On Fri, Jul 16, 2021 at 06:28:06PM -0700, Junio C Hamano wrote: > >> I then just never got to picking it up again, I'll probably re-roll it & > >> make it a part of this series, then we can remove this whole OID != NULL > >> case and will be sure nothing fishy's going on. > > > > Yeah, that sounds like a good path forward. I do think the patch under > > discussion here is probably the right thing to do. But it becomes all > > the more obvious if lock_ref_oid_basic() ends up dropping that parameter > > entirely. > > OK, so what's the final verdict on this step? It is unfortunate > that when Ævar took over a topic from Han-Wen, this patch has been > inserted as the very first step before the patches in the series, so > until we know we are happy with it, it takes several other patches > hostage. I just read through v2. Modulo a few small nits (mostly typos, but a few commit message suggestions), it looks good to me. I agree it's a lot to stick in front of another set of patches, but I think in this case we can proceed quickly enough to make it worth doing in the order Ævar suggests. -Peff
