fix(edit-history): bounded image probe, linear path scan (review of #693)

- probeLocalPaths: every agent-printed candidate goes through
  probePathKind() first (no pastCap, bulk scan), sequentially; only a path
  that answered 'file' gets realpath + stat. A path on a mount that stopped
  answering no longer hangs the sync, a queued sync or remove().
- PATH_TOKEN_RE: branch 4 starts only at a token boundary (`:` excluded so
  URLs stay ignored), branch 1 stops at `:`. Linear in the line length
  (30,000 `#` took ~3 s before).
- Load older yields to the event loop between the files it parses.
- GET /edits answers unsupported without loading a store.
- CLAUDE.md paragraph trimmed to the rules; probe and regex rules added to
  architecture-invariants#edit-history-panel.

Tests: stall test through the real sync and probe with a stat that never
settles; time-budget tests with 30,000-character inputs and a 139-result
base64 chunk; route test asserts no store load for unsupported sessions.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
RandalixandClaude Opus 5.5 committed 2026-11-03 09:46:23 +01:00
1 parent bf2a705705
commit de43657676
10 files changed
+229 -35

No files matched your search

+10 -4
View File
@@ -104,13 +104,19 @@ const TOKEN_CHAR = '[^\\s\'"`<>|;&()=,:*?]';
const PATH_TOKEN_RE = new RegExp(
[
// 1. file:// URL (Codex: `Saved to: file:///…`) — may contain spaces up to the extension.
`file://(/[^\\n]*?\\.${IMAGE_EXT})${EXT_END}`,
// No `:` inside: it keeps the lazy scan from running to the line end from every
// `file://` of a line that repeats it (quadratic otherwise).
`file://(/[^\\n:]*?\\.${IMAGE_EXT})${EXT_END}`,
// 2./3. Quoted paths, spaces allowed.
`"([^"\\n]*\\.${IMAGE_EXT})"`,
`'([^'\\n]*\\.${IMAGE_EXT})'`,
// 4. Unquoted token. The lookbehind keeps a match from starting inside a URL, a glob
// or a longer name; the name before the extension must not be empty.
`(?<![\\w:/.~*?$-])((?:${TOKEN_CHAR}*/)?[^\\s'"\`<>|;&()=,:*?/]+\\.${IMAGE_EXT})${EXT_END}`,
// 4. Unquoted token. A match starts only at a token boundary (line start or a character
// that ends a token), never inside a URL (`:` is not a boundary), a glob or a longer
// name. Anchoring at boundaries keeps this linear: a lookbehind that only excluded
// some characters let a match start at every other character of a long token
// (`####…`, a base64 blob) and rescan the token from each, quadratic in the line
// length. The name before the extension must not be empty.
`(?<=^|[\\s'"\`<>|;&()=,*?])((?:${TOKEN_CHAR}*/)?[^\\s'"\`<>|;&()=,:*?/]+\\.${IMAGE_EXT})${EXT_END}`,
].join('|'),
'gi'
);