Re: [PATCH 003/194] object-store: move packed_git and packed_git_mru to object store

René Scharfe [off-list ref] writes:

A short semantic patch with a limited time of usefulness and possible
side-effects can easily be included in a commit message, of course..

Yeah, I think that is Jonathan's favourite approach as well, and I
do not have problem with that.  This transition spatch is unlike the
"avoid strbuf_addf when we do not have to use it" one, in that there
is a definite cut-over where the old way becomes an invalid C code
and compilers can help us spot unconverted places (as opposed to a
valid but unwanted way that can come into the codebase after one
round of code cleaning is done, which is a useful thing to catch
with the "make coccicheck" target).