On Thu, Aug 01, 2019 at 01:24:17PM -0400, Jeff Hostetler wrote: > > By the way, I wondered why trace2's existing config reading did not > > cause us to segfault because of this. It is because it invented the > > "very early config" function which always ignores some config sources > > (working around this problem, but also making it weirdly unlike most > > other config). > > Yes, I added the "very early config" to try to work around some of > the chicken-n-egg problems. I can't say that I was completely happy > with having to do that. I meant to comment a little further on this earlier. While I do think it's unfortunate to have yet another set of special rules, I think what you ended up with is probably the least-bad thing. This trace2 config read is happening so early (even after my movement patch) that it makes me nervous to do anything at all complicated. Not just for performance reasons, which you cited in the original commit, but also for correctness and even security reasons. This is code we're running for every single binary before we even hit main(), and it's probably a good thing that we are not loading .git/config values from a potentially untrusted repository (e.g., upload-pack is supposed to be safe to run in an untrusted repo). -Peff
