dschmidt commented on PR #3037: URL: https://github.com/apache/tika/pull/3037#issuecomment-5339262672
Thanks for the follow-up pass. These five are all judgement calls rather than defects, so I'm leaving them as-is: 1. **Re-opening the file per preview** – the per-preview `skip` is an `lseek`, not a read, and previews are now de-duplicated and capped at 32, so reusing a channel wouldn't buy anything measurable. 2. **Immutable config** – the setters are required by the JSON config deserializer, and this mirrors `PSDParser`/`PDFParser` exactly (the config *reference* is already `final`). Diverging here would make this parser inconsistent with its siblings. 3. **Non-positive `maxPreviewLengthBytes`** – `PSDParserConfig.maxDataLengthBytes` is unvalidated for the same reason; kept consistent. 4. **Resource-name gaps** – numbering is intentionally by position in the file (`thumbnail-N` = Nth preview), which is stable regardless of any `shouldParseEmbedded` filtering. 5. **Magic comments** – `MM.+`/`II+.` follows the existing `MM.*`/`II*.` shorthand a couple of lines above. Happy to reconsider any of these if you'd prefer otherwise. -- 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]
