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

Re: [PATCH 02/12] wrapper.c: add a new function unlink_or_msg

From
Junio C Hamano <gitster@pobox.com>
Date
Jul 22, 2014, 17:56 UTC
Message-ID
<xmqq4my9gtvj.fsf@gitster.dls.corp.google.com>
In-Reply-To
<CAL=YDWmHcy+Kf+gJLHyFK7bVjnD+bk7rX22jqHVFmTXoHDCEhQ@mail.gmail.com>
Ronnie Sahlberg <sahlberg@google.com> writes:
Show 21 quoted lines
> One reason for the former could be if there are problems with multiple
> refs in a single transaction.
> It would be very annoying to have to do
> $ git <some command>
>    error: ref foo has a problem
>
> $ <run command to fix the problem>
> $ git <some sommand>     (try again)
>    error: ref bar has a problem
> ...
>
> And it might be more userfriendly if that instead would be
> $ git <some command>
>    error: ref foo has a problem
>    error: ref bar has a problem
>    ...
>
> And get all the failures in one go instead of having to iterate.
> ...
> I personally do not know yet which approach is the best but would like
> to keep the door open for the "try all and fail at the end".

Yes, and often it is useful (e.g. we allow to push multiple and then show the result for individual refs). But that still does not show a need for accumulating the error messages to strbuf, does it?

Previous: Ronnie SahlbergNext: Ronnie Sahlberg
Message 8 of 26 in “Use ref transactions part 3”
  1. 00/12 Use ref transactions part 3Ronnie Sahlberg, Jul 16, 2014
  2. 01/12 wrapper.c: simplify warn_if_unremovableRonnie Sahlberg, Jul 16, 2014
  3. Junio C HamanoJul 18, 2014
  4. 02/12 wrapper.c: add a new function unlink_or_msgRonnie Sahlberg, Jul 16, 2014
  5. Junio C HamanoJul 18, 2014
  6. Junio C HamanoJul 18, 2014
  7. Ronnie SahlbergJul 22, 2014
  8. Junio C HamanoJul 22, 2014
  9. 03/12 refs.c: add an err argument to delete_ref_looseRonnie Sahlberg, Jul 16, 2014
  10. 04/12 refs.c: pass the ref log message to _create/delete/update instead of _commitRonnie Sahlberg, Jul 16, 2014
  11. 05/12 refs.c: pass NULL as *flags to read_ref_fullRonnie Sahlberg, Jul 16, 2014
  12. Junio C HamanoJul 18, 2014
  13. Ronnie SahlbergJul 22, 2014
  14. Ronnie SahlbergJul 22, 2014
  15. 06/12 refs.c: move the check for valid refname to lock_ref_sha1_basicRonnie Sahlberg, Jul 16, 2014
  16. Junio C HamanoJul 18, 2014
  17. 07/12 refs.c: call lock_ref_sha1_basic directly from commitRonnie Sahlberg, Jul 16, 2014
  18. 08/12 refs.c: pass a skip list to name_conflict_fnRonnie Sahlberg, Jul 16, 2014
  19. 09/12 refs.c: propagate any errno==ENOTDIR from _commit back to the callersRonnie Sahlberg, Jul 16, 2014
  20. 10/12 fetch.c: change s_update_ref to use a ref transactionRonnie Sahlberg, Jul 16, 2014
  21. 11/12 refs.c: make write_ref_sha1 staticRonnie Sahlberg, Jul 16, 2014
  22. 12/12 refs.c: fix handling of badly named refsRonnie Sahlberg, Jul 16, 2014
  23. Junio C HamanoJul 22, 2014
  24. Ronnie SahlbergJul 22, 2014
  25. Ronnie SahlbergJul 22, 2014
  26. Junio C HamanoJul 22, 2014

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.