On Mon, Nov 11, 2024 at 02:42:11AM -0800, karthik nayak wrote: > Patrick Steinhardt <ps@xxxxxx> writes: > > > The "files" backend has implemented special logic when committing > > the first transactions in an otherwise empty ref store: instead of > > writing all refs as separate loose files, it instead knows to write them > > all into a "packed-refs" file directly. This is significantly more > > efficient than having to write each of the refs as separate "loose" ref. > > > > The only user of this optimization is git-clone(1), which only uses this > > mechanism to write regular refs. Consequently, the implementation does > > not know how to handle both symbolic and root refs. While fine in the > > context of git-clone(1), this keeps us from using the mechanism in more > > cases. > > > > Adapt the logic to also support symbolic and root refs by using a second > > transaction that we use for all of the refs that need to be written as > > loose refs. > > > > The patch looks good. I was wondering if another way would be to just > add symref and root ref support to packed-refs. But that might be a > bigger undertaking than what we're doing here. > > But thinking about it, seems like we can do that in a backwards > compatible way too. I don't think you can without introducing a new version of the format, mostly because you also have to think about clients that don't support the new format. So all of this would be a big undertaking, and given that we have the reftable backend nowadays I don't think it'd be all that valuable. Patrick
