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

Re: [PATCH v2] commit: check result of resolve_ref_unsafe

From
Jeff King <peff@peff.net>
Date
Oct 19, 2017, 17:44 UTC
Message-ID
<20171019174452.hd3c47ocducddvgr@sigill.intra.peff.net>
In-Reply-To
<5fa1f5c6-249e-2aa9-5e9f-c00ebe2c0d9d@samsung.com>
On Thu, Oct 19, 2017 at 12:36:50PM +0300, Andrey Okoshkin wrote:
Show 14 quoted lines
> Add check of the resolved HEAD reference while printing of a commit summary.
> resolve_ref_unsafe() may return NULL pointer if underlying calls of lstat()
> or
> open() fail in files_read_raw_ref().
> Such situation can be caused by race: file becomes inaccessible to this
> moment.
> 
> Signed-off-by: Andrey Okoshkin <a.okoshkin@samsung.com>
> ---
> Thank you for your review.
> 
> Changes since the previous patch:
> * BUG is replaced with die, message;
> * Message is changed.
Thanks, this looks good to me. One other possible minor improvement:
>  	head = resolve_ref_unsafe("HEAD", 0, junk_oid.hash, NULL);
> +	if (!head)
> +		die(_("unable to resolve HEAD after creating commit"));

Should we use die_errno() here to report the value of errno? I think resolve_ref_unsafe() should set it consistently (even an internal problem, like an illegally-formatted refname, yields EINVAL).

I grepped the code base looking for other instances of the same problem, and found four of them. Patches to follow.

Unlike this one, I ended up quietly returning an error in most cases. The individual commit messages discuss the reasoning for each case, but I do wonder if we ought to simply die() in each case out of an abundance of caution (either the repo has a broken symref, or some weird filesystem error occurred, but either way it may be best not to continue). I dunno.

These are all independent, so can be applied in any order or combination with respect to each other and to your patch.

  [1/4]: test-ref-store: avoid passing NULL to printf
  [2/4]: remote: handle broken symrefs
  [3/4]: log: handle broken HEAD in decoration check
  [4/4]: worktree: handle broken symrefs in find_shared_symref()
 builtin/remote.c          | 2 +-
 log-tree.c                | 2 +-
 t/helper/test-ref-store.c | 2 +-
 worktree.c                | 3 ++-
 4 files changed, 5 insertions(+), 4 deletions(-)
-Peff
Previous: Andrey OkoshkinNext: Jeff King
Message 7 of 20 in “commit: check result of resolve_ref_unsafe”
  1. commit: check result of resolve_ref_unsafeAndrey Okoshkin, Oct 18, 2017
  2. Jeff KingOct 18, 2017
  3. Junio C HamanoOct 19, 2017
  4. Jeff KingOct 19, 2017
  5. Andrey OkoshkinOct 19, 2017
  6. commit: check result of resolve_ref_unsafeAndrey Okoshkin, Oct 19, 2017
  7. Jeff KingOct 19, 2017
  8. 1/4 test-ref-store: avoid passing NULL to printfJeff King, Oct 19, 2017
  9. 2/4 remote: handle broken symrefsJeff King, Oct 19, 2017
  10. Jeff KingOct 19, 2017
  11. 3/4 log: handle broken HEAD in decoration checkJeff King, Oct 19, 2017
  12. 4/4 worktree: handle broken symrefs in find_shared_symref()Jeff King, Oct 19, 2017
  13. Eric SunshineOct 21, 2017
  14. Jeff KingOct 21, 2017
  15. Junio C HamanoOct 22, 2017
  16. Andrey OkoshkinOct 20, 2017
  17. commit: check result of resolve_ref_unsafeAndrey Okoshkin, Oct 20, 2017
  18. commit: check result of resolve_ref_unsafeAndrey Okoshkin, Oct 20, 2017
  19. Jeff KingOct 21, 2017
  20. Junio C HamanoOct 22, 2017

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.