git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH v2] http*: cleanup slot->local after fclose

From
Junio C Hamano <gitster@pobox.com>
Date
May 31, 2009, 20:21 UTC
Message-ID
<7vws7xqazi.fsf@alter.siamese.dyndns.org>
In-Reply-To
<20090531175413.962a55c3.rctay89@gmail.com>
Tay Ray Chuan <rctay89@gmail.com> writes:
> I tested this on top of the first 4 patches in 'rc/http-push' in 'pu',
> applied on 'maint'. [1] I wonder if this should instead be queued for
> 'pu' [2], since the issue only occurs there, although, conceivably, it
> *could* happen without those patches in 'pu'.
That "*could*" leaves me puzzled even more.  Could you clarify?

Do you mean "earlier, nobody tried to fclose(slot->local) twice, but with the four patches there are codepaths that do so, which triggered the bug reported by Clemens"? But then the issue couldn't have happened without those patches.

If existing code always knew when slot->local is and is not valid without having to rely on its NULLness, and that was the reason why nobody tried to fclose(slot->local) twice, then I'd agree that it is a bug waiting to happen, and stuffing NULL to slot->local after the code fclose() it would be a good clean-up.

Previous: Tay Ray ChuanNext: Tay Ray Chuan
Message 12 of 15 in “http-push: send out fetch requests on queue”
  1. 3/3 http-push: send out fetch requests on queueTay Ray Chuan, Apr 24, 2009
  2. Clemens BuchacherMay 30, 2009
  3. Tay Ray ChuanMay 30, 2009
  4. Clemens BuchacherMay 30, 2009
  5. Tay Ray ChuanMay 30, 2009
  6. Tay Ray ChuanMay 30, 2009
  7. http*: cleanup slot->local after fcloseTay Ray Chuan, May 30, 2009
  8. Clemens BuchacherMay 30, 2009
  9. Junio C HamanoMay 31, 2009
  10. Tay Ray ChuanMay 31, 2009
  11. http*: cleanup slot->local after fcloseTay Ray Chuan, May 31, 2009
  12. Junio C HamanoMay 31, 2009
  13. Tay Ray ChuanJun 1, 2009
  14. Junio C HamanoJun 1, 2009
  15. Tay Ray ChuanJun 2, 2009

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.