Skip to content

merge: upstream安全コミット11件の取り込み#83

Merged
MocA-Love merged 11 commits into
mainfrom
merge/upstream-safe-picks
Apr 5, 2026
Merged

merge: upstream安全コミット11件の取り込み#83
MocA-Love merged 11 commits into
mainfrom
merge/upstream-safe-picks

Conversation

@MocA-Love
Copy link
Copy Markdown
Owner

Summary

upstream (superset-sh/superset) から安全なコミット11件をcherry-pickで取り込み。

確実に安全 (5件)

おそらく安全 (6件)

残りの要検討コミット(別PRで対応予定)

  • N6: VS Code fuzzy scorer(ハイブリッド方式で検索エンジンのみ差し替え)
  • G2→N8→N10→N14: ターミナル系チェーン(ライフサイクル分離、terminalId、Host Service)
  • G3→N11: ファイルツリー系(WorkspaceSidebar、Alert API変更)
  • N1: GitHub polling簡素化(フォークのSyncServiceと非互換のためスキップ予定)

Test plan

  • デスクトップアプリのビルド確認
  • ターミナル起動・TUIアプリ(Claude Code)の表示確認
  • ファイルビューアの開閉・再利用動作確認
  • GitHub統合ページの表示確認
  • Markdownペインの表示確認

MocA-Love and others added 11 commits April 5, 2026 09:35
upstream 538f306 — marketing/package.json脆弱性パッチ
…ion (superset-sh#3121)

These per-tool hooks add overhead without providing value for Codex.
Only SessionStart, UserPromptSubmit, and Stop are needed.
…rset-sh#3054)

The `emulator.onData` callback only forwarded terminal query responses
(DA1, DSR) to the PTY subprocess when `attachedClients.size === 0`.
When a renderer client is attached (the normal case when viewing a
terminal), the headless emulator's DA1 response was silently dropped.

This caused fish shell to wait 10 seconds for a DA1 response that
never arrived, then print:
  "warning: fish could not read response to Primary Device Attribute
  query after waiting for 10 seconds"

The previous fix in superset-sh#3030 removed the `shellReadyState === "pending"`
guard but left the `attachedClients.size === 0` check intact. The
renderer's xterm also generates DA1 responses, but those go through
`write()` which drops all escape sequences during shell init — so
neither path delivered the response to fish.

Remove the `attachedClients.size === 0` condition so the headless
emulator always forwards query responses to the subprocess. This is
safe because `sendWriteToSubprocess` writes directly to the PTY via
IPC, bypassing the renderer's write path entirely.

Fixes superset-sh#3028

Co-authored-by: zombopanda <1810282+zombopanda@users.noreply.github.com>
Co-authored-by: Kiet Ho <hoakiet98@gmail.com>
)

* fix(desktop): reduce line-height in markdown renderer pane styles

Tighten line-height and spacing in default and tufte markdown styles for a more compact reading experience in panes.

* Lint
…n mode (superset-sh#3093)

Decouples the reuse-existing-pane lookup from the openInNewTab flag so
that an already-open file is always located first. The pane is still
only reused in-place when openInNewTab is false, preventing duplicate
panes when the user's file-open mode is set to "new-tab".
* Lint

* chore(desktop): bump version to 1.4.7
Patch upgrade to latest 40.x stable to fix 4 Dependabot security alerts:
protocol handler injection, second-instance IPC OOB read, use-after-free
in fullscreen/download callbacks, and registry key injection.
Remove the PostHog feature flag gate and desktop paywall from the GitHub
integration so it is accessible to all plan tiers. Update billing plans
to reflect GitHub as a free feature and replace GitHub with Slack in the
pro features paywall preview.
…set-sh#3174)

* Reuse dnd backend

* Remove provider from v2 workspace

* Lint

* fix: await async closeTab in bulk close callbacks

* docs(panes): document DndProvider requirement for consumers
…pps (superset-sh#3154)

* fix(desktop): send correct terminal dimensions after attach for TUI apps

When opening a terminal via preset (Ctrl+1), the PTY was spawned with
stale dimensions because fitAddon.fit() ran before the container had
its final layout. Most CLIs handle the late SIGWINCH from the debounced
ResizeObserver, but ink-based TUI apps like Claude Code commit to the
initial width during their first render cycle, resulting in a narrow UI.

After a successful createOrAttach, schedule a requestAnimationFrame to
re-fit and send corrected dimensions to the PTY. The resize is only
sent if dimensions actually changed, making it a no-op when the initial
size was already correct.

* refactor(terminal): remove stale dimension handling comments

Removed outdated comments regarding terminal dimension handling after attach, as the logic has been updated to ensure correct dimensions are sent to TUI apps. This cleanup enhances code readability and maintains focus on the current implementation.

* refactor(terminal): streamline terminal dimension handling logic

Reintroduced the requestAnimationFrame logic for updating terminal dimensions after attach, ensuring accurate dimensions are sent to TUI apps. This change enhances the responsiveness of terminal resizing and maintains consistency in the terminal's display behavior.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants