Is there a clever reason why __floatdisf and __floatundisf are unimplemented?

[ketsuban]

Looking at the code it seems like they should be really easy - all the functions in this group use a dead-simple macro. Is it just that nobody got around to doing it, or are these functions special somehow (e.g. problems a la rust-lang/rust#10184)?

[kevinaboos]

I also need these functions, specifically __floatundisf. I'm happy to contribute them to this crate for x86_64, but currently I have no idea where to begin. If someone can point me in the right direction as to how to implement them, I can get started.

[alexcrichton]

Sure yeah! Currently we mostly just port the implementations from C in the compiler-rt project directly to Rust (then add some tests and such). You can find the compiler-rt intrinsics in compiler-rt/lib/builtins typically, and I normally grep around for the symbol name to find where it's defined

[ketsuban]

Back when I first created this issue I forked the repo and added these two intrinsics using __floatdidf and __floatundidf as a template, purely to address my personal need. I didn't make a pull request because I was cognizant of the fact I was engaging in cargo-cult programming, but now there's eyes on this issue I'll make sure everyone knows about it even if only so someone can tell me I was doing it wrong and should feel bad.

[alexcrichton]

Oh we'd love a PR! Right now the tests hopefully cover enough cases to ensure that it's implemented correctly, and otherwise cargo-culting from compiler-rt is par for the course :)