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

Re: [PATCH v3 2/3] git-push: clean up some of the output from git push --porcelain

From
Jeff King <peff@peff.net>
Date
Feb 12, 2010, 00:03 UTC
Message-ID
<20100212000302.GB21930@coredump.intra.peff.net>
In-Reply-To
<7vwryjpdb6.fsf@alter.siamese.dyndns.org>
On Thu, Feb 11, 2010 at 03:23:09PM -0800, Junio C Hamano wrote:
Show 12 quoted lines
> > I would say yes. The advice output should all be on stderr, and the
> > porcelain output should all be on stdout. So there is no parsing
> > conflict. And stderr either goes to /dev/null (because the porcelain is
> > not terminal-based, or doesn't want the terminal screwed up), in which
> > case the advice does nothing, or stderr goes to the terminal (because
> > the porcelain is some simple script), in which case the message is
> > probably something the user would want to see.
> 
> Hmm, if we make it a rule to send all the advice messages to the standard
> error stream, wouldn't the logical conclusion be that the hunk you are
> commenting is unnecessary?  "push --porcelain" should give advice just as
> usual (but we would simply send the advice to the standard error stream).

Yes and no. The advice it is giving is specifically related to things you can't see, so it is not really as helpful as it could be. And in that sense, squelching may be positive.

But it seems that the thread has come to the conclusion that we're still going to have messages on stderr. There is not much point in suppressing one particular message if there is a bunch of other cruft. Either the user should see stuff on stderr, or they shouldn't, and the caller can decide by redirecting stderr. So yes, squelching that advice probably should just be dropped.

-Peff
Previous: Junio C HamanoNext: Larry D'Anna
Message 36 of 45 in “git-push: fix an error message so it goes to stderr”
  1. 1/3 git-push: fix an error message so it goes to stderrLarry D'Anna, Feb 8, 2010
  2. 2/3 git-push: clean up some of the output from git push --porcelainLarry D'Anna, Feb 8, 2010
  3. Junio C HamanoFeb 8, 2010
  4. Junio C HamanoFeb 8, 2010
  5. Jeff KingFeb 8, 2010
  6. Larry D'AnnaFeb 8, 2010
  7. Junio C HamanoFeb 8, 2010
  8. Larry D'AnnaFeb 8, 2010
  9. git-push: clean up some of the output from git pushLarry D'Anna, Feb 8, 2010
  10. Junio C HamanoFeb 8, 2010
  11. Larry D'AnnaFeb 8, 2010
  12. Junio C HamanoFeb 8, 2010
  13. git-push: fix the documentation to explain all the status flagsLarry D'Anna, Feb 8, 2010
  14. Junio C HamanoFeb 9, 2010
  15. Junio C HamanoFeb 9, 2010
  16. Junio C HamanoFeb 9, 2010
  17. Larry D'AnnaFeb 9, 2010
  18. Junio C HamanoFeb 9, 2010
  19. Larry D'AnnaFeb 9, 2010
  20. Larry D'AnnaFeb 9, 2010
  21. Junio C HamanoFeb 9, 2010
  22. Larry D'AnnaFeb 9, 2010
  23. Shawn O. PearceFeb 9, 2010
  24. Nicolas PitreFeb 9, 2010
  25. Shawn O. PearceFeb 9, 2010
  26. Junio C HamanoFeb 9, 2010
  27. Junio C HamanoFeb 9, 2010
  28. Shawn O. PearceFeb 10, 2010
  29. Larry D'AnnaFeb 9, 2010
  30. Jeff KingFeb 10, 2010
  31. Larry D'AnnaFeb 10, 2010
  32. Tay Ray ChuanFeb 10, 2010
  33. Junio C HamanoFeb 8, 2010
  34. Jeff KingFeb 10, 2010
  35. Junio C HamanoFeb 11, 2010
  36. Jeff KingFeb 12, 2010
  37. 3/3 git-push: make git push --dry-run --porcelain exit with status 0 even if updates will be rejectedLarry D'Anna, Feb 8, 2010
  38. Junio C HamanoFeb 8, 2010
  39. Larry D'AnnaFeb 8, 2010
  40. Junio C HamanoFeb 9, 2010
  41. Larry D'AnnaFeb 10, 2010
  42. 4/4 git-push: make git push --dry-run --porcelain exit with status 0 even if updates will be rejectedLarry D'Anna, Feb 10, 2010
  43. Larry D'AnnaFeb 15, 2010
  44. Junio C HamanoFeb 15, 2010
  45. Junio C HamanoFeb 8, 2010

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.