fix(cli): keep whisper CJK tokens as separate timed words (#6058)

* fix(cli): keep whisper CJK tokens as separate timed words

whisper.cpp marks word boundaries with a leading space, but Chinese and
Japanese tokens never carry one, so parseWhisperCpp merged every CJK token
into the previous word, across segments too. A 9.5-minute Mandarin episode
came out as two "words" of about 280 s each, leaving captions and
sentence-building nothing to split.

A token now starts its own word when it crosses a CJK boundary, full-width
punctuation still attaches to the word before it, and cues end at 。!?.

* fix(ci): capture package name in release-age exclusion
This commit is contained in:
Xuanru Li
2026-10-22 10:44:19 +00:00
committed by GitHub
parent f5873b8089
commit 3a0caeeb2d
3 changed files with 107 additions and 16 deletions
+1 -1
View File
@@ -479,7 +479,7 @@ function writeConsumerFixture(packDir, packedWorkspaces) {
export const CONSUMER_MIN_RELEASE_AGE_SECONDS = 24 * 60 * 60;
const BLOCKED_BY_AGE =
/No version matching "([^"]+)" found for specifier "[^"]*" \(blocked by minimum-release-age/g;
/No version matching "[^"]+" found for specifier "([^"]+)" \(blocked by minimum-release-age/g;
export function installConsumer(dir, { args = [], env = process.env } = {}) {
const install = () =>