Turso is a distributed SQLite database built on libSQL, offering HTTP-based
connectivity with auth token authentication. This change adds full support:
- New DatabaseType::Turso and PoolKind::Turso variants
- TursoClient driver using libSQL HTTP pipeline API (/v2/pipeline)
- Bearer token auth via password or url_params (auth_token=)
- Multi-statement batch pipeline for transactional integrity
- Standalone BEGIN/COMMIT/ROLLBACK treated as no-ops (pipeline is auto-commit)
- Full metadata support: tables, columns, indexes, foreign keys, triggers, DDL
- Frontend: connection form, SQL completion (SQLite syntax), capability sets
- Built-in databases (SQLite/Turso/DuckDB/RQLite/Access) placed first in selector
- Unit tests (15) + integration tests against live libsql-server (12)
- Single connection pool, skips TCP probe
Closes#948
MongoDB shell syntax like db.collection.find({}) was incorrectly
parsed as SQL, causing a "sql parser error" diagnostic in the editor.
Added MongoDB guards in shouldRunSqlSemanticDiagnostics and
refreshSemanticDiagnostics, matching existing Elasticsearch logic.
Also fix a pre-existing TiDB cloud URL param ordering test failure.
- Add Navicat numeric ConnType codes (1-9) to typeMap for correct DB type mapping
- Add port-based inference fallback (6379→Redis, 27017→MongoDB, etc.)
- Expand typeMap with ClickHouse, Snowflake, KingbaseES, GaussDB, OceanBase
- Skip unrecognized connections instead of silently falling back to MySQL
- Extend hexToRgba() to handle rgba/rgb input passthrough
- Add pipette button with native color picker + text input in popover
- Custom color and preset colors are mutually exclusive
* feat(grid): add layer preview for geometry columns with Leaflet map
* fix: restore dev:backend script with env vars and cargo watch
* chore: update pnpm-lock.yaml for leaflet and dom-to-image-more dependencies
---------
Co-authored-by: dqn <someonegdeng@qq.com>
Frontend: add regex-based MongoDB URL parser to handle multi-host URIs
that the WHATWG URL parser rejects. Backend: separate
server_selection_timeout from connect_timeout for multi-host URIs to
prevent topology discovery from being cancelled by tokio timeout.
DOM mode: replace box-shadow with outline to avoid stacking with row border-b.
Canvas mode: draw row border before selected cell overlay so selection uses
uniform color on all four sides; clamp top edge to prevent canvas clipping.
coercePostgresArrayValue() only handled JSON [...] format but editor
outputs PG {...} format. When the user entered and exited edit mode
without changes, the string "{1,2,3}" failed to parse back to an
array, creating a false dirty detection via strict equality.
Add parsePostgresArrayText() recursive descent parser and return
oldValue reference when parsed result is equivalent.
- Route SELECT * FROM <index> through /_search for simple browses; aggregates/projections stay on /_sql
- Drive completion columns from /_mapping (flattened dotted fields, multi-fields)
- Switch editor completion to SQL when the active statement starts with SELECT/WITH on ES
- Skip generic SQL semantic diagnostics for ES (hyphenated wildcard indices caused 500s)
- Wire ES into server-side pagination plan; derive completion context from line blocks
- Fix result-grid pagination exceeding in-memory result sets (allRowsLoaded)
The data grid header only showed column names, and column types were
available only via tableMeta (open-table view). Arbitrary query results
(e.g. `select * from pg_depend`) therefore showed no type at all, which
is exactly the case the reporter hit.
Backend: add `column_types` to QueryResult (serde-default, backward
compatible) and populate it for the native drivers where the type is
readily available — PostgreSQL, MySQL, SQL Server, ClickHouse. Other
drivers leave it empty for now (no behavior change); schemaless stores
(Mongo/Redis/ES) have no column types.
Frontend: render a type row under each column name in the grid header,
color-coded by type. The type is resolved from tableMeta first (richer,
includes precision) and falls back to the query result's column_types by
index. Add a `showColumnTypesInHeader` setting (default on) and keep the
msgpack tab-result cache compatible. The source-selection logic is
extracted to lib/dataGridColumnType.ts with unit tests.
Co-authored-by: vrustx <vrustx@vrustxdeMac-mini.local>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The sidebar already supports multi-select, so deletion now resolves the context-clicked selected connection set before confirming. Connection removal gained a batched path to persist once while pruning sidebar layout, pins, active state, connection errors, and stale selection for every removed connection.
Constraint: No new dependencies and existing single-connection delete behavior must remain available.\nRejected: Delete every selected tree item regardless of type | mixed tree selections could unexpectedly remove connections while table/object deletion has separate SQL confirmation behavior.\nConfidence: high\nScope-risk: narrow\nTested: pnpm fmt; pnpm test; pnpm typecheck; pnpm lint; pnpm build\nNot-tested: Manual desktop UI interaction against a running Tauri window
Co-authored-by: caisin <caisin@caisins-Mac-mini.local>
When editing SQL with concrete tables already referenced (a FROM clause,
a "table." qualifier, or an INSERT column list), column completions only
carried `computeBoost + keyBoost (0/500)` — lower than keyword boosts
(1200-1900) — so the table's own columns were interleaved among keywords
instead of ranking at the top where the user expects them.
Give columns a relevance boost (+2000) in these referenced-table contexts
so they rank above plain keywords. Added a unit test asserting columns
outrank keywords when a table is referenced.
Co-authored-by: vrustx <vrustx@vrustxdeMac-mini.local>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Address the follow-up UI feedback on the data detail dialogs:
- cell detail: toggle between raw value and formatted JSON inside a single
value area (via a "format JSON" button) instead of stacking the value,
formatted value and formatted JSON blocks, removing the redundancy.
- kill horizontal scrollbars: add break-words to the formatted-value block,
and constrain row/column detail value columns with w-full max-w-0 so long
unbroken values wrap instead of overflowing the table.
- row detail: column name uses break-words instead of break-all to stop long
names from wrapping character-by-character.
- unify footer button layout across the three detail dialogs (bulk-copy
actions on the left, copy-name on the right).
- add a search box to row & column detail to filter fields by name / value /
row number (filterDataGridDetailFields + unit tests).
- i18n: add detailSearchPlaceholder / detailSearchNoMatch for all 6 locales.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Database connection context menus now show a copy action when the connection has enabled transport layers. The action starts or reuses the ordered transport chain, copies the final local forwarding port, and keeps runtime config cached so later disconnect cleanup can stop the created tunnel.
Constraint: Local testing sometimes needs external tools to connect through DBX's generated proxy/tunnel endpoint
Rejected: Copy full host:port endpoint | user specifically asked for the port and existing tunnel host is always localhost
Rejected: Require an active pool before copying | local test workflows may need the tunnel before opening the DBX connection
Confidence: high
Scope-risk: moderate
Tested: pnpm test -- packages/app-tests/connectionTransport.test.ts (ran full app-tests: 862 passed)
Tested: pnpm typecheck
Tested: pnpm exec oxlint --vue-plugin apps/desktop/src/components/sidebar/TreeItem.vue apps/desktop/src/lib/connectionTransport.ts apps/desktop/src/lib/api.ts apps/desktop/src/lib/tauri.ts apps/desktop/src/lib/http.ts
Tested: cargo check -p dbx-core -p dbx-web
Tested: cargo check -p dbx
Tested: cargo fmt --check --all
Co-authored-by: caisin <caisin@caisins-Mac-mini.local>
t8y2