Compare commits

..

9 commits

Author SHA1 Message Date
Claude
ca995df664
Add busy_timeout_ms and journal_mode to expected /-/settings.json
Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 06:18:42 +00:00
Claude
f6ded9af75
Document how Datasette manages transactions
Adds an internals documentation section describing the task-equals-
transaction model for write connections, the autocommit read path, and
the cross-project constraint that Datasette and sqlite-utils both rely
on the legacy sqlite3 transaction handling - Python 3.12+ autocommit=
style connections are not supported and any migration would need to be
coordinated across both projects.

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 06:11:39 +00:00
Claude
3713d6c0d5
Add opt-in journal_mode setting, enable WAL on persistent internal DB
New journal_mode setting lets deployments opt mutable database files
into WAL mode (or delete/truncate/persist), applied on the write
connection. WAL is paired with PRAGMA synchronous=NORMAL. Datasette
does not change the journal mode of database files by default.

Also fixes an inconsistency: a persistent internal database passed via
--internal now gets WAL enabled, matching the temporary internal
database default (which was moved to a temp disk file specifically so
it could use WAL).

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 06:10:15 +00:00
Claude
14815cb092
Add busy_timeout_ms setting
The SQLite busy timeout was previously an implicit policy - every
connection inherited the sqlite3 driver's silent 5 second default. It
is now an explicit, documented setting passed as timeout= to every
sqlite3.connect() call. The default remains 5000ms.

This matters for deployments where external processes write to the
same database files Datasette is serving.

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 06:06:57 +00:00
Claude
5cec9c9faa
Stop running sqlite-utils plugins on Datasette connections
Wrapping a connection in sqlite_utils.Database() runs sqlite-utils
plugins' prepare_connection hooks against it by default. Datasette's
write API views and introspection helpers now pass execute_plugins=False
(matching what utils/internal_db.py already did), so third-party
sqlite-utils plugins no longer touch Datasette's connections.

Also apply PRAGMA recursive_triggers=on in Datasette._prepare_connection
so every connection gets consistent trigger semantics - previously only
the write connection got it, as a side effect of the first sqlite-utils
based write.

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 06:03:42 +00:00
Claude
a3f8b440e6
Rebuild internal catalog for a database in a single atomic write task
populate_schema_tables() previously ran six separate transactions per
database: one task deleting that database's catalog rows followed by
five execute_write_many inserts. Readers of the internal database could
observe the intermediate state where a database had no catalog rows.

Schema details are still collected on a read connection of the target
database first; the delete-and-reinsert now happens inside one
execute_write_fn task so the rebuild is atomic.

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 05:59:20 +00:00
Claude
26d326c709
Write API atomicity regression tests, remove manual transaction in alter
Adds regression tests confirming the JSON write API is atomic per
request now that write tasks open an explicit transaction: /db/-/create
with failing initial rows creates no table, a failing operation in
/db/table/-/alter rolls back earlier operations, and insert with
"return": true rolls back all rows if one fails.

Also removes the "with operation_conn:" block from the alter endpoint -
write functions run inside the task transaction and should not manage
transactions themselves (that context manager would commit the task
transaction early on success).

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 05:57:32 +00:00
Claude
1ff4e67b79
Make execute_write_script transactional, matching its documentation
execute_write_script() was documented as running inside a transaction
but actually passed transaction=False and used conn.executescript(),
which commits each statement as it executes - a failing script could
half-apply.

Scripts are now split into complete statements (via
sqlite3.complete_statement) and executed one at a time inside the task
transaction, so a failing script applies nothing. Scripts containing
statements that cannot run in a transaction (VACUUM, ATTACH, DETACH,
PRAGMA) or that manage transactions themselves (BEGIN, COMMIT,
SAVEPOINT etc) keep the previous executescript() autocommit behavior.

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 05:54:02 +00:00
Claude
84fb4221ab
Open write task transactions explicitly with BEGIN IMMEDIATE
Write tasks with transaction=True previously relied on the sqlite3
driver's implicit BEGIN, which only fires on the first raw
data-modifying statement. sqlite-utils 4.0 write methods found no open
transaction, committed their own work mid-task, and Datasette's
commit/rollback at task end was a no-op - so a failing write function
could leave partial writes permanently committed.

The write thread (and the non-threaded write path) now executes BEGIN
IMMEDIATE before invoking each transaction=True task, commits when it
returns and rolls back if it raises. sqlite-utils methods nest inside
that transaction as savepoints, restoring task-level atomicity for
every write path.

execute_write() now detects statements SQLite refuses to run inside a
transaction (VACUUM, ATTACH, DETACH, PRAGMA) and runs those in
autocommit mode, preserving previous behavior for e.g. trusted canned
queries that run VACUUM.

Refs #2831

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N76afGMhBRQk528VF1LTpR
2026-07-09 05:50:26 +00:00
34 changed files with 1085 additions and 1234 deletions

View file

@ -220,6 +220,16 @@ SETTINGS = (
"Number of threads in the thread pool for executing SQLite queries",
),
Setting("sql_time_limit_ms", 1000, "Time limit for a SQL query in milliseconds"),
Setting(
"busy_timeout_ms",
5000,
"How long SQLite waits for a locked database file before giving up",
),
Setting(
"journal_mode",
"",
"Journal mode to set on mutable database files, e.g. wal - leave blank to use each file's existing mode",
),
Setting(
"default_facet_size", 30, "Number of values to return for requested facets"
),
@ -478,7 +488,11 @@ class Datasette:
if internal is None:
self._internal_database = Database(self, is_temp_disk=True)
else:
self._internal_database = Database(self, path=internal, mode="rwc")
# WAL for the same reason as the temporary internal database:
# the catalog can be read while it is being rewritten
self._internal_database = Database(
self, path=internal, mode="rwc", enable_wal=True
)
self._internal_database.name = INTERNAL_DB_NAME
self.cache_headers = cache_headers
@ -753,7 +767,19 @@ class Datasette:
# Compare schema versions to see if we should skip it
if schema_version == current_schema_versions.get(database_name):
continue
await populate_schema_tables(internal_db, db, schema_version)
placeholders = "(?, ?, ?, ?)"
values = [database_name, str(db.path), db.is_memory, schema_version]
if db.path is None:
placeholders = "(?, null, ?, ?)"
values = [database_name, db.is_memory, schema_version]
await internal_db.execute_write(
"""
INSERT OR REPLACE INTO catalog_databases (database_name, path, is_memory, schema_version)
VALUES {}
""".format(placeholders),
values,
)
await populate_schema_tables(internal_db, db)
@property
def urls(self):
@ -1553,6 +1579,10 @@ class Datasette:
conn.execute("SELECT load_extension(?)", [extension])
if self.setting("cache_size_kb"):
conn.execute(f"PRAGMA cache_size=-{self.setting('cache_size_kb')}")
# Consistent trigger semantics on every connection - sqlite-utils
# would otherwise enable this on just the write connection, as a
# side effect of the first sqlite-utils based write
conn.execute("PRAGMA recursive_triggers=on")
# pylint: disable=no-member
if database != INTERNAL_DB_NAME:
pm.hook.prepare_connection(conn=conn, database=database, datasette=self)
@ -2563,7 +2593,7 @@ class Datasette:
JsonDataView.as_view(
self,
"plugins.json",
self._plugins,
lambda request: {"plugins": self._plugins(request)},
needs_request=True,
),
r"/-/plugins(\.(?P<format>json))?$",

View file

@ -5,6 +5,7 @@ import inspect
import os
from pathlib import Path
import queue
import re
import sqlite_utils
import sys
import tempfile
@ -17,7 +18,6 @@ from .utils import (
detect_fts,
detect_primary_keys,
detect_spatialite,
escape_sqlite,
get_all_foreign_keys,
get_outbound_foreign_keys,
md5_not_usedforsecurity,
@ -36,6 +36,8 @@ EXECUTE_WRITE_RETURNING_LIMIT = 10
AttachedDatabase = namedtuple("AttachedDatabase", ("seq", "name", "file"))
ALLOWED_JOURNAL_MODES = {"delete", "truncate", "persist", "wal"}
class DatasetteClosedError(RuntimeError):
"""Raised when using a Datasette or Database instance after close()."""
@ -43,6 +45,77 @@ class DatasetteClosedError(RuntimeError):
_SHUTDOWN = object()
# Statements that SQLite refuses to execute inside a transaction. These run
# without the task transaction - a single statement needs no extra atomicity.
_STATEMENTS_DISALLOWED_IN_TRANSACTION = {"vacuum", "attach", "detach", "pragma"}
_FIRST_KEYWORD_RE = re.compile(
r"^(?:\s+|--[^\n]*(?:\n|$)|/\*.*?\*/)*([a-zA-Z]+)", re.DOTALL
)
def _can_execute_in_transaction(sql):
match = _FIRST_KEYWORD_RE.match(sql)
if match is None:
return True
return match.group(1).lower() not in _STATEMENTS_DISALLOWED_IN_TRANSACTION
# Scripts that manage their own transactions also cannot run inside the
# task transaction
_SCRIPT_STATEMENTS_DISALLOWED_IN_TRANSACTION = _STATEMENTS_DISALLOWED_IN_TRANSACTION | {
"begin",
"commit",
"end",
"rollback",
"savepoint",
"release",
}
def _script_can_execute_in_transaction(statements):
for statement in statements:
match = _FIRST_KEYWORD_RE.match(statement)
if (
match is not None
and match.group(1).lower() in _SCRIPT_STATEMENTS_DISALLOWED_IN_TRANSACTION
):
return False
return True
def _iter_sql_statements(sql):
# Split a multi-statement SQL string into complete statements using
# sqlite3.complete_statement()
statement = []
for char in sql:
statement.append(char)
statement_sql = "".join(statement).strip()
if statement_sql and sqlite3.complete_statement(statement_sql):
yield statement_sql
statement = []
remainder = "".join(statement).strip()
if remainder:
yield remainder
def _run_write_in_transaction(conn, fn):
# Open the transaction explicitly instead of relying on the sqlite3
# driver's implicit BEGIN, which only fires on the first data-modifying
# statement. With a transaction genuinely open, sqlite-utils write
# methods nest inside it as savepoints instead of committing their own
# transactions mid-task: https://github.com/simonw/datasette/issues/2831
conn.execute("BEGIN IMMEDIATE")
try:
result = fn(conn)
except Exception:
if conn.in_transaction:
conn.rollback()
raise
if conn.in_transaction:
conn.commit()
return result
class Database:
# For table counts stop at this many rows:
@ -58,6 +131,7 @@ class Database:
memory_name=None,
mode=None,
is_temp_disk=False,
enable_wal=False,
):
self.name = None
self._thread_local_id = f"x{self._thread_local_id_counter}"
@ -69,6 +143,8 @@ class Database:
self.is_memory = is_memory
self.memory_name = memory_name
self.is_temp_disk = is_temp_disk
self.enable_wal = enable_wal or is_temp_disk
self._wal_enabled = False
if memory_name is not None:
self.is_memory = True
if is_temp_disk:
@ -77,10 +153,7 @@ class Database:
self.path = temp_path
self.is_mutable = True
self.mode = "rwc"
self._wal_enabled = False
atexit.register(self._cleanup_temp_file)
else:
self._wal_enabled = False
self.cached_hash = None
self.cached_size = None
self._cached_table_counts = None
@ -139,6 +212,10 @@ class Database:
extra_kwargs = {}
if write:
extra_kwargs["isolation_level"] = "IMMEDIATE"
# An explicit busy timeout policy rather than the driver's silent
# 5 second default - matters when external processes write to the
# same database files
extra_kwargs["timeout"] = self.ds.setting("busy_timeout_ms") / 1000
if self.memory_name:
uri = "file:{}?mode=memory&cache=shared".format(self.memory_name)
conn = sqlite3.connect(
@ -166,9 +243,23 @@ class Database:
f"file:{self.path}{qs}", uri=True, check_same_thread=False, **extra_kwargs
)
self._all_file_connections.append(conn)
if self.is_temp_disk and not self._wal_enabled:
if self.enable_wal and not self._wal_enabled:
conn.execute("PRAGMA journal_mode=WAL")
self._wal_enabled = True
if write and self.is_mutable and not self.enable_wal:
journal_mode = self.ds.setting("journal_mode")
if journal_mode:
if journal_mode not in ALLOWED_JOURNAL_MODES:
raise ValueError(
"journal_mode setting must be one of: {}".format(
", ".join(sorted(ALLOWED_JOURNAL_MODES))
)
)
conn.execute("PRAGMA journal_mode={}".format(journal_mode))
if journal_mode == "wal":
# The standard WAL pairing - fewer fsyncs, application
# level consistency guarantees are unchanged
conn.execute("PRAGMA synchronous=NORMAL")
return conn
def close(self):
@ -247,7 +338,6 @@ class Database:
request=None,
return_all=False,
returning_limit=EXECUTE_WRITE_RETURNING_LIMIT,
transaction=True,
):
self._check_not_closed()
if returning_limit < 0:
@ -261,19 +351,30 @@ class Database:
with trace("sql", database=self.name, sql=sql.strip(), params=params):
results = await self.execute_write_fn(
_inner, block=block, request=request, transaction=transaction
_inner,
block=block,
transaction=_can_execute_in_transaction(sql),
request=request,
)
return results
async def execute_write_script(self, sql, block=True, request=None):
self._check_not_closed()
statements = list(_iter_sql_statements(sql))
transaction = _script_can_execute_in_transaction(statements)
def _inner(conn):
return conn.executescript(sql)
if transaction:
# Execute statements one at a time so they run inside the
# task transaction - conn.executescript() would commit it
for statement in statements:
conn.execute(statement)
else:
return conn.executescript(sql)
with trace("sql", database=self.name, sql=sql.strip(), executescript=True):
results = await self.execute_write_fn(
_inner, block=block, transaction=False, request=request
_inner, block=block, transaction=transaction, request=request
)
return results
@ -351,9 +452,7 @@ class Database:
self._write_connection = self.connect(write=True)
self.ds._prepare_connection(self._write_connection, self.name)
if transaction:
with self._write_connection:
self._write_connection.execute("BEGIN IMMEDIATE")
result = fn(self._write_connection)
result = _run_write_in_transaction(self._write_connection, fn)
else:
result = fn(self._write_connection)
else:
@ -481,9 +580,7 @@ class Database:
else:
try:
if task.transaction:
with conn:
conn.execute("BEGIN IMMEDIATE")
result = task.fn(conn)
result = _run_write_in_transaction(conn, task.fn)
else:
result = task.fn(conn)
except Exception as e:
@ -609,7 +706,7 @@ class Database:
try:
table_count = (
await self.execute(
f"select count(*) from (select * from {escape_sqlite(table)} limit {self.count_limit + 1})",
f"select count(*) from (select * from [{table}] limit {self.count_limit + 1})",
custom_time_limit=limit,
)
).rows[0][0]
@ -680,7 +777,7 @@ class Database:
def column_details(conn):
# Returns {column_name: (type, is_unique)}
db = sqlite_utils.Database(conn)
db = sqlite_utils.Database(conn, execute_plugins=False)
columns = db[table].columns_dict
indexes = db[table].indexes
details = {}

View file

@ -96,10 +96,6 @@ class ConfigPermissionProcessor:
"""Evaluate an allow block against the current actor."""
if allow_block is None:
return None
# Values passed using ``-s permissions.* 1`` or ``0`` are parsed as
# integers, but should retain the CLI's boolean 1/0 behavior.
if isinstance(allow_block, int) and allow_block in (0, 1):
return bool(allow_block)
return actor_matches_allow(self.actor, allow_block)
def is_in_restriction_allowlist(

View file

@ -85,7 +85,7 @@ class Facet:
self.database = database
# For foreign key expansion. Can be None for e.g. stored SQL queries:
self.table = table
self.sql = sql or f"select * from {escape_sqlite(table)}"
self.sql = sql or f"select * from [{table}]"
self.params = params or []
self.table_config = table_config
# row_count can be None, in which case we calculate it ourselves:

View file

@ -6,20 +6,8 @@
padding: 1.5em;
margin-bottom: 2em;
}
.permission-form form {
max-width: 60rem;
}
.permission-form-grid {
display: grid;
gap: 1.5rem;
grid-template-columns: repeat(2, minmax(0, 1fr));
}
.permission-form-result {
margin-top: 1rem;
max-width: 60rem;
}
.form-section {
margin-bottom: 1.25em;
margin-bottom: 1em;
}
.form-section label {
display: block;
@ -27,51 +15,22 @@
font-weight: bold;
}
.form-section input[type="text"],
.form-section input[type="number"],
.form-section select,
.permission-textarea {
background-color: #fff;
border: 1px solid #aaa;
border-radius: 4px;
box-sizing: border-box;
box-shadow: inset 0 1px 2px rgba(0, 0, 0, 0.08);
color: #222;
font-family: inherit;
font-size: 1rem;
line-height: 1.4;
max-width: none;
width: 100%;
}
.form-section input[type="text"] {
height: 3rem;
padding: 0.6rem 0.75rem;
}
.form-section input[type="number"] {
height: 3rem;
max-width: 7rem;
padding: 0.6rem 0.75rem;
}
.form-section select {
height: 3rem;
padding: 0.6rem 0.75rem;
}
.permission-textarea {
font-family: monospace;
min-height: 12rem;
padding: 0.75rem;
resize: vertical;
width: 100%;
max-width: 500px;
padding: 0.5em;
box-sizing: border-box;
border: 1px solid #ccc;
border-radius: 3px;
}
.form-section input[type="text"]:focus,
.form-section input[type="number"]:focus,
.form-section select:focus,
.permission-textarea:focus {
.form-section select:focus {
outline: 2px solid #0066cc;
border-color: #0066cc;
box-shadow: 0 0 0 3px rgba(0, 102, 204, 0.18);
outline: none;
}
.form-section small {
display: block;
margin-top: 0.45em;
margin-top: 0.3em;
color: #666;
}
.form-actions {
@ -183,9 +142,4 @@
text-align: center;
color: #666;
}
@media only screen and (max-width: 576px) {
.permission-form-grid {
grid-template-columns: minmax(0, 1fr);
}
}
</style>

View file

@ -44,10 +44,10 @@
</style>
<nav class="permissions-debug-tabs">
<a href="{{ urls.path('-/check') }}{{ query_string }}" {% if current_tab == "check" %}class="active"{% endif %}>Explain</a>
<a href="{{ urls.path('-/allowed') }}{{ query_string }}" {% if current_tab == "allowed" %}class="active"{% endif %}>Access map</a>
<a href="{{ urls.path('-/rules') }}{{ query_string }}" {% if current_tab == "rules" %}class="active"{% endif %}>Rule explorer</a>
<a href="{{ urls.path('-/permissions') }}" {% if current_tab == "permissions" %}class="active"{% endif %}>Activity</a>
<a href="{{ urls.path('-/permissions') }}" {% if current_tab == "permissions" %}class="active"{% endif %}>Playground</a>
<a href="{{ urls.path('-/check') }}{{ query_string }}" {% if current_tab == "check" %}class="active"{% endif %}>Check</a>
<a href="{{ urls.path('-/allowed') }}{{ query_string }}" {% if current_tab == "allowed" %}class="active"{% endif %}>Allowed</a>
<a href="{{ urls.path('-/rules') }}{{ query_string }}" {% if current_tab == "rules" %}class="active"{% endif %}>Rules</a>
<a href="{{ urls.path('-/actions') }}" {% if current_tab == "actions" %}class="active"{% endif %}>Actions</a>
<a href="{{ urls.path('-/allow-debug') }}" {% if current_tab == "allow_debug" %}class="active"{% endif %}>Allow debug</a>
</nav>

View file

@ -3,11 +3,29 @@
{% block title %}Debug allow rules{% endblock %}
{% block extra_head %}
{% include "_permission_ui_styles.html" %}
<style>
textarea {
height: 10em;
width: 95%;
box-sizing: border-box;
padding: 0.5em;
border: 2px dotted black;
}
.two-col {
display: inline-block;
width: 48%;
}
.two-col label {
width: 48%;
}
p.message-warning {
white-space: pre-wrap;
}
@media only screen and (max-width: 576px) {
.two-col {
width: 100%;
}
}
</style>
{% endblock %}
@ -20,28 +38,24 @@ p.message-warning {
<p>Use this tool to try out different actor and allow combinations. See <a href="https://docs.datasette.io/en/stable/authentication.html#defining-permissions-with-allow-blocks">Defining permissions with "allow" blocks</a> for documentation.</p>
<div class="permission-form">
<form class="core" action="{{ urls.path('-/allow-debug') }}" method="get">
<div class="permission-form-grid">
<div class="form-section">
<label for="allow-block">Allow block</label>
<textarea class="permission-textarea" id="allow-block" name="allow">{{ allow_input }}</textarea>
</div>
<div class="form-section">
<label for="allow-actor">Actor</label>
<textarea class="permission-textarea" id="allow-actor" name="actor">{{ actor_input }}</textarea>
</div>
</div>
<div class="form-actions">
<button type="submit" class="submit-btn">Apply allow block to actor</button>
</div>
</form>
<form class="core" action="{{ urls.path('-/allow-debug') }}" method="get" style="margin-bottom: 1em">
<div class="two-col">
<p><label>Allow block</label></p>
<textarea name="allow">{{ allow_input }}</textarea>
</div>
<div class="two-col">
<p><label>Actor</label></p>
<textarea name="actor">{{ actor_input }}</textarea>
</div>
<div style="margin-top: 1em;">
<input type="submit" value="Apply allow block to actor">
</div>
</form>
{% if error %}<p class="message-warning permission-form-result">{{ error }}</p>{% endif %}
{% if error %}<p class="message-warning">{{ error }}</p>{% endif %}
{% if result == "True" %}<p class="message-info permission-form-result">Result: allow</p>{% endif %}
{% if result == "True" %}<p class="message-info">Result: allow</p>{% endif %}
{% if result == "False" %}<p class="message-error permission-form-result">Result: deny</p>{% endif %}
</div>
{% if result == "False" %}<p class="message-error">Result: deny</p>{% endif %}
{% endblock %}

View file

@ -49,7 +49,7 @@
<div class="form-section">
<label for="page_size">Page size:</label>
<input type="number" id="page_size" name="_size" value="50" min="1" max="200">
<input type="number" id="page_size" name="_size" value="50" min="1" max="200" style="max-width: 100px;">
<small>Number of results per page (max 200)</small>
</div>

View file

@ -1,6 +1,6 @@
{% extends "base.html" %}
{% block title %}Explain a permission decision{% endblock %}
{% block title %}Permission Check{% endblock %}
{% block extra_head %}
<script src="{{ static('json-format-highlight-1.0.1.js') }}"></script>
@ -13,35 +13,29 @@
border-radius: 5px;
}
#output.allowed {
background-color: #f3fbf4;
background-color: #e8f5e9;
border: 2px solid #4caf50;
}
#output.denied {
background-color: #fff7f7;
background-color: #ffebee;
border: 2px solid #f44336;
}
#output h2 {
margin-top: 0;
}
#output h3 {
margin-bottom: 0.5em;
}
#output .result-badge,
.effect-badge,
.rule-status {
#output .result-badge {
display: inline-block;
padding: 0.2em 0.5em;
padding: 0.3em 0.8em;
border-radius: 3px;
font-weight: bold;
font-size: 1.1em;
}
#output .allowed-badge,
.effect-allow {
background-color: #2e7d32;
#output .allowed-badge {
background-color: #4caf50;
color: white;
}
#output .denied-badge,
.effect-deny {
background-color: #c62828;
#output .denied-badge {
background-color: #f44336;
color: white;
}
.details-section {
@ -54,130 +48,70 @@
.details-section dd {
margin-left: 1em;
}
.explanation-section {
background: rgba(255, 255, 255, 0.75);
border: 1px solid #ddd;
border-radius: 4px;
margin-top: 1em;
padding: 0 1em 1em;
}
.rules-table {
border-collapse: collapse;
width: 100%;
}
.rules-table th,
.rules-table td {
border-bottom: 1px solid #ddd;
padding: 0.5em;
text-align: left;
vertical-align: top;
}
.rule-status {
background: #e8f5e9;
color: #1b5e20;
}
.rule-ignored {
background: #eee;
color: #555;
font-weight: normal;
}
.requirement-allowed {
color: #1b5e20;
}
.requirement-denied {
color: #b71c1c;
}
@media only screen and (max-width: 576px) {
.rules-table,
.rules-table tbody,
.rules-table tr,
.rules-table td {
display: block;
}
.rules-table thead {
display: none;
}
.rules-table td::before {
content: attr(data-label) ": ";
font-weight: bold;
}
}
</style>
{% endblock %}
{% block content %}
<h1>Explain a permission decision</h1>
<h1>Permission check</h1>
{% set current_tab = "check" %}
{% include "_permissions_debug_tabs.html" %}
<p>Test an actor, action and resource. The result explains which rules matched, which specificity level won, and whether actor restrictions or required actions changed the verdict.</p>
<p>Use this tool to test permission checks for the current actor. It queries the <code>/-/check.json</code> API endpoint.</p>
{% if request.actor %}
<p>Current actor: <strong>{{ request.actor.get("id", "anonymous") }}</strong></p>
{% else %}
<p>Current actor: <strong>anonymous (not logged in)</strong></p>
{% endif %}
<div class="permission-form">
<form id="check-form" method="get" action="{{ urls.path('-/check') }}">
<form id="check-form" method="get" action="{{ urls.path("-/check") }}">
<div class="form-section">
<label for="actor">Actor JSON:</label>
<textarea class="permission-textarea" id="actor" name="actor">{{ actor_json }}</textarea>
<small>Use <code>null</code> for an anonymous actor. This actor is simulated; it does not change who you are signed in as.</small>
</div>
<div class="form-section">
<label for="action">Action:</label>
<label for="action">Action (permission name):</label>
<select id="action" name="action" required>
<option value="">Select an action...</option>
{% for action in actions %}
<option value="{{ action.name }}">{{ action.name }}{% if action.description %} — {{ action.description }}{% endif %}</option>
{% for action_name in sorted_actions %}
<option value="{{ action_name }}">{{ action_name }}</option>
{% endfor %}
</select>
<small id="action-help">The operation to evaluate</small>
<small>The permission action to check</small>
</div>
<div class="form-section" id="parent-section">
<label for="parent">Parent resource:</label>
<div class="form-section">
<label for="parent">Parent resource (optional):</label>
<input type="text" id="parent" name="parent" placeholder="e.g., database name">
<small>The database or other parent resource</small>
<small>For database-level permissions, specify the database name</small>
</div>
<div class="form-section" id="child-section">
<label for="child">Child resource:</label>
<input type="text" id="child" name="child" placeholder="e.g., table or query name">
<small>The table, query or other child resource</small>
<div class="form-section">
<label for="child">Child resource (optional):</label>
<input type="text" id="child" name="child" placeholder="e.g., table name">
<small>For table-level permissions, specify the table name (requires parent)</small>
</div>
<div class="form-actions">
<button type="submit" class="submit-btn" id="submit-btn">Explain decision</button>
<button type="submit" class="submit-btn" id="submit-btn">Check Permission</button>
</div>
</form>
</div>
<div id="output" style="display: none;">
<h2>Result: <span class="result-badge" id="result-badge"></span></h2>
<p id="result-summary"></p>
<dl class="details-section">
<dt>Actor:</dt>
<dd><code id="result-actor"></code></dd>
<dt>Action:</dt>
<dd><code id="result-action"></code></dd>
<dt>Resource:</dt>
<dd><code id="result-resource"></code></dd>
<dd id="result-action"></dd>
<dt>Resource Path:</dt>
<dd id="result-resource"></dd>
<dt>Actor ID:</dt>
<dd id="result-actor"></dd>
<div id="additional-details"></div>
</dl>
<section class="explanation-section">
<h3>Matching rules</h3>
<div id="matching-rules"></div>
</section>
<section class="explanation-section" id="restrictions-section">
<h3>Actor restrictions</h3>
<div id="restriction-results"></div>
</section>
<section class="explanation-section" id="requirements-section">
<h3>Required actions</h3>
<div id="requirement-results"></div>
</section>
<details style="margin-top: 1em;">
<summary style="cursor: pointer; font-weight: bold;">Raw JSON response</summary>
<pre id="raw-json" style="margin-top: 1em; padding: 1em; background-color: #f5f5f5; border: 1px solid #ddd; border-radius: 3px; overflow-x: auto;"></pre>
@ -185,134 +119,152 @@
</div>
<script>
const actions = Object.fromEntries({{ actions|tojson }}.map(action => [action.name, action]));
const form = document.getElementById('check-form');
const output = document.getElementById('output');
const submitBtn = document.getElementById('submit-btn');
const actionSelect = document.getElementById('action');
function updateResourceFields() {
const action = actions[actionSelect.value];
document.getElementById('parent-section').style.display = action && action.takes_parent ? 'block' : 'none';
document.getElementById('child-section').style.display = action && action.takes_child ? 'block' : 'none';
let help = action && action.description ? action.description : 'The operation to evaluate';
if (action && action.also_requires) {
help += `; also requires ${action.also_requires}`;
}
document.getElementById('action-help').textContent = help;
}
async function performCheck() {
submitBtn.disabled = true;
submitBtn.textContent = 'Explaining...';
const params = new URLSearchParams(new FormData(form));
submitBtn.textContent = 'Checking...';
const formData = new FormData(form);
const params = new URLSearchParams();
for (const [key, value] of formData.entries()) {
if (value) {
params.append(key, value);
}
}
try {
const response = await fetch('{{ urls.path("-/check.json") }}?' + params.toString(), {
headers: {'Accept': 'application/json'}
method: 'GET',
headers: {
'Accept': 'application/json',
}
});
const data = await response.json();
if (response.ok) {
displayResult(data);
} else {
displayError(data);
}
} catch (error) {
displayError({error: error.message});
alert('Error: ' + error.message);
} finally {
submitBtn.disabled = false;
submitBtn.textContent = 'Explain decision';
submitBtn.textContent = 'Check Permission';
}
}
// Populate form on initial load
(function() {
const params = populateFormFromURL();
const action = params.get('action');
if (action) {
performCheck();
}
})();
function displayResult(data) {
output.style.display = 'block';
// Set badge and styling
const resultBadge = document.getElementById('result-badge');
output.className = data.allowed ? 'allowed' : 'denied';
resultBadge.className = `result-badge ${data.allowed ? 'allowed-badge' : 'denied-badge'}`;
resultBadge.textContent = data.allowed ? 'ALLOWED ✓' : 'DENIED ✗';
document.getElementById('result-summary').textContent = data.explanation.summary;
document.getElementById('result-actor').textContent = data.actor === null ? 'anonymous' : JSON.stringify(data.actor);
document.getElementById('result-action').textContent = data.action;
document.getElementById('result-resource').textContent = data.resource.path;
displayRules(data.explanation);
displayRestrictions(data.explanation.restrictions);
displayRequirements(data.explanation.required_actions);
if (data.allowed) {
output.className = 'allowed';
resultBadge.className = 'result-badge allowed-badge';
resultBadge.textContent = 'ALLOWED ✓';
} else {
output.className = 'denied';
resultBadge.className = 'result-badge denied-badge';
resultBadge.textContent = 'DENIED ✗';
}
// Basic details
document.getElementById('result-action').textContent = data.action || 'N/A';
document.getElementById('result-resource').textContent = data.resource?.path || '/';
document.getElementById('result-actor').textContent = data.actor_id || 'anonymous';
// Additional details
const additionalDetails = document.getElementById('additional-details');
additionalDetails.innerHTML = '';
if (data.reason !== undefined) {
const dt = document.createElement('dt');
dt.textContent = 'Reason:';
const dd = document.createElement('dd');
dd.textContent = data.reason || 'N/A';
additionalDetails.appendChild(dt);
additionalDetails.appendChild(dd);
}
if (data.source_plugin !== undefined) {
const dt = document.createElement('dt');
dt.textContent = 'Source Plugin:';
const dd = document.createElement('dd');
dd.textContent = data.source_plugin || 'N/A';
additionalDetails.appendChild(dt);
additionalDetails.appendChild(dd);
}
if (data.used_default !== undefined) {
const dt = document.createElement('dt');
dt.textContent = 'Used Default:';
const dd = document.createElement('dd');
dd.textContent = data.used_default ? 'Yes' : 'No';
additionalDetails.appendChild(dt);
additionalDetails.appendChild(dd);
}
if (data.depth !== undefined) {
const dt = document.createElement('dt');
dt.textContent = 'Depth:';
const dd = document.createElement('dd');
dd.textContent = data.depth;
additionalDetails.appendChild(dt);
additionalDetails.appendChild(dd);
}
// Raw JSON
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data);
}
function displayRules(explanation) {
const container = document.getElementById('matching-rules');
if (!explanation.matched_rules.length) {
container.innerHTML = '<p>No rules matched. Datasette denies access when there is no matching rule.</p>';
return;
}
let html = '<table class="rules-table"><thead><tr><th>Effect</th><th>Scope</th><th>Source</th><th>Reason</th><th>Role in decision</th></tr></thead><tbody>';
for (const rule of explanation.matched_rules) {
const status = rule.decisive
? '<span class="rule-status">Decisive</span>'
: `<span class="rule-status rule-ignored">${escapeHtml(rule.ignored_because)}</span>`;
html += '<tr>';
html += `<td data-label="Effect"><span class="effect-badge effect-${rule.effect}">${rule.effect.toUpperCase()}</span></td>`;
html += `<td data-label="Scope">${escapeHtml(rule.scope)}</td>`;
html += `<td data-label="Source"><code>${escapeHtml(rule.source || 'unknown')}</code></td>`;
html += `<td data-label="Reason">${escapeHtml(rule.reason || 'No reason supplied')}</td>`;
html += `<td data-label="Role in decision">${status}</td>`;
html += '</tr>';
}
container.innerHTML = html + '</tbody></table>';
}
function displayRestrictions(restrictions) {
const section = document.getElementById('restrictions-section');
const container = document.getElementById('restriction-results');
section.style.display = restrictions.length ? 'block' : 'none';
container.innerHTML = restrictions.map(restriction => {
const className = restriction.allowed ? 'requirement-allowed' : 'requirement-denied';
const verdict = restriction.allowed ? 'INCLUDED ✓' : 'EXCLUDED ✗';
return `<p class="${className}"><strong>${verdict}</strong> by <code>${escapeHtml(restriction.source || 'unknown')}</code>: ${escapeHtml(restriction.reason)}</p>`;
}).join('');
}
function displayRequirements(requirements) {
const section = document.getElementById('requirements-section');
const container = document.getElementById('requirement-results');
section.style.display = requirements.length ? 'block' : 'none';
container.innerHTML = requirements.map(requirement => {
const className = requirement.allowed ? 'requirement-allowed' : 'requirement-denied';
const verdict = requirement.allowed ? 'ALLOWED ✓' : 'DENIED ✗';
return `<p class="${className}"><strong>${escapeHtml(requirement.action)}: ${verdict}</strong> — ${escapeHtml(requirement.summary)}</p>`;
}).join('');
// Scroll to output
output.scrollIntoView({ behavior: 'smooth', block: 'nearest' });
}
function displayError(data) {
output.style.display = 'block';
output.className = 'denied';
const resultBadge = document.getElementById('result-badge');
resultBadge.className = 'result-badge denied-badge';
resultBadge.textContent = 'ERROR';
document.getElementById('result-summary').textContent = data.error || 'Unknown error';
document.getElementById('result-actor').textContent = '—';
document.getElementById('result-action').textContent = '—';
document.getElementById('result-resource').textContent = '—';
document.getElementById('matching-rules').innerHTML = '';
document.getElementById('restrictions-section').style.display = 'none';
document.getElementById('requirements-section').style.display = 'none';
document.getElementById('result-action').textContent = 'N/A';
document.getElementById('result-resource').textContent = 'N/A';
document.getElementById('result-actor').textContent = 'N/A';
const additionalDetails = document.getElementById('additional-details');
additionalDetails.innerHTML = '<dt>Error:</dt><dd>' + (data.error || 'Unknown error') + '</dd>';
document.getElementById('raw-json').innerHTML = jsonFormatHighlight(data);
output.scrollIntoView({ behavior: 'smooth', block: 'nearest' });
}
form.addEventListener('submit', event => {
event.preventDefault();
performCheck();
});
actionSelect.addEventListener('change', updateResourceFields);
// Disable child input if parent is empty
const parentInput = document.getElementById('parent');
const childInput = document.getElementById('child');
(function initializeFromUrl() {
const params = populateFormFromURL();
updateResourceFields();
if (params.get('action')) {
performCheck();
childInput.addEventListener('focus', () => {
if (!parentInput.value) {
alert('Please specify a parent resource first before adding a child resource.');
parentInput.focus();
}
})();
});
</script>
{% endblock %}

View file

@ -1,6 +1,6 @@
{% extends "base.html" %}
{% block title %}Permission activity{% endblock %}
{% block title %}Debug permissions{% endblock %}
{% block extra_head %}
{% include "_permission_ui_styles.html" %}
@ -20,45 +20,60 @@
.check-action, .check-when, .check-result {
font-size: 1.3em;
}
textarea {
height: 10em;
width: 95%;
box-sizing: border-box;
padding: 0.5em;
border: 2px dotted black;
}
.two-col {
display: inline-block;
width: 48%;
}
.two-col label {
width: 48%;
}
@media only screen and (max-width: 576px) {
.two-col {
width: 100%;
}
}
</style>
{% endblock %}
{% block content %}
<h1>Permission activity</h1>
<h1>Permission playground</h1>
{% set current_tab = "permissions" %}
{% include "_permissions_debug_tabs.html" %}
<h2>Raw simulator</h2>
<p>This form runs a hypothetical permission check and returns its raw explanation JSON. Use the <a href="{{ urls.path('-/check') }}">Explain tool</a> for a visual explanation of the same decision.</p>
<p>This tool lets you simulate an actor and a permission check for that actor.</p>
<div class="permission-form">
<form action="{{ urls.path('-/permissions') }}" id="debug-post" method="post">
<div class="permission-form-grid">
<div>
<div class="form-section">
<label for="activity-actor">Actor</label>
<textarea class="permission-textarea" id="activity-actor" name="actor">{% if actor_input %}{{ actor_input }}{% else %}{"id": "root"}{% endif %}</textarea>
</div>
<div class="two-col">
<div class="form-section">
<label>Actor</label>
<textarea name="actor">{% if actor_input %}{{ actor_input }}{% else %}{"id": "root"}{% endif %}</textarea>
</div>
<div>
<div class="form-section">
<label for="permission">Action</label>
<select name="permission" id="permission">
{% for permission in permissions %}
<option value="{{ permission.name }}">{{ permission.name }}</option>
{% endfor %}
</select>
</div>
<div class="form-section">
<label for="resource_1">Parent</label>
<input type="text" id="resource_1" name="resource_1" placeholder="e.g., database name">
</div>
<div class="form-section">
<label for="resource_2">Child</label>
<input type="text" id="resource_2" name="resource_2" placeholder="e.g., table name">
</div>
</div>
<div class="two-col" style="vertical-align: top">
<div class="form-section">
<label for="permission">Action</label>
<select name="permission" id="permission">
{% for permission in permissions %}
<option value="{{ permission.name }}">{{ permission.name }}</option>
{% endfor %}
</select>
</div>
<div class="form-section">
<label for="resource_1">Parent</label>
<input type="text" id="resource_1" name="resource_1" placeholder="e.g., database name">
</div>
<div class="form-section">
<label for="resource_2">Child</label>
<input type="text" id="resource_2" name="resource_2" placeholder="e.g., table name">
</div>
</div>
<div class="form-actions">
@ -110,7 +125,7 @@ debugPost.addEventListener('submit', function(ev) {
});
</script>
<h2>Recent permission checks</h2>
<h1>Recent permissions checks</h1>
<p>
{% if filter != "all" %}<a href="?filter=all">All</a>{% else %}<strong>All</strong>{% endif %},

View file

@ -37,7 +37,7 @@
<div class="form-section">
<label for="page_size">Page size:</label>
<input type="number" id="page_size" name="_size" value="50" min="1" max="200">
<input type="number" id="page_size" name="_size" value="50" min="1" max="200" style="max-width: 100px;">
<small>Number of results per page (max 200)</small>
</div>

View file

@ -636,7 +636,7 @@ def detect_primary_keys(conn, table):
def get_outbound_foreign_keys(conn, table):
infos = conn.execute(f"PRAGMA foreign_key_list({escape_sqlite(table)})").fetchall()
infos = conn.execute(f"PRAGMA foreign_key_list([{table}])").fetchall()
fks = []
for info in infos:
if info is not None:

View file

@ -252,62 +252,88 @@ async def _build_single_action_sql(
]
)
# Continue with the cascading logic.
# Aggregate the RULES by cascade level (small), rather than grouping
# base x rules (which scales with the number of resources).
def _agg(select_key, where, group_by):
parts = [
f" SELECT {select_key}",
" MAX(CASE WHEN allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN allow = 1 THEN 1 ELSE 0 END) AS any_allow,",
" json_group_array(CASE WHEN allow = 0 THEN source_plugin || ': ' || reason END) AS deny_reasons,",
" json_group_array(CASE WHEN allow = 1 THEN source_plugin || ': ' || reason END) AS allow_reasons",
f" FROM all_rules WHERE {where}",
]
if group_by:
parts.append(f" GROUP BY {group_by}")
return parts
# Continue with the cascading logic
query_parts.extend(
["child_agg AS ("]
+ _agg(
"parent, child,",
"parent IS NOT NULL AND child IS NOT NULL",
"parent, child",
)
+ ["),", "parent_agg AS ("]
+ _agg("parent,", "parent IS NOT NULL AND child IS NULL", "parent")
+ ["),", "global_agg AS ("]
+ _agg("", "parent IS NULL AND child IS NULL", None)
+ ["),"]
[
"child_lvl AS (",
" SELECT b.parent, b.child,",
" MAX(CASE WHEN ar.allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN ar.allow = 1 THEN 1 ELSE 0 END) AS any_allow,",
" json_group_array(CASE WHEN ar.allow = 0 THEN ar.source_plugin || ': ' || ar.reason END) AS deny_reasons,",
" json_group_array(CASE WHEN ar.allow = 1 THEN ar.source_plugin || ': ' || ar.reason END) AS allow_reasons",
" FROM base b",
" LEFT JOIN all_rules ar ON ar.parent = b.parent AND ar.child = b.child",
" GROUP BY b.parent, b.child",
"),",
"parent_lvl AS (",
" SELECT b.parent, b.child,",
" MAX(CASE WHEN ar.allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN ar.allow = 1 THEN 1 ELSE 0 END) AS any_allow,",
" json_group_array(CASE WHEN ar.allow = 0 THEN ar.source_plugin || ': ' || ar.reason END) AS deny_reasons,",
" json_group_array(CASE WHEN ar.allow = 1 THEN ar.source_plugin || ': ' || ar.reason END) AS allow_reasons",
" FROM base b",
" LEFT JOIN all_rules ar ON ar.parent = b.parent AND ar.child IS NULL",
" GROUP BY b.parent, b.child",
"),",
"global_lvl AS (",
" SELECT b.parent, b.child,",
" MAX(CASE WHEN ar.allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN ar.allow = 1 THEN 1 ELSE 0 END) AS any_allow,",
" json_group_array(CASE WHEN ar.allow = 0 THEN ar.source_plugin || ': ' || ar.reason END) AS deny_reasons,",
" json_group_array(CASE WHEN ar.allow = 1 THEN ar.source_plugin || ': ' || ar.reason END) AS allow_reasons",
" FROM base b",
" LEFT JOIN all_rules ar ON ar.parent IS NULL AND ar.child IS NULL",
" GROUP BY b.parent, b.child",
"),",
]
)
# Add anonymous decision logic if needed
if include_is_private:
def _anon_agg(select_key, where, group_by):
parts = [
f" SELECT {select_key}",
" MAX(CASE WHEN allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN allow = 1 THEN 1 ELSE 0 END) AS any_allow",
f" FROM anon_rules WHERE {where}",
]
if group_by:
parts.append(f" GROUP BY {group_by}")
return parts
query_parts.extend(
["anon_child_agg AS ("]
+ _anon_agg(
"parent, child,",
"parent IS NOT NULL AND child IS NOT NULL",
"parent, child",
)
+ ["),", "anon_parent_agg AS ("]
+ _anon_agg("parent,", "parent IS NOT NULL AND child IS NULL", "parent")
+ ["),", "anon_global_agg AS ("]
+ _anon_agg("", "parent IS NULL AND child IS NULL", None)
+ ["),"]
[
"anon_child_lvl AS (",
" SELECT b.parent, b.child,",
" MAX(CASE WHEN ar.allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN ar.allow = 1 THEN 1 ELSE 0 END) AS any_allow",
" FROM base b",
" LEFT JOIN anon_rules ar ON ar.parent = b.parent AND ar.child = b.child",
" GROUP BY b.parent, b.child",
"),",
"anon_parent_lvl AS (",
" SELECT b.parent, b.child,",
" MAX(CASE WHEN ar.allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN ar.allow = 1 THEN 1 ELSE 0 END) AS any_allow",
" FROM base b",
" LEFT JOIN anon_rules ar ON ar.parent = b.parent AND ar.child IS NULL",
" GROUP BY b.parent, b.child",
"),",
"anon_global_lvl AS (",
" SELECT b.parent, b.child,",
" MAX(CASE WHEN ar.allow = 0 THEN 1 ELSE 0 END) AS any_deny,",
" MAX(CASE WHEN ar.allow = 1 THEN 1 ELSE 0 END) AS any_allow",
" FROM base b",
" LEFT JOIN anon_rules ar ON ar.parent IS NULL AND ar.child IS NULL",
" GROUP BY b.parent, b.child",
"),",
"anon_decisions AS (",
" SELECT",
" b.parent, b.child,",
" CASE",
" WHEN acl.any_deny = 1 THEN 0",
" WHEN acl.any_allow = 1 THEN 1",
" WHEN apl.any_deny = 1 THEN 0",
" WHEN apl.any_allow = 1 THEN 1",
" WHEN agl.any_deny = 1 THEN 0",
" WHEN agl.any_allow = 1 THEN 1",
" ELSE 0",
" END AS anon_is_allowed",
" FROM base b",
" JOIN anon_child_lvl acl ON b.parent = acl.parent AND (b.child = acl.child OR (b.child IS NULL AND acl.child IS NULL))",
" JOIN anon_parent_lvl apl ON b.parent = apl.parent AND (b.child = apl.child OR (b.child IS NULL AND apl.child IS NULL))",
" JOIN anon_global_lvl agl ON b.parent = agl.parent AND (b.child = agl.child OR (b.child IS NULL AND agl.child IS NULL))",
"),",
]
)
# Final decisions
@ -316,28 +342,31 @@ async def _build_single_action_sql(
"decisions AS (",
" SELECT",
" b.parent, b.child,",
" -- Cascading permission logic: child -> parent -> global, DENY beats ALLOW at each level",
" -- Cascading permission logic: child → parent → global, DENY beats ALLOW at each level",
" -- Priority order:",
" -- 1. Child-level deny 2. Child-level allow",
" -- 3. Parent-level deny 4. Parent-level allow",
" -- 5. Global-level deny 6. Global-level allow",
" -- 1. Child-level deny (most specific, blocks access)",
" -- 2. Child-level allow (most specific, grants access)",
" -- 3. Parent-level deny (intermediate, blocks access)",
" -- 4. Parent-level allow (intermediate, grants access)",
" -- 5. Global-level deny (least specific, blocks access)",
" -- 6. Global-level allow (least specific, grants access)",
" -- 7. Default deny (no rules match)",
" CASE",
" WHEN ca.any_deny = 1 THEN 0",
" WHEN ca.any_allow = 1 THEN 1",
" WHEN pa.any_deny = 1 THEN 0",
" WHEN pa.any_allow = 1 THEN 1",
" WHEN ga.any_deny = 1 THEN 0",
" WHEN ga.any_allow = 1 THEN 1",
" WHEN cl.any_deny = 1 THEN 0",
" WHEN cl.any_allow = 1 THEN 1",
" WHEN pl.any_deny = 1 THEN 0",
" WHEN pl.any_allow = 1 THEN 1",
" WHEN gl.any_deny = 1 THEN 0",
" WHEN gl.any_allow = 1 THEN 1",
" ELSE 0",
" END AS is_allowed,",
" CASE",
" WHEN ca.any_deny = 1 THEN ca.deny_reasons",
" WHEN ca.any_allow = 1 THEN ca.allow_reasons",
" WHEN pa.any_deny = 1 THEN pa.deny_reasons",
" WHEN pa.any_allow = 1 THEN pa.allow_reasons",
" WHEN ga.any_deny = 1 THEN ga.deny_reasons",
" WHEN ga.any_allow = 1 THEN ga.allow_reasons",
" WHEN cl.any_deny = 1 THEN cl.deny_reasons",
" WHEN cl.any_allow = 1 THEN cl.allow_reasons",
" WHEN pl.any_deny = 1 THEN pl.deny_reasons",
" WHEN pl.any_allow = 1 THEN pl.allow_reasons",
" WHEN gl.any_deny = 1 THEN gl.deny_reasons",
" WHEN gl.any_allow = 1 THEN gl.allow_reasons",
" ELSE '[]'",
" END AS reason",
]
@ -345,34 +374,21 @@ async def _build_single_action_sql(
if include_is_private:
query_parts.append(
" , CASE WHEN ("
"CASE"
" WHEN aca.any_deny = 1 THEN 0"
" WHEN aca.any_allow = 1 THEN 1"
" WHEN apa.any_deny = 1 THEN 0"
" WHEN apa.any_allow = 1 THEN 1"
" WHEN aga.any_deny = 1 THEN 0"
" WHEN aga.any_allow = 1 THEN 1"
" ELSE 0 END"
") = 0 THEN 1 ELSE 0 END AS is_private"
" , CASE WHEN ad.anon_is_allowed = 0 THEN 1 ELSE 0 END AS is_private"
)
query_parts.extend(
[
" FROM base b",
" LEFT JOIN child_agg ca ON ca.parent = b.parent AND ca.child = b.child",
" LEFT JOIN parent_agg pa ON pa.parent = b.parent",
" CROSS JOIN global_agg ga",
" JOIN child_lvl cl ON b.parent = cl.parent AND (b.child = cl.child OR (b.child IS NULL AND cl.child IS NULL))",
" JOIN parent_lvl pl ON b.parent = pl.parent AND (b.child = pl.child OR (b.child IS NULL AND pl.child IS NULL))",
" JOIN global_lvl gl ON b.parent = gl.parent AND (b.child = gl.child OR (b.child IS NULL AND gl.child IS NULL))",
]
)
if include_is_private:
query_parts.extend(
[
" LEFT JOIN anon_child_agg aca ON aca.parent = b.parent AND aca.child = b.child",
" LEFT JOIN anon_parent_agg apa ON apa.parent = b.parent",
" CROSS JOIN anon_global_agg aga",
]
query_parts.append(
" JOIN anon_decisions ad ON b.parent = ad.parent AND (b.child = ad.child OR (b.child IS NULL AND ad.child IS NULL))"
)
query_parts.append(")")
@ -384,28 +400,8 @@ async def _build_single_action_sql(
restriction_intersect = "\nINTERSECT\n".join(
f"SELECT * FROM ({sql})" for sql in restriction_sqls
)
# Decompose by NULL-pattern so the final filter can use pure-equality
# EXISTS lookups (satisfiable via automatic indexes) instead of a
# correlated OR-scan over the whole list.
query_parts.extend(
[
",",
"restriction_list AS (",
f" {restriction_intersect}",
"),",
"restriction_exact AS (",
" SELECT parent, child FROM restriction_list WHERE parent IS NOT NULL AND child IS NOT NULL",
"),",
"restriction_parent_any AS (",
" SELECT DISTINCT parent FROM restriction_list WHERE parent IS NOT NULL AND child IS NULL",
"),",
"restriction_child_any AS (",
" SELECT DISTINCT child FROM restriction_list WHERE parent IS NULL AND child IS NOT NULL",
"),",
"restriction_all AS (",
" SELECT 1 AS matched FROM restriction_list WHERE parent IS NULL AND child IS NULL LIMIT 1",
")",
]
[",", "restriction_list AS (", f" {restriction_intersect}", ")"]
)
# Final SELECT
@ -420,11 +416,10 @@ async def _build_single_action_sql(
# Add restriction filter if there are restrictions
if restriction_sqls:
query_parts.append("""
AND (
EXISTS (SELECT 1 FROM restriction_all)
OR EXISTS (SELECT 1 FROM restriction_parent_any r WHERE r.parent = decisions.parent)
OR EXISTS (SELECT 1 FROM restriction_child_any r WHERE r.child = decisions.child)
OR EXISTS (SELECT 1 FROM restriction_exact r WHERE r.parent = decisions.parent AND r.child = decisions.child)
AND EXISTS (
SELECT 1 FROM restriction_list r
WHERE (r.parent = decisions.parent OR r.parent IS NULL)
AND (r.child = decisions.child OR r.child IS NULL)
)""")
# Add parent filter if specified
@ -678,239 +673,3 @@ async def check_permission_for_resource(
child=child,
)
return results[action]
async def explain_permission_for_resource(
*,
datasette: "Datasette",
actor: dict | None,
action: str,
parent: str | None,
child: str | None,
) -> dict:
"""Explain a permission decision for one action and resource.
This is intended for Datasette's permission debugging tools. It uses the
same ``permission_resources_sql`` hook results and the same resolution
rules as :func:`check_permissions_for_actions`, but also returns the
matching rules, actor restriction results and ``also_requires`` chain.
The returned dictionary is part of Datasette's unstable debugging API.
"""
action_obj = datasette.actions.get(action)
if action_obj is None:
raise ValueError(f"Unknown action: {action}")
explanation = await _explain_single_action(
datasette=datasette,
actor=actor,
action=action,
parent=parent,
child=child,
)
required_actions = []
if action_obj.also_requires:
required = await explain_permission_for_resource(
datasette=datasette,
actor=actor,
action=action_obj.also_requires,
parent=parent,
child=child,
)
required_actions.append(required)
explanation["required_actions"] = required_actions
explanation["allowed"] = bool(
explanation["rule_allowed"]
and explanation["restriction_allowed"]
and all(required["allowed"] for required in required_actions)
)
explanation["summary"] = _permission_explanation_summary(explanation)
return explanation
async def _explain_single_action(
*,
datasette: "Datasette",
actor: dict | None,
action: str,
parent: str | None,
child: str | None,
) -> dict:
"""Return matching rules and restrictions for a single action."""
from datasette.utils.permissions import SKIP_PERMISSION_CHECKS
permission_sqls = await gather_permission_sql_from_hooks(
datasette=datasette,
actor=actor,
action=action,
)
if permission_sqls is SKIP_PERMISSION_CHECKS:
return {
"action": action,
"rule_allowed": True,
"restriction_allowed": True,
"winning_scope": "global",
"matched_rules": [
{
"scope": "global",
"effect": "allow",
"source": "skip_permission_checks",
"reason": "Permission checks were explicitly skipped",
"decisive": True,
"ignored_because": None,
}
],
"restrictions": [],
}
db = datasette.get_internal_database()
matched_rules = []
restrictions = []
for permission_sql in permission_sqls:
params = dict(permission_sql.params or {})
parent_param = _unused_parameter_name(params, "_explain_parent")
params[parent_param] = parent
child_param = _unused_parameter_name(params, "_explain_child")
params[child_param] = child
if permission_sql.sql:
rows = await db.execute(
f"""
SELECT parent, child, allow, reason
FROM ({permission_sql.sql}) AS permission_rules
WHERE (parent IS NULL OR parent = :{parent_param})
AND (child IS NULL OR child = :{child_param})
""",
params,
)
for row in rows:
specificity = (
2
if row["child"] is not None
else 1 if row["parent"] is not None else 0
)
matched_rules.append(
{
"scope": ("resource", "parent", "global")[2 - specificity],
"effect": "allow" if row["allow"] else "deny",
"source": permission_sql.source,
"reason": row["reason"],
"_specificity": specificity,
}
)
if permission_sql.restriction_sql:
restriction_row = (
await db.execute(
f"""
SELECT EXISTS(
SELECT 1 FROM ({permission_sql.restriction_sql}) AS restriction_rules
WHERE (parent IS NULL OR parent = :{parent_param})
AND (child IS NULL OR child = :{child_param})
) AS resource_is_in_allowlist
""",
params,
)
).first()
restriction_allowed = bool(restriction_row[0])
restrictions.append(
{
"source": permission_sql.source,
"allowed": restriction_allowed,
"reason": params.get("deny")
or (
"Resource is included in this restriction allowlist"
if restriction_allowed
else "Resource is not included in this restriction allowlist"
),
}
)
matched_rules.sort(
key=lambda rule: (
-rule["_specificity"],
0 if rule["effect"] == "deny" else 1,
rule["source"] or "",
rule["reason"] or "",
)
)
if matched_rules:
winning_specificity = matched_rules[0]["_specificity"]
winning_rules = [
rule
for rule in matched_rules
if rule["_specificity"] == winning_specificity
]
rule_allowed = not any(rule["effect"] == "deny" for rule in winning_rules)
winning_scope = winning_rules[0]["scope"]
else:
winning_specificity = None
rule_allowed = False
winning_scope = None
for rule in matched_rules:
specificity = rule.pop("_specificity")
if specificity != winning_specificity:
rule["decisive"] = False
rule["ignored_because"] = "A more specific rule matched"
elif not rule_allowed and rule["effect"] == "allow":
rule["decisive"] = False
rule["ignored_because"] = "A deny rule matched at the same scope"
else:
rule["decisive"] = True
rule["ignored_because"] = None
return {
"action": action,
"rule_allowed": rule_allowed,
"restriction_allowed": all(
restriction["allowed"] for restriction in restrictions
),
"winning_scope": winning_scope,
"matched_rules": matched_rules,
"restrictions": restrictions,
}
def _unused_parameter_name(params: dict, preferred: str) -> str:
"""Return a SQL parameter name that is not already in ``params``."""
candidate = preferred
suffix = 2
while candidate in params:
candidate = f"{preferred}_{suffix}"
suffix += 1
return candidate
def _permission_explanation_summary(explanation: dict) -> str:
denied_requirement = next(
(
required
for required in explanation["required_actions"]
if not required["allowed"]
),
None,
)
if denied_requirement:
return (
f"Denied because {explanation['action']} also requires "
f"{denied_requirement['action']}, which was denied."
)
if not explanation["matched_rules"]:
return "Denied because no permission rule matched this actor and resource."
if not explanation["rule_allowed"]:
return (
f"Denied by a {explanation['winning_scope']}-level rule. "
"Deny rules take precedence over allow rules at the same scope."
)
if not explanation["restriction_allowed"]:
return (
"Denied because the resource is not included in the actor's restrictions."
)
return f"Allowed by the matching {explanation['winning_scope']}-level rule."

View file

@ -3,7 +3,7 @@ import textwrap
from sqlite_utils import Database as SQLiteUtilsDatabase
from sqlite_utils import Migrations
from datasette.utils import escape_sqlite, table_column_details
from datasette.utils import table_column_details
INTERNAL_DB_SCHEMA_TABLES = {
"catalog_databases",
@ -180,7 +180,7 @@ async def init_internal_db(db):
await db.execute_write_fn(apply_migrations, transaction=False)
async def populate_schema_tables(internal_db, db, schema_version):
async def populate_schema_tables(internal_db, db):
database_name = db.name
tables = (await db.execute("select * from sqlite_master WHERE type = 'table'")).rows
@ -213,7 +213,7 @@ async def populate_schema_tables(internal_db, db, schema_version):
for column in columns
)
foreign_keys = conn.execute(
f"PRAGMA foreign_key_list({escape_sqlite(table_name)})"
f"PRAGMA foreign_key_list([{table_name}])"
).fetchall()
foreign_keys_to_insert.extend(
{
@ -222,9 +222,7 @@ async def populate_schema_tables(internal_db, db, schema_version):
}
for foreign_key in foreign_keys
)
indexes = conn.execute(
f"PRAGMA index_list({escape_sqlite(table_name)})"
).fetchall()
indexes = conn.execute(f"PRAGMA index_list([{table_name}])").fetchall()
indexes_to_insert.extend(
{
**{"database_name": database_name, "table_name": table_name},
@ -248,45 +246,38 @@ async def populate_schema_tables(internal_db, db, schema_version):
indexes_to_insert,
) = await db.execute_fn(collect_info)
def replace_catalog(conn):
# Delete child rows before their catalog_tables parents so this also
# works if a prepare_connection plugin enables foreign key enforcement.
for table in (
"catalog_columns",
"catalog_foreign_keys",
"catalog_indexes",
"catalog_views",
"catalog_tables",
):
conn.execute(
"DELETE FROM {} WHERE database_name = ?".format(table),
[database_name],
)
def update_catalog(conn):
# Delete and reinsert as one write task so internal database readers
# never observe a database with missing catalog rows
# https://github.com/simonw/datasette/issues/2831
conn.execute(
"""
INSERT OR REPLACE INTO catalog_databases (
database_name, path, is_memory, schema_version
) VALUES (?, ?, ?, ?)
""",
[
database_name,
str(db.path) if db.path is not None else None,
db.is_memory,
schema_version,
],
"DELETE FROM catalog_tables WHERE database_name = ?", [database_name]
)
conn.execute(
"DELETE FROM catalog_views WHERE database_name = ?", [database_name]
)
conn.execute(
"DELETE FROM catalog_columns WHERE database_name = ?", [database_name]
)
conn.execute(
"DELETE FROM catalog_foreign_keys WHERE database_name = ?",
[database_name],
)
conn.execute(
"DELETE FROM catalog_indexes WHERE database_name = ?", [database_name]
)
conn.executemany(
"""
INSERT INTO catalog_tables (database_name, table_name, rootpage, sql)
values (?, ?, ?, ?)
""",
""",
tables_to_insert,
)
conn.executemany(
"""
INSERT INTO catalog_views (database_name, view_name, rootpage, sql)
values (?, ?, ?, ?)
""",
""",
views_to_insert,
)
conn.executemany(
@ -296,7 +287,7 @@ async def populate_schema_tables(internal_db, db, schema_version):
) VALUES (
:database_name, :table_name, :cid, :name, :type, :notnull, :default_value, :is_pk, :hidden
)
""",
""",
columns_to_insert,
)
conn.executemany(
@ -306,7 +297,7 @@ async def populate_schema_tables(internal_db, db, schema_version):
) VALUES (
:database_name, :table_name, :id, :seq, :table, :from, :to, :on_update, :on_delete, :match
)
""",
""",
foreign_keys_to_insert,
)
conn.executemany(
@ -316,8 +307,8 @@ async def populate_schema_tables(internal_db, db, schema_version):
) VALUES (
:database_name, :table_name, :seq, :name, :unique, :origin, :partial
)
""",
""",
indexes_to_insert,
)
await internal_db.execute_write_fn(replace_catalog)
await internal_db.execute_write_fn(update_catalog)

View file

@ -1,2 +1,2 @@
__version__ = "1.0a37"
__version__ = "1.0a36"
__version_info__ = tuple(__version__.split("."))

View file

@ -643,15 +643,8 @@ class QueryView(View):
ok = None
redirect_url = None
try:
execute_write_kwargs = {"request": request}
if stored_query.is_trusted:
analysis = await db.analyze_sql(stored_query.sql, params_for_query)
if any(
operation.operation == "vacuum" for operation in analysis.operations
):
execute_write_kwargs["transaction"] = False
cursor = await db.execute_write(
stored_query.sql, params_for_query, **execute_write_kwargs
stored_query.sql, params_for_query, request=request
)
# success message can come from on_success_message or on_success_message_sql
message = None

View file

@ -749,7 +749,9 @@ class RowDeleteView(BaseView):
# Delete table
def delete_row(conn):
sqlite_utils.Database(conn)[resolved.table].delete(resolved.pk_values)
sqlite_utils.Database(conn, execute_plugins=False)[resolved.table].delete(
resolved.pk_values
)
try:
await resolved.db.execute_write_fn(delete_row, request=request)
@ -830,7 +832,7 @@ class RowUpdateView(BaseView):
return Response.error(["Permission denied for alter-table"], 403)
def update_row(conn):
sqlite_utils.Database(conn)[resolved.table].update(
sqlite_utils.Database(conn, execute_plugins=False)[resolved.table].update(
resolved.pk_values, update, alter=alter
)

View file

@ -600,7 +600,7 @@ class PermissionRulesView(BaseView):
async def _check_permission_for_actor(ds, action, parent, child, actor):
"""Shared logic for checking and explaining a permission decision."""
"""Shared logic for checking permissions. Returns a dict with check results."""
if action not in ds.actions:
return error_body(f"Unknown action: {action}", 404), 404
@ -629,28 +629,15 @@ async def _check_permission_for_actor(ds, action, parent, child, actor):
allowed = await ds.allowed(action=action, resource=resource_obj, actor=actor)
from datasette.utils.actions_sql import explain_permission_for_resource
explanation = await explain_permission_for_resource(
datasette=ds,
actor=actor,
action=action,
parent=parent,
child=child,
)
response = {
"ok": True,
"unstable": UNSTABLE_API_MESSAGE,
"action": action,
"allowed": bool(allowed),
"actor": actor,
"resource": {
"parent": parent,
"child": child,
"path": _resource_path(parent, child),
},
"explanation": explanation,
}
if actor and "id" in actor:
@ -668,25 +655,11 @@ class PermissionCheckView(BaseView):
as_format = request.url_vars.get("format")
if not as_format:
actions = [
{
"name": action.name,
"description": action.description,
"takes_parent": action.takes_parent,
"takes_child": action.takes_child,
"also_requires": action.also_requires,
}
for action in sorted(
self.ds.actions.values(), key=lambda action: action.name
)
]
return await self.render(
["debug_check.html"],
request,
{
"actions": actions,
"actor_json": request.args.get("actor")
or json.dumps(request.actor, indent=2),
"sorted_actions": sorted(self.ds.actions.keys()),
"has_debug_permission": True,
},
)
@ -698,18 +671,9 @@ class PermissionCheckView(BaseView):
parent = request.args.get("parent")
child = request.args.get("child")
actor = request.actor
actor_json = request.args.get("actor")
if actor_json is not None:
try:
actor = json.loads(actor_json)
except json.JSONDecodeError as ex:
return Response.error(f"Invalid actor JSON: {ex}", 400)
if actor is not None and not isinstance(actor, dict):
return Response.error("actor must be a JSON object or null", 400)
response, status = await _check_permission_for_actor(
self.ds, action, parent, child, actor
self.ds, action, parent, child, request.actor
)
return Response.json(response, status=status)

View file

@ -1143,7 +1143,7 @@ class TableInsertView(BaseView):
row_pk_values_for_later = [tuple(row[pk] for pk in pks) for row in rows]
def insert_or_upsert_rows(conn):
table = sqlite_utils.Database(conn)[table_name]
table = sqlite_utils.Database(conn, execute_plugins=False)[table_name]
kwargs = {}
if upsert:
kwargs = {
@ -1407,7 +1407,7 @@ class TableDropView(BaseView):
# Drop table
def drop_table(conn):
sqlite_utils.Database(conn)[table_name].drop()
sqlite_utils.Database(conn, execute_plugins=False)[table_name].drop()
await db.execute_write_fn(drop_table, request=request)
await self.ds.track_event(

View file

@ -889,11 +889,13 @@ class TableCreateView(BaseView):
initial_schema = None
if table_exists:
initial_schema = await db.execute_fn(
lambda conn: sqlite_utils.Database(conn)[table_name].schema
lambda conn: sqlite_utils.Database(conn, execute_plugins=False)[
table_name
].schema
)
def create_table(conn):
db_for_write = sqlite_utils.Database(conn)
db_for_write = sqlite_utils.Database(conn, execute_plugins=False)
table = db_for_write[table_name]
if rows:
table.insert_all(
@ -1187,7 +1189,9 @@ class TableAlterView(BaseView):
before_schema = _table_schema_from_conn(conn, table_name)
def apply_operations(operation_conn):
db_for_write = sqlite_utils.Database(operation_conn)
db_for_write = sqlite_utils.Database(
operation_conn, execute_plugins=False
)
table = db_for_write[table_name]
current_table_name = table_name
@ -1261,62 +1265,62 @@ class TableAlterView(BaseView):
elif operation.op == "set_foreign_keys":
foreign_keys = [fk.tuple for fk in args.foreign_keys]
with operation_conn:
for column in add_columns:
not_null_default = None
if column.not_null:
if "default_expr" in column.model_fields_set:
not_null_default = _default_expression_sql(
column.default_expr
)
else:
not_null_default = _literal_default(
db_for_write, column.default
)
table.add_column(
column.name,
column.type,
not_null_default=not_null_default,
)
# The write task transaction makes these operations atomic
for column in add_columns:
not_null_default = None
if column.not_null:
if "default_expr" in column.model_fields_set:
not_null_default = _default_expression_sql(
column.default_expr
)
else:
not_null_default = _literal_default(
db_for_write, column.default
)
table.add_column(
column.name,
column.type,
not_null_default=not_null_default,
)
should_transform = any(
(
types,
rename,
drop,
not_null,
defaults,
column_order is not None,
pk is not SQLITE_UTILS_DEFAULT,
add_foreign_keys,
drop_foreign_keys,
foreign_keys is not None,
should_transform = any(
(
types,
rename,
drop,
not_null,
defaults,
column_order is not None,
pk is not SQLITE_UTILS_DEFAULT,
add_foreign_keys,
drop_foreign_keys,
foreign_keys is not None,
)
)
if should_transform:
table.transform(
types=types or None,
rename=rename or None,
drop=drop or None,
pk=pk,
not_null=not_null or None,
defaults=defaults or None,
column_order=column_order,
add_foreign_keys=add_foreign_keys or None,
drop_foreign_keys=drop_foreign_keys or None,
foreign_keys=foreign_keys,
)
if (
rename_table_to is not None
and rename_table_to != current_table_name
):
operation_conn.execute(
"alter table {} rename to {}".format(
escape_sqlite(current_table_name),
escape_sqlite(rename_table_to),
)
)
if should_transform:
table.transform(
types=types or None,
rename=rename or None,
drop=drop or None,
pk=pk,
not_null=not_null or None,
defaults=defaults or None,
column_order=column_order,
add_foreign_keys=add_foreign_keys or None,
drop_foreign_keys=drop_foreign_keys or None,
foreign_keys=foreign_keys,
)
if (
rename_table_to is not None
and rename_table_to != current_table_name
):
operation_conn.execute(
"alter table {} rename to {}".format(
escape_sqlite(current_table_name),
escape_sqlite(rename_table_to),
)
)
current_table_name = rename_table_to
current_table_name = rename_table_to
return current_table_name, _table_schema_from_conn(
operation_conn, current_table_name

View file

@ -45,7 +45,7 @@ Using the "root" actor
Datasette currently leaves almost all forms of authentication to plugins - `datasette-auth-github <https://github.com/simonw/datasette-auth-github>`__ for example.
The one exception is the "root" account, which you can sign into while using Datasette on your local machine. The root user starts with **all permissions**: Datasette contributes a global allow rule for every action. More specific deny rules can still override that global rule.
The one exception is the "root" account, which you can sign into while using Datasette on your local machine. The root user has **all permissions** - they can perform any action regardless of other permission rules.
The ``--root`` flag is designed for local development and testing. When you start Datasette with ``--root``, the root user automatically receives every permission, including:
@ -84,12 +84,12 @@ Click on that link and then visit ``http://127.0.0.1:8001/-/actor`` to confirm t
Permissions
===========
Datasette's permissions system is built around SQL queries. Datasette and its plugins construct SQL queries to resolve the list of resources that an actor cas access.
The key question the permissions system answers is this:
Is this **actor** allowed to perform this **action**, optionally against this particular **resource**?
Every permission decision can be understood in terms of those three values. Datasette implements the decisions using SQL, but you do not need to understand the generated SQL to configure or debug permissions.
**Actors** are :ref:`described above <authentication_actor>`.
An **action** is a string describing the action the actor would like to perform. A full list is :ref:`provided below <actions>` - examples include ``view-table`` and ``execute-sql``.
@ -138,51 +138,7 @@ This configuration will deny access to everyone except the user with ``id`` of `
How permissions are resolved
----------------------------
Permission rules describe an effect (``allow`` or ``deny``) at one of three levels:
``resource``
A specific child resource, such as the ``analytics/sales`` table.
``parent``
A parent resource, such as the ``analytics`` database. A parent rule also applies to its child resources.
``global``
Every resource for that action.
Datasette resolves matching rules from most specific to least specific:
#. Resource rules take precedence over parent and global rules.
#. Parent rules take precedence over global rules.
#. If both allow and deny rules match at the same level, deny takes precedence.
#. If no rule matches, access is denied.
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.
.. list-table:: Permission rule examples
:header-rows: 1
* - Matching rules
- Result
- Explanation
* - Global allow
- Allow
- The global rule is the most specific matching rule.
* - Global allow, parent deny
- Deny
- The parent rule is more specific.
* - Parent deny, resource allow
- Allow
- The resource rule is more specific.
* - Resource allow and resource deny
- Deny
- Deny takes precedence at the same level.
* - No matching rules
- Deny
- Permissions default to deny when no rule applies.
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.
Datasette performs checks using :ref:`datasette_allowed`, which accepts keyword arguments for ``action``, ``resource`` and an optional ``actor``.
Datasette performs permission checks using the internal :ref:`datasette_allowed`, method 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.
@ -193,12 +149,12 @@ resources were allowed or denied. The combined sources are:
* ``allow`` blocks configured in :ref:`datasette.yaml <authentication_permissions_config>`.
* :ref:`Actor restrictions <authentication_cli_create_token_restrict>` encoded into the actor dictionary or API token.
* The "root" user rule when ``--root`` (or :attr:`Datasette.root_enabled <datasette.app.Datasette.root_enabled>`) is active. This is a global allow rule, so a more specific configuration deny can override it.
* The "root" user shortcut when ``--root`` (or :attr:`Datasette.root_enabled <datasette.app.Datasette.root_enabled>`) is active, replying ``True`` to all permission chucks unless configuration rules deny them at a more specific level.
* Any additional SQL provided by plugins implementing :ref:`plugin_hook_permission_resources_sql`.
Actor restrictions are applied after the allow/deny rules. They act as an additional allowlist: a restriction can remove access but cannot grant access that the actor did not already have. See :ref:`authentication_cli_create_token_restrict`.
Some actions have dependencies on other actions. These are evaluated as an ``AND`` condition. For example, ``execute-sql`` also requires ``view-database``: both decisions must be allowed for the final result to be allowed.
Datasette evaluates the SQL to determine if the requested ``resource`` is
included. Explicit deny rules returned by configuration or plugins will block
access even if other rules allowed it.
.. _authentication_permissions_allow:
@ -1189,21 +1145,11 @@ The debug tool at ``/-/permissions`` is available to any actor with the ``permis
datasette -s permissions.permissions-debug true data.db
The permission debug tools answer four different questions:
The page shows the permission checks that have been carried out by the Datasette instance.
Why was this decision allowed or denied?
Use :ref:`PermissionCheckView`. It shows every matching rule, identifies the winning specificity level, applies actor restrictions and evaluates any required actions.
It also provides an interface for running hypothetical permission checks against a hypothetical actor. This is a useful way of confirming that your configured permissions work in the way you expect.
Which resources can the current actor access?
Use :ref:`AllowedResourcesView` to view an access map for a selected action.
Which raw rules did Datasette and its plugins contribute?
Use :ref:`PermissionRulesView` to inspect the rules before they are resolved into decisions.
Which checks has this Datasette instance performed recently?
Use ``/-/permissions`` to view recent permission activity.
These tools are designed to help administrators and plugin authors understand and confirm the effective permissions configuration.
This is designed to help administrators and plugin authors understand exactly how permission checks are being carried out, in order to effectively configure Datasette's permission system.
These debug endpoints are exempt from the :ref:`JSON API stability promise <json_api_stability>` - their JSON shapes may change in future releases.
@ -1238,20 +1184,11 @@ This endpoint requires the ``permissions-debug`` permission.
Permission check view
---------------------
The ``/-/check`` endpoint evaluates and explains a single actor, action and resource decision. The explanation includes:
* Every matching allow and deny rule, with its source and reason.
* The winning resource, parent or global scope.
* Rules ignored because a more specific rule matched, or because a deny won at the same scope.
* Actor restriction allowlists that included or excluded the resource.
* Additional actions required by the requested action.
* An explicit default-deny explanation when no rule matched.
The ``/-/check`` endpoint evaluates a single action/resource pair and returns information indicating whether the access was allowed along with diagnostic information.
This endpoint provides an interactive HTML form interface. Add ``.json`` to the URL path (e.g. ``/-/check.json?action=view-instance``) to get the raw JSON response instead.
Pass ``?action=`` to specify the action to check, and optional ``?parent=`` and ``?child=`` parameters to specify the resource. The interactive form also accepts actor JSON, allowing a hypothetical actor to be tested without signing in as that actor. The JSON endpoint accepts the same value using the ``actor`` query string parameter. Use ``actor=null`` to represent an anonymous actor.
This endpoint requires the ``permissions-debug`` permission. The hypothetical actor is used only for the decision being explained; access to the debug tool is checked against the actor who is actually signed in.
Pass ``?action=`` to specify the action to check, and optional ``?parent=`` and ``?child=`` parameters to specify the resource.
.. _authentication_ds_actor:

View file

@ -4,19 +4,20 @@
Changelog
=========
.. _v1_0_a37:
.. _v_unreleased:
1.0a37 (2026-07-14)
-------------------
Unreleased
----------
Performance improvement for SQL-backed permission checks, plus an improved permission debugging interface.
- SQL used to resolve permission checks now aggregates permission rules before joining them to resources, improving performance on instances with large schemas. (:issue:`2832`)
- The :ref:`PermissionCheckView` permission debugger now explains why a decision was allowed or denied, including the matching rules. The interactive form can also test a hypothetical actor supplied as JSON, and the :ref:`permissions documentation <authentication_permissions_explained>` now describes resolution rules in more detail. (:issue:`2841`)
- :ref:`db.execute_write(sql, ..., transaction=True) <database_execute_write>` has a new ``transaction=`` parameter, which can be set to ``False`` for statements such as ``VACUUM`` that cannot run inside a transaction. Write tasks now start their transactions using ``BEGIN IMMEDIATE``, which also ensures that writes are rolled back if the task fails. (:issue:`2831`)
- Refreshing a database's schema in Datasette's internal catalog is now performed as a single atomic operation. (:issue:`2831`)
- Fixed schema introspection, table pages, facets and table counts for tables with names containing a ``]`` character. Thanks, `TowyTowy <https://github.com/TowyTowy>`__. (:issue:`2431`, :pr:`2846`)
- ``/-/plugins.json`` once again returns a top-level JSON array of plugin objects, reverting the object envelope introduced in 1.0a36. This should fix a large number of trivial test failures in existing plugins. (:issue:`2842`, :pr:`2843`)
- Write functions run via ``await db.execute_write_fn()`` now execute inside an explicitly opened ``BEGIN IMMEDIATE`` transaction, committed when the function returns or rolled back if it raises. Previously the transaction was only opened implicitly by the first raw data-modifying statement, which meant writes made through sqlite-utils committed independently mid-task - a function that used sqlite-utils and then failed could leave those writes permanently committed. sqlite-utils write methods now nest inside the task transaction as savepoints, so a failing write function rolls back everything it did. Functions run with ``transaction=True`` should no longer manage transactions themselves - use ``transaction=False`` for manual transaction control. (:issue:`2831`)
- ``await db.execute_write()`` detects statements that SQLite cannot execute inside a transaction - ``VACUUM``, ``ATTACH``, ``DETACH`` and ``PRAGMA`` - and runs them in autocommit mode instead. (:issue:`2831`)
- ``await db.execute_write_script()`` is now transactional, matching its documentation: if any statement in the script fails, none of its statements are applied. Scripts containing statements that cannot run inside a transaction, or that manage transactions themselves, fall back to the previous ``conn.executescript()`` autocommit behavior. (:issue:`2831`)
- The JSON write API is now atomic per request: ``/db/-/create`` with initial rows, multi-operation ``/db/table/-/alter`` calls and inserts using ``"return": true`` now either fully apply or roll back entirely if any part fails. Previously a failure part way through could leave earlier writes from the same request permanently committed. (:issue:`2831`)
- Rebuilding the internal database catalog for a database is now a single atomic write. Previously the rebuild used six separate transactions, so queries against the internal database could observe a database with missing catalog rows while a rebuild was in progress. (:issue:`2831`)
- sqlite-utils plugins no longer have their ``prepare_connection()`` hooks executed against Datasette's database connections - use Datasette's own :ref:`prepare_connection() <plugin_hook_prepare_connection>` plugin hook to customize connections. ``PRAGMA recursive_triggers=on`` is now applied consistently to every connection Datasette opens - previously it was enabled just on the write connection, as a side effect of the first sqlite-utils based write. (:issue:`2831`)
- New :ref:`setting_busy_timeout_ms` setting controlling how long SQLite waits for a locked database file before giving up, previously hard-wired to the ``sqlite3`` driver's silent 5 second default. This matters when external processes write to the same database files Datasette is serving. (:issue:`2831`)
- New :ref:`setting_journal_mode` setting for opting mutable database files into SQLite's `WAL mode <https://www.sqlite.org/wal.html>`__ (or another journal mode), allowing reads and writes to proceed concurrently. Datasette does not change the journal mode of database files by default. (:issue:`2831`)
- A persistent internal database specified with ``--internal`` now has WAL mode enabled, matching the behavior of the default temporary internal database. (:issue:`2831`)
.. _v1_0_a36:

View file

@ -251,6 +251,11 @@ These can be passed to ``datasette serve`` using ``datasette serve --setting nam
executing SQLite queries (default=3)
sql_time_limit_ms Time limit for a SQL query in milliseconds
(default=1000)
busy_timeout_ms How long SQLite waits for a locked database file
before giving up (default=5000)
journal_mode Journal mode to set on mutable database files,
e.g. wal - leave blank to use each file's
existing mode (default=)
default_facet_size Number of values to return for requested facets
(default=30)
facet_time_limit_ms Time limit for calculating a requested facet

View file

@ -2023,8 +2023,8 @@ Example usage:
.. _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)
--------------------------------------------------------------------------------------------------------
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.
@ -2059,18 +2059,18 @@ If you need to retrieve every row returned by a statement, pass ``return_all=Tru
If you pass ``block=False`` this behavior changes to "fire and forget" - queries will be added to the write queue and executed in a separate thread while your code can continue to do other things. The method will return a UUID representing the queued task.
Each call to ``execute_write()`` will be executed inside a transaction. Pass
``transaction=False`` for statements such as ``VACUUM`` that cannot run inside
a transaction.
Each call to ``execute_write()`` will be executed inside a transaction, with the exception of statements that SQLite does not allow to run inside a transaction: ``VACUUM``, ``ATTACH``, ``DETACH`` and ``PRAGMA``. Those statements are executed in autocommit mode instead.
.. _database_execute_write_script:
await db.execute_write_script(sql, block=True)
----------------------------------------------
Like ``execute_write()`` but can be used to send multiple SQL statements in a single string separated by semicolons, using the ``sqlite3`` `conn.executescript() <https://docs.python.org/3/library/sqlite3.html#sqlite3.Cursor.executescript>`__ method.
Like ``execute_write()`` but can be used to send multiple SQL statements in a single string separated by semicolons.
Each call to ``execute_write_script()`` will be executed inside a transaction.
Each call to ``execute_write_script()`` will be executed inside a transaction - if any statement fails, none of the statements will be applied.
The exception is scripts that include statements which SQLite cannot execute inside a transaction - ``VACUUM``, ``ATTACH``, ``DETACH``, ``PRAGMA`` - or that manage transactions themselves using ``BEGIN``, ``COMMIT``, ``ROLLBACK``, ``SAVEPOINT`` or ``RELEASE``. Those scripts are executed using the ``sqlite3`` `conn.executescript() <https://docs.python.org/3/library/sqlite3.html#sqlite3.Cursor.executescript>`__ method instead, where each statement is committed as it executes.
.. _database_execute_write_many:
@ -2153,7 +2153,11 @@ The value returned from ``await database.execute_write_fn(...)`` will be the ret
If your function raises an exception that exception will be propagated up to the ``await`` line.
By default your function will be executed inside a transaction. You can pass ``transaction=False`` to disable this behavior, though if you do that you should be careful to manually apply transactions - ideally using the ``with conn:`` pattern, or you may see ``OperationalError: database table is locked`` errors.
By default your function will be executed inside a transaction. Datasette executes ``BEGIN IMMEDIATE`` on the write connection before calling your function, then commits the transaction when your function returns - or rolls it back if your function raises an exception. Nothing your function writes will be visible to other connections until that final commit.
Because the transaction is already open when your function is called, write methods from libraries such as `sqlite-utils <https://sqlite-utils.datasette.io/>`__ will nest their work inside it (sqlite-utils uses savepoints) rather than committing independently, so an exception rolls back everything the function did.
Your function should not manage transactions itself when ``transaction=True`` - do not execute ``BEGIN`` or call ``conn.commit()`` or ``conn.rollback()`` on the connection. If you need to manage transactions manually, pass ``transaction=False`` - ideally using the ``with conn:`` pattern, or you may see ``OperationalError: database table is locked`` errors.
If you specify ``block=False`` the method becomes fire-and-forget, queueing your function to be executed and then allowing your code after the call to ``.execute_write_fn()`` to continue running while the underlying thread waits for an opportunity to run your function. A UUID representing the queued task will be returned. Any exceptions in your code will be silently swallowed.
@ -2176,6 +2180,17 @@ Functions run using ``execute_isolated_fn()`` share the same queue as ``execute_
The return value of the function will be returned by this method. Any exceptions raised by the function will be raised out of the ``await`` line as well.
.. _internals_database_transactions:
How Datasette manages transactions
----------------------------------
Each mutable database has a single write connection, owned by a dedicated write thread that executes queued write tasks one at a time. Each task submitted with ``transaction=True`` (the default) runs inside its own transaction: Datasette executes ``BEGIN IMMEDIATE`` before invoking the task, commits if it completes and rolls back if it raises. One write task equals one transaction, and writes made by a task only become visible to readers when the task completes.
Read connections never open transactions - each read statement executes in autocommit mode with its own implicit SQLite read snapshot.
Datasette's write connections rely on the legacy transaction handling of Python's ``sqlite3`` module (connections opened with ``isolation_level=`` rather than the ``autocommit=`` parameter introduced in Python 3.12). This is a deliberate constraint shared with `sqlite-utils <https://sqlite-utils.datasette.io/>`__, which refuses connections created with ``autocommit=True`` or ``autocommit=False`` - any future migration to the modern transaction API would need to be coordinated across both projects. Plugins should not change the transaction control mode of connections Datasette passes to them.
.. _database_close:
db.close()

View file

@ -80,15 +80,18 @@ Shows a list of currently installed plugins and their versions. `Plugins example
.. code-block:: json
[
{
"name": "datasette_cluster_map",
"static": true,
"templates": false,
"version": "0.10",
"hooks": ["extra_css_urls", "extra_js_urls", "extra_body_script"]
}
]
{
"ok": true,
"plugins": [
{
"name": "datasette_cluster_map",
"static": true,
"templates": false,
"version": "0.10",
"hooks": ["extra_css_urls", "extra_js_urls", "extra_body_script"]
}
]
}
Add ``?all=1`` to include details of the default plugins baked into Datasette.

View file

@ -103,6 +103,39 @@ You can optionally set a lower time limit for an individual query using the ``?_
This would set the time limit to 100ms for that specific query. This feature is useful if you are working with databases of unknown size and complexity - a query that might make perfect sense for a smaller table could take too long to execute on a table with millions of rows. By setting custom time limits you can execute queries "optimistically" - e.g. give me an exact count of rows matching this query but only if it takes less than 100ms to calculate.
.. _setting_busy_timeout_ms:
busy_timeout_ms
~~~~~~~~~~~~~~~
How long SQLite should wait when a database file is locked by another connection or process before giving up with a ``database is locked`` error, in milliseconds. The default is 5000 (five seconds), matching the default used by Python's ``sqlite3`` module.
This mostly matters when other processes write to the same database files that Datasette is serving - a common pattern is a separate script (using `sqlite-utils <https://sqlite-utils.datasette.io/>`__ or similar) that periodically updates a database while Datasette serves it. A larger value makes Datasette more patient with long write transactions from those processes::
datasette mydatabase.db --setting busy_timeout_ms 10000
.. _setting_journal_mode:
journal_mode
~~~~~~~~~~~~
`Journal mode <https://www.sqlite.org/pragma.html#pragma_journal_mode>`__ to set on mutable database files. Leave blank (the default) to use whatever journal mode each database file already uses.
Setting this to ``wal`` enables SQLite's `Write-Ahead Logging <https://www.sqlite.org/wal.html>`__ mode, which allows reads and writes to proceed concurrently - in the default rollback journal mode each commit blocks readers, and long reads block the writer::
datasette mydatabase.db --setting journal_mode wal
When ``wal`` is selected Datasette also sets ``PRAGMA synchronous=NORMAL`` on the write connection, the standard pairing for WAL which reduces the number of ``fsync`` operations without weakening application-level consistency guarantees.
Other allowed values are ``delete``, ``truncate`` and ``persist``.
Things to be aware of before enabling WAL:
- WAL mode is persistent - it is recorded in the database file and stays in effect when other tools open the same file later.
- SQLite creates ``-wal`` and ``-shm`` files alongside the database file.
- WAL does not work reliably on network filesystems such as NFS.
- The directory containing the database must be writable by Datasette.
.. _setting_max_returned_rows:
max_returned_rows

View file

@ -1,6 +1,6 @@
from datasette.app import Datasette
from datasette.plugins import DEFAULT_PLUGINS
from datasette.utils import UNSTABLE_API_MESSAGE, escape_sqlite, tilde_encode
from datasette.utils import UNSTABLE_API_MESSAGE
from datasette.utils.sqlite import sqlite_version
from datasette.version import __version__
from .fixtures import make_app_client, EXPECTED_PLUGINS
@ -614,13 +614,13 @@ async def test_plugins_json(ds_client):
response = await ds_client.get("/-/plugins.json")
# Filter out TrackEventPlugin
actual_plugins = sorted(
[p for p in response.json() if p["name"] != "TrackEventPlugin"],
[p for p in response.json()["plugins"] if p["name"] != "TrackEventPlugin"],
key=lambda p: p["name"],
)
assert EXPECTED_PLUGINS == actual_plugins
# Try with ?all=1
response = await ds_client.get("/-/plugins.json?all=1")
names = {p["name"] for p in response.json()}
names = {p["name"] for p in response.json()["plugins"]}
assert names.issuperset(p["name"] for p in EXPECTED_PLUGINS)
assert names.issuperset(DEFAULT_PLUGINS)
@ -694,6 +694,8 @@ async def test_settings_json(ds_client):
"max_insert_rows": 100,
"max_post_body_bytes": 2 * 1024 * 1024,
"sql_time_limit_ms": 200,
"busy_timeout_ms": 5000,
"journal_mode": "",
"allow_download": True,
"allow_signed_tokens": True,
"max_signed_tokens_ttl": 0,
@ -930,37 +932,6 @@ async def test_tilde_encoded_database_names(db_name):
assert response2.status_code == 200
@pytest.mark.asyncio
@pytest.mark.parametrize("table_name", ("[foo]", "foo]", "[foo]/bar"))
async def test_table_with_reserved_characters_in_name(table_name):
# Table names containing characters such as "]" that cannot be escaped
# using SQLite [bracket] quoting used to break schema introspection and
# the table page - https://github.com/simonw/datasette/issues/2431
ds = Datasette()
db = ds.add_memory_database("test_reserved_table_names")
await db.execute_write(
"create table {} (id integer primary key, name text)".format(
escape_sqlite(table_name)
)
)
await db.execute_write(
"insert into {} (id, name) values (1, 'one')".format(escape_sqlite(table_name))
)
# Schema introspection (populate_schema_tables) must not crash:
db_response = await ds.client.get("/test_reserved_table_names.json")
assert db_response.status_code == 200
tables = {t["name"]: t for t in db_response.json()["tables"]}
assert tables[table_name]["count"] == 1
# And the table page itself must load and return the row:
table_response = await ds.client.get(
"/test_reserved_table_names/{}.json?_shape=array".format(
tilde_encode(table_name)
)
)
assert table_response.status_code == 200
assert table_response.json() == [{"id": 1, "name": "one"}]
@pytest.mark.asyncio
@pytest.mark.parametrize(
"config,expected",

View file

@ -2717,3 +2717,103 @@ async def test_create_using_alter_against_existing_table(
insert_rows_event = ds_write._tracked_events[1]
assert insert_rows_event.name == "insert-rows"
assert insert_rows_event.num_rows == 1
@pytest.mark.asyncio
async def test_create_table_with_failing_rows_is_atomic(ds_write):
# https://github.com/simonw/datasette/issues/2831
# If inserting the initial rows fails, the create table should
# be rolled back as well
token = write_token(ds_write)
response = await ds_write.client.post(
"/data/-/create",
json={
"table": "atomic_create",
"rows": [{"id": 1, "name": "one"}, {"id": 1, "name": "dupe"}],
"pk": "id",
},
headers=_headers(token),
)
assert response.status_code == 400
assert not await ds_write.get_database("data").table_exists("atomic_create")
@pytest.mark.asyncio
async def test_alter_table_with_failing_operation_is_atomic(ds_write):
# https://github.com/simonw/datasette/issues/2831
# If a later operation fails, earlier operations should be rolled back
token = write_token(ds_write)
response = await ds_write.client.post(
"/data/docs/-/alter",
json={
"operations": [
{"op": "add_column", "args": {"name": "new_col", "type": "text"}},
# Fails - column "title" already exists
{"op": "add_column", "args": {"name": "title", "type": "text"}},
]
},
headers=_headers(token),
)
assert response.status_code == 400
columns = await ds_write.get_database("data").table_columns("docs")
assert "new_col" not in columns
@pytest.mark.asyncio
async def test_insert_with_return_failing_row_is_atomic(ds_write):
# https://github.com/simonw/datasette/issues/2831
# Insert with "return": true runs one insert per row - a failure
# part way through should roll back the earlier rows
token = write_token(ds_write)
response = await ds_write.client.post(
"/data/docs/-/insert",
json={
"rows": [
{"id": 1, "title": "one"},
{"id": 2, "title": "two"},
{"id": 2, "title": "dupe"},
],
"return": True,
},
headers=_headers(token),
)
assert response.status_code == 400
count = (
await ds_write.get_database("data").execute("select count(*) from docs")
).single_value()
assert count == 0
@pytest.mark.asyncio
async def test_write_api_does_not_run_sqlite_utils_plugins(ds_write):
# https://github.com/simonw/datasette/issues/2831
# sqlite-utils plugins should not have their prepare_connection hooks
# executed against Datasette's connections
import sqlite_utils.plugins
from sqlite_utils import hookimpl
prepared = []
class TrackingPlugin:
@hookimpl
def prepare_connection(self, conn):
prepared.append(conn)
sqlite_utils.plugins.pm.register(TrackingPlugin(), name="datasette-test-tracking")
try:
token = write_token(ds_write)
response = await ds_write.client.post(
"/data/docs/-/insert",
json={"rows": [{"id": 1, "title": "one"}]},
headers=_headers(token),
)
assert response.status_code == 201
response = await ds_write.client.post(
"/data/docs/1/-/update",
json={"update": {"title": "two"}},
headers=_headers(token),
)
assert response.status_code == 200
assert prepared == []
finally:
sqlite_utils.plugins.pm.unregister(name="datasette-test-tracking")

View file

@ -109,7 +109,7 @@ def test_settings(config_dir_client):
def test_plugins(config_dir_client):
response = config_dir_client.get("/-/plugins.json")
assert 200 == response.status
plugins = response.json
plugins = response.json["plugins"]
assert "hooray.py" in {p["name"] for p in plugins}
assert "non_py_file.txt" not in {p["name"] for p in plugins}
assert "mypy_cache" not in {p["name"] for p in plugins}

View file

@ -334,3 +334,51 @@ async def test_orphan_stale_catalog_child_entries_removed(tmp_path):
assert response.status_code == 200
ds2.close()
@pytest.mark.asyncio
async def test_populate_schema_tables_is_a_single_write_task(tmp_path):
# https://github.com/simonw/datasette/issues/2831
# The catalog rebuild should be one atomic write task, so readers of the
# internal database never observe a database with deleted-but-not-yet-
# reinserted catalog rows
from datasette.app import Datasette
from datasette.utils.internal_db import populate_schema_tables
path = str(tmp_path / "data.db")
conn = sqlite3.connect(path)
conn.execute("create table t (id integer primary key, name text)")
conn.execute("create view v as select * from t")
conn.close()
ds = Datasette([path])
await ds.invoke_startup()
await ds.refresh_schemas()
internal_db = ds.get_internal_database()
write_fns = []
original_execute_write_fn = internal_db.execute_write_fn
async def counting_execute_write_fn(fn, *args, **kwargs):
write_fns.append(fn)
return await original_execute_write_fn(fn, *args, **kwargs)
internal_db.execute_write_fn = counting_execute_write_fn
try:
await populate_schema_tables(internal_db, ds.get_database("data"))
finally:
internal_db.execute_write_fn = original_execute_write_fn
assert len(write_fns) == 1
# And the catalog should be correct
tables = await internal_db.execute(
"select table_name from catalog_tables where database_name = 'data'"
)
assert [row["table_name"] for row in tables.rows] == ["t"]
views = await internal_db.execute(
"select view_name from catalog_views where database_name = 'data'"
)
assert [row["view_name"] for row in views.rows] == ["v"]
columns = await internal_db.execute(
"select name from catalog_columns where database_name = 'data' and table_name = 't'"
)
assert {row["name"] for row in columns.rows} == {"id", "name"}

View file

@ -112,6 +112,129 @@ async def test_execute_fn_transaction_false():
await db.execute_write_fn(run, transaction=False)
@pytest.mark.asyncio
async def test_execute_write_fn_wraps_sqlite_utils_writes_in_transaction():
# https://github.com/simonw/datasette/issues/2831
# sqlite-utils write methods commit their own transactions unless one is
# already open - the write thread must open one before running each task
# so that a failing task rolls back everything, including those writes.
datasette = Datasette(memory=True)
db = datasette.add_memory_database("test_txn_sqlite_utils")
await db.execute_write("create table t (a integer)")
def failing_task(conn):
sqlite_utils.Database(conn)["t"].insert_all({"a": i} for i in range(5))
assert conn.in_transaction
raise ValueError("boom")
with pytest.raises(ValueError):
await db.execute_write_fn(failing_task)
count = (await db.execute("select count(*) from t")).single_value()
assert count == 0
# The write connection should be back in autocommit mode
assert (
await db.execute_write_fn(lambda conn: conn.in_transaction, transaction=False)
) is False
# And a transaction=True task should see a transaction already open
assert (await db.execute_write_fn(lambda conn: conn.in_transaction)) is True
@pytest.mark.asyncio
async def test_execute_write_statements_disallowed_in_transaction(tmp_path):
# VACUUM (and ATTACH/DETACH/PRAGMA) cannot run inside a transaction, so
# execute_write() must run them outside one
# https://github.com/simonw/datasette/issues/2831
path = str(tmp_path / "test.db")
setup_conn = sqlite3.connect(path)
setup_conn.execute("create table t (a integer)")
setup_conn.close()
datasette = Datasette([path])
db = datasette.get_database("test")
await db.execute_write("vacuum")
await db.execute_write(" -- a comment\n VACUUM")
# But regular DML statements still run inside a transaction
from datasette.database import _can_execute_in_transaction
assert _can_execute_in_transaction("insert into t values (1)")
assert _can_execute_in_transaction("with x as (select 1) insert into t select 1")
assert not _can_execute_in_transaction("vacuum")
assert not _can_execute_in_transaction("/* hi */ PRAGMA optimize")
@pytest.mark.asyncio
async def test_execute_write_fn_sqlite_utils_integrity_error_rolls_back_task():
# https://github.com/simonw/datasette/issues/2831
datasette = Datasette(memory=True)
db = datasette.add_memory_database("test_txn_integrity")
await db.execute_write("create table t (id integer primary key)")
def two_inserts(conn):
table = sqlite_utils.Database(conn)["t"]
table.insert({"id": 1})
table.insert({"id": 1}) # IntegrityError
with pytest.raises(sqlite3.IntegrityError):
await db.execute_write_fn(two_inserts)
count = (await db.execute("select count(*) from t")).single_value()
assert count == 0
@pytest.mark.asyncio
async def test_execute_write_script_is_transactional():
# https://github.com/simonw/datasette/issues/2831
# A script that fails part way through should apply none of its statements
datasette = Datasette(memory=True)
db = datasette.add_memory_database("test_write_script_txn")
with pytest.raises(sqlite3.OperationalError):
await db.execute_write_script(
"create table one (id integer primary key);\n"
"insert into one (id) values (1);\n"
"insert into no_such_table (id) values (2);"
)
assert "one" not in await db.table_names()
@pytest.mark.asyncio
async def test_execute_write_script_with_transaction_unsafe_statements(tmp_path):
# Scripts containing statements that cannot run inside a transaction
# (VACUUM, BEGIN etc) still execute, using the previous autocommit behavior
path = str(tmp_path / "test.db")
sqlite3.connect(path).close()
datasette = Datasette([path])
db = datasette.get_database("test")
await db.execute_write_script("create table t (id integer primary key);\nvacuum;")
assert "t" in await db.table_names()
await db.execute_write_script("begin;\ninsert into t (id) values (1);\ncommit;")
count = (await db.execute("select count(*) from t")).single_value()
assert count == 1
@pytest.mark.asyncio
async def test_execute_write_fn_writes_invisible_to_readers_until_task_ends(tmp_path):
# https://github.com/simonw/datasette/issues/2831
path = str(tmp_path / "test.db")
setup_conn = sqlite3.connect(path)
setup_conn.execute("create table t (a integer)")
setup_conn.close()
datasette = Datasette([path])
db = datasette.get_database("test")
def insert_and_check(conn):
sqlite_utils.Database(conn)["t"].insert_all({"a": i} for i in range(5))
# A separate read-only connection must not see the rows mid-task
reader = sqlite3.connect("file:{}?mode=ro".format(path), uri=True)
try:
return reader.execute("select count(*) from t").fetchone()[0]
finally:
reader.close()
mid_task_count = await db.execute_write_fn(insert_and_check)
assert mid_task_count == 0
# After the task commits the rows are visible
count = (await db.execute("select count(*) from t")).single_value()
assert count == 5
@pytest.mark.parametrize(
"tables,exists",
(
@ -719,41 +842,6 @@ async def test_execute_write_fn_exception(db):
await db.execute_write_fn(write_fn)
@pytest.mark.asyncio
@pytest.mark.parametrize("num_sql_threads", (0, 1))
async def test_execute_write_fn_sqlite_utils_transaction(tmp_path, num_sql_threads):
# A write inside a failing Datasette task must never become visible or
# survive the rollback. Exercise both the synchronous and writer-thread
# paths against a file-backed database so a second connection can observe
# committed state independently.
db_path = tmp_path / "test.db"
sqlite3.connect(db_path).close()
ds = Datasette([str(db_path)], settings={"num_sql_threads": num_sql_threads})
db = ds.get_database("test")
await db.execute_write("create table items (id integer primary key)")
# This reader is used inside the write callback, which may run on another
# thread, but it is never accessed concurrently.
reader = sqlite3.connect(db_path, check_same_thread=False)
def insert_then_fail(conn):
# Datasette must open the outer transaction before sqlite-utils writes.
assert conn.in_transaction
sqlite_utils.Database(conn)["items"].insert({"id": 1})
# If sqlite-utils committed its own transaction, this would return 1.
assert reader.execute("select count(*) from items").fetchone()[0] == 0
# Simulate a later step failing after the sqlite-utils write succeeded.
raise ValueError("deliberate")
try:
with pytest.raises(ValueError, match="deliberate"):
await db.execute_write_fn(insert_then_fail)
# The outer transaction must roll back the sqlite-utils write as well.
assert reader.execute("select count(*) from items").fetchone()[0] == 0
finally:
reader.close()
db.close()
@pytest.mark.asyncio
@pytest.mark.parametrize("param_name", ["conn", "connection", "db", "c"])
async def test_execute_write_fn_accepts_any_single_param_name(db, param_name):
@ -1214,3 +1302,107 @@ async def test_database_close_is_idempotent(tmpdir):
# Second call should be a no-op, not raise
db.close()
ds._internal_database.close()
@pytest.mark.asyncio
async def test_recursive_triggers_enabled_on_all_connections(tmp_path):
# https://github.com/simonw/datasette/issues/2831
# Previously recursive_triggers was only enabled on the write connection,
# and only as a side effect of the first sqlite-utils based write - so
# trigger semantics could differ between connections
path = str(tmp_path / "test.db")
sqlite3.connect(path).close()
datasette = Datasette([path])
db = datasette.get_database("test")
write_value = await db.execute_write_fn(
lambda conn: conn.execute("PRAGMA recursive_triggers").fetchone()[0],
transaction=False,
)
read_value = await db.execute_fn(
lambda conn: conn.execute("PRAGMA recursive_triggers").fetchone()[0]
)
assert write_value == 1
assert read_value == 1
@pytest.mark.asyncio
async def test_busy_timeout_ms_setting(tmp_path):
# https://github.com/simonw/datasette/issues/2831
# The SQLite busy timeout should be an explicit, configurable policy
# instead of the sqlite3 driver's inherited 5 second default
path = str(tmp_path / "test.db")
sqlite3.connect(path).close()
datasette = Datasette([path], settings={"busy_timeout_ms": 250})
db = datasette.get_database("test")
read_value = await db.execute_fn(
lambda conn: conn.execute("PRAGMA busy_timeout").fetchone()[0]
)
write_value = await db.execute_write_fn(
lambda conn: conn.execute("PRAGMA busy_timeout").fetchone()[0],
transaction=False,
)
assert read_value == 250
assert write_value == 250
@pytest.mark.asyncio
async def test_busy_timeout_ms_default(tmp_path):
# Default matches the sqlite3 driver's historical 5 second default
path = str(tmp_path / "test.db")
sqlite3.connect(path).close()
datasette = Datasette([path])
db = datasette.get_database("test")
read_value = await db.execute_fn(
lambda conn: conn.execute("PRAGMA busy_timeout").fetchone()[0]
)
assert read_value == 5000
@pytest.mark.asyncio
async def test_journal_mode_setting_applies_wal(tmp_path):
# https://github.com/simonw/datasette/issues/2831
# Opt-in WAL support for mutable database files - not the default
path = str(tmp_path / "test.db")
sqlite3.connect(path).close()
datasette = Datasette([path], settings={"journal_mode": "wal"})
db = datasette.get_database("test")
mode = await db.execute_write_fn(
lambda conn: conn.execute("PRAGMA journal_mode").fetchone()[0],
transaction=False,
)
assert mode == "wal"
# WAL is paired with synchronous=NORMAL (1) on the write connection
synchronous = await db.execute_write_fn(
lambda conn: conn.execute("PRAGMA synchronous").fetchone()[0],
transaction=False,
)
assert synchronous == 1
@pytest.mark.asyncio
async def test_journal_mode_defaults_to_leaving_files_alone(tmp_path):
path = str(tmp_path / "test.db")
sqlite3.connect(path).close()
datasette = Datasette([path])
db = datasette.get_database("test")
mode = await db.execute_write_fn(
lambda conn: conn.execute("PRAGMA journal_mode").fetchone()[0],
transaction=False,
)
assert mode == "delete"
@pytest.mark.asyncio
async def test_persistent_internal_database_gets_wal(tmp_path):
# https://github.com/simonw/datasette/issues/2831
# The temporary internal database enables WAL - a persistent one passed
# via --internal should get the same treatment
internal_path = str(tmp_path / "internal.db")
datasette = Datasette(memory=True, internal=internal_path)
await datasette.invoke_startup()
internal_db = datasette.get_internal_database()
mode = await internal_db.execute_write_fn(
lambda conn: conn.execute("PRAGMA journal_mode").fetchone()[0],
transaction=False,
)
assert mode == "wal"

View file

@ -457,20 +457,6 @@ async def test_permissions_debug(ds_client, filter_):
assert checks == expected_checks
@pytest.mark.asyncio
@pytest.mark.parametrize(
"permissions_debug,expected_status",
(
(1, 200),
(0, 403),
),
)
async def test_permissions_debug_numeric_boolean(permissions_debug, expected_status):
ds = Datasette(config={"permissions": {"permissions-debug": permissions_debug}})
response = await ds.client.get("/-/permissions")
assert response.status_code == expected_status
@pytest.mark.asyncio
@pytest.mark.parametrize(
"actor,allow,expected_fragment",
@ -762,12 +748,7 @@ async def test_actor_restricted_permissions(
}
if actor.get("id"):
expected["actor_id"] = actor["id"]
data = response.json()
for key, value in expected.items():
assert data[key] == value
assert data["actor"] == actor
assert data["explanation"]["allowed"] is expected_result
assert data["explanation"]["summary"]
assert response.json() == expected
PermConfigTestCase = collections.namedtuple(
@ -1753,8 +1734,6 @@ async def test_permission_check_view_requires_debug_permission():
data = response.json()
assert data["action"] == "view-instance"
assert data["allowed"] is True
assert data["explanation"]["allowed"] is True
assert data["explanation"]["summary"]
@pytest.mark.asyncio
@ -1780,211 +1759,6 @@ async def test_permission_check_view_query_actions(action):
}
@pytest.mark.asyncio
async def test_permission_check_explains_specificity_for_hypothetical_actor():
ds = Datasette(
config={
"permissions": {"view-table": {"id": "alice"}},
"databases": {
"analytics": {
"permissions": {"view-table": False},
"tables": {
"public": {"permissions": {"view-table": {"id": "alice"}}}
},
}
},
}
)
ds.root_enabled = True
await ds.invoke_startup()
def path_for(child):
return "/-/check.json?" + urllib.parse.urlencode(
{
"action": "view-table",
"parent": "analytics",
"child": child,
"actor": json.dumps({"id": "alice"}),
}
)
public_response = await ds.client.get(path_for("public"), actor={"id": "root"})
assert public_response.status_code == 200
public = public_response.json()
assert public["actor"] == {"id": "alice"}
assert public["allowed"] is True
assert public["explanation"]["allowed"] is True
assert public["explanation"]["winning_scope"] == "resource"
public_rules = public["explanation"]["matched_rules"]
assert any(
rule["scope"] == "resource" and rule["effect"] == "allow" and rule["decisive"]
for rule in public_rules
)
assert any(
rule["scope"] == "parent"
and rule["effect"] == "deny"
and rule["ignored_because"] == "A more specific rule matched"
for rule in public_rules
)
private_response = await ds.client.get(path_for("private"), actor={"id": "root"})
assert private_response.status_code == 200
private = private_response.json()
assert private["allowed"] is False
assert private["explanation"]["allowed"] is False
assert private["explanation"]["winning_scope"] == "parent"
assert private["explanation"]["summary"].startswith("Denied by a parent-level rule")
@pytest.mark.asyncio
async def test_permission_check_explains_deny_wins_at_same_scope():
ds = Datasette(config={"permissions": {"view-table": {"id": "someone-else"}}})
ds.root_enabled = True
await ds.invoke_startup()
path = "/-/check.json?" + urllib.parse.urlencode(
{
"action": "view-table",
"parent": "analytics",
"child": "users",
"actor": json.dumps({"id": "alice"}),
}
)
response = await ds.client.get(path, actor={"id": "root"})
assert response.status_code == 200
data = response.json()
assert data["allowed"] is False
assert data["explanation"]["winning_scope"] == "global"
rules = data["explanation"]["matched_rules"]
assert any(rule["effect"] == "deny" and rule["decisive"] for rule in rules)
assert any(
rule["effect"] == "allow"
and rule["ignored_because"] == "A deny rule matched at the same scope"
for rule in rules
)
@pytest.mark.asyncio
async def test_permission_check_explains_default_deny():
ds = Datasette()
ds.root_enabled = True
await ds.invoke_startup()
path = "/-/check.json?" + urllib.parse.urlencode(
{
"action": "insert-row",
"parent": "analytics",
"child": "users",
"actor": json.dumps({"id": "alice"}),
}
)
response = await ds.client.get(path, actor={"id": "root"})
assert response.status_code == 200
data = response.json()
assert data["allowed"] is False
explanation = data["explanation"]
assert explanation["allowed"] is False
assert explanation["matched_rules"] == []
assert explanation["winning_scope"] is None
assert explanation["summary"] == (
"Denied because no permission rule matched this actor and resource."
)
@pytest.mark.asyncio
async def test_permission_check_explains_actor_restrictions():
ds = Datasette()
ds.root_enabled = True
await ds.invoke_startup()
restricted_actor = {
"id": "alice",
"_r": {"r": {"analytics": {"public": ["vt"]}}},
}
path = "/-/check.json?" + urllib.parse.urlencode(
{
"action": "view-table",
"parent": "analytics",
"child": "private",
"actor": json.dumps(restricted_actor),
}
)
response = await ds.client.get(path, actor={"id": "root"})
assert response.status_code == 200
data = response.json()
assert data["allowed"] is False
explanation = data["explanation"]
assert explanation["rule_allowed"] is True
assert explanation["restriction_allowed"] is False
assert explanation["allowed"] is False
assert explanation["restrictions"]
assert any(
restriction["allowed"] is False for restriction in explanation["restrictions"]
)
assert "actor's restrictions" in explanation["summary"]
@pytest.mark.asyncio
async def test_permission_check_explains_required_actions():
from datasette import hookimpl
from datasette.permissions import PermissionSQL
class StoreQueryPermissions:
@hookimpl
def permission_resources_sql(self, actor, action):
if not actor or actor.get("id") != "alice":
return None
if action == "store-query":
return PermissionSQL(
sql="SELECT 'analytics' AS parent, NULL AS child, 1 AS allow, 'alice can store queries' AS reason"
)
if action == "execute-sql":
return PermissionSQL(
sql="SELECT 'analytics' AS parent, NULL AS child, 0 AS allow, 'alice cannot execute SQL' AS reason"
)
ds = Datasette()
ds.root_enabled = True
await ds.invoke_startup()
ds.pm.register(StoreQueryPermissions(), name="store-query-test")
path = "/-/check.json?" + urllib.parse.urlencode(
{
"action": "store-query",
"parent": "analytics",
"actor": json.dumps({"id": "alice"}),
}
)
response = await ds.client.get(path, actor={"id": "root"})
assert response.status_code == 200
data = response.json()
assert data["allowed"] is False
explanation = data["explanation"]
assert explanation["rule_allowed"] is True
assert explanation["required_actions"][0]["action"] == "execute-sql"
assert explanation["required_actions"][0]["allowed"] is False
assert explanation["summary"] == (
"Denied because store-query also requires execute-sql, which was denied."
)
@pytest.mark.asyncio
async def test_permission_check_hypothetical_actor_validation():
ds = Datasette()
ds.root_enabled = True
await ds.invoke_startup()
response = await ds.client.get(
"/-/check.json?action=view-instance&actor=not-json",
actor={"id": "root"},
)
assert response.status_code == 400
assert response.json()["error"].startswith("Invalid actor JSON:")
response = await ds.client.get(
"/-/check.json?action=view-instance&actor=%5B%5D",
actor={"id": "root"},
)
assert response.status_code == 400
assert response.json()["error"] == "actor must be a JSON object or null"
@pytest.mark.asyncio
async def test_root_allow_block_with_table_restricted_actor():
"""

View file

@ -1482,7 +1482,7 @@ async def test_plugin_is_installed():
datasette.pm.register(DummyPlugin(), name="DummyPlugin")
response = await datasette.client.get("/-/plugins.json")
assert response.status_code == 200
installed_plugins = {p["name"] for p in response.json()}
installed_plugins = {p["name"] for p in response.json()["plugins"]}
assert "DummyPlugin" in installed_plugins
finally:

View file

@ -2,8 +2,8 @@
Tests for the canonical JSON success envelope.
Every JSON object returned by a Datasette endpoint on success should include
"ok": true. /-/plugins intentionally returns a top-level array instead, while
/-/databases and /-/actions use the object envelope.
"ok": true. (Endpoints that return a top-level array are being converted to
objects separately - see /-/plugins, /-/databases, /-/actions.)
"""
import pytest
@ -79,16 +79,17 @@ async def test_permissions_post_success_has_ok_true(ds_envelope):
@pytest.mark.asyncio
async def test_plugins_json_is_array(ds_client):
async def test_plugins_json_is_object(ds_client):
response = await ds_client.get("/-/plugins.json")
assert response.status_code == 200
data = response.json()
assert isinstance(data, list)
assert all(isinstance(plugin, dict) for plugin in data)
assert set(data.keys()) == {"ok", "plugins"}
assert data["ok"] is True
assert isinstance(data["plugins"], list)
# ?all=1 should include Datasette's default plugins in the same shape
response_all = await ds_client.get("/-/plugins.json?all=1")
all_plugins = response_all.json()
assert len(all_plugins) > len(data)
all_plugins = response_all.json()["plugins"]
assert len(all_plugins) > len(data["plugins"])
@pytest.mark.asyncio