On Wed, Jul 18, 2018 at 04:31:52PM -0400, Jeff King wrote: > Since Stefan's patch logically undoes e1111cef23, I think that's why he > put in the comment to move back to the old name. > > Personally, I do not find one name any more informative than the other, > and would be happy to leave it as-is (dropping the comment). > > But I'm also fine with following through on the "do". According to > c3c36d7de2, that was waiting for a calmer time in the code base. I guess > the best way to find out is to write the patch and see how terribly it > conflicts with pu. :) It turns out there are no conflicts right now, aside from the patch I just sent. And perhaps your commit-graph work is going to add another reference, if you take my suggestion in the other thread. ;) So I remain ambivalent, but here is the patch to do so, with mine now on top (the only difference is s/read/check/ in the variable name). But note that while changing this, I noticed a leftover bit from c3c36d7de2 that should be dealt with in either case. I put that at the front of the series. [1/3]: check_replace_refs: fix outdated comment [2/3]: check_replace_refs: rename to read_replace_refs [3/3]: add core.usereplacerefs config option Documentation/config.txt | 5 +++++ builtin/fsck.c | 2 +- builtin/index-pack.c | 2 +- builtin/pack-objects.c | 2 +- builtin/prune.c | 2 +- builtin/replace.c | 2 +- builtin/unpack-objects.c | 2 +- builtin/upload-pack.c | 2 +- cache.h | 6 ++---- config.c | 5 +++++ environment.c | 4 ++-- git.c | 2 +- log-tree.c | 2 +- replace-object.c | 2 +- replace-object.h | 2 +- t/t6050-replace.sh | 6 ++++++ 16 files changed, 31 insertions(+), 17 deletions(-) -Peff
