Repository navigation
fix(desktop): compact AskToolCard height, merge actions into header - #1292
Merged
Merged
Conversation
Keep the layout regression assertion aligned with the reduced option viewport so the suite guards the current compact-card behavior.
Owner
|
CI 当前失败项已定位并已在 PR 分支更新: 另有一项合入前测试缺口:这次把“跳过 / 下一题 / 提交”按钮从卡片底部移入头部,改变了 DOM 与键盘 Tab 顺序。当前已有测试是源码/CSS 契约,没有覆盖用户经键盘或点击选择选项后推进问题、提交答案的组件交互路径。请补一条组件级交互测试,确认新的操作顺序下仍可完成选择、跳过和提交,再考虑合入。 |
The header absorbed every confirmation affordance, changing DOM and Tab order. Mount the production AskToolCard in real Chromium against the real store slice and walk select/next/submit, skip/submit, decline-all and a custom answer, asserting the header Tab order and exact resolutions.
Contributor
Author
|
交互测试缺口已补上(commit a6e3b3d),另把最新 main 合入:
|
* feat(storage): add safe custom data migration Move desktop and Chromium state through a cold, verified migration so host and browser data stay consistent across drives. Keep the original roots as backups until explicit cleanup, and limit cache clearing to rebuildable paths. * docs(adr): index custom storage decision Keep the architecture decision discoverable and satisfy the docs index check. * fix(i18n): keep Korean release notes aligned Place the storage migration highlight in the current release so every locale has the same versioned highlight count. * test(desktop): follow startup initialization ownership Assert application identity and early error handlers where the startup split installs them, while preserving the logger wiring check in the composition root.
…pted-in default model A spawn without a modelKey fell back to the provider default even when that model was never enabled for AI delegation, so a plugin worker could run on a model the Task gate would refuse for the same operation (issue vastsa#1183). The fallback now only picks a model that carries the availableForSubagents opt-in; with none enabled the spawn fails with MODEL_NOT_CONFIGURED instead of silently spending an unapproved model.
The Pi 0.99.1 migration replaced the source-aware limit resolution in enrichProvider's binding mapping with the raw catalog number, so a user who pinned a 1M context window on a relay model saw the published 128k value overwrite it on every read (issue vastsa#1176, spec §9.1: a user limit is never replaced). resolveBindingLimits + modelConfigWithBinding are restored for the mapping, while catalog-sourced windows keep following published corrections.
fix(desktop): stop session-collaboration spawn from spending the un opted-in default model
Resolve catalog and user token limits through the provenance helper while keeping model capability shaping out of provider enrichment. This retains hand-typed model overrides and the existing enrichment boundary.
…mits fix(desktop): keep user-pinned model limits across provider enrichment
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
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.
压缩提问确认卡(AskTool)高度,把确认操作并入头部空白区,聊天输出不再被大面积遮挡。
改动(2 个文件,纯 UI,行为不变):
验证: