mirror of
https://github.com/stablyai/orca.git
synced 2026-09-27 08:02:35 +00:00
Jinwoo-H
85d5ef0b8a
Main is at35005fb65c, one commit past the sha in the brief (`35005fb65c`, a `src/main` pi change, unrelated to mobile). Two conflicts, both against PR B (C7.10 item B, mermaid on the page), plus two counts the merge falsified. `mobile/.oxlintrc.json`: both ignore entries kept, this branch's generated page factory and PR B's generated page engine. `mobile/.gitignore` auto-merged carrying both artifacts; all six `*.generated.ts` on disk after the four postinstall generators run are exactly its six entries, with no strays. `config/scripts/mobile-web-app-session-terminal-closure.test.mjs`: the conflict is the docstring's reading, and both sides' explanations are kept. The number is re-measured on the merged tree rather than summed, which is what the sum would have hidden: main35005fb65cmodules 4324 local 974 merged modules 4284 local 934 (-40, -40) Main's own recorded 4323 was againstec82173130and main has drifted a module since, so the merged total is one above this branch's -40 added to PR B's +3. Both sides read with `mobileWebAppRouteClosure(SESSION_ROUTE)` and the four generators run first, main's in a scratch worktree detached at the same sha and removed afterwards. `MODULES_WITH_MERMAID` is renamed `SESSION_ROUTE_MODULES`, because the number is now neither side's alone. The byte reading is not re-measured and stays anchored atec82173130, since that function returns no byte total and a figure produced here would be a different computation. `mobile/src/test-support/census-source-files.ts` (PR B's walker) still excludes this branch's page factory: the predicate is the `.generated.ts` suffix, which the artifact's name carries. Verified on the real tree, not only in scratch — the walk over `mobile/src` returns 2,189 files with no `.generated.ts` among them while the artifact is on disk, and the same artifact copied to `.bundled.ts` comes back in the walk at 2,190. Its docstring said five artifacts and its test said a sixth would be the failure; both are one higher now, so both are corrected. Claude-Session: https://claude.ai/code/session_01JNnE9qzUZMMnqpZWCqM3nb