On Sun, Oct 01, 2023 at 09:39:09PM -0500, Eric W. Biederman wrote:

This addresses all of the known test failures from v1 of this set of
changes.  In particular I have reworked commit_tree_extended which
was flagged by smatch, -Werror=array-bounds, and the leak detector.

One functional bug was fixed in repo_for_each_abbrev where it was
mistakenly displaying too many ambiguous oids.

I am posting this so that people review and testing of this patchset
won't be distracted by the known and fixed issues.

Thanks! I've reviewed this patch series up to patch 7.

I think the most important question mark I currently have is scalability
of the proposed object mapping format. The complexity to load the object
mappings is currently O(nlogn) and requires reading a file that is
easily hundreds of megabytes or even gigabytes in size.

I have a feeling that this needs to be addressed before such an object
mapping would be feasible for production use, or otherwise it would
incur too high a cost to be useful. I'm afraid that this will make the
whole series more complex to implement -- I'm sorry about that.

I've added comments and ideas regarding this issue on patch 6 and 7. It
could totally be that I'm missing the obvious though, or that my ideas
suck. Please don't hesitate to point that out if that is the case.

Patrick