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

Re: [PATCH v3 08/19] write_or_die: raise SIGPIPE when we get EPIPE

From
Jonathan Nieder <jrnieder@gmail.com>
Date
Feb 20, 2013, 21:51 UTC
Message-ID
<20130220215043.GA24236@google.com>
In-Reply-To
<20130220200136.GH25647@sigill.intra.peff.net>
Jeff King wrote:
Show 19 quoted lines
> The write_or_die function will always die on an error,
> including EPIPE. However, it currently treats EPIPE
> specially by suppressing any error message, and by exiting
> with exit code 0.
>
> Suppressing the error message makes some sense; a pipe death
> may just be a sign that the other side is not interested in
> what we have to say. However, exiting with a successful
> error code is not a good idea, as write_or_die is frequently
> used in cases where we want to be careful about having
> written all of the output, and we may need to signal to our
> caller that we have done so (e.g., you would not want a push
> whose other end has hung up to report success).
>
> This distinction doesn't typically matter in git, because we
> do not ignore SIGPIPE in the first place. Which means that
> we will not get EPIPE, but instead will just die when we get
> a SIGPIPE. But it's possible for a default handler to be set
> by a parent process,

Not so much "default" as "insane inherited", as in the example of old versions of Python's subprocess.Popen.

I suspect this used exit(0) instead of raise(SIGPIPE) in the first place to work around a bash bug (too much verbosity about SIGPIPE). If any programs still have that kind of bug, I'd rather put pressure on them to fix it by *not* working around it. So the basic idea here looks good to me.

[...]
Show 12 quoted lines
> --- a/write_or_die.c
> +++ b/write_or_die.c
> @@ -1,5 +1,15 @@
>  #include "cache.h"
>  
> +static void check_pipe(int err)
> +{
> +	if (err == EPIPE) {
> +		signal(SIGPIPE, SIG_DFL);
> +		raise(SIGPIPE);
> +		/* Should never happen, but just in case... */
> +		exit(141);
How about
		die("BUG: another thread changed SIGPIPE handling behind my back!");
to make it easier to find and fix such problems?

Thanks, Jonathan

Previous: Jeff KingNext: Jeff King
Message 10 of 40 in “[PATCHv3 0/19] pkt-line cleanups and fixes”
  1. Jeff KingFeb 20, 2013
  2. 01/19 upload-pack: use get_sha1_hex to parse "shallow" linesJeff King, Feb 20, 2013
  3. 02/19 upload-pack: do not add duplicate objects to shallow listJeff King, Feb 20, 2013
  4. 03/19 upload-pack: remove packet debugging harnessJeff King, Feb 20, 2013
  5. 04/19 fetch-pack: fix out-of-bounds buffer offset in get_ackJeff King, Feb 20, 2013
  6. 05/19 send-pack: prefer prefixcmp over memcmp in receive_statusJeff King, Feb 20, 2013
  7. 06/19 upload-archive: do not copy repo nameJeff King, Feb 20, 2013
  8. 07/19 upload-archive: use argv_array to store client argumentsJeff King, Feb 20, 2013
  9. 08/19 write_or_die: raise SIGPIPE when we get EPIPEJeff King, Feb 20, 2013
  10. Jonathan NiederFeb 20, 2013
  11. Jeff KingFeb 20, 2013
  12. Jonathan NiederFeb 20, 2013
  13. Jeff KingFeb 20, 2013
  14. Jonathan NiederFeb 20, 2013
  15. Jeff KingFeb 20, 2013
  16. Junio C HamanoFeb 20, 2013
  17. [BUG] MSVC: error box when interrupting `gitlog` by quitting lessMarat Radchenko, Mar 28, 2014
  18. Marat RadchenkoMar 28, 2014
  19. Jeff KingMar 28, 2014
  20. Marat RadchenkoMar 28, 2014
  21. Jeff KingMar 28, 2014
  22. Johannes SixtMar 28, 2014
  23. MSVC: link in invalidcontinue.obj for better POSIX compatibilityMarat Radchenko, Mar 28, 2014
  24. Junio C HamanoMar 28, 2014
  25. Marat RadchenkoMar 28, 2014
  26. Junio C HamanoMar 28, 2014
  27. MSVC: link in invalidcontinue.obj for better POSIX compatibilityMarat Radchenko, Mar 28, 2014
  28. Junio C HamanoMar 28, 2014
  29. 09/19 pkt-line: move a misplaced commentJeff King, Feb 20, 2013
  30. 10/19 pkt-line: drop safe_write functionJeff King, Feb 20, 2013
  31. 11/19 pkt-line: provide a generic reading function with optionsJeff King, Feb 20, 2013
  32. 12/19 pkt-line: teach packet_read_line to chomp newlinesJeff King, Feb 20, 2013
  33. 13/19 pkt-line: move LARGE_PACKET_MAX definition from sidebandJeff King, Feb 20, 2013
  34. 14/19 pkt-line: provide a LARGE_PACKET_MAX static bufferJeff King, Feb 20, 2013
  35. 15/19 pkt-line: share buffer/descriptor reading implementationJeff King, Feb 20, 2013
  36. Eric SunshineFeb 22, 2013
  37. 16/19 teach get_remote_heads to read from a memory bufferJeff King, Feb 20, 2013
  38. 17/19 remote-curl: pass buffer straight to get_remote_headsJeff King, Feb 20, 2013
  39. 18/19 remote-curl: move ref-parsing code up in fileJeff King, Feb 20, 2013
  40. 19/19 remote-curl: always parse incoming refsJeff King, Feb 20, 2013

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.