Environment variable iterator overrides

[Nemo157]

For case #24214, additional iterators mentioned in #24214 (comment).

Overrides std::env::{ Vars, VarsOs }::{ count, nth, last} to proxy through to their inner iterator (would be nice if there were some iterator wrapper trait that meant these didn't need to be implemented every time an iterator was wrapped...).

Overrides std::sys::unix::os::Env::{ count, nth, last } to proxy through to the inner vec::IntoIter. Should make count faster, and future proofs in case last or nth get optimised implementations in the future.

Overrides std::sys::windows::Env::{ count, nth, last } to not allocate OsStrings when not needed. Done by defining a sub-iterator that returns the raw pointers and lengths, this uses the default implementations as they should be pretty much equivalent to manually implemented versions, then in the outer iterator does the allocation with the result.

I don't have a machine I can actually test the windows changes on, with some hackery (https://gist.github.com/Nemo157/0d1f02b08c132f1a00da) I managed to get the type checker at least running to ensure it should compile. It would be nice if there were some way to test the higher level platform specific code. As far as I can tell testing windows::Env on unix would work fine if the infrastructure was setup for it.

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @brson (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[alexcrichton]

To me when I initially read this it wasn't clear why the separation here was needed, could you add a comment explaining why these are separate?

[alexcrichton]

Could you add some tests as well which exercise these functions?

[Nemo157]

@alexcrichton done.

Again I could only type check the tests as I don't have a windows machine setup for rust dev.

I hope using the type alias for sys::windows::os::Env is ok? I couldn't figure out any other way to make sure the tests had the correct lifetime and use slices between the iterators without either having to add a unused lifetime parameter to the sys::unix::os::Env or change the public API.

[alexcrichton]

Ah that may be an extra layer or two of indirection which may not be worth it, perhaps just Env could be used which uses the inner copy with 'static? I'd also be fine if EnvSlices just returned &'static [u16] and documenting that it's unsafe and the lifetime needs to be contained elsewhere.

Also, could you move the tests so they cover both the Unix and Windows cases? That should also help exercise the implementation in terms of what's actually being run with env::vars. I'd recommend adding the test to run-pass to ensure it's isolated from everything else going on.

[alexcrichton]

This may or may not pass on Windows due to #31259

[Nemo157]

Damn, that is... annoying 😦

[alexcrichton]

Ok, this looks good to me, could you also squash the commits together?

[alexcrichton]

@bors: r+ 52d506e

Thanks!

[bors]

⌛ Testing commit 52d506e with merge 9e16bea...

[alexcrichton]

@bors: retry force

[bors]

⌛ Testing commit 52d506e with merge ce41c0e...

[bors]

💔 Test failed - auto-linux-64-x-android-t

[alexcrichton]

Closing due to inactivity, but feel free to reopen with tests fixed!