")
assert (await db.execute("select value from items")).single_value() == 1
finally:
ds.close()
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_execute_write_block_true(db): async def test_execute_write_block_true(db):
result = await db.execute_write( result = await db.execute_write(
@ -705,6 +760,33 @@ async def test_execute_write_fn_block_false(db):
assert isinstance(task_id, uuid.UUID) assert isinstance(task_id, uuid.UUID)
@pytest.mark.asyncio
@pytest.mark.parametrize("disable_threads", (False, True))
async def test_execute_write_fn_block_false_returns_uuid(tmp_path, disable_threads):
# block=False is documented to return "a UUID representing the queued task".
# With num_sql_threads=0 there is no write thread, so the non-threaded branch
# has to satisfy the same contract as the threaded one.
settings = {"num_sql_threads": 0} if disable_threads else {}
ds = Datasette([], memory=True, settings=settings)
await ds.invoke_startup()
db = ds.add_memory_database("test_block_false")
await db.execute_write(
"create table if not exists t (id integer primary key, v text)"
)
def write_fn(conn):
conn.execute("insert into t (v) values ('a')")
# Returns None, like most write functions.
task_id = await db.execute_write_fn(write_fn, block=False)
assert isinstance(task_id, uuid.UUID)
# Distinct per call, so a caller can tell two queued tasks apart.
second = await db.execute_write_fn(write_fn, block=False)
assert isinstance(second, uuid.UUID)
assert second != task_id
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_execute_write_fn_block_true(db): async def test_execute_write_fn_block_true(db):
def write_fn(conn): def write_fn(conn):

View file

@ -1,4 +1,4 @@
import httpx import httpx2
import pytest import pytest
import pytest_asyncio import pytest_asyncio
@ -43,7 +43,7 @@ async def datasette_with_permissions():
async def test_client_methods(datasette, method, path, expected_status): async def test_client_methods(datasette, method, path, expected_status):
client_method = getattr(datasette.client, method) client_method = getattr(datasette.client, method)
response = await client_method(path) response = await client_method(path)
assert isinstance(response, httpx.Response) assert isinstance(response, httpx2.Response)
assert response.status_code == expected_status assert response.status_code == expected_status
# Try that again using datasette.client.request # Try that again using datasette.client.request
response2 = await datasette.client.request(method, path) response2 = await datasette.client.request(method, path)
@ -63,7 +63,7 @@ async def test_client_post(datasette, prefix):
"message": "A message", "message": "A message",
}, },
) )
assert isinstance(response, httpx.Response) assert isinstance(response, httpx2.Response)
assert response.status_code == 302 assert response.status_code == 302
assert "ds_messages" in response.cookies assert "ds_messages" in response.cookies
finally: finally:
@ -135,7 +135,7 @@ async def test_skip_permission_checks_all_methods(datasette_with_permissions, me
response = await client_method("/test_db.json", skip_permission_checks=True) response = await client_method("/test_db.json", skip_permission_checks=True)
# We don't check status code since some methods might not be allowed, # We don't check status code since some methods might not be allowed,
# but we verify the request doesn't fail due to permissions # but we verify the request doesn't fail due to permissions
assert isinstance(response, httpx.Response) assert isinstance(response, httpx2.Response)
@pytest.mark.asyncio @pytest.mark.asyncio
@ -340,7 +340,7 @@ async def test_actor_parameter_all_http_methods(datasette, method):
client_method = getattr(datasette.client, method) client_method = getattr(datasette.client, method)
# Just verify no TypeError about unexpected 'actor' kwarg # Just verify no TypeError about unexpected 'actor' kwarg
response = await client_method("/", actor={"id": "root"}) response = await client_method("/", actor={"id": "root"})
assert isinstance(response, httpx.Response) assert isinstance(response, httpx2.Response)
@pytest.mark.asyncio @pytest.mark.asyncio

344
tests/test_lifespan.py Normal file
View file

@ -0,0 +1,344 @@
"""
Tests for wiring Datasette startup (setup_db table counts + invoke_startup)
into the ASGI lifespan protocol.
These exercise Datasette._startup_sequence() via three different callers:
- AsgiLifespan, by hand-driving lifespan.startup messages (no HTTP request)
- AsgiRunOnFirstRequest, the fallback for hosts that never send lifespan
events (this is what DatasetteClient / plain httpx2.ASGITransport uses)
- Both at once, to prove startup hooks run at most once
"""
import asyncio
import contextlib
import sqlite3
import httpx2
import pytest
from datasette import hookimpl
from datasette.app import Datasette
from datasette.database import Database
from datasette.plugins import pm
async def _drive_lifespan_startup(app):
"""Send a single lifespan.startup message into app's ASGI lifespan loop
and return the list of messages sent back - without ever sending
lifespan.shutdown. Mirrors what a real server does: after startup
completes it parks waiting for the next event. We cancel that wait
once we've observed the startup response, rather than closing the
Datasette instance down with a shutdown message.
"""
messages_sent = []
startup_responded = asyncio.Event()
delivered = False
async def receive():
nonlocal delivered
if not delivered:
delivered = True
return {"type": "lifespan.startup"}
# No further messages: block until the task is cancelled below,
# same as a real server parked waiting for lifespan.shutdown.
await asyncio.Event().wait()
async def send(message):
messages_sent.append(message)
startup_responded.set()
task = asyncio.create_task(app({"type": "lifespan"}, receive, send))
try:
await asyncio.wait_for(startup_responded.wait(), timeout=5)
finally:
task.cancel()
with contextlib.suppress(asyncio.CancelledError):
await task
return messages_sent
@pytest.mark.asyncio
async def test_lifespan_startup_runs_before_any_request():
ds = Datasette(memory=True)
assert ds._startup_invoked is False
app = ds.app()
messages = await _drive_lifespan_startup(app)
assert {"type": "lifespan.startup.complete"} in messages
assert ds._startup_invoked is True
# Internal catalog tables should be populated too, entirely without an
# HTTP request having been made.
internal_db = ds.get_internal_database()
databases = await internal_db.execute("select * from catalog_databases")
assert len(databases.rows) >= 1
@pytest.mark.asyncio
async def test_lifespan_startup_failure_reports_lifespan_startup_failed():
class RaisingStartupPlugin:
__name__ = "RaisingStartupPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
raise RuntimeError("boom from startup hook")
return inner
ds = Datasette(memory=True)
pm.register(RaisingStartupPlugin(), name="raising_startup_plugin")
try:
app = ds.app()
messages = await _drive_lifespan_startup(app)
finally:
pm.unregister(name="raising_startup_plugin")
assert messages == [
{"type": "lifespan.startup.failed", "message": "boom from startup hook"}
]
# The exception happened before invoke_startup() got to the end of its
# body, so startup is not considered to have completed.
assert ds._startup_invoked is False
@pytest.mark.asyncio
async def test_startup_runs_exactly_once_across_lifespan_and_first_request():
call_count = {"n": 0}
class CountingStartupPlugin:
__name__ = "CountingStartupPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
call_count["n"] += 1
return inner
ds = Datasette(memory=True)
pm.register(CountingStartupPlugin(), name="counting_startup_plugin")
try:
# Build the ASGI app once, the way a real deployment does - and
# reuse the SAME app instance for both the lifespan drive and the
# HTTP requests below, since a fresh ds.app() call would reset the
# AsgiRunOnFirstRequest fallback's state.
app = ds.app()
messages = await _drive_lifespan_startup(app)
assert {"type": "lifespan.startup.complete"} in messages
assert call_count["n"] == 1
# A first HTTP request (as if the host never sent lifespan events,
# or lifespan already ran) should not run the hook again.
transport = httpx2.ASGITransport(app=app)
async with httpx2.AsyncClient(
transport=transport, base_url="http://localhost"
) as client:
response1 = await client.get("/-/versions.json")
assert response1.status_code == 200
# ... nor should a second, repeat request.
response2 = await client.get("/-/versions.json")
assert response2.status_code == 200
finally:
pm.unregister(name="counting_startup_plugin")
assert call_count["n"] == 1
@pytest.mark.asyncio
async def test_no_lifespan_first_request_still_triggers_startup():
# Pin today's behavior: a client that never drives ASGI lifespan events
# at all (like httpx2.ASGITransport, which DatasetteClient uses) still
# gets startup armed by the AsgiRunOnFirstRequest fallback.
ds = Datasette(memory=True)
assert ds._startup_invoked is False
app = ds.app()
transport = httpx2.ASGITransport(app=app)
async with httpx2.AsyncClient(
transport=transport, base_url="http://localhost"
) as client:
response = await client.get("/-/versions.json")
assert response.status_code == 200
assert ds._startup_invoked is True
internal_db = ds.get_internal_database()
databases = await internal_db.execute("select * from catalog_databases")
assert len(databases.rows) >= 1
@pytest.mark.asyncio
async def test_datasette_client_first_request_triggers_startup():
# Same as above, but through the real DatasetteClient (ds.client) that
# plugins and tests actually use, to confirm nothing regressed there.
ds = Datasette(memory=True)
assert ds._startup_invoked is False
response = await ds.client.get("/-/versions.json")
assert response.status_code == 200
assert ds._startup_invoked is True
@pytest.mark.asyncio
async def test_concurrent_first_requests_all_wait_for_slow_startup():
call_count = {"n": 0}
class SlowStartupPlugin:
__name__ = "SlowStartupPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
call_count["n"] += 1
await asyncio.sleep(0.2)
return inner
ds = Datasette(memory=True)
pm.register(SlowStartupPlugin(), name="slow_startup_plugin")
try:
app = ds.app()
transport = httpx2.ASGITransport(app=app)
async with httpx2.AsyncClient(
transport=transport, base_url="http://localhost"
) as client:
responses = await asyncio.gather(
*[client.get("/-/versions.json") for _ in range(10)]
)
finally:
pm.unregister(name="slow_startup_plugin")
# Every one of the 10 simultaneous first requests must have blocked
# until startup actually finished, not raced ahead of it.
assert all(response.status_code == 200 for response in responses)
assert call_count["n"] == 1
assert ds._startup_invoked is True
@pytest.mark.asyncio
async def test_setup_db_still_runs_when_invoke_startup_ran_first(tmp_path, monkeypatch):
# Regression test: `datasette serve` (cli.py _serve_async) calls
# ds.invoke_startup() directly, before uvicorn ever sends a
# lifespan.startup event that drives _startup_sequence(). If
# _startup_sequence()'s fast path only checked `_startup_invoked`, it
# would see startup already done and skip the immutable-database
# table-count precompute (setup_db) entirely - a silent regression
# versus main, where AsgiRunOnFirstRequest ran setup_db unconditionally
# on request #1.
db_path = tmp_path / "immutable.db"
conn = sqlite3.connect(str(db_path))
conn.execute("create table t (id integer primary key)")
conn.commit()
conn.close()
ds = Datasette([], immutables=[str(db_path)])
call_count = {"n": 0}
original_table_counts = Database.table_counts
async def counting_table_counts(self, *args, **kwargs):
call_count["n"] += 1
return await original_table_counts(self, *args, **kwargs)
monkeypatch.setattr(Database, "table_counts", counting_table_counts)
# Simulate the CLI path: invoke_startup() runs directly and completes
# BEFORE _startup_sequence() ever gets a chance to run setup_db.
await ds.invoke_startup()
assert ds._startup_invoked is True
assert call_count["n"] == 0
# The lifespan/first-request path (or the CLI itself, per the fix)
# calling the shared entry point afterwards must still precompute
# table counts for immutable databases.
await ds._startup_sequence()
assert call_count["n"] == 1
assert ds._setup_db_done is True
# Idempotency: a second call must not recompute.
await ds._startup_sequence()
assert call_count["n"] == 1
@pytest.mark.asyncio
async def test_asgi_wrapper_runs_after_startup_fallback_path():
class AssertStartupPlugin:
__name__ = "AssertStartupPlugin"
@hookimpl
def asgi_wrapper(self, datasette):
def wrap(app):
async def check_startup(scope, receive, send):
if scope["type"] == "http":
assert (
datasette._startup_invoked is True
), "asgi_wrapper saw an http scope before startup completed"
await app(scope, receive, send)
return check_startup
return wrap
ds = Datasette(memory=True)
pm.register(AssertStartupPlugin(), name="assert_startup_plugin")
try:
assert ds._startup_invoked is False
app = ds.app()
transport = httpx2.ASGITransport(app=app)
async with httpx2.AsyncClient(
transport=transport, base_url="http://localhost"
) as client:
response = await client.get("/-/versions.json")
assert response.status_code == 200
finally:
pm.unregister(name="assert_startup_plugin")
assert ds._startup_invoked is True
@pytest.mark.asyncio
async def test_short_circuit_wrapper_no_longer_defers_startup():
# Middleware that returns a response before getting to the rest of
# Datasette should still cause _startup_invoked=True
class ShortCircuitPlugin:
__name__ = "ShortCircuitPlugin"
@hookimpl
def asgi_wrapper(self, datasette):
def wrap(app):
async def forbidden(scope, receive, send):
if scope["type"] != "http":
await app(scope, receive, send)
return
await send(
{
"type": "http.response.start",
"status": 403,
"headers": [[b"content-type", b"text/plain"]],
}
)
await send(
{
"type": "http.response.body",
"body": b"Forbidden",
}
)
return forbidden
return wrap
ds = Datasette(memory=True)
pm.register(ShortCircuitPlugin(), name="short_circuit_plugin")
try:
assert ds._startup_invoked is False
app = ds.app()
transport = httpx2.ASGITransport(app=app)
async with httpx2.AsyncClient(
transport=transport, base_url="http://localhost"
) as client:
response = await client.get("/-/versions.json")
assert response.status_code == 403
finally:
pm.unregister(name="short_circuit_plugin")
assert ds._startup_invoked is True

View file

@ -1,4 +1,5 @@
from pathlib import Path from pathlib import Path
from unittest import mock
import pytest import pytest
@ -20,6 +21,29 @@ def has_compiled_ext():
return False return False
@pytest.mark.parametrize("load_fails", (False, True))
def test_load_extension_is_disabled(load_fails):
ds = Datasette(sqlite_extensions=[COMPILED_EXTENSION_PATH])
connection = mock.Mock()
if load_fails:
connection.load_extension.side_effect = RuntimeError
if load_fails:
with pytest.raises(RuntimeError):
ds._prepare_connection(connection, "data")
else:
ds._prepare_connection(connection, "data")
# Extensions are loaded using the Python API, never via SQL
assert connection.load_extension.mock_calls == [
mock.call(COMPILED_EXTENSION_PATH),
]
assert connection.enable_load_extension.mock_calls == [
mock.call(True),
mock.call(False),
]
@pytest.mark.asyncio @pytest.mark.asyncio
@pytest.mark.skipif(not has_compiled_ext(), reason="Requires compiled ext.c") @pytest.mark.skipif(not has_compiled_ext(), reason="Requires compiled ext.c")
async def test_load_extension_default_entrypoint(): async def test_load_extension_default_entrypoint():
@ -64,3 +88,20 @@ async def test_load_extension_multiple_entrypoints():
response = await ds.client.get("/_memory/-/query.json?_shape=arrays&sql=select+c()") response = await ds.client.get("/_memory/-/query.json?_shape=arrays&sql=select+c()")
assert response.status_code == 200 assert response.status_code == 200
assert response.json()["rows"][0][0] == "c" assert response.json()["rows"][0][0] == "c"
@pytest.mark.asyncio
@pytest.mark.skipif(not has_compiled_ext(), reason="Requires compiled ext.c")
async def test_sql_cannot_load_additional_extension():
ds = Datasette(sqlite_extensions=[COMPILED_EXTENSION_PATH])
response = await ds.client.get(
"/_memory/-/query.json",
params={
"sql": "select load_extension(:path, :entrypoint)",
"path": COMPILED_EXTENSION_PATH,
"entrypoint": "sqlite3_ext_b_init",
},
)
assert response.status_code == 400
assert response.json()["error"] == "not authorized"

View file

@ -494,3 +494,31 @@ async def test_execute_sql_requires_view_database():
) )
finally: finally:
ds.pm.unregister(plugin) ds.pm.unregister(plugin)
@pytest.mark.asyncio
@pytest.mark.parametrize("path", ["/-/allowed", "/-/allowed.json?action=view-table"])
async def test_allowed_requires_view_instance(path):
"""
GHSA-hp2x-vx2r-6vxg: /-/allowed should be gated like its /-/rules sibling.
An actor who is denied view-instance gets 403 from / and /-/rules, but
/-/allowed (HTML and JSON) currently returns 200 to the same actor.
"""
ds = Datasette(config={"allow": {"id": "alice"}})
await ds.invoke_startup()
db = ds.add_memory_database("live")
await db.execute_write("CREATE TABLE IF NOT EXISTS t (id INTEGER PRIMARY KEY)")
await ds.refresh_schemas()
assert (await ds.client.get("/")).status_code == 403
assert (await ds.client.get("/-/rules.json?action=view-table")).status_code == 403
response = await ds.client.get(path)
assert response.status_code == 403
# Alice is still allowed
response = await ds.client.get(
path, cookies={"ds_actor": ds.client.actor_cookie({"id": "alice"})}
)
assert response.status_code == 200

View file

@ -520,6 +520,7 @@ def view_instance_client():
"/-/plugins", "/-/plugins",
"/-/settings", "/-/settings",
"/-/threads", "/-/threads",
"/-/tasks",
"/-/databases", "/-/databases",
"/-/permissions", "/-/permissions",
"/-/messages", "/-/messages",

View file

@ -5,7 +5,7 @@ import subprocess
import sys import sys
import time import time
import httpx import httpx2
import pytest import pytest
from datasette.fixtures import write_fixture_database from datasette.fixtures import write_fixture_database
@ -34,11 +34,11 @@ def wait_for_server(process, url, timeout=30):
f"stderr:\n{stderr}" f"stderr:\n{stderr}"
) )
try: try:
response = httpx.get(url, timeout=1.0) response = httpx2.get(url, timeout=1.0)
if response.status_code < 500: if response.status_code < 500:
return return
last_error = f"HTTP {response.status_code}: {response.text[:200]}" last_error = f"HTTP {response.status_code}: {response.text[:200]}"
except httpx.HTTPError as ex: except httpx2.HTTPError as ex:
last_error = repr(ex) last_error = repr(ex)
time.sleep(0.1) time.sleep(0.1)
if process.poll() is None: if process.poll() is None:
@ -108,6 +108,11 @@ def write_playwright_database(db_path):
conn = sqlite3.connect(db_path) conn = sqlite3.connect(db_path)
try: try:
conn.executescript(""" conn.executescript("""
create table count_numbers (id integer primary key);
with recursive sequence(id) as (
select 1 union all select id + 1 from sequence where id < 10002
)
insert into count_numbers select id from sequence;
create table projects ( create table projects (
id integer primary key, id integer primary key,
title text not null, title text not null,
@ -336,7 +341,7 @@ def project_rows(datasette_server, **filters):
"_shape": "objects", "_shape": "objects",
**{key: str(value) for key, value in filters.items()}, **{key: str(value) for key, value in filters.items()},
} }
response = httpx.get(f"{datasette_server}data/projects.json", params=params) response = httpx2.get(f"{datasette_server}data/projects.json", params=params)
response.raise_for_status() response.raise_for_status()
return response.json()["rows"] return response.json()["rows"]
@ -348,7 +353,7 @@ def project_row(datasette_server, pk):
def binary_file_blob(datasette_server, pk): def binary_file_blob(datasette_server, pk):
response = httpx.get( response = httpx2.get(
f"{datasette_server}data/binary_files/{pk}.blob", f"{datasette_server}data/binary_files/{pk}.blob",
params={"_blob_column": "data"}, params={"_blob_column": "data"},
) )
@ -369,7 +374,7 @@ def bulk_default_rows(datasette_server, **filters):
"_shape": "objects", "_shape": "objects",
**{key: str(value) for key, value in filters.items()}, **{key: str(value) for key, value in filters.items()},
} }
response = httpx.get(f"{datasette_server}data/bulk_defaults.json", params=params) response = httpx2.get(f"{datasette_server}data/bulk_defaults.json", params=params)
response.raise_for_status() response.raise_for_status()
return response.json()["rows"] return response.json()["rows"]
@ -379,7 +384,7 @@ def upsert_item_rows(datasette_server, **filters):
"_shape": "objects", "_shape": "objects",
**{key: str(value) for key, value in filters.items()}, **{key: str(value) for key, value in filters.items()},
} }
response = httpx.get(f"{datasette_server}data/upsert_items.json", params=params) response = httpx2.get(f"{datasette_server}data/upsert_items.json", params=params)
response.raise_for_status() response.raise_for_status()
return response.json()["rows"] return response.json()["rows"]
@ -473,7 +478,7 @@ def test_create_table_flow(page, datasette_server):
page.wait_for_url("**/data/playwright_created") page.wait_for_url("**/data/playwright_created")
assert "playwright_created" in page.locator("h1").inner_text() assert "playwright_created" in page.locator("h1").inner_text()
response = httpx.get( response = httpx2.get(
f"{datasette_server}data/playwright_created.json?_extra=columns,column_types" f"{datasette_server}data/playwright_created.json?_extra=columns,column_types"
) )
response.raise_for_status() response.raise_for_status()
@ -487,7 +492,7 @@ def test_create_table_flow(page, datasette_server):
assert data["column_types"] == { assert data["column_types"] == {
"metadata": {"type": "json", "config": None}, "metadata": {"type": "json", "config": None},
} }
schema_response = httpx.get( schema_response = httpx2.get(
f"{datasette_server}data/-/query.json", f"{datasette_server}data/-/query.json",
params={ params={
"sql": ( "sql": (
@ -603,7 +608,7 @@ def test_create_table_from_data_flow(page, datasette_server):
dialog.locator(".table-create-save").click() dialog.locator(".table-create-save").click()
page.wait_for_url("**/data/playwright_from_data") page.wait_for_url("**/data/playwright_from_data")
response = httpx.get( response = httpx2.get(
f"{datasette_server}data/playwright_from_data.json?_shape=objects" f"{datasette_server}data/playwright_from_data.json?_shape=objects"
) )
response.raise_for_status() response.raise_for_status()
@ -639,7 +644,7 @@ def test_create_table_from_csv_keeps_numeric_type_when_values_are_blank(
dialog.locator(".table-create-save").click() dialog.locator(".table-create-save").click()
page.wait_for_url("**/data/playwright_numeric_blanks") page.wait_for_url("**/data/playwright_numeric_blanks")
response = httpx.get( response = httpx2.get(
f"{datasette_server}data/playwright_numeric_blanks.json?_shape=objects" f"{datasette_server}data/playwright_numeric_blanks.json?_shape=objects"
) )
response.raise_for_status() response.raise_for_status()
@ -648,7 +653,7 @@ def test_create_table_from_csv_keeps_numeric_type_when_values_are_blank(
{"name": "B", "score": None}, {"name": "B", "score": None},
] ]
schema_response = httpx.get( schema_response = httpx2.get(
f"{datasette_server}data/-/query.json", f"{datasette_server}data/-/query.json",
params={ params={
"sql": ( "sql": (
@ -856,7 +861,7 @@ def test_alter_table_flow(page, datasette_server):
columns = [] columns = []
for _ in range(20): for _ in range(20):
response = httpx.get(f"{datasette_server}data/projects.json?_extra=columns") response = httpx2.get(f"{datasette_server}data/projects.json?_extra=columns")
response.raise_for_status() response.raise_for_status()
columns = response.json()["columns"] columns = response.json()["columns"]
if "status" in columns: if "status" in columns:
@ -1604,3 +1609,45 @@ def test_delete_row_flow_removes_row(page, datasette_server):
page.locator(".row-mutation-status", has_text="Deleted row 1").wait_for() page.locator(".row-mutation-status", has_text="Deleted row 1").wait_for()
page.locator('tr[data-row="1"]').wait_for(state="detached") page.locator('tr[data-row="1"]').wait_for(state="detached")
assert project_rows(datasette_server, id=1) == [] assert project_rows(datasette_server, id=1) == []
@pytest.mark.playwright
def test_count_all(page, datasette_server):
page.goto(datasette_server + "data/count_numbers?id__gt=1&_sort=id")
assert page.locator(".table-count").inner_text() == "10,000+ rows"
with page.expect_response("**/count_numbers/-/count?*") as response:
page.get_by_role("button", name="count all", exact=True).click()
assert response.value.request.method == "POST"
assert response.value.request.post_data is None
assert "content-type" not in response.value.request.headers
assert response.value.json() == {"ok": True, "count": 10001}
page.wait_for_function(
'document.querySelector(".table-count").textContent === "10,001 rows"'
)
assert page.locator(".count-all").count() == 0
assert "id" in page.locator("h3").first.inner_text()
@pytest.mark.playwright
def test_count_all_error_retry(page, datasette_server):
page.goto(datasette_server + "data/count_numbers?id__gt=1")
page.route(
"**/count_numbers/-/count?*",
lambda route: route.fulfill(
status=400,
content_type="application/json",
body=json.dumps({"ok": False, "errors": ["Count query timed out"]}),
),
)
button = page.get_by_role("button", name="count all", exact=True)
button.click()
page.wait_for_function(
'document.querySelector(".count-error").textContent === "Count query timed out"'
)
assert button.is_enabled()
page.unroute("**/count_numbers/-/count?*")
button.click()
page.wait_for_function(
'document.querySelector(".table-count").textContent === "10,001 rows"'
)
assert page.locator(".count-error").inner_text() == ""

View file

@ -53,6 +53,13 @@ def test_hook_jump_items_sql():
assert "jump_items_sql" in dir(pm.hook) assert "jump_items_sql" in dir(pm.hook)
def test_hook_shutdown():
# Detailed behavior (ordering against background-task cancellation and
# close(), idempotency, exception handling, sync vs async support) is
# covered in tests/test_shutdown.py.
assert "shutdown" in dir(pm.hook)
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_hook_plugins_dir_plugin_prepare_connection(ds_client): async def test_hook_plugins_dir_plugin_prepare_connection(ds_client):
response = await ds_client.get( response = await ds_client.get(
@ -1300,7 +1307,8 @@ async def test_hook_filters_from_request(ds_client):
ds_client.ds.pm.register(ReturnNothingPlugin(), name="ReturnNothingPlugin") ds_client.ds.pm.register(ReturnNothingPlugin(), name="ReturnNothingPlugin")
response = await ds_client.get("/fixtures/facetable?_nothing=1") response = await ds_client.get("/fixtures/facetable?_nothing=1")
assert "0 rows\n where NOTHING" in response.text summary = Soup(response.text, "html.parser").select_one(".table-summary")
assert summary.get_text(" ", strip=True) == "0 rows where NOTHING"
json_response = await ds_client.get("/fixtures/facetable.json?_nothing=1") json_response = await ds_client.get("/fixtures/facetable.json?_nothing=1")
assert json_response.json()["rows"] == [] assert json_response.json()["rows"] == []
ds_client.ds.pm.unregister(name="ReturnNothingPlugin") ds_client.ds.pm.unregister(name="ReturnNothingPlugin")

View file

@ -0,0 +1,140 @@
"""Policy and compatibility coverage for PR #76, run against the fixed checkout."""
import uuid
import pytest
from datasette.app import Datasette
from datasette.resources import TableResource
from datasette.utils.sqlite import sqlite3, sqlite_derived_table_dependencies
@pytest.mark.parametrize("vocab_name", ["words", "name USING fts4aux", 'quoted"name'])
@pytest.mark.parametrize(
"module,arguments",
[
("fts5vocab", "'Search,Index', 'row'"),
("fts5vocab", "'SEARCH,INDEX', 'col'"),
("fts5vocab", "'Search,Index', 'instance'"),
("fts4aux", "'Search,Index'"),
],
)
def test_vocabulary_dependency_identity(module, arguments, vocab_name):
conn = sqlite3.connect(":memory:")
try:
fts = "fts5" if module == "fts5vocab" else "fts4"
conn.execute(f'create virtual table "Search,Index" using {fts}(body)')
quoted_name = '"' + vocab_name.replace('"', '""') + '"'
conn.execute(
f"create virtual table {quoted_name} USING /* module */ {module}({arguments})"
)
assert sqlite_derived_table_dependencies(conn)[vocab_name] == "Search,Index"
finally:
conn.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("module", ["fts5", "fts4"])
@pytest.mark.parametrize("external_content", [False, True], ids=["one-hop", "two-hop"])
@pytest.mark.parametrize(
"source_allowed,vocab_allowed", [(False, True), (True, False), (True, True)]
)
async def test_vocabulary_immediate_source_permissions(
module, external_content, source_allowed, vocab_allowed
):
ds = Datasette(
memory=True,
config={
"databases": {
"data": {
"tables": {
"search": {
"permissions": {
"view-table": (
{"id": "reader"} if source_allowed else False
)
}
},
"words": {"permissions": {"view-table": vocab_allowed}},
}
}
}
},
)
db = ds.add_memory_database(uuid.uuid4().hex, name="data")
await db.execute_write("create table documents(body text)")
options = "body, content='documents'" if external_content else "body"
await db.execute_write(f"create virtual table search using {module}({options})")
definition = (
"fts5vocab('SEARCH', 'row')" if module == "fts5" else "fts4aux('SEARCH')"
)
await db.execute_write(f"create virtual table words using {definition}")
await ds.invoke_startup()
try:
actor = {"id": "reader"}
expected = source_allowed and vocab_allowed and not external_content
for name in ("words", "WORDS"):
assert (
await ds.allowed(
action="view-table",
resource=TableResource("data", name),
actor=actor,
)
is expected
)
resources = await ds.allowed_resources(
"view-table", parent="data", actor=actor, include_is_private=True
)
words = [r for r in resources.resources if r.child == "words"]
assert bool(words) is expected
if expected:
assert words[0].private
assert not await ds.allowed(
action="view-table", resource=TableResource("data", "words")
)
# Dropping the source invalidates dependency metadata and remains denied.
await db.execute_write("drop table search")
assert not await ds.allowed(
action="view-table", resource=TableResource("data", "words"), actor=actor
)
finally:
ds.close()
@pytest.mark.parametrize(
"module,definition",
[
("fts5", "fts5vocab('main', 'search', 'row')"),
("fts4", "fts4aux('main', 'search')"),
],
)
def test_cross_schema_vocabulary_is_unresolved(module, definition):
conn = sqlite3.connect(":memory:")
try:
conn.execute(f"create virtual table search using {module}(body)")
conn.execute(f"create virtual table temp.words using {definition}")
# Cross-schema ownership is not representable by the current map.
# The source is itself derived, so the immediate-source policy denies it.
assert (
sqlite_derived_table_dependencies(conn, schema="temp")["words"] == "words"
)
finally:
conn.close()
@pytest.mark.parametrize(
"definition",
[
"""CREATE VIRTUAL TABLE"words"USING"fts5vocab"('search', 'row')""",
"""CREATE VIRTUAL TABLE[words]USING[fts5vocab]('search', 'row')""",
"""CREATE VIRTUAL TABLE`words`USING`fts5vocab`('search', 'row')""",
],
)
def test_vocabulary_quoted_token_boundaries(definition):
conn = sqlite3.connect(":memory:")
try:
conn.execute("create virtual table search using fts5(body)")
conn.execute(definition)
assert sqlite_derived_table_dependencies(conn)["words"] == "search"
finally:
conn.close()

View file

@ -0,0 +1,113 @@
"""Statistics access policy and plugin replacement coverage for PR #76."""
import uuid
import pytest
from datasette import hookimpl
from datasette.app import Datasette
from datasette.permissions import PermissionSQL
from datasette.resources import TableResource
@pytest.mark.asyncio
@pytest.mark.parametrize("scope", [None, "global", "database", "table", "root"])
async def test_statistics_denied_despite_allow_rules(scope):
config = {"databases": {"data": {"tables": {"sqlite_stat1": {}}}}}
grant = {"view-table": True}
if scope == "global":
config["permissions"] = grant
elif scope == "database":
config["databases"]["data"]["permissions"] = grant
elif scope == "table":
config["databases"]["data"]["tables"]["sqlite_stat1"]["permissions"] = grant
ds = Datasette(memory=True, config=config)
ds.root_enabled = scope == "root"
actor = {"id": "root"} if scope == "root" else {"id": "reader"}
db = ds.add_memory_database(uuid.uuid4().hex, name="data")
await db.execute_write("create table items(value text)")
await db.execute_write("create index items_value on items(value)")
await db.execute_write("insert into items values ('example')")
await db.execute_write("analyze")
await ds.invoke_startup()
try:
assert "view-sqlite-statistics" not in ds.actions
for name in ("sqlite_stat1", "SQLITE_STAT1"):
assert not await ds.allowed(
action="view-table", resource=TableResource("data", name), actor=actor
)
for suffix in ("", ".json", ".csv"):
assert (
await ds.client.get(f"/data/sqlite_stat1{suffix}", actor=actor)
).status_code == 403
resources = await ds.allowed_resources("view-table", parent="data", actor=actor)
assert "sqlite_stat1" not in {r.child for r in resources.resources}
assert "items" in {r.child for r in resources.resources}
finally:
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize(
"table", ["sqlite_stat1", "sqlite_stat2", "sqlite_stat3", "sqlite_stat4"]
)
@pytest.mark.parametrize("default_deny", [False, True])
async def test_statistics_names_denied(table, default_deny):
ds = Datasette(memory=True, default_deny=default_deny)
ds.root_enabled = True
await ds.invoke_startup()
try:
for name in (table, table.upper()):
assert not await ds.allowed(
action="view-table",
resource=TableResource("_memory", name),
actor={"id": "root"},
)
finally:
ds.close()
@pytest.mark.asyncio
async def test_plugin_can_replace_statistics_policy():
class ReplacementPolicy:
@hookimpl
def permission_resources_sql(self, action, actor):
if action == "view-table":
return PermissionSQL(
sql="SELECT 'data' AS parent, 'sqlite_stat1' AS child, :statistics_allowed AS allow, 'custom statistics policy' AS reason",
params={"statistics_allowed": int(actor == {"id": "reader"})},
)
ds = Datasette(memory=True)
db = ds.add_memory_database(uuid.uuid4().hex, name="data")
await db.execute_write("create table items(value text)")
await db.execute_write("analyze")
await ds.invoke_startup()
name = "datasette.default_permissions.sqlite_statistics"
original = ds.pm.unregister(name=name)
assert original is not None
replacement = ReplacementPolicy()
ds.pm.register(replacement, name="test-replacement-statistics-policy")
try:
actor = {"id": "reader"}
assert await ds.allowed(
action="view-table",
resource=TableResource("data", "sqlite_stat1"),
actor=actor,
)
assert not await ds.allowed(
action="view-table", resource=TableResource("data", "sqlite_stat1")
)
resources = await ds.allowed_resources(
"view-table", parent="data", actor=actor, include_is_private=True
)
stats = [r for r in resources.resources if r.child == "sqlite_stat1"]
assert len(stats) == 1 and stats[0].private
assert (
await ds.client.get("/data/sqlite_stat1.json", actor=actor)
).status_code == 200
assert (await ds.client.get("/data/sqlite_stat1.json")).status_code == 403
finally:
ds.pm.unregister(replacement)
ds.pm.register(original, name=name)
ds.close()

View file

@ -3248,74 +3248,6 @@ async def test_execute_write_create_table_uses_create_table_permission():
assert not await db.table_exists("should_not_exist") assert not await db.table_exists("should_not_exist")
@pytest.mark.asyncio
async def test_execute_write_create_view_uses_create_view_permission():
ds = Datasette(
memory=True,
default_deny=True,
config={
"permissions": {
"insert-row": {"id": "row-writer"},
"update-row": {"id": "row-writer"},
},
"databases": {
"data": {
"permissions": {
"view-database": {"id": ["creator", "row-writer"]},
"execute-write-sql": {"id": ["creator", "row-writer"]},
"create-view": {"id": "creator"},
}
}
},
},
)
db = ds.add_memory_database("execute_write_create_view", name="data")
await db.execute_write("create table dogs (id integer primary key, name text)")
await ds.invoke_startup()
analysis_response = await ds.client.get(
"/data/-/execute-write/analyze",
actor={"id": "creator"},
params={"sql": "create view dog_names as select id, name from dogs"},
)
allowed_response = await ds.client.post(
"/data/-/execute-write",
actor={"id": "creator"},
json={"sql": "create view dog_names as select id, name from dogs"},
)
row_permission_response = await ds.client.post(
"/data/-/execute-write",
actor={"id": "row-writer"},
json={"sql": "create view should_not_exist as select id from dogs"},
)
assert analysis_response.status_code == 200
analysis_data = analysis_response.json()
assert analysis_data["ok"] is True
assert analysis_data["execute_disabled"] is False
assert analysis_data["analysis_rows"] == [
{
"operation": "create",
"database": "data",
"table": "dog_names",
"required_permission": "create-view",
"source": None,
"allowed": True,
}
]
assert allowed_response.status_code == 200
assert allowed_response.json()["ok"] is True
assert allowed_response.json()["message"] == "Query executed"
assert await db.view_exists("dog_names")
assert row_permission_response.status_code == 403
assert row_permission_response.json()["errors"] == [
"Permission denied: need create-view on data"
]
assert not await db.view_exists("should_not_exist")
@pytest.mark.parametrize( @pytest.mark.parametrize(
( (
"database_name", "database_name",

View file

@ -246,3 +246,114 @@ async def test_table_not_exists(schema_ds):
response = await schema_ds.client.get("/schema_public_db/nonexistent/-/schema.md") response = await schema_ds.client.get("/schema_public_db/nonexistent/-/schema.md")
assert response.status_code == 404 assert response.status_code == 404
assert "not found" in response.text.lower() assert "not found" in response.text.lower()
@pytest_asyncio.fixture(scope="module")
async def schema_table_perms_ds():
"""
A database that is viewable by anonymous users, but with one table
locked down using the documented per-table lockdown recipe:
a table-level allow block combined with allow_sql: false.
"""
ds = Datasette(
config={
"databases": {
"schema_table_perms_db": {
"allow_sql": False,
"tables": {"employee_salaries": {"allow": {"id": "root"}}},
}
}
}
)
db = ds.add_memory_database("schema_table_perms_db")
await db.execute_write(
"CREATE TABLE IF NOT EXISTS public_posts (id INTEGER PRIMARY KEY, title TEXT)"
)
await db.execute_write(
"CREATE TABLE IF NOT EXISTS employee_salaries "
"(id INTEGER PRIMARY KEY, ssn TEXT, salary_usd INTEGER)"
)
await db.execute_write(
"CREATE INDEX IF NOT EXISTS idx_employee_salaries_ssn ON employee_salaries(ssn)"
)
await db.execute_write(
"CREATE TRIGGER IF NOT EXISTS trg_employee_salaries "
"AFTER INSERT ON employee_salaries BEGIN SELECT 1; END"
)
return ds
@pytest.mark.asyncio
async def test_schema_table_perms_controls(schema_table_perms_ds):
"""Sanity check: the locked down table really is denied to anonymous users."""
ds = schema_table_perms_ds
for path in (
"/schema_table_perms_db/employee_salaries.json",
"/schema_table_perms_db/employee_salaries/-/schema.json",
"/schema_table_perms_db/-/query.json?sql=select+*+from+employee_salaries",
):
response = await ds.client.get(path)
assert response.status_code == 403, path
response = await ds.client.get("/schema_table_perms_db.json")
assert response.status_code == 200
assert "employee_salaries" not in response.text
@pytest.mark.asyncio
@pytest.mark.parametrize(
"base_url",
["/-/schema", "/schema_table_perms_db/-/schema"],
)
@pytest.mark.parametrize("format_ext", ["json", "md", ""])
async def test_schema_parent_views_hide_denied_tables(
schema_table_perms_ds, base_url, format_ext
):
"""
GHSA-926p-cw2f-643h: /-/schema and /db/-/schema must not disclose the DDL
of tables the actor is denied view-table on, including indexes and
triggers that belong to those tables.
"""
url = base_url + (f".{format_ext}" if format_ext else "")
# Anonymous: allowed table visible, denied table (and its columns,
# index and trigger) absent
response = await schema_table_perms_ds.client.get(url)
assert response.status_code == 200
assert "public_posts" in response.text
assert "employee_salaries" not in response.text
assert "ssn" not in response.text
assert "salary_usd" not in response.text
assert "idx_employee_salaries_ssn" not in response.text
assert "trg_employee_salaries" not in response.text
# root can see everything
response = await schema_table_perms_ds.client.get(url, actor={"id": "root"})
assert response.status_code == 200
assert "public_posts" in response.text
assert "CREATE TABLE employee_salaries" in response.text
assert "idx_employee_salaries_ssn" in response.text
assert "trg_employee_salaries" in response.text
@pytest.mark.asyncio
@pytest.mark.parametrize(
"object_name", ["idx_employee_salaries_ssn", "trg_employee_salaries"]
)
@pytest.mark.parametrize("format_ext", ["json", "md", ""])
async def test_table_schema_does_not_serve_objects_of_denied_table(
schema_table_perms_ds, object_name, format_ext
):
"""
Related to GHSA-926p-cw2f-643h: /db/<name>/-/schema looks up sqlite_master
by name without restricting to tables/views, so requesting the name of an
index or trigger that belongs to a denied table serves its DDL. The
view-table check runs against the index/trigger name, which is not a
restricted table, so it passes.
"""
url = f"/schema_table_perms_db/{object_name}/-/schema"
if format_ext:
url += f".{format_ext}"
response = await schema_table_perms_ds.client.get(url)
assert response.status_code in (403, 404)
assert "employee_salaries" not in response.text
assert "ssn" not in response.text

318
tests/test_shutdown.py Normal file
View file

@ -0,0 +1,318 @@
"""
Tests for the shutdown(datasette) plugin hook and Datasette.invoke_shutdown().
Order under test: plugin `shutdown` hooks run first (while background tasks
are still alive) -> supervised background tasks are cancelled and drained
(fixed 5s grace) -> databases are closed. Hook exceptions are logged, never
propagated, and never skip a later step. The whole sequence is idempotent,
so a second lifespan.shutdown message (or any other second caller) is a
no-op.
"""
import asyncio
import contextlib
import logging
import pytest
from datasette import hookimpl
from datasette.app import Datasette
from datasette.plugins import pm
async def _drive_lifespan(app, messages):
"""Drive a single ASGI lifespan connection against `app`, feeding
`messages` to it in order via receive(). Returns once `app` itself
returns, which - per AsgiLifespan - happens after it has processed a
lifespan.shutdown message and sent back .complete or .failed. Returns
the list of messages `app` sent via send().
`messages` must end with a lifespan.shutdown (or a startup that fails)
or this will hang forever waiting for a message that never comes,
since AsgiLifespan only returns after handling shutdown.
"""
sent = []
idx = 0
async def receive():
nonlocal idx
if idx < len(messages):
message = messages[idx]
idx += 1
return message
# Real servers park here waiting for lifespan.shutdown; nothing
# left to deliver in this test, so just block - the caller is
# expected to have already gotten what it needs via a preceding
# lifespan.shutdown in `messages`.
await asyncio.Event().wait()
async def send(message):
sent.append(message)
await app({"type": "lifespan"}, receive, send)
return sent
@pytest.mark.asyncio
async def test_shutdown_hook_runs_before_task_cancellation_then_closes():
# Ordering proof: the shutdown hook observes the background task still
# "running" (hooks run BEFORE cancellation); after the whole lifespan
# drive completes, the task is "cancelled" and the Datasette instance
# is closed.
events = []
task_state_seen_in_hook = {}
async def bg_task(datasette):
await asyncio.Event().wait()
class LifecyclePlugin:
__name__ = "LifecyclePlugin"
@hookimpl
def startup(self, datasette):
async def inner():
datasette.add_background_task(bg_task, name="bg-task")
return inner
@hookimpl
def shutdown(self, datasette):
async def inner():
events.append("shutdown-hook-ran")
handle = datasette._background_tasks.tasks()[0]
task_state_seen_in_hook["state"] = handle.state
return inner
ds = Datasette(memory=True)
pm.register(LifecyclePlugin(), name="lifecycle_plugin")
try:
app = ds.app()
messages = await _drive_lifespan(
app, [{"type": "lifespan.startup"}, {"type": "lifespan.shutdown"}]
)
finally:
pm.unregister(name="lifecycle_plugin")
assert {"type": "lifespan.startup.complete"} in messages
assert {"type": "lifespan.shutdown.complete"} in messages
assert events == ["shutdown-hook-ran"]
assert task_state_seen_in_hook["state"] == "running"
handle = ds._background_tasks.tasks()[0]
assert handle.state == "cancelled"
assert ds._closed is True
assert ds._shutdown_invoked is True
@pytest.mark.asyncio
async def test_second_lifespan_shutdown_does_not_double_invoke():
call_count = {"n": 0}
class CountingShutdownPlugin:
__name__ = "CountingShutdownPlugin"
@hookimpl
def shutdown(self, datasette):
async def inner():
call_count["n"] += 1
return inner
ds = Datasette(memory=True)
pm.register(CountingShutdownPlugin(), name="counting_shutdown_plugin")
try:
app = ds.app()
messages = await _drive_lifespan(
app, [{"type": "lifespan.startup"}, {"type": "lifespan.shutdown"}]
)
assert {"type": "lifespan.shutdown.complete"} in messages
assert call_count["n"] == 1
# A second, separate lifespan connection (a misbehaving host, or a
# second embedder driving the same Datasette instance) sends
# lifespan.shutdown again. AsgiLifespan itself has no memory of
# the earlier connection, so this exercises invoke_shutdown()'s
# own idempotency guard, not anything ASGI-layer.
messages2 = await _drive_lifespan(app, [{"type": "lifespan.shutdown"}])
assert {"type": "lifespan.shutdown.complete"} in messages2
assert call_count["n"] == 1
finally:
pm.unregister(name="counting_shutdown_plugin")
assert ds._closed is True
@pytest.mark.asyncio
async def test_raising_shutdown_hook_is_logged_and_does_not_block_the_rest(caplog):
events = []
async def bg_task(datasette):
await asyncio.Event().wait()
class RaisingShutdownPlugin:
__name__ = "RaisingShutdownPlugin"
@hookimpl
def shutdown(self, datasette):
async def inner():
raise RuntimeError("boom from shutdown hook")
return inner
class WellBehavedPlugin:
__name__ = "WellBehavedPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
datasette.add_background_task(bg_task, name="bg-task")
return inner
@hookimpl
def shutdown(self, datasette):
async def inner():
events.append("well-behaved-ran")
return inner
ds = Datasette(memory=True)
pm.register(RaisingShutdownPlugin(), name="raising_shutdown_plugin")
pm.register(WellBehavedPlugin(), name="well_behaved_plugin")
try:
app = ds.app()
with caplog.at_level(logging.ERROR, logger="datasette"):
messages = await _drive_lifespan(
app, [{"type": "lifespan.startup"}, {"type": "lifespan.shutdown"}]
)
finally:
pm.unregister(name="raising_shutdown_plugin")
pm.unregister(name="well_behaved_plugin")
# The exception must never turn into lifespan.shutdown.failed - it's
# swallowed inside invoke_shutdown(), logged, and teardown continues.
assert {"type": "lifespan.shutdown.complete"} in messages
assert events == ["well-behaved-ran"]
assert "shutdown hook failed" in caplog.text
assert "boom from shutdown hook" in caplog.text
handle = ds._background_tasks.tasks()[0]
assert handle.state == "cancelled"
assert ds._closed is True
@pytest.mark.asyncio
async def test_sync_shutdown_hook_variant_works():
events = []
class SyncShutdownPlugin:
__name__ = "SyncShutdownPlugin"
@hookimpl
def shutdown(self, datasette):
# Deliberately not returning a coroutine/callable - a plain
# sync hookimpl, same as `def startup(datasette): ...` is
# supported via await_me_maybe.
events.append("sync-shutdown-ran")
ds = Datasette(memory=True)
pm.register(SyncShutdownPlugin(), name="sync_shutdown_plugin")
try:
app = ds.app()
messages = await _drive_lifespan(
app, [{"type": "lifespan.startup"}, {"type": "lifespan.shutdown"}]
)
finally:
pm.unregister(name="sync_shutdown_plugin")
assert {"type": "lifespan.shutdown.complete"} in messages
assert events == ["sync-shutdown-ran"]
assert ds._closed is True
@pytest.mark.asyncio
async def test_shutdown_logs_stragglers_that_outlive_the_grace_period(
caplog, monkeypatch
):
async def stubborn(datasette):
with contextlib.suppress(asyncio.CancelledError):
await asyncio.sleep(10)
# Swallowing CancelledError above and returning normally simulates
# a task that ignores cancellation for longer than the grace
# period - same shape as test_background_tasks.py's equivalent
# test.
await asyncio.sleep(10)
class StubbornTaskPlugin:
__name__ = "StubbornTaskPlugin"
@hookimpl
def startup(self, datasette):
async def inner():
datasette.add_background_task(stubborn, name="stubborn-task")
return inner
ds = Datasette(memory=True)
pm.register(StubbornTaskPlugin(), name="stubborn_task_plugin")
# invoke_shutdown() calls self._background_tasks.cancel_all(grace=5.0)
# with a grace hardcoded in app.py, per the ticket. Rather than
# sleeping for a real 5s in this test, monkeypatch the
# BackgroundTaskSupervisor *instance's* cancel_all to a wrapper that
# ignores the caller-supplied grace and substitutes a small one - this
# is the cleanest seam because it requires no production-code changes
# (no grace= setting/attribute to add) and leaves invoke_shutdown's
# own code under test untouched.
real_cancel_all = ds._background_tasks.cancel_all
async def fast_cancel_all(grace=5.0):
return await real_cancel_all(grace=0.1)
monkeypatch.setattr(ds._background_tasks, "cancel_all", fast_cancel_all)
try:
app = ds.app()
sent = []
queue = asyncio.Queue()
startup_complete = asyncio.Event()
async def receive():
return await queue.get()
async def send(message):
sent.append(message)
if message.get("type") == "lifespan.startup.complete":
startup_complete.set()
task = asyncio.create_task(app({"type": "lifespan"}, receive, send))
await queue.put({"type": "lifespan.startup"})
await asyncio.wait_for(startup_complete.wait(), timeout=5)
# Let the stubborn background task actually start running and
# reach its CancelledError-suppressing sleep before shutdown
# cancels it - a task cancelled before it has ever run its first
# step never enters that block at all, so it would finish
# cancelling immediately instead of behaving like a straggler.
await asyncio.sleep(0.05)
with caplog.at_level(logging.WARNING, logger="datasette.background_tasks"):
await queue.put({"type": "lifespan.shutdown"})
await asyncio.wait_for(task, timeout=5)
assert {"type": "lifespan.shutdown.complete"} in sent
assert "stubborn-task" in caplog.text
assert ds._closed is True
finally:
pm.unregister(name="stubborn_task_plugin")
# Clean up: the stubborn task ignored cancellation and is still
# sleeping past the shrunk grace period; actually cancel and await
# it now that the test has made its assertions, so it doesn't leak
# past the end of the test.
handles = ds._background_tasks.tasks()
if handles and handles[0].task and not handles[0].task.done():
handles[0].task.cancel()
with contextlib.suppress(asyncio.CancelledError):
await handles[0].task

View file

@ -208,11 +208,12 @@ def test_custom_params(stored_write_client):
) )
def test_stored_query_pages_no_vary_header(stored_write_client): def test_stored_query_pages_vary_by_credentials(stored_write_client):
# These pages no longer embed per-cookie CSRF tokens, so they must not # Even without per-cookie CSRF tokens, anonymous pages must not be reused
# set Vary: Cookie - they should be cacheable across users. # for authenticated users whose permissions or navigation can differ.
assert "vary" not in stored_write_client.get("/data").headers for path in ("/data", "/data/update_name"):
assert "vary" not in stored_write_client.get("/data/update_name").headers response = stored_write_client.get(path)
assert response.headers["vary"] == "Cookie, Authorization"
def test_json_post_body(stored_write_client): def test_json_post_body(stored_write_client):

View file

@ -59,6 +59,7 @@ async def test_success_object_has_ok_true(ds_client, path):
"/-/rules.json?action=view-instance", "/-/rules.json?action=view-instance",
"/-/check.json?action=view-instance", "/-/check.json?action=view-instance",
"/-/threads.json", "/-/threads.json",
"/-/tasks.json",
), ),
) )
async def test_permission_debug_success_has_ok_true(ds_envelope, path): async def test_permission_debug_success_has_ok_true(ds_envelope, path):

View file

@ -619,7 +619,10 @@ def test_searchmode(table_metadata, querystring, expected_rows):
], ],
), ),
( (
"/fixtures/searchable_view.json?_shape=arrays&_search=weasel&_fts_table=searchable_fts&_fts_pk=pk", (
"/fixtures/searchable_view_configured_by_metadata.json"
"?_shape=arrays&_search=weasel&_fts_table=searchable_fts&_fts_pk=pk"
),
[[2, "terry dog", "sara weasel", "puma"]], [[2, "terry dog", "sara weasel", "puma"]],
), ),
], ],
@ -1778,3 +1781,34 @@ async def test_next_url_included_by_default(ds_client):
data = response.json() data = response.json()
assert data["next"] is None assert data["next"] is None
assert data["next_url"] is None assert data["next_url"] is None
@pytest.mark.asyncio
async def test_table_through_requires_view_table_on_through_table():
# GHSA-53fc-rhfg-h7qp issue 3: ?_through= runs a sub-select against the
# caller-supplied through table, so the actor must be allowed to view it.
# Otherwise it is an equality oracle over any column of a denied table.
from datasette.app import Datasette
ds = Datasette(
memory=True,
config={"databases": {"data": {"tables": {"salaries": {"allow": False}}}}},
)
db = ds.add_memory_database("table_through_denied", name="data")
await db.execute_write("create table people (id integer primary key, name text)")
await db.execute_write(
"create table salaries (id integer primary key, "
"person_id integer references people(id), note text)"
)
await db.execute_write("insert into people values (1, 'alice'), (2, 'bob')")
await db.execute_write("insert into salaries values (1, 1, 'TOPSECRET-A')")
await ds.invoke_startup()
# Sanity: anonymous cannot read salaries directly
assert (await ds.client.get("/data/salaries.json")).status_code == 403
response = await ds.client.get(
"/data/people.json?_shape=array"
'&_through={"table":"salaries","column":"note","value":"TOPSECRET-A"}'
)
assert response.status_code == 403, response.text

View file

@ -270,7 +270,8 @@ async def test_empty_search_parameter_gets_removed(ds_client):
async def test_searchable_view_persists_fts_table(ds_client): async def test_searchable_view_persists_fts_table(ds_client):
# The search form should persist ?_fts_table as a hidden field # The search form should persist ?_fts_table as a hidden field
response = await ds_client.get( response = await ds_client.get(
"/fixtures/searchable_view?_fts_table=searchable_fts&_fts_pk=pk" "/fixtures/searchable_view_configured_by_metadata"
"?_fts_table=searchable_fts&_fts_pk=pk"
) )
inputs = Soup(response.text, "html.parser").find("form").find_all("input") inputs = Soup(response.text, "html.parser").find("form").find_all("input")
hiddens = [i for i in inputs if i["type"] == "hidden"] hiddens = [i for i in inputs if i["type"] == "hidden"]

View file

@ -0,0 +1,420 @@
"""Table permission identities must agree with SQLite identifier resolution."""
import uuid
from unittest.mock import AsyncMock
import pytest
from datasette import hookimpl
from datasette.app import Datasette
from datasette.default_permissions import restrictions_allow_action
from datasette.permissions import Action, PermissionSQL, _permission_check_cache
from datasette.resources import QueryResource, TableResource
from datasette.utils.actions_sql import explain_permission_for_resource
from datasette.utils.asgi import Forbidden
from datasette.utils.permissions import gather_permission_sql_from_hooks
@pytest.mark.asyncio
@pytest.mark.parametrize("kind", ["table", "view"])
@pytest.mark.parametrize("spelling", ["Inventory", "inventory", "INVENTORY"])
@pytest.mark.parametrize("allowed", [False, True])
@pytest.mark.parametrize("rule_spelling", ["Inventory", "iNvEnToRy"])
async def test_table_permission_identity(
kind, spelling, allowed, rule_spelling, monkeypatch
):
ds = Datasette(
config={
"permissions": {"view-table": not allowed, "insert-row": not allowed},
"databases": {
"data": {
"tables": {
rule_spelling: {
"permissions": {
"view-table": allowed,
"insert-row": allowed,
}
}
}
}
},
}
)
db = ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
cache_token = _permission_check_cache.set({})
try:
await db.execute_write(
"create table Inventory (id integer primary key)"
if kind == "table"
else "create view Inventory as select 1 as id"
)
await ds.invoke_startup()
# Identity matching needs no target-schema lookup. Derived-table
# permissions may still check the schema version. All spellings and
# API entry points should share the existing permission result cache.
target_execute = AsyncMock(wraps=db.execute)
monkeypatch.setattr(db, "execute", target_execute)
internal_execute = AsyncMock(wraps=ds.get_internal_database().execute)
monkeypatch.setattr(ds.get_internal_database(), "execute", internal_execute)
resource = TableResource("data", spelling)
assert await ds.allowed_many(
actions=["view-table", "insert-row"], resource=resource
) == {"view-table": allowed, "insert-row": allowed}
assert await ds.allowed(action="view-table", resource=resource) is allowed
assert await ds.check_visibility(None, "view-table", resource) == (
allowed,
False,
)
if allowed:
await ds.ensure_permission(action="view-table", resource=resource)
else:
with pytest.raises(Forbidden):
await ds.ensure_permission(action="view-table", resource=resource)
assert resource.child == spelling # Do not mutate caller-owned resources.
for variant in ("Inventory", "inventory", "INVENTORY"):
assert (
await ds.allowed(
action="view-table", resource=TableResource("data", variant)
)
is allowed
)
assert internal_execute.await_count == 1
assert all(
call.args[0] == "PRAGMA schema_version"
for call in target_execute.await_args_list
)
assert all(key[3] == "inventory" for key in _permission_check_cache.get())
finally:
_permission_check_cache.reset(cache_token)
ds.close()
@pytest.mark.asyncio
async def test_other_permission_identities_are_preserved():
ds = Datasette(
config={
"databases": {
"data": {
"tables": {
"Äpfel": {"permissions": {"view-table": False}},
"Future": {"permissions": {"view-table": False}},
},
"queries": {
"Report": {
"sql": "select 1",
"permissions": {"view-query": False},
},
"report": {
"sql": "select 1",
"permissions": {"view-query": True},
},
},
}
}
}
)
db = ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
try:
await db.execute_write('create table "Äpfel" (id integer primary key)')
await db.execute_write('create table "äpfel" (id integer primary key)')
await db.execute_write("create table Report (id integer primary key)")
await ds.invoke_startup()
# SQLite folds ASCII identifier casing, not Unicode casing.
for name, expected in [
("ÄPFEL", False),
("äPFEL", True),
("Future", False),
("future", False),
]:
assert (
await ds.allowed(
action="view-table", resource=TableResource("data", name)
)
is expected
)
# Query names remain case-sensitive even when a table has the same name.
for name, expected in [("Report", False), ("report", True)]:
assert (
await ds.allowed(
action="view-query", resource=QueryResource("data", name)
)
is expected
)
finally:
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("allow", [True, False, {"id": "reader"}])
async def test_table_listings_and_explanations(allow):
ds = Datasette(
config={
"databases": {
"data": {
"tables": {
"inventory": {"permissions": {"view-table": allow}},
}
}
}
}
)
db = ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
try:
await db.execute_write("create table Inventory (id integer primary key)")
await db.execute_write("create view InventoryView as select id from Inventory")
await ds.invoke_startup()
for actor in (None, {"id": "reader"}):
expected = allow is True or (isinstance(allow, dict) and actor == allow)
explanation = await explain_permission_for_resource(
datasette=ds,
actor=actor,
action="view-table",
parent="data",
child="INVENTORY",
)
assert explanation["allowed"] is expected
assert explanation["winning_scope"] == "resource"
assert any(
"data/inventory" in rule["reason"]
for rule in explanation["matched_rules"]
)
page = await ds.allowed_resources(
"view-table",
actor,
parent="data",
include_is_private=True,
include_reasons=True,
limit=1,
)
resources = [resource async for resource in page.all()]
matching = [r for r in resources if r.child == "Inventory"]
assert bool(matching) is expected
assert len(matching) <= 1
if matching:
assert matching[0].private is isinstance(allow, dict)
assert any(r.child == "InventoryView" for r in resources)
finally:
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("deny_first", [True, False])
async def test_case_variant_rules_deny_wins(deny_first):
rules = [("inventory", False), ("INVENTORY", True)]
if not deny_first:
rules.reverse()
ds = Datasette(
config={
"databases": {
"data": {
"tables": {
name: {"permissions": {"view-table": allow}}
for name, allow in rules
}
}
}
}
)
db = ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
try:
await db.execute_write("create table Inventory (id integer primary key)")
await ds.invoke_startup()
assert not await ds.allowed(
action="view-table", resource=TableResource("data", "Inventory")
)
assert not (
await ds.allowed_resources(
"view-table", parent="data", include_is_private=True
)
).resources
explanation = await explain_permission_for_resource(
datasette=ds,
actor=None,
action="view-table",
parent="data",
child="Inventory",
)
assert not explanation["allowed"]
assert any(
rule["effect"] == "allow" and not rule["decisive"]
for rule in explanation["matched_rules"]
)
assert any(
rule["effect"] == "deny" and rule["decisive"]
for rule in explanation["matched_rules"]
)
finally:
ds.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("config_style", ["allow", "permissions"])
@pytest.mark.parametrize("allowed", [True, False])
async def test_case_variant_token_restrictions(config_style, allowed):
table_config = (
{"allow": allowed}
if config_style == "allow"
else {"permissions": {"view-table": allowed}}
)
ds = Datasette(
config={"databases": {"data": {"tables": {"Inventory": table_config}}}}
)
db = ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
actor = {"id": "reader", "_r": {"r": {"data": {"inventory": ["vt"]}}}}
try:
await db.execute_write("create table Inventory (id integer primary key)")
await ds.invoke_startup()
assert restrictions_allow_action(
ds, actor["_r"], "view-table", ("data", "INVENTORY")
)
assert not restrictions_allow_action(
ds, actor["_r"], "view-table", ("Data", "Inventory")
)
assert (
await ds.allowed(
action="view-table",
resource=TableResource("data", "INVENTORY"),
actor=actor,
)
is allowed
)
page = await ds.allowed_resources("view-table", actor, parent="data")
assert [(r.parent, r.child) for r in page.resources] == (
[("data", "Inventory")] if allowed else []
)
explanation = await explain_permission_for_resource(
datasette=ds,
actor=actor,
action="view-table",
parent="data",
child="Inventory",
)
assert explanation["restriction_allowed"]
assert explanation["allowed"] is allowed
finally:
ds.close()
@pytest.mark.asyncio
async def test_plugin_restriction_intersection_and_dependencies():
class Plugin:
@hookimpl
def register_actions(self, datasette):
return [
Action(
name="inspect-inventory",
description="Inspect inventory",
resource_class=TableResource,
also_requires="view-table",
)
]
@hookimpl
def permission_resources_sql(self, action):
if action not in ("view-table", "inspect-inventory"):
return None
return [
PermissionSQL(
sql="SELECT 'data' AS parent, 'INVENTORY' AS child, 1 AS allow, 'inventory grant' AS reason",
restriction_sql="SELECT 'data' AS parent, 'inventory' AS child",
),
PermissionSQL(
restriction_sql="SELECT 'data' AS parent, 'InVeNtOrY' AS child"
),
]
ds = Datasette(default_deny=True)
ds.pm.register(Plugin(), name="identity-test")
db = ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
try:
await db.execute_write("create table Inventory (id integer primary key)")
await db.execute_write("create table Other (id integer primary key)")
await ds.invoke_startup()
for action in ("view-table", "inspect-inventory"):
assert await ds.allowed(
action=action, resource=TableResource("data", "Inventory")
)
assert not await ds.allowed(
action=action, resource=TableResource("data", "Other")
)
resources = (
await ds.allowed_resources(
action, parent="data", include_is_private=True
)
).resources
assert [r.child for r in resources] == ["Inventory"]
explanation = await explain_permission_for_resource(
datasette=ds,
actor=None,
action=action,
parent="data",
child="Inventory",
)
assert explanation["allowed"]
assert all(item["allowed"] for item in explanation["restrictions"])
finally:
ds.pm.unregister(name="identity-test")
ds.close()
@pytest.mark.asyncio
async def test_shared_plugin_rule_keeps_query_identity_and_original_sql():
shared = PermissionSQL(
sql="SELECT 'data' AS parent, 'Inventory' AS child, 0 AS allow, 'shared deny' AS reason"
)
original_sql = shared.sql
class Plugin:
@hookimpl
def permission_resources_sql(self, action):
if action in ("view-table", "view-query"):
return shared
ds = Datasette(
config={
"databases": {
"data": {
"queries": {
"Inventory": "select 1",
"inventory": "select 1",
}
}
}
}
)
ds.add_memory_database("identity_" + uuid.uuid4().hex, name="data")
ds.pm.register(Plugin(), name="identity-test")
try:
await ds.invoke_startup()
for _ in range(2):
await gather_permission_sql_from_hooks(
datasette=ds, actor=None, action="view-table"
)
assert shared.sql == original_sql
assert not await ds.allowed(
action="view-table", resource=TableResource("data", "inventory")
)
assert await ds.allowed(
action="view-query", resource=QueryResource("data", "inventory")
)
assert not await ds.allowed(
action="view-query", resource=QueryResource("data", "Inventory")
)
assert await ds.allowed(
action="view-table", resource=TableResource("Data", "Inventory")
)
assert restrictions_allow_action(
ds,
{"r": {"data": {"Inventory": ["vq"]}}},
"view-query",
("data", "Inventory"),
)
assert not restrictions_allow_action(
ds,
{"r": {"data": {"Inventory": ["vq"]}}},
"view-query",
("data", "inventory"),
)
finally:
ds.pm.unregister(name="identity-test")
ds.close()

View file

@ -0,0 +1,122 @@
"""
Tests for the /-/tasks introspection endpoint.
/-/tasks exposes datasette._background_tasks (see tests/test_background_tasks.py
for the supervisor machinery itself) the same way /-/threads exposes threading
internals: gated behind the permissions-debug permission, JSON-only.
"""
import asyncio
import contextlib
import functools
import pytest
from datasette.app import Datasette
async def example_task(datasette):
pass
class ExampleWorker:
async def run(self, datasette):
pass
async def __call__(self, datasette):
pass
@pytest.mark.asyncio
@pytest.mark.parametrize(
"func, qualified_name",
[
(example_task, "example_task"),
(functools.partial(example_task), "example_task"),
(ExampleWorker().run, "ExampleWorker.run"),
(ExampleWorker(), "ExampleWorker.__call__"),
],
)
async def test_task_function_path(func, qualified_name):
ds = Datasette(memory=True)
ds.root_enabled = True
handle = ds.add_background_task(func, name="custom-name")
try:
response = await ds.client.get("/-/tasks.json", actor={"id": "root"})
assert response.status_code == 200
task = response.json()["tasks"][0]
assert task["name"] == "custom-name"
assert task["function"] == f"{__name__}.{qualified_name}"
assert handle.function == task["function"]
assert "plugin" not in task
await handle.task
html = await ds.client.get("/-/tasks", actor={"id": "root"})
assert html.status_code == 200
assert task["function"] in html.text
finally:
await ds.invoke_shutdown()
@pytest.mark.asyncio
async def test_tasks_requires_permissions_debug():
ds = Datasette(memory=True)
ds.root_enabled = True
denied = await ds.client.get("/-/tasks.json")
assert denied.status_code == 403
allowed = await ds.client.get("/-/tasks.json", actor={"id": "root"})
assert allowed.status_code == 200
data = allowed.json()
assert data["ok"] is True
assert "tasks" in data
assert "launched" in data
@pytest.mark.asyncio
async def test_running_and_crashed_task_states():
ds = Datasette(memory=True)
ds.root_enabled = True
async def long_running(datasette):
await asyncio.Event().wait()
async def crashing_task(datasette):
raise RuntimeError("kaboom")
long_handle = ds.add_background_task(long_running, name="long-runner")
crash_handle = ds.add_background_task(crashing_task, name="crashing_task")
await ds.start_background_tasks()
# Let the crashing_task run to completion and its done-callback (which sets
# handle.state = "crashed") actually fire before we read state back out.
await asyncio.wait_for(
asyncio.gather(crash_handle.task, return_exceptions=True), timeout=5
)
await asyncio.sleep(0)
try:
response = await ds.client.get("/-/tasks.json", actor={"id": "root"})
assert response.status_code == 200
data = response.json()
assert data["launched"] is True
by_name = {t["name"]: t for t in data["tasks"]}
assert by_name["long-runner"]["state"] == "running"
assert by_name["long-runner"]["exception"] is None
assert by_name["long-runner"]["started_at"] is not None
crashed = by_name["crashing_task"]
assert crashed["function"] == (
f"{__name__}.test_running_and_crashed_task_states.<locals>.crashing_task"
)
assert crashed["state"] == "crashed"
assert crashed["exception"] is not None
assert isinstance(crashed["exception"], str)
assert "kaboom" in crashed["exception"]
assert "RuntimeError" in crashed["exception"]
finally:
long_handle.cancel()
with contextlib.suppress(asyncio.CancelledError):
await long_handle.task

View file

@ -16,6 +16,7 @@ from datasette.app import Datasette
from datasette.utils.asgi import Request from datasette.utils.asgi import Request
from datasette.utils.sqlite import ( from datasette.utils.sqlite import (
sqlite3, sqlite3,
sqlite_derived_table_dependencies,
sqlite_hidden_table_names, sqlite_hidden_table_names,
sqlite_table_type, sqlite_table_type,
supports_returning, supports_returning,
@ -369,6 +370,46 @@ def test_sqlite_hidden_table_names_hides_multiline_content_fts_table():
conn.close() conn.close()
def test_sqlite_derived_table_dependencies():
conn = utils.sqlite3.connect(":memory:")
try:
conn.executescript("""
create table docs(id integer primary key, body text);
create virtual table external_fts5 using fts5(
body, content='docs', content_rowid='id'
);
create virtual table internal_fts5 using fts5(body);
create virtual table contentless_fts5 using fts5(body, content='');
create virtual table external_fts4 using fts4(body, content="docs");
create virtual table internal_fts4 using fts4(body);
create virtual table contentless_fts4 using fts4(body, content="");
create table [docs, archive](body text);
create virtual table commented_fts5 using fts5(
body, tokenize='porter unicode61',
/* Comments and commas in quoted values must not confuse parsing. */
content='docs, archive'
);
create virtual table boxes using rtree(id, minx, maxx, miny, maxy);
""")
dependencies = sqlite_derived_table_dependencies(conn)
assert dependencies["external_fts5"] == "docs"
assert dependencies["external_fts4"] == "docs"
assert dependencies["commented_fts5"] == "docs, archive"
assert "contentless_fts5" not in dependencies
assert "contentless_fts4" not in dependencies
assert dependencies["internal_fts5_content"] == "internal_fts5"
assert dependencies["internal_fts4_content"] == "internal_fts4"
assert dependencies["external_fts5_data"] == "external_fts5"
assert dependencies["external_fts4_segments"] == "external_fts4"
assert dependencies["boxes_node"] == "boxes"
assert dependencies["boxes_parent"] == "boxes"
assert dependencies["boxes_rowid"] == "boxes"
finally:
conn.close()
@pytest.mark.parametrize( @pytest.mark.parametrize(
"url,expected", "url,expected",
[ [

View file

@ -439,7 +439,7 @@ def test_analyze_attached_database_tables(conn):
} }
def test_analyze_clears_authorizer_on_error(): def test_analyze_disables_authorizer_on_error():
class FakeConnection: class FakeConnection:
def __init__(self): def __init__(self):
self.authorizers = [] self.authorizers = []
@ -455,4 +455,5 @@ def test_analyze_clears_authorizer_on_error():
with pytest.raises(sqlite3.OperationalError): with pytest.raises(sqlite3.OperationalError):
analyze_sql_tables(conn, "bad SQL") analyze_sql_tables(conn, "bad SQL")
assert conn.authorizers[-1] is None final_authorizer = conn.authorizers[-1]
assert final_authorizer is None or final_authorizer() == sqlite3.SQLITE_OK