fix(desktop): expose full usage request values - #3676
Conversation
Generated-by: OpenAI Codex
Astro-Han
left a comment
There was a problem hiding this comment.
Scope
First independent full review of exact head ed70f7f34265fec39f456226bb913f83937d5012.
I covered the request-table rendering change, the upstream settings:usageStats → SettingsStore.usageStats → stored session metadata path, all usage row shapes (model, tool, numeric, and interactive session cells), the long-tail Storybook coverage, and the exact-head CI result.
Findings
No P0–P3 findings in the covered scope.
The change keeps the existing UsageStats data authority intact and only changes how rows are mapped into Astryx Table: primitive cells use the table's default renderer (so truncate-mode tooltip behavior applies), while React elements such as session buttons retain a custom renderer. The fixed request-column widths and long-content story assertions cover the intended visibility regression without changing the underlying usage values or privacy boundary. I found no new exposure of request IDs, provider response fields, or secrets.
Verification
- Exact-head CI
testrun32694397282, job97333713315: terminal success. - Changed-file Biome check: passed.
git diff --check: passed.
Evidence gap
The PR adds behavioral Storybook play assertions for width, full long-target text, and tooltip content, but no screenshot or visual artifact is attached. I record the missing visual artifact as an acceptance-evidence gap, not a code finding.
There was a problem hiding this comment.
Approving at head ed70f7f34265fec39f456226bb913f83937d5012.
This head already had a full independent review covering the request-table rendering change, the settings:usageStats to SettingsStore.usageStats to stored session metadata path, every usage row shape, and the Storybook coverage — with no P0-P3 findings and no open threads. That review was published as a comment rather than an approval, which is why the PR sat blocked; this converts it.
I also re-checked integration with the latest main before approving. Main has since touched settings-pages.stories.tsx, so the overlap is real, but a trial merge is clean and after merging neither usage-settings-page.tsx nor settings-pages.stories.tsx produces any renderer typecheck error. Hosted test is SUCCESS on this exact head.
简体中文
在 head ed70f7f34265fec39f456226bb913f83937d5012 上给出 approve。
这个 head 之前已经有过一次完整的独立复审,覆盖了 request 表格渲染改动、settings:usageStats → SettingsStore.usageStats → 会话元数据的整条路径、所有 usage 行形态以及 Storybook 覆盖,结论是 0 个 P0–P3,也没有未决线程。当时那次是以普通 comment 发布的,没有转成 approve,所以这个 PR 一直卡在 blocked——这次把它补上。
approve 之前我另外复核了它和最新 main 的集成:main 后来动过 settings-pages.stories.tsx,确实有重叠,但试合并干净,合并后 usage-settings-page.tsx 和 settings-pages.stories.tsx 都没有 renderer 类型检查错误。这个精确 head 上的 test 是 SUCCESS。
|
LGTM, merging. 简体中文LGTM,合并中。 |
Summary
UsageLongTailStorybook interaction covering full timestamps and overflow tooltipsFixes #3675
Verification
npm --workspace @maka/desktop run typechecknpx biome check apps/desktop/src/renderer/settings/usage-settings-page.tsx apps/desktop/stories/settings/settings-pages.stories.tsxnpm --workspace @maka/desktop run build:renderergit show --check HEADUsageLongTailstory in Chinese and Englishbefore


after
AI use
Select exactly one:
Tool(s) and scope: OpenAI Codex diagnosed the table rendering issue, implemented the fix and regression coverage, ran verification, and drafted the issue and PR text. The human contributor of record reviewed and owns the contribution.
Checklist
Does this PR entail a change in behavior?