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

Re: [PATCH 05/16] midx: introduce `--checksum-only` for incremental MIDX writes

From
Taylor Blau <me@ttaylorr.com>
Date
Apr 2, 2026, 22:51 UTC
Message-ID
<ac7y7t+Qgm0htU3c@nand.local>
In-Reply-To
<20260330231547.GE41843@coredump.intra.peff.net>
On Mon, Mar 30, 2026 at 07:15:47PM -0400, Jeff King wrote:
Show 7 quoted lines
> Your goal makes sense, but that is not at all what I would have expected
> a "--checksum-only" option to do. From the name I expected it to just
> compute or verify checksums and not write anything to disk. I dunno.
> Maybe it should be "--no-write-chain-file" or something?
>
> This is perhaps on the border of bike-shedding, and maybe the fact that
> it is an option of "multi-pack-index write" would be enough to clarify.
Heh :-).

I think that it definitely qualifies as bike-shedding, but I'm glad you brought it up. I hadn't considered the name being potentially confusing, but now that you mention it I can't un-see it.

Calling this something like '--no-write-chain-file' feels like a good choice to me, and doesn't paint us into a corner if we want to later add an option that does something like you describe above.

(As an aside, we already have 'git multi-pack-index verify', which is probably the closest thing to what you're describing, so perhaps the concern is lessened by that fact. But even if we already have that functionality laying around, I think that "--no-write-chain-file" is a more descriptive name for the command-line flag than "--checksum-only".)

Show 18 quoted lines
> > @@ -239,6 +254,15 @@ static int cmd_multi_pack_index_compact(int argc, const char **argv,
> >  	if (argc != 2)
> >  		usage_with_options(builtin_multi_pack_index_compact_usage,
> >  				   options);
> > +
> > +	if (opts.flags & MIDX_WRITE_CHECKSUM_ONLY &&
> > +	    !(opts.flags & MIDX_WRITE_INCREMENTAL)) {
> > +		error(_("cannot use %s without %s"),
> > +		      "--checksum-only", "--incremental");
> > +		usage_with_options(builtin_multi_pack_index_compact_usage,
> > +				   options);
> > +	}
> > +
>
> OK. It's sad that we have to duplicate these bits between the write and
> compact operations, but I'm not sure how much sense it would make to
> factor out common options (both how much work it is, and how much
> duplication we could actually drop in practice).

I agree. I looked at unifying these a little further, but decided against it as it ended up making things more complex than it was worth IMHO.

Show 23 quoted lines
> >  	if (ctx.incremental) {
> > -		FILE *chainf = fdopen_lock_file(&lk, "w");
> >  		struct strbuf final_midx_name = STRBUF_INIT;
> >  		struct multi_pack_index *m = ctx.base_midx;
> >
> > -		if (!chainf) {
> > -			error_errno(_("unable to open multi-pack-index chain file"));
> > -			goto cleanup;
> > +		if (!(opts->flags & MIDX_WRITE_CHECKSUM_ONLY)) {
> > +			FILE *chainf = fdopen_lock_file(&lk, "w");
> > +			if (!chainf) {
> > +				error_errno(_("unable to open multi-pack-index chain file"));
> > +				goto cleanup;
> > +			}
> >  		}
>
> OK, and this is the actual write which is skipped. I guess this could
> also be conditional on is_lock_file_locked(&lk), assuming we initialize
> it with LOCK_INIT. In some ways that seems easier to analyze to me,
> since you don't have to realize that "lk" is only active when
> CHECKSUM_ONLY is not set. But I don't think there's another reason we'd
> choose not to take a lock file, so I think it may be six of one and half
> a dozen of the other.

Yeah, I think this is more or less equivalent. In some ways I like the explicitness of the conditional as it is currently written, but I think your suggestion reads clearer.

It ends up producing nice code like:
    if (is_lock_file_locked(&lk)) {
        FILE *chainf = fdopen_lock_file(&lk, "w");
        /* ... */
    }

which is pleasing to me in the sense that it reads like "if the lockfile is held, then do something with it".

I think there is a small risk of coupling the state of the lockfile with the MIDX_NO_WRITE_CHAIN bit in the flags, but it's easily mitigated by an ASSERT(opts->flags & MIDX_NO_WRITE_CHAIN) in the else branch of the conditional above.

There are a few such conditionals, and repeatedly checking that the flag bits are set is wasteful, so I check it only once in the big "if (ctx->incremental)" block, and then once again in the top-level write_midx_internal() function before we do the final commit and cleanup.

Thanks, Taylor

Previous: Jeff KingNext: Taylor Blau
Message 19 of 113 in “repack: incremental MIDX/bitmap-based repacking”
  1. 00/16 repack: incremental MIDX/bitmap-based repackingTaylor Blau, Mar 29, 2026
  2. 01/16 midx-write: handle noop writes when converting incremental chainsTaylor Blau, Mar 29, 2026
  3. Jeff KingMar 30, 2026
  4. Taylor BlauMar 31, 2026
  5. 02/16 midx: use `string_list` for retained MIDX filesTaylor Blau, Mar 29, 2026
  6. Jeff KingMar 30, 2026
  7. Taylor BlauMar 31, 2026
  8. 03/16 strvec: introduce `strvec_init_alloc()`Taylor Blau, Mar 29, 2026
  9. Jeff KingMar 30, 2026
  10. 04/16 midx: use `strvec` for `keep_hashes`Taylor Blau, Mar 29, 2026
  11. Jeff KingMar 30, 2026
  12. Taylor BlauMar 31, 2026
  13. Taylor BlauMar 31, 2026
  14. Jeff KingMar 31, 2026
  15. Taylor BlauApr 1, 2026
  16. Jeff KingApr 1, 2026
  17. 05/16 midx: introduce `--checksum-only` for incremental MIDX writesTaylor Blau, Mar 29, 2026
  18. Jeff KingMar 30, 2026
  19. Taylor BlauApr 2, 2026
  20. 06/16 midx: support custom `--base` for incremental MIDX writesTaylor Blau, Mar 29, 2026
  21. Jeff KingApr 7, 2026
  22. Taylor BlauApr 14, 2026
  23. 07/16 repack: track the ODB source via existing_packsTaylor Blau, Mar 29, 2026
  24. Jeff KingApr 7, 2026
  25. Taylor BlauApr 14, 2026
  26. 08/16 midx: expose `midx_layer_contains_pack()`Taylor Blau, Mar 29, 2026
  27. Jeff KingApr 7, 2026
  28. 09/16 repack-midx: factor out `repack_prepare_midx_command()`Taylor Blau, Mar 29, 2026
  29. 10/16 repack-midx: extract `repack_fill_midx_stdin_packs()`Taylor Blau, Mar 29, 2026
  30. Jeff KingApr 7, 2026
  31. 11/16 repack-geometry: prepare for incremental MIDX repackingTaylor Blau, Mar 29, 2026
  32. Jeff KingApr 7, 2026
  33. Elijah NewrenApr 16, 2026
  34. Taylor BlauApr 21, 2026
  35. 12/16 builtin/repack.c: convert `--write-midx` to an `OPT_CALLBACK`Taylor Blau, Mar 29, 2026
  36. Jeff KingApr 7, 2026
  37. 13/16 packfile: ensure `close_pack_revindex()` frees in-memory revindexTaylor Blau, Mar 29, 2026
  38. Jeff KingApr 7, 2026
  39. 14/16 repack: implement incremental MIDX repackingTaylor Blau, Mar 29, 2026
  40. Elijah NewrenApr 16, 2026
  41. Taylor BlauApr 21, 2026
  42. 15/16 repack: introduce `--write-midx=incremental`Taylor Blau, Mar 29, 2026
  43. Elijah NewrenApr 16, 2026
  44. Taylor BlauApr 21, 2026
  45. 16/16 repack: allow `--write-midx=incremental` without `--geometric`Taylor Blau, Mar 29, 2026
  46. Taylor BlauApr 14, 2026
  47. 00/16 repack: incremental MIDX/bitmap-based repackingTaylor Blau, Apr 21, 2026
  48. 01/16 midx-write: handle noop writes when converting incremental chainsTaylor Blau, Apr 21, 2026
  49. 02/16 midx: use `strset` for retained MIDX filesTaylor Blau, Apr 21, 2026
  50. 03/16 midx: build `keep_hashes` array in orderTaylor Blau, Apr 21, 2026
  51. 04/16 midx: use `strvec` for `keep_hashes`Taylor Blau, Apr 21, 2026
  52. 05/16 midx: introduce `--no-write-chain-file` for incremental MIDX writesTaylor Blau, Apr 21, 2026
  53. 06/16 midx: support custom `--base` for incremental MIDX writesTaylor Blau, Apr 21, 2026
  54. 07/16 repack: track the ODB source via existing_packsTaylor Blau, Apr 21, 2026
  55. 08/16 midx: expose `midx_layer_contains_pack()`Taylor Blau, Apr 21, 2026
  56. 09/16 repack-midx: factor out `repack_prepare_midx_command()`Taylor Blau, Apr 21, 2026
  57. 10/16 repack-midx: extract `repack_fill_midx_stdin_packs()`Taylor Blau, Apr 21, 2026
  58. Jeff KingApr 29, 2026
  59. Taylor BlauApr 29, 2026
  60. 11/16 repack-geometry: prepare for incremental MIDX repackingTaylor Blau, Apr 21, 2026
  61. 12/16 builtin/repack.c: convert `--write-midx` to an `OPT_CALLBACK`Taylor Blau, Apr 21, 2026
  62. 13/16 packfile: ensure `close_pack_revindex()` frees in-memory revindexTaylor Blau, Apr 21, 2026
  63. 14/16 repack: implement incremental MIDX repackingTaylor Blau, Apr 21, 2026
  64. Jeff KingApr 29, 2026
  65. Taylor BlauApr 29, 2026
  66. Jeff KingApr 29, 2026
  67. Taylor BlauApr 29, 2026
  68. 15/16 repack: introduce `--write-midx=incremental`Taylor Blau, Apr 21, 2026
  69. Taylor BlauApr 21, 2026
  70. 16/16 repack: allow `--write-midx=incremental` without `--geometric`Taylor Blau, Apr 21, 2026
  71. Elijah NewrenApr 22, 2026
  72. Jeff KingApr 29, 2026
  73. 00/16 repack: incremental MIDX/bitmap-based repackingTaylor Blau, Apr 30, 2026
  74. 01/16 midx-write: handle noop writes when converting incremental chainsTaylor Blau, Apr 30, 2026
  75. 02/16 midx: use `strset` for retained MIDX filesTaylor Blau, Apr 30, 2026
  76. 03/16 midx: build `keep_hashes` array in orderTaylor Blau, Apr 30, 2026
  77. 04/16 midx: use `strvec` for `keep_hashes`Taylor Blau, Apr 30, 2026
  78. 05/16 midx: introduce `--no-write-chain-file` for incremental MIDX writesTaylor Blau, Apr 30, 2026
  79. 06/16 midx: support custom `--base` for incremental MIDX writesTaylor Blau, Apr 30, 2026
  80. 07/16 repack: track the ODB source via existing_packsTaylor Blau, Apr 30, 2026
  81. 08/16 midx: expose `midx_layer_contains_pack()`Taylor Blau, Apr 30, 2026
  82. 09/16 repack-midx: factor out `repack_prepare_midx_command()`Taylor Blau, Apr 30, 2026
  83. SZEDER GáborMay 13, 2026
  84. 10/16 repack-midx: extract `repack_fill_midx_stdin_packs()`Taylor Blau, Apr 30, 2026
  85. 11/16 repack-geometry: prepare for incremental MIDX repackingTaylor Blau, Apr 30, 2026
  86. 12/16 builtin/repack.c: convert `--write-midx` to an `OPT_CALLBACK`Taylor Blau, Apr 30, 2026
  87. 13/16 packfile: ensure `close_pack_revindex()` frees in-memory revindexTaylor Blau, Apr 30, 2026
  88. 14/16 repack: implement incremental MIDX repackingTaylor Blau, Apr 30, 2026
  89. 15/16 repack: introduce `--write-midx=incremental`Taylor Blau, Apr 30, 2026
  90. Jeff KingMay 13, 2026
  91. 16/16 repack: allow `--write-midx=incremental` without `--geometric`Taylor Blau, Apr 30, 2026
  92. Jeff KingMay 1, 2026
  93. 00/16 repack: incremental MIDX/bitmap-based repackingTaylor Blau, May 19, 2026
  94. 01/16 midx-write: handle noop writes when converting incremental chainsTaylor Blau, May 19, 2026
  95. 02/16 midx: use `strset` for retained MIDX filesTaylor Blau, May 19, 2026
  96. 03/16 midx: build `keep_hashes` array in orderTaylor Blau, May 19, 2026
  97. 04/16 midx: use `strvec` for `keep_hashes`Taylor Blau, May 19, 2026
  98. 05/16 midx: introduce `--no-write-chain-file` for incremental MIDX writesTaylor Blau, May 19, 2026
  99. 06/16 midx: support custom `--base` for incremental MIDX writesTaylor Blau, May 19, 2026
  100. SZEDER GáborJun 12, 2026
  101. Junio C HamanoJun 12, 2026
  102. Taylor BlauJun 12, 2026
  103. Taylor BlauJun 12, 2026
  104. 07/16 repack: track the ODB source via existing_packsTaylor Blau, May 19, 2026
  105. 08/16 midx: expose `midx_layer_contains_pack()`Taylor Blau, May 19, 2026
  106. 09/16 repack-midx: factor out `repack_prepare_midx_command()`Taylor Blau, May 19, 2026
  107. 10/16 repack-midx: extract `repack_fill_midx_stdin_packs()`Taylor Blau, May 19, 2026
  108. 11/16 repack-geometry: prepare for incremental MIDX repackingTaylor Blau, May 19, 2026
  109. 12/16 builtin/repack.c: convert `--write-midx` to an `OPT_CALLBACK`Taylor Blau, May 19, 2026
  110. 13/16 packfile: ensure `close_pack_revindex()` frees in-memory revindexTaylor Blau, May 19, 2026
  111. 14/16 repack: implement incremental MIDX repackingTaylor Blau, May 19, 2026
  112. 15/16 repack: introduce `--write-midx=incremental`Taylor Blau, May 19, 2026
  113. 16/16 repack: allow `--write-midx=incremental` without `--geometric`Taylor Blau, May 19, 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.