v0.7.5: bundled PG tools, schema CRUD, object search, copy-as-DDL, object dependencies (#11)

* chore: bump version to 0.7.5 (Task 1.1)

* feat(db): pure object DDL/search/depend builders (Task 1.2)

* feat(models): ObjectSearchHit, DependencyInfo, PgToolPaths, tool source fields (Task 1.3)

* feat(backup): bundle-aware pg tool resolution, system-first fallback (Task 2.1)

* feat(objects): schema CRUD commands + integration test (Task 2.2)

* feat(objects): search_objects command (current-schema, all types) (Task 2.3)

* feat(objects): get_object_ddl for all browsable types (Task 2.4)

* feat(objects): pg_depend object dependencies + schema contents (Task 2.5)

* feat(ipc): register object management commands (Task 3.1)

* feat(ipc): frontend wrappers + types for object management (Task 3.2)

* feat(build): declare bundled pg_tools resources (Task 3.3)

* test(capabilities): lock objects capability PG-only for search/ddl/dependencies (Task 3.4)

* feat(ui): Cmd+K object search palette in DB viewer (Task 4.1)

* feat(ui): schema CRUD menu + dependency dialog (Task 4.2)

* feat(ui): copy-as-DDL + dependency view context menu in Objects (Task 4.3)

* feat(ui): dependency check before table drop + bundled-tool status (Task 4.4)

* docs: v0.7.5 release notes + status table + bundled-tools (Task 5.1)

* ci: build/verify bundled pg client tools per platform before tauri build (Task 5.2)

* fix(objects): report pg_rewrite dependencies as the dependent view (pg_class)

pg_depend records view dependencies via the view's internal rewrite rule
(classid = pg_rewrite). The user-facing dependent object is the VIEW itself,
so map that classid to pg_class (name still resolved through ev_class).
Fixes the live-PG integration test object_dependencies_for_table_includes_view
and makes the dependency dialog readable for view dependents.

* test: add idempotent PG integration-test seed script

Seeds the first PG test db (GRIDLINE_TEST_SRC) with the objects the
#[ignore] integration tests assert against: users (+users_id_seq),
products (3 rows), orders (3 rows), order_summary view, audit_log +
get_user functions, user_role enum. Idempotent: DROP + recreate.

* fix(build): regenerate multi-resolution icons + roadmap Keychain item

- icon.ico previously contained a single 16x16 frame (Windows scaled it
  up -> blurry taskbar/start-menu icon). Regenerated from the 512px source
  via 'tauri icon': ICO now has 16/24/32/48/64/256 frames, icns has full
  @2x coverage up to 1024px, PNGs re-rendered from the same source.
- Added icons/icon.png (512px) to bundle.icon so Linux hicolor installs
  a high-DPI entry.
- ROADMAP: log the 'Enable Keychain' toggle (currently a form-only
  placeholder) under Next up -> Connection & credentials.

* docs(readme): dynamic version-free download links + 3-col table

- release.yml: set releaseAssetNamePattern to '[name]_[platform]_[arch][setup][ext]'
  (version-free), so asset filenames are stable across releases:
  Gridline_darwin_aarch64.dmg, Gridline_windows_x64-setup.exe,
  Gridline_linux_amd64.deb, Gridline_linux_x86_64.rpm, etc.
- README: both download tables now 3-column (OS | Architecture | Download)
  and link via GitHub's releases/latest/download/<file> redirect — they
  always point at the newest published release, no per-release edits.
- AGENTS.md: releases checklist updated — download links stay version-free.

* docs(readme): platform-per-column download table (macOS | Windows | Linux)

Table now mirrors the release layout: one column per OS with a logo row
and a download-links row underneath. Links stay version-free via
releases/latest/download/<file> (releaseAssetNamePattern in release.yml).

* docs(readme): revert download table to clean image-less 3-col layout

Logo row was hard to read on GitHub dark mode; the plain OS | Architecture
| Download table is cleaner and still uses dynamic releases/latest links.
This commit is contained in:
2026-08-05 20:46:36 +08:00
committed by GitHub
parent cbc54119ed
commit 04e8ed3300
67 changed files with 2084 additions and 108 deletions
+18 -7
View File
@@ -6,6 +6,15 @@ This file is the **source of truth** for what Gridline is building. [AGENTS.md](
---
## ✅ Shipped (0.7.5)
- **Bundled PG client tools** — static `pg_dump`/`pg_restore`/`psql` ship with the app (system-first, bundled fallback) so backup/restore/sync work with no separate install.
- **Schema CRUD** — create/rename/drop schemas from the object tree; CASCADE drop with typed-name confirm + dependency warning.
- **Global object search** — Cmd+K palette in the DB viewer, current-schema scope, all object types; results open a table tab or jump to the Objects view.
- **Copy as DDL for any object** — `CREATE` DDL for every browsable type (tables via pg_dump; pg_get_*def passthrough; sequences/enums/extensions/views synthesized).
- **Object dependencies** — `pg_depend` "what depends on this?" view, shown before destructive drops (table drop, schema drop).
- **Version bump** 0.7.0 → **0.7.5**.
## ✅ Shipped (0.7.0)
- **New Connection screen revamp** — a single progressive flow: connection-string input + 2-column provider tab grid → expands into the full configuration form (label, tags/env/folder, General + SSH·SSL tabs). Removes the simple/detailed toggle.
@@ -36,18 +45,15 @@ Gridline can already **browse** every PostgreSQL object type (functions, trigger
- Staged through the changes queue (with confirmation) — never a surprise DDL
- **Out of scope this release:** the MySQL "Objects" view stays deferred (see In the queue) — this iteration is PostgreSQL-only
### Shipping alongside (companions)
- **Schema CRUD** — create/rename/drop schemas from the object tree
- **Global object search** — Cmd+K-style search across tables, functions, triggers, sequences, and extensions by name
- **Copy as DDL for any object** — `CREATE FUNCTION` / `CREATE TRIGGER` / … via the same DDL-generation used for edit/drop previews (also covers the "Table structure export" queue item)
- **Object dependencies** — `pg_depend`-based "what depends on this object?" view, shown before drops so nothing breaks silently
### Admin follow-up (after object management)
- **PostgreSQL users/roles + grants management** — create roles and set privileges from a UI (DB Pro has this at 0% on their roadmap — a differentiator to hold)
- **Maintenance actions** — right-click table → VACUUM / ANALYZE / REINDEX
### Connection & credentials
- **Wire up the "Enable Keychain" toggle** — currently a form-only placeholder: the flag is submitted and stored with the connection record, but the Rust backend never reads it and the frontend store unconditionally calls `saveConnectionPassword`. Decide the intended behavior (e.g. off = store the password with the connection record / don't persist at all, on = OS keychain as today) and implement the conditional path + migration for existing records.
## 📋 In the queue
### Full Redis Support
@@ -128,6 +134,11 @@ Deferred from 0.7.0, slated for this bucket:
## ✅ Shipped
- Bundled PG client tools — `pg_dump`/`pg_restore`/`psql` shipped with the app, system-first with bundled fallback (v0.7.5)
- Schema CRUD — create/rename/drop schemas with CASCADE + dependency warning (v0.7.5)
- Global object search — Cmd+K across all object types in the current schema (v0.7.5)
- Copy as DDL for any object (v0.7.5)
- Object dependencies — `pg_depend` view before destructive drops (v0.7.5)
- Tauri 2.0 + React 19 + TypeScript 5.8 project shell
- PostgreSQL and SQLite browse/query support
- Full MySQL DB viewer — connect, browse, query, edit + changes queue (v0.7.0)