Jeff King <peff@xxxxxxxx> writes: > This was noticed by Coverity. The leak isn't new, but I think was > "re-found" by Coverity because some nearby code did an unrelated > s/sha1/oid/ change, throwing off its heuristics. > > I also checked whether this was in Dscho's big pack of Coverity fixups > from earlier today, and it's not. > > The first one is the actual fix. The second one is pure cleanup, but I > think worth doing. The final one is arguably just churn, and maybe > people even like the end result less. I'm OK to drop it. > > [1/3]: am: fix commit buffer leak in get_commit_info() > [2/3]: am: simplify allocations in get_commit_info() > [3/3]: am: shorten ident_split variable name in get_commit_info() All looked sensible. Thanks.