Re: [PATCH v6 0/6] [GSoC] bitmap: integrate a lookup table extension to the bitmap format

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Aug 22, 2022, 14:42 UTC
Message-ID: <80852679-rsso-rp45-q328-99n36q0639sq@tzk.qr>
In-Reply-To: <xmqqlerkj5f9.fsf@gitster.g>

Hi Junio,

On Fri, 19 Aug 2022, Junio C Hamano wrote:

Show 20 quoted lines

> "Abhradeep Chakraborty via GitGitGadget" <gitgitgadget@gmail.com>
> writes:
>
> > When parsing the .bitmap file, git loads all the bitmaps one by one even if
> > some of the bitmaps are not necessary. We can remove this overhead by
> > loading only the necessary bitmaps. A look up table extension can solve this
> > issue.
> >
> > Changes since v5:
> >
> > As the failure in the test case is not due to this code, I think it makes no
> > sense to delay the patch further.
> >
> >  * The performance test changes were not accurate as the second
> >    test_bitmap_cases call using the repo built for the previous call. This
> >    version fixes that.
> >  * Taylor suggested some minor changes. Those are addressed in this version.
>
> The discussion on v5 was quite active, but we haven't seen any
> traffic on this round.  Is everybody happy with what we see here?

The part of the lively discussion in which I participated exclusively focused on the failed CI runs and trying to get to the bottom of this bug.

Taylor contributed <cover.1660944574.git.me@ttaylorr.com> to address the bug. While he seems grateful for my help, I am honestly puzzled because I lack too much knowledge about the code to have been of assistance in any meaningful way.

My participation in this thread should not be mistaken for a review: I am woefully unfamiliar with the bitmap design (let alone code) and would therefore not _dare_ to offer anything that I would claim is a code review.

Ciao, Dscho