Allow the Datasette subprocess more time to become reachable under CI load.
If startup still times out, terminate the process and include stdout and stderr in the assertion so future failures have useful diagnostics.
Keep copied spreadsheet templates aligned with the single-row insert columns while still previewing auto primary keys in bulk insert rows.
Treat invalid JSON during live bulk-insert validation as unfinished input so Firefox does not expose browser-specific parse errors while typing.
Render omitted auto integer primary key values as muted auto text in the bulk insert preview so generated IDs are easier to understand.
Expose the auto primary key flag in insert UI metadata and cover the preview behavior in Playwright.
Expose conflict handling in the bulk insert UI when pasted rows include primary key columns, and route update-or-insert selections through the existing upsert API.
Add live validation for bulk textarea column errors and Playwright coverage for upsert permissions, conflict options, and validation behavior.
Refs shttps://github.com/simonw/datasette/pull/2813#issuecomment-4878320713
Expose whether the actor can insert rows in the create-table page data, then use that flag to hide and guard the create-table-from-data path. This avoids offering a flow that fails after preview for create-table-only actors.
Refs https://github.com/simonw/datasette/pull/2813#issuecomment-4878146376
Convert blank delimited cells to null when create-table-from-data inferred an integer or float column. This keeps sqlite-utils from treating mixed numeric and empty-string values as text.
Add a Playwright regression covering CSV table creation with a blank numeric value.
New create-view and drop-view actions.
Also fix a related bug in analyze_sql_tables(): SQLite's authorizer fires a spurious SQLITE_DELETE callback against the view name when a view is dropped (the same thing it does for dropped tables), which was incorrectly surfaced as a delete-row requirement on the view. Broaden the existing drop-table-delete suppression to cover dropped views too.
Closes#2819
Works entirely using existing JSON API.
User can paste in CSV, TSV or JSON (or drop a text file on).
They get a preview of the rows to be imported, then they
are imported in max_insert_rows batches.
Add a collapsed rename-table section to the alter table modal and include rename_table operations in the review/apply flow.
Redirect to the renamed table URL after applying changes and cover the review text in Playwright.
Share default value controls between the create and alter table dialogs and expose create-table default expressions to the frontend.
Add create-table not-null/default handling and align the shared foreign key picker behavior across both dialogs.
Include current foreign key metadata in the alter table page data and allow the foreign-key-targets endpoint to be read by actors with alter-table permission for a specific table.
Add API and HTML data tests for the new alter-table foreign key support.
- Add foreignKeyTargetsPath to create table page data
- Filter hidden tables from database-level foreign key target results
- Update JSON API docs and tests for filtered targets
Returns a list of tables with a single primary key, and for each one
the name of that primary key column and its SQLite type affinity.
This will be used by the create table UI to suggest foreign keys.
- Add add_foreign_key, drop_foreign_key, and set_foreign_keys operations.
- Validate flat fk_table and fk_column arguments with Pydantic.
- Document the API and cover inferred primary-key and validation cases.
- Add fk_table and optional fk_column support to create-table columns.
- Validate create-table requests with Pydantic while preserving existing errors.
- Document the API and cover inferred primary-key and validation cases.
Refs https://github.com/simonw/datasette/pull/2789#issuecomment-4733544452
- Extract reusable helpers for database and table action permission preloading.
- Precompute those permissions before building table-page HTML data.
- Document the default table actions plugin.
- Use a per-process socket path for the UDS test fixture.
- Clean up stale socket files before and after the fixture runs.
- Close the HTTP client and wait for the Datasette subprocess to exit.
