Edward Thomson
Edward Thomson
I left it implicit - please let me be explicit about it: I realize that the sheer number of people using a single API in one product is _my_ problem,...
@dahlbyk Yeah, I think that's relatively accurate. I don't think that there's necessarily a lot of work here or that it's a particularly hard problem. :)
I'd rather move a null check to the beginning of the function? `GIT_ASSERT_ARG(value)`?
I was heartened when I read this: > Let's address this mess and return the pseudoref terminology back to its original intent: a ref that sometimes behave like a ref,...
> So in some very specific situations they behave like refs, but they really aren't refs because their semantics are just so special. These aren't *semantically* special, they're *structurally* special....
What's the right behavior here, do you think? Should we try to automatically downgrade when `mmap` fails? 🤔
Interesting, and surprising. I've not seen this before. The failure is that it's not getting a 0 exit code from the `env.cmd` tool that we bundle to check that we...
Surprising. I haven't had time to take a look yet; does it work with any backend for you? 🤔
That's so weird. Thanks for the details.
This does seem unexpected and certainly may be a bug.