fix(web-shell): land on the split's first pane when a shrink folds the split - #7000
Merged
Conversation
…e split When the viewport shrinks below the large-screen breakpoint the split view folds back to the single chat. If that chat had no session of its own — the common case when the split was entered from the Session Overview or a `?split=a,b` link — the user was stranded on an empty "new chat" instead of the split's first (leftmost) pane. This restores a fallback that was tried and then superseded during #6746: the fold-time re-point to the first pane was dropped because it wiped the git branch and changed the session+URL of a chat that already had a session. Re-add it but guarded on the *empty* chat (`!currentSessionId`) — an empty chat has no git branch or session to preserve — so it fixes the empty-new-chat case without regressing the branch-preservation case. Best-effort and standalone-split-only, as before. Covered by two tests: lands on the first pane when the outer chat is empty, and keeps the existing session (no re-point) when it isn't.
Collaborator
|
Thanks for the PR! Template looks good ✓ — all required sections present including bilingual summary. Problem: Observed regression with solid evidence. The split-to-single-chat fold strands users on an empty "new chat" when entered from Session Overview or a Direction: Aligned. This is a targeted fix to the Web Shell split-view fold behavior — a purely client-side UX concern. The Size: 13 production lines (App.tsx) + 99 test lines (App.test.tsx). Not applicable to core module protection. Approach: Scope is exactly right — one conditional block in the fold handler, with the guard on the precise invariant (empty chat vs. own session). Error handling is best-effort with a swallowed catch, which is appropriate here. The two tests directly validate both branches of the guard. No scope creep, no drive-by changes. Clean. Moving on to code review. 🔍 中文说明感谢贡献！ 模板完整 ✓ — 所有必需章节齐全，含双语说明。 问题：已观测到的回归，证据充分。从 Session Overview 或 方向：对齐。这是 Web Shell 分屏折叠行为的定向修复，纯客户端 UX 问题。 规模：13 行生产代码（App.tsx）+ 99 行测试代码（App.test.tsx）。不涉及核心模块保护。 方案：范围完全正确 — 折叠处理函数中一个条件块，守卫放在精确的不变量上（空 chat vs. 有自己的会话）。错误处理是尽力而为的 catch 吞掉，在这里是合适的。两个测试直接验证了守卫的两个分支。没有范围蔓延，没有顺手改动。干净。 进入代码审查 🔍 — Qwen Code · qwen3.7-max Reviewed at |
Collaborator
🖼️ web-shell visual previewRendered against a mock daemon (no real backend): the PR base vs this PR head Screenshots · before / after
Full-resolution recordings (.webm) are attached to the workflow run. — Qwen Code · web-shell visuals |
Collaborator
Code ReviewIndependent proposal: In the Comparison with the diff: The PR's approach is exactly what I would have done. The guard on No critical blockers. No AGENTS.md violations. No scope creep. Reuse check: TestingThis is a browser viewport-resize behavior (React Full test suite (with fix): Controlled revert (fix removed, same test): The controlled revert confirms the test is meaningful — without the one-block fix, 中文说明代码审查独立方案： 在 与 diff 的对比： PR 的方案和我想的完全一样。 无关键阻塞。无 AGENTS.md 违反。无范围蔓延。 复用检查： 测试这是浏览器视口缩放行为（React 完整测试套件（含修复）：105/105 通过 ✅ — Qwen Code · qwen3.7-max Reviewed at |
Collaborator
|
Confidence: 5/5 — Clean across every stage; would merge without hesitation. This is exactly the kind of PR the gate is designed to let through quickly: a focused, minimal fix to a real observed regression, with precise guards, meaningful tests, and clear reasoning about why the previous attempts failed and why this reconciliation works. The core insight — re-point to the first pane only when the chat is empty ( 13 production lines, 99 test lines, 2 files. Nothing to cut, nothing to add. 中文说明置信度：5/5 — 每个阶段都干净；毫不犹豫地合并。 这正是门禁设计来快速放行的那种 PR：对一个已观测到的真实回归的精准、最小修复，有精确的守卫、有意义的测试，以及对为什么之前尝试失败、为什么这次调和有效的清晰推理。 核心洞察——仅在 chat 为空时（ 13 行生产代码，99 行测试代码，2 个文件。没有什么要砍的，也没有什么要加的。 — Qwen Code · qwen3.7-max Reviewed at |
qwen-code-ci-bot
approved these changes
Jul 16, 2026
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
LGTM, looks ready to ship. ✅
Contributor
Code Coverage Summary
CLI Package - Full Text ReportCore Package - Full Text ReportFor detailed HTML reports, please see the 'coverage-reports-22.x-ubuntu-latest' artifact from the main CI run. |
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
Reviewed. Not reviewed: chunk 1 — no agent reported covering these; nobody read them.
— qwen3.7-max via Qwen Code /review
ytahdn
approved these changes
Jul 16, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this PR does
Restores the behavior where narrowing the browser past the large-screen breakpoint — which folds the split view back to a single chat — lands on the split's first (leftmost) pane instead of an empty "new chat", when the chat being folded to has no session of its own. A chat that already has its own session is left untouched, exactly as before.
Why it's needed
When the viewport shrinks below the large-screen breakpoint, the split view auto-folds to the single chat. If the split was entered straight from the Session Overview or a
?split=a,blink, the underlying chat has no session of its own, so a bare fold strands the user on an empty "new chat" instead of the sessions they were just looking at.This exact fallback was implemented once and then deliberately superseded during #6746: re-pointing the chat to the first pane on fold wiped the git branch and changed the session+URL of a chat that did have a session, so that PR shipped "fold without touching the connection" — which reintroduced the empty-new-chat symptom. The right reconciliation is to re-point only when the chat is empty (
!currentSessionId): an empty chat has no git branch or session/URL to preserve, so it fixes the empty-new-chat case without regressing the branch-preservation case that motivated dropping it.Reviewer Test Plan
How to verify
?split=a,blink in a fresh tab) so the underlying single chat has no session of its own.Automated (from
packages/web-shell):npm testis green (105 files / 1672 tests). Two new tests inApp.test.tsxdrive the realAppcomponent: one assertsloadSession(firstPane)fires when the folded chat is empty, the other asserts it does NOT fire when the chat already has a session. Controlled revert (below) shows the empty-chat test fails onmainand passes with the fix.npm run typecheck/lint/format:checkclean.Evidence (Before & After)
Controlled revert — same test, same
Appcomponent, only the one-block fix toggled. Without it the shrink never re-attaches a session (stranding a blank new-chat); with it the folded chat lands on the split's first pane, while the guard leaves a chat that already has its own session untouched:Tested on
Environment (optional)
Local:
npm test/npm run typecheck/npm run lint/npm run format:checkinpackages/web-shell, plus a controlled-revert experiment (git stash push App.tsx→ run the empty-chat test → it fails →git stash pop→ it passes).Risk & Scope
loadSessionfailure (e.g. a non-primary-workspace first pane the single connection can't own) simply leaves the empty chat, i.e. today's behavior.!currentSessionIdguard and covered by a dedicated test, but only in the standalone (uncontrolled) split path.Linked Issues
N/A — reported regression: the narrow-fold shows a new-session window instead of the leftmost pane.
中文说明
这个 PR 做了什么
恢复这样的行为：当浏览器拉窄、越过大屏断点导致分屏折叠回单个 chat 时，如果被折叠到的那个 chat 自己没有会话，就落到分屏的第一个（最左侧）pane，而不是一个空的「新建会话」。如果那个 chat 本来就有自己的会话，则保持不变，和之前完全一样。
为什么需要
当视口缩小到大屏断点以下时，分屏会自动折叠回单个 chat。如果分屏是从 Session Overview 或
?split=a,b链接进来的，底层 chat 并没有自己的会话，于是折叠后就把用户丢在一个空的「新建会话」上，而不是他们刚才在看的会话。这个 fallback 其实实现过一次，但在 #6746 里被有意替换掉了：折叠时把 chat 重新指向第一个 pane 会抹掉一个「本来有会话」的 chat 的 git branch、并改变它掉回去的 session+URL，所以那个 PR 最终改成了「折叠时不动连接」——空会话的症状也就回来了。正确的调和方式是：仅当 chat 为空时（
!currentSessionId）才重新指向第一个 pane——空 chat 没有 git branch、也没有要保留的 session/URL，因此既修好空会话的情况，又不回归当初促使去掉它的「保留 branch」情况。复现验证方式
?split=a,b链接），使底层单 chat 没有自己的会话。自动化（在
packages/web-shell下）：npm test全绿（105 文件 / 1672 测试）。App.test.tsx里新增两个测试，驱动真实App组件：一个断言空 chat 折叠时会触发loadSession(firstPane)，另一个断言 chat 已有会话时不会触发。下方的「controlled revert」显示空会话测试在main上失败、加了修复后通过。typecheck/lint/format:check均干净。证据（Before & After）
Controlled revert——同一个测试、同一个
App组件，只切换这一处修复。没有它时，缩窄从不重新挂上会话（把用户丢在空白新 chat）；有了它，折叠后的 chat 落到分屏第一个 pane，而守卫让「已有自己会话」的 chat 保持不动（上图）。风险与范围
loadSession失败（例如单连接无法拥有的非主 workspace 第一个 pane）就保持空 chat，即今天的行为。!currentSessionId守卫保住了它，并有专门测试覆盖，但仅在独立分屏路径上。关联 Issue
N/A —— 回归上报：窄屏折叠显示了新建会话窗口，而不是最左侧 pane。