Jeff King <p...@peff.net> writes: > On Thu, Mar 16, 2017 at 03:23:59PM -0700, Junio C Hamano wrote: > >> I am wondering if we should queue another one for .travis.yml on top >> to force use of USE_SHA1DC=YesPlease during the tests. I expect >> that we'd be encouraging its use for ordinary users without any >> specific needs in the release notes in 2.13 release. > > I don't think it would buy us much. There's not really any way for this > build to interact with the rest of the code in any interesting way, so > either it works as a SHA-1 implementation or it doesn't. If you just > want it exercised, I'll say that it's powering all of github.com right > now. > > I did wonder if we should ship with it as the default (instead of > openssl). It's definitely slower, but maybe widespread safety is a good > thing. OTOH, I think we have a fair bit of time before we see any > real-life collisions, just given the time and expense of generating > them.
My .travis.yml suggestion was about testing with SHA1DC in preparation for making it the default. That would give us another incentive to keep an eye on its performance, too, before we make it the default in Makefile, at which time the forced selection in the travis configuration can be removed. Thanks.
