Re: [PATCH v2] add test to demonstrate that shallow recursive clones fail

[Stefan Beller <sbeller@google.com>]

On Wed, Nov 18, 2015 at 1:11 PM, Jeff King <peff@peff.net> wrote:
> On Wed, Nov 18, 2015 at 01:32:36PM +0100, Duy Nguyen wrote:
>
>> Yeah I think that was the 10k commits in Shawn's mail: the number of
>> commits we may have to walk until we hit a reachability bitmap. It
>> looks like C Git will create a bitmap every 5k commits, not 10k,
>> though, if I read the code correctly. The point is reachability test
>> with the presence of pack bitmaps is not O(n) anymore. Which is
>> probably good enough for now.
>
> There are some pathological cases, though. I hit one recently that still
> took 40s to do "rev-list --objects --all --use-bitmap-index" (it's 80s
> without bitmaps).  The problem is that it has over 20,000 refs. We try
> to put a bitmap at the tip of each ref, but it's tough when there are
> that many.

+Terry, who did optimize the JGit implementation for bitmaps,
as we also had a "lots of refs" hoarder repo, which underperformed
before.

>
> I suspect there's room for improvement in the commit selection in such
> cases. That code hasn't really been tweaked since it was originally
> written, and repositories like that are extreme outliers.
>
> -Peff