SEZ9 commented on PR #12299:
URL: https://github.com/apache/seatunnel/pull/12299#issuecomment-5903186076

   @CryoThrust thanks for the merge-readiness summary on `2209591b1a9c`.
   
   Before I sign off, I'd like to close the loop on the points from my earlier 
review, since I can't see from the thread that they've been addressed on this 
head. Could you confirm the status of each, or point me to where they were 
resolved?
   
   1. **Charset consistency** – the same log file is decoded via the 
platform-default `readFileToStr` when it's within the limit but as UTF-8 when 
it's over the limit. Both paths should use the same explicit charset.
   2. **Truncation marker** – a truncated response currently carries no 
indication that the head of the file was dropped, so the Web UI iframe and API 
consumers can't distinguish a full log from a tail.
   3. **Compatibility record** – the default 64 MB tail changes the existing 
behaviour of both the v1 and v2 log-content endpoints for every deployment. 
Please add an entry to `incompatible-changes.md`.
   4. **v1 REST docs** – the v1 endpoints are now bounded too, but 
`rest-api-v1.md` (en/zh) is not updated; the option is documented only on the 
v2 page.
   5. **Docs wording** – the note hard-codes "last 64 MB" rather than referring 
to the configured value, and the zh page has a stray double backtick right 
above the new note.
   6. **Duplicated conversion** – the MB-to-bytes conversion and the `<=0` 
sentinel are repeated verbatim in `LogBaseServlet` and `LogService`; a single 
shared helper would be cleaner.
   7. **Memory bound** – `tailFromLineStart` copies the retained window a 
second time, so per request the tail is materialised several times on the heap 
and the actual bound is looser than the PR claims. If you'd rather keep that as 
a follow-up, please at least adjust the PR description so it doesn't overstate 
the guarantee.
   
   Items 1–5 are the ones I'd like fixed in this PR; 6 and 7 I'm fine deferring 
if you'd prefer, but please say so explicitly. Once those are in (or you point 
me to where they already landed), I'll do the final pass.
   
   <!-- streview-comment:1420 -->


-- 
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]

Reply via email to