Context
Commit 0a2a0feb (chore: biome --write --unsafe across the repo + two manual fixups) landed inside the architecture-ralph branch (PR #998) as "cycle 17" of the loop. It's a 100-file style sweep: noUnusedImports, useImportType, useOptionalChain. It is not architecture work — it's noise inside a PR that is otherwise about volatility-based decomposition and framework extraction.
Mixing the style sweep into the architecture PR makes review harder. Two of biome's unsafe rewrites also narrowed return types and needed manual ?? false / ?? null patches — visible in the commit but harder to spot when reading 100 files of formatting changes.
What to do
Re-do the biome cleanup as its own dedicated PR off master (or whatever the relevant base is after #998 merges). Scope:
biome check . --write --unsafe at the current master HEAD.- Verify the two known unsafe-rewrite traps (or any new ones biome surfaces):
packages/integrations/codex/src/transcript.ts — ?? false after optional-chain collapse if the function's return type is bare boolean.packages/client/src/canvas/dock/WorkspaceGrid.tsx — ?? null after optional-chain collapse if the function's return type is T | null.
- Run
just check + just test-unit + just smoke before push.
- One commit, scoped title (e.g.
chore: biome --write --unsafe), no architecture changes interleaved.
Why this didn't get reverted out of #998
A direct git revert 0a2a0feb against the post-#998 tree produces content conflicts: the cycle-17 changes have been further modified by 10 subsequent commits (renames, the master merge, the @kolu/solid-xterm@0.2 reshape that rewrote Terminal.tsx). The cleanest revert path requires un-doing legitimate later work too, which would be worse than the original mixing.
#998's description flags 0a2a0feb as the style commit so reviewers can skip it; this issue tracks doing it properly afterwards.
Acceptance
Context
Commit
0a2a0feb(chore: biome --write --unsafe across the repo + two manual fixups) landed inside the architecture-ralph branch (PR #998) as "cycle 17" of the loop. It's a 100-file style sweep:noUnusedImports,useImportType,useOptionalChain. It is not architecture work — it's noise inside a PR that is otherwise about volatility-based decomposition and framework extraction.Mixing the style sweep into the architecture PR makes review harder. Two of biome's unsafe rewrites also narrowed return types and needed manual
?? false/?? nullpatches — visible in the commit but harder to spot when reading 100 files of formatting changes.What to do
Re-do the biome cleanup as its own dedicated PR off
master(or whatever the relevant base is after #998 merges). Scope:biome check . --write --unsafeat the currentmasterHEAD.packages/integrations/codex/src/transcript.ts—?? falseafter optional-chain collapse if the function's return type is bareboolean.packages/client/src/canvas/dock/WorkspaceGrid.tsx—?? nullafter optional-chain collapse if the function's return type isT | null.just check+just test-unit+just smokebefore push.chore: biome --write --unsafe), no architecture changes interleaved.Why this didn't get reverted out of #998
A direct
git revert 0a2a0febagainst the post-#998 tree produces content conflicts: the cycle-17 changes have been further modified by 10 subsequent commits (renames, the master merge, the@kolu/solid-xterm@0.2reshape that rewroteTerminal.tsx). The cleanest revert path requires un-doing legitimate later work too, which would be worse than the original mixing.#998's description flags
0a2a0febas the style commit so reviewers can skip it; this issue tracks doing it properly afterwards.Acceptance
masterdoing the biome sweep.just check,just test-unit,just smoke).