Re: [PATCH v5 3/6] pack-bitmap-write: learn pack.writeBitmapLookupTable and add tests

[Abhradeep Chakraborty <chakrabortyabhradeep79@gmail.com>]

On Fri, Jul 29, 2022 at 12:52 AM Johannes Schindelin
<Johannes.Schindelin@gmx.de> wrote:
> That's quite a large a change, and unfortunately I pinpointed a flake to
> this patch when running with GIT_TEST_DEFAULT_HASH=sha256. The symptom is
> this:

Hi Dscho, sorry for this long delay in response. I was quite busy for
3-4 days in hostel room shifting. So, I couldn't work properly during
this time.

> -- snip --
> [...]
> + diff -u expect.normalized actual.normalized
> + rm -f expect.normalized actual.normalized
> ok 317 - enumerate --objects (full bitmap, other)
>
> expecting success of 5326.318 'bitmap --objects handles non-commit objects (full bitmap, other)':
>                 git rev-list --objects --use-bitmap-index $branch tagged-blob >actual &&
>                 grep $blob actual
>
> + git rev-list --objects --use-bitmap-index other tagged-blob
> + grep bff4ed5e839bd73e821f78b45a7fa34208aa85596535ec8e9ac5eab477ca6f81 actual
> bff4ed5e839bd73e821f78b45a7fa34208aa85596535ec8e9ac5eab477ca6f81
> ok 318 - bitmap --objects handles non-commit objects (full bitmap, other)
>
> expecting success of 5326.319 'clone from bitmapped repository':
>                 rm -fr clone.git &&
>                 git clone --no-local --bare . clone.git &&
>                 git rev-parse HEAD >expect &&
>                 git --git-dir=clone.git rev-parse HEAD >actual &&
>                 test_cmp expect actual
>
> + rm -fr clone.git
> + git clone --no-local --bare . clone.git
> Cloning into bare repository 'clone.git'...
> remote: Enumerating objects: 756, done.
> remote: Counting objects: 100% (754/754), done.
> remote: Compressing objects: 100% (281/281), done.
> remote: Total 756 (delta 245), reused 740 (delta 234), pack-reused 2
> Receiving objects: 100% (756/756), 77.50 KiB | 8.61 MiB/s, done.
> fatal: REF_DELTA at offset 221 already resolved (duplicate base 4d332072f161629ffe4652ecd3ce377ef88447bec73f05ab0f3515f98bd061cf?)
> fatal: fetch-pack: invalid index-pack output
> error: last command exited with $?=128
> not ok 319 - clone from bitmapped repository
> #
> #                       rm -fr clone.git &&
> #                       git clone --no-local --bare . clone.git &&
> #                       git rev-parse HEAD >expect &&
> #                       git --git-dir=clone.git rev-parse HEAD >actual &&
> #                       test_cmp expect actual
> #
> 1..319
> -- snap --
>
> On a hunch, I ran this through valgrind (took a while) but it did not
> point out the problem.
>
> Again, this is only with SHA-256 (and somewhat flaky), it passes every
> time with SHA-1. Maybe you can reproduce on your side with that
> information?

Yeah, I can reproduce it on my side. But I am sure it is not related
to the lookup table implementation code. Because when I swap the order
of calling  `test_midx_bitmap_cases "pack.writeBitmapLookupTable"` and
`test_midx_bitmap_cases` (in t5326-multi-pack-bitmaps.sh), in that
case, the error is being generated in  `test_midx_bitmap_cases` call.
Generally speaking, the error is always being generated in the second
call.

For now, my understanding says that there is something fishy in the
test script. I am still not able to figure out the problem here. But
let me further investigate.

If anyone has some idea about what could be the culprit, I will be
very happy to know.

Thanks :)