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

Re: [PATCH 2/2] fetch: clobber existing tags with --prune-tags

From
Orgad Shaneh <orgads@gmail.com>
Date
Mar 26, 2026, 07:57 UTC
Message-ID
<CAGHpTBJKaTURMJmJ6W4iiCAy3-M2YWb48rF8GMoEPCfpGcE=QA@mail.gmail.com>
In-Reply-To
<xmqq1phdavik.fsf@gitster.g>
On Sat, Mar 21, 2026 at 8:27 PM Junio C Hamano <gitster@pobox.com> wrote:
Show 8 quoted lines
>
> "Orgad Shaneh via GitGitGadget" <gitgitgadget@gmail.com> writes:
>
> > From: Orgad Shaneh <orgad.shaneh@audiocodes.com>
> >
> > This was documented but not implemented.
>
> And instead what did the command do / how did the command behave?

Even when passing --prune-tags or setting fetch.pruneTags, git fetch fails to force-update tags, resulting in the following error:

! [rejected]        testTag    -> testTag  (would clobber existing tag)

Furthermore, when a tag clobber is rejected, the entire fetch operation fails, preventing the update of other refs and even new tags.

Show 10 quoted lines
> It also is curious when it was "broken".  It could be that it was
> broken from day one when 97716d21 (fetch: add a --prune-tags option
> and fetch.pruneTags config, 2018-02-09) was written, c1a7902f (Merge
> branch 'ab/fetch-prune', 2018-03-06) merged it, and Git 2.17 was
> shipped with it, but knowing Ævar (by the way where is he these
> days???), I somehow doubt it.
>
> And unless it was broken from day one, we must find out if the
> change in behaviour was deliberate, in which case it would be the
> doucmentation and not the implementation that needs fixing.
I dug into the feature history and found:
1. The flag was added in 97716d21 (fetch: add a --prune-tags option
   and fetch.pruneTags config, 2018-02-09) as you mentioned.
2. Tag clobbering protection was added later in 0bc8d71b (fetch: stop
   clobbering existing tags without --force, 2018-08-31).

It appears the --prune-tags flag was overlooked in that later commit, either in its behavior or its documentation.

Since there is currently no way to suppress this error globally, if you consider the current behavior acceptable and prefer to fix the documentation, I suggest adding a fetch.force config to allow clobbering.

Show 20 quoted lines
> >                       git -C ../child1 fetch origin '+refs/tags/*:refs/tags/*' &&
> >                       git tag $tag_args testTag HEAD^ &&
> >                       test_must_fail git -C ../child1 fetch origin tag testTag &&
> > -                     git -C ../child1 fetch --force origin tag testTag
> > +                     git -C ../child1 fetch --force origin tag testTag &&
> > +                     git tag $tag_args testTag HEAD &&
> > +                     test_must_fail git -C ../child1 fetch origin tag testTag &&
> > +                     git -C ../child1 fetch --prune-tags origin tag testTag &&
> > +                     git tag $tag_args testTag HEAD^ &&
> > +                     test_must_fail git -C ../child1 fetch origin tag testTag &&
> > +                     git -C ../child1 -c fetch.prunetags=true fetch origin tag testTag
> >               )
> >       "
> >  }
>
> None of the steps we see in the added test do not seem to check that
> --prune-tags does clobber existing tag that no longer exists on the
> other side.  It only checks the "git fetch" command exits with
> status 0, but does not see if the tag actually went away after the
> operation is done.

In these tests, the tag is being replaced rather than deleted. Existing tests for the pruning mechanism itself are located in t/t5510-fetch.sh.

Would you like me to add a check for the tag content itself to verify the update? I suppose I should do the same for the existing test cases in that block as well.

- Orgad
Previous: Junio C HamanoNext: Orgad Shaneh
Message 6 of 9 in “fetch: clobber existing tags with --prune-tags”
  1. 0/2 fetch: clobber existing tags with --prune-tagsOrgad Shaneh via GitGitGadget, Feb 15, 2026
  2. 1/2 fetch: add a test for --force flagOrgad Shaneh via GitGitGadget, Feb 15, 2026
  3. 2/2 fetch: clobber existing tags with --prune-tagsOrgad Shaneh via GitGitGadget, Feb 15, 2026
  4. Orgad ShanehMar 18, 2026
  5. Junio C HamanoMar 21, 2026
  6. Orgad ShanehMar 26, 2026
  7. Orgad ShanehSep 13, 2026
  8. Junio C HamanoSep 13, 2026
  9. Orgad ShanehOct 6, 2026

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.