Repository navigation
Conversation
Subagents can list multiple fallback models, but one thinking setting cannot express the right request for each alternative. Keep the existing flat pin format and clamp each request to the target model.
|
补充 CI 排查:#1013 的两项失败均可在当前上游
这两处均不属于本 PR 的备用模型功能,故未混入无关修改。Rust 检查和最新基线检查已经通过。 |
Reuse the upstream token-only fix in ui-kit.css; preserve the existing 4px checkbox radius while satisfying the style-token check.
Integrate upstream main at 35d6903 into the existing subagent fallback thinking feature branch. Preserve PR history and incorporate upstream style-token, documentation-locale, and architecture fixes.
Use the existing unit line-height token and move the light model menu's three shadow layers into the shared theme tokens. Add regressions for the unchanged line height and shadow values.
Keep the upstream model-menu shadow token when merging the per-fallback thinking feature. Limit the import-page contract test to its module so unrelated settings cards cannot make the JS gate fail.
GitHub's pull-request webhook base SHA can lag behind the synthetic merge checkout when main advances. Compare newly added source files with the merge candidate's first parent instead, so upstream files cannot block an unrelated PR.
muzimu217
left a comment
There was a problem hiding this comment.
核心功能(每个备用模型独立思考强度档位、默认 inherit、切换时按目标模型能力调整)本身清晰有用,验证段也扎实(e2e 23/23、针对性测试+构建+typecheck、Windows 隔离说明如实)。一个主要意见:
PR 范围与正文严重不符,疑似分支叠加。 正文只描述"备用模型思考强度"一件事,但文件清单里有大量与本功能无关的内容:mid-autumn-egg 整个特性(assets 资源、scene/timeline/sampling、preferences、overlay、store)、taskbar-unread-badge/overlay、generated-images 一族、notification-inbox、ci.yml 改动,以及多份无关 spec(21-image-generation、NAV、ui-design-system 等),合计 +5561/−310。如果这些是其它在飞工作叠进来的:按本仓库"一个 PR 只交一件事"的强惯例(维护者此前对多个关注点的 PR 一律要求拆分),建议把本特性 rebase 到干净 main 上单独出——功能本身不大(SubagentFallbackModels/SubagentEditorSheet + model_fallbacks.rs + ADR 0307 + 23 号 spec),拆出来过审会快很多,叠加的部分各自走各自的 PR。
拆分时记得 spec 只保留真正属于本特性的部分(03-runtime/23 与相关 en/zh 成对项),e2e-plan 条目同步核对。
# Conflicts: # apps/desktop/src/components/settings/SubagentFallbackModels.tsx # apps/desktop/test/settings-import-page.test.mjs # docs/adr/README.md
Restore the CI architecture-base change, its doc note and test, and the composer model-menu token test to upstream main; they belong to separate fixes, not per-fallback thinking.
|
感谢 review @muzimu217,之前的 diff 确实混进了无关的改动。我已经合并了最新的 main,并移除了所有不属于本功能的内容(CI architecture-base 的改动及其测试和文档说明,以及 composer 模型菜单 token 的测试)。现在 PR 是 26 个文件、+262/−29,只包含「每个备用模型单独设置思考强度」:shared 和 host-core 中 |
Bring the per-fallback thinking branch up to date with upstream main so PR vastsa#1013 no longer conflicts. The only conflict was in SubagentFallbackModels.tsx, where main now resolves fallback row labels through subagentModelDisplay() to keep provider names and disabled / unavailable status. Rows still parse the `pin|level` entry first, then use main's display resolution for the pin. The fallback user-path test from main assumed the move-up button was the first button in each row and that the add-fallback trigger was already in view. Per-fallback thinking menus add a button and make rows taller, so the test now finds move-up by its accessible name and scrolls the add trigger into view before clicking, as a person would.
背景
子智能体可以配置多个备用模型,但目前只能为子智能体整体设置思考强度。
改动
inherit或标准思考强度档位;默认显示inherit。验证
ee31195f0;Base main:21dcd3ddb。test:e2e:subagent-models和完整test:e2e通过(23/23;无密钥的实时模型用例跳过);test:e2e:subagents因 Windows 隔离限制未运行。