On Mon, Apr 30, 2018 at 08:03:12PM +0200, Duy Nguyen wrote: > On Tue, Apr 24, 2018 at 1:39 AM, brian m. carlson > wrote: > > [0] I can synthesize blobs, trees, and commits, but things are currently > > totally broken, which is, I suppose, to be expected. > > Yup. I was tired and bored so I went playing with the new hash. > Writing and reading blobs (with hash-object/cat-file) were relatively > easy after fixing up fill_sha1_path and get_oid_basic). Then I worked > my way up to update-index/ls-files so that I could make trees with > write-tree. And I hit the first road block: struct ondisk_cache_entry > hard codes hash size so I would need to re-organize the code for more > flexibility (or even redesign the file format if I want to keep byte > alignment). Eck... > > I guess I'll be helping review this series instead :D Yeah, I have code to fix that, but it's ugly. You can see the work on part2 and part3 of the test fixes, plus the fixes for all of that stuff on my object-id-part14 branch. -- brian m. carlson: Houston, Texas, US OpenPGP: https://keybase.io/bk2204