Jeff King wrote: > This series fixes a regression in v2.11.1 where we might read past the > end of an mmap'd buffer. It was introduced in cf3c635210,
The above information is super helpful. Can it go in one of the commit messages? > but I didn't > base the patch on there, for a few reasons: > > 1. There's a trivial conflict when merging up (because of > git_open_noatime() becoming just git_open() in the inerim). > > 2. The reproduction advice relies on our SANITIZE Makefile knob, which > didn't exist back then. > > 3. The second patch does not apply there because we don't have > warn_on_fopen_errors(). Though admittedly it could be applied > separately after merging up; it's just a clean-up on top. Even this part could go in a commit message, but it's fine for it not to. Thanks, Jonathan
