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

Re: [PATCH] MIDX: revert the default version to v1

From
Taylor Blau <me@ttaylorr.com>
Date
Apr 16, 2026, 21:12 UTC
Message-ID
<aeFQvu4iqJAQMjCy@nand.local>
In-Reply-To
<20260416200659.GB1887222@coredump.intra.peff.net>
On Thu, Apr 16, 2026 at 04:06:59PM -0400, Jeff King wrote:
> This looks fine to me, and you can add my S-o-b if you want. But let me
> propose a slight alternative that reduces the test churn and may make
> things easier going forward.

Same here, I agree with the discussion earlier in the thread about what the right short- and medium-term solutions are, and I think that including Junio's patch is a good approach.

If we take that, please also feel free to add my Acked-by, or Reviewed-by, or similar.

Show 13 quoted lines
> -- >8 --
> Subject: [PATCH] MIDX: revert the default version to v1
>
> We introduced midx version 2 in b2ec8e90c2 (midx: do not require packs
> to be sorted in lexicographic order, 2026-02-24) and now write it by
> default. The rationale was that older versions should ignore the v2 midx
> and fall back to using the packs (just like we do for other midx
> errors). Unfortunately this is not the case, as we have a hard die()
> when we see an unknown midx version.
>
> As a result, writing a midx with Git 2.54-rc2 puts the repository into a
> state that is unusable with Git 2.53. And this midx write may happen
> behind the scenes as part of normal operations, like fetch.

I'm not sure if it's worth mentioning, but I think that it's reasonable to say "Git 2.53 and earlier", with the implied lower bound being which version first introduced the MIDX. I think it's equally fine as-is, though.

Show 19 quoted lines
> Let's switch back to writing v1 by default to avoid regressing the case
> where multiple versions of Git are used on the same repository.
>
> There is one gotcha, though: the v2 format is required for some new
> features, like midx compaction, and running "git multi-pack-index
> compact" will complain when asked to write a v1 index. The user must set
> midx.version to "2" to make the feature work.
>
> So instead of always using v1, we'll base the default on whether the
> requested feature requires v2. That does mean that running midx
> compaction will create a repository that can't be read by older versions
> of Git. But we never do that by default; only people experimenting with
> the new feature will be affected.
>
> We have to adjust the test expectation in t5319, since it will now
> generate v1 files. And our "auto-select v2" is covered by the tests in
> t5335, which continue to check that compaction works without having to
> set midx.version manually (and also explicitly check that asking for v1
> with compaction reports the problem).

I think that the test fallout that Junio's patch necessitates isn't all that bad, and in some sense I think the "write version 1 usually, but version 2 if the feature requires it" is a little magical. That being said, anyone doing things that would require a v2 MIDX likely already understand what the trade-offs are, so in that sense I think that this is less magical and more "do the sensible thing by default".

I don't have strong feelings either way and would be fine with either. I think if anything I have a vague preference towards the approach taken here, but either would be fine with me.

Show 5 quoted lines
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> I have a feeling there are probably some gaps in v2 testing in t5319,
> since we are no longer using v2 for the bulk of the tests. IMHO that is
> OK to sort out post-release.

Perhaps, although you could make the opposite argument for v1 MIDXs when we previously switched the default to write v2 MIDXs. The format differs only in the version field, and the ordering constraints on the packs within the MIDX. That area and the compatibility issues here are the "interesting" parts to test IMHO.

>  Documentation/git-multi-pack-index.adoc | 3 +++
>  midx-write.c                            | 4 +++-
>  t/t5319-multi-pack-index.sh             | 2 +-
>  3 files changed, 7 insertions(+), 2 deletions(-)

The patch itself looks as expected to me. Thanks for working on it, and sorry again for the mess here.

Thanks, Taylor

Previous: Jeff KingNext: Jeff King
Message 27 of 30 in “[ANNOUNCE] Git v2.54.0-rc2”
  1. Junio C HamanoApr 14, 2026
  2. MIDX woes, was Re: [ANNOUNCE] Git v2.54.0-rc2Johannes Schindelin, Apr 15, 2026
  3. Junio C HamanoApr 15, 2026
  4. Jeff KingApr 16, 2026
  5. Jeff KingApr 16, 2026
  6. Derrick StoleeApr 16, 2026
  7. Junio C HamanoApr 16, 2026
  8. Taylor BlauApr 16, 2026
  9. Derrick StoleeApr 19, 2026
  10. Junio C HamanoApr 20, 2026
  11. Taylor BlauApr 16, 2026
  12. Jeff KingApr 16, 2026
  13. Junio C HamanoApr 16, 2026
  14. Junio C HamanoApr 16, 2026
  15. Jeff KingApr 16, 2026
  16. Junio C HamanoApr 16, 2026
  17. Jeff KingApr 16, 2026
  18. Junio C HamanoApr 17, 2026
  19. MIDX: revert the default version to v1Junio C Hamano, Apr 16, 2026
  20. Junio C HamanoApr 16, 2026
  21. Junio C HamanoApr 16, 2026
  22. Taylor BlauApr 16, 2026
  23. Jeff KingApr 16, 2026
  24. Junio C HamanoApr 16, 2026
  25. Jeff KingApr 16, 2026
  26. Jeff KingApr 16, 2026
  27. Taylor BlauApr 16, 2026
  28. Jeff KingApr 16, 2026
  29. Junio C HamanoApr 17, 2026
  30. Elijah NewrenApr 17, 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.