AP24110011695 commented on PR #3566: URL: https://github.com/apache/tika/pull/3566#issuecomment-6743583406
Thank you for the definitive analysis, @tballison — withdrawing this PR. You're right on all three counts: the guard only relocated the junk line into the extracted content, the test couldn't see the problem (cues() collects p contents and the last assertion counted only data-start divs), and on the Jinja path the text wrongly inherited the previous cue's data-start. Main's behavior — drop the unparseable timing string, keep the text in a fresh div, never credit text to the wrong cue — is the correct contract, and your explanation of the WebVTT parsing algorithm (a block whose timings fail to parse is dropped, not degraded to text) makes that clear. Withdrawing as the fix premise was wrong. Thank you for the detailed review — it was a better lesson in the parser's contract than the PR was a contribution. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
