Lldb cleanups

[richo]

While poking at rust in lldb I found a few nits to clean up.

[alexcrichton]

r? @michaelwoerister

[michaelwoerister]

This looks like it would put the type name inside the braces instead of in front of them.

[michaelwoerister]

Apart from the comments above this looks good to me. If you fix these, please run make check-stage1-debuginfo-lldb to make sure that no tests get broken by the changes.

[richo]

I'm fixing the issues now. The test harness looks broken on master. Poking now.

[richo]

This seems like it's the same predicate as has_field_names, but I haven't dug far enough or thought hard enough about it to verify. If it is, this can be cleaned up further by just removing the $(type_name)s from the second template. Is that accurate?

[michaelwoerister]

No, it's a bit different. Tuple structs have a type name but no field names, as in struct MyInt(int). The same is true for tuple-like enum variants.

[richo]

Oh, as opposed to a type alias for a tuple type, eg type Foobar = (int, f64)?

[richo]

Ignore that, it's just namespaced enums fall out. Patch will go out on this PR in a sec.

[richo]

Ok, the non ignored tests are all green, and the compiletest harness works (is there a reason this doesn't happen in CI? something something bors doesn't have debuggers handy?)

r? @michaelwoerister

[michaelwoerister]

Thanks, @richo. This looks good to me now.
It's strange though that you had to do those changes in runtest.rs. The

#[cfg(not(stage0))]
use self::TargetLocation::*;

part at the top should have taken care of that. @alexcrichton, any idea what that's about?

[alexcrichton]

Yeah that's fine, the stage0 compiler inserted that import automatically (essentially), so it's flagged as unused, but the stage1+ compiler doesn't do that, so we have to add it explicitly.

[michaelwoerister]

Yeah, but even with the explicit use statement, these changes in runtest.rs were apparently needed for it to compile. Am I missing something here?

[alexcrichton]

Oh now that I wouldn't expect!

[alexcrichton]

It may just be some stage0/stage1 weirdness though, I wouldn't worry too too much about it.

[michaelwoerister]

Thanks Alex! I'll try to give bors the OK on the PR.