3 Commits

Author SHA1 Message Date
debont80 4492eee054 Self-review: drop the formatter re-export shim, import @shared/format directly
The prior commit kept a pass-through `export { fmtBytes, fmtSpeed, fmtEta }` in
main/lib/formatters.ts purely so existing importers wouldn't have to change —
which re-exported fmtSpeed (zero importers, dead) and left download.ts re-
exporting fmtBytes/fmtEta with no consumer at all, undercutting the "one home"
goal.

Now the real consumers import from the canonical module directly:
- probe.ts and the download.test import from @shared/format.
- formatters.ts keeps only its internal `import { fmtBytes }` (for parseProgress)
  and no longer re-exports.
- download.ts re-exports only parseProgress (its own function); the dead
  fmtBytes/fmtEta re-export is removed.

No behavior change. typecheck + 248 tests + eslint + prettier green.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-30 21:10:11 -04:00
debont80 3d09174c16 H4 + H2 formatter half: carry raw numbers across the progress boundary
Closes H4 and the remaining formatter half of H2.

- New @shared/format.ts is the single home for fmtBytes/fmtSpeed/fmtEta, imported
  by both main and renderer. main/lib/formatters.ts re-exports them.
- DownloadProgress now carries raw speedBytesPerSec/etaSeconds instead of
  pre-formatted speed/eta strings. main's parseProgress emits the raw numbers.
- The renderer stores the raw numbers on DownloadItem; QueueItem formats them for
  per-item display, and summarizeQueue sums/maxes them directly. The lossy
  string->number->string round-trip in queueStats (parseSpeed / parseEtaSeconds /
  a local formatSpeed / formatEta) is deleted, along with a duplicate fmtEta in
  store/downloads.ts.
- Per-item and aggregate speed now use the same 1024-based scale; the aggregate
  previously used a 1000-based formatter (a latent inconsistency).

Tests updated for the raw-number contract (parseProgress, summarizeQueue).
typecheck + 248 tests + eslint + prettier green.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-30 21:03:14 -04:00
debont80 8ab85da67e Fix H5/H6/M21/L37-40/L47/L55/L68/L72/L76/L91/L98/L108/L165/L166 audit items
H5: HistoryEntry now stores formatId+formatHasAudio; redownload passes them
    back to addFromUrl so the original yt-dlp format is reused, not guessed.
    Compound quality labels stripped to the preset prefix for old history.
L72: thumbnail now passed in redownload so re-queued rows keep their art.
H6: TerminalView log capped at MAX_LOG_LINES=2000 to prevent unbounded growth.
M21: --audio-quality omitted for lossless audioFormats (flac/wav).
L166: fmtEta hour rollover fixed in all 3 locations -- 2h00:00 not 120:00.
L165: queueStats formatSpeed precision aligned with fmtBytes.
L55: QueueItem suppresses sizeLabel when already in probed-format quality label.
L47: probeMeta null sends empty meta event so Resolving clears via SR6.
L68: notifiedBackground resets on window show for subsequent close-while-downloading.
L76: dup warning falls back to URL when title is still a placeholder.
L91: SettingsView onFormatSelect uses indexOf instead of unsafe tuple cast.
L98: Embed chapters field gets a hint text.
L108: BrowserWindow created with explicit title AeroFetch.
L37: pure formatters extracted to src/main/lib/formatters.ts and unit-tested.
L38: buildArgs test covers webm thumbnail suppression.
L39: CROP_SQUARE_PPA test is structural not exact-string.
L40: looksLikeUrl/looksLikeSingleVideo extracted to src/renderer/src/lib/urlHelpers.ts.

typecheck + 242 tests + eslint + prettier green.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-30 13:29:35 -04:00