closure handling for NLL
This branch overhauls how we handle closures and universally quantified regions in the NLL code. The goal is to lay the groundwork for "region erasure" by the HIR-based type checker, as well as to avoid "crazy hacks" when it comes to closures. This means that when we type-check a closure, we cannot make use of *any* of the precise values of free regions in its signature, since those are inferred by the HIR type-checker. Therefore, the new code handles closures by:
- Creating fresh regions for every free region that appears in the closure's type, which now includes both its signature and the types of all upvars.
- This means that the closure is type-checked without knowing about the connections.
- When we encounter some region relationship that we can't locally verify, we propagate it to the closure's creator.
- During MIR typeck, the closure creators then validates those relationships.
For a good example and explanation, see e.g. the test `src/test/nll/closure-requirements/propagate-despite-same-free-region.rs`.
Upcoming changes in the next NLL PR (not included in this PR in order to keep it manageable):
- Improvements to the MIR type checker so that it handles a lot of stuff presently overlooked
- Refactor how we store region values to make it more efficient, better encapsulated
- Propagate "type-outlives" relationships like `T: 'a` in a similar fashion to the one covered in this PR (still in the works, but close)
- Improvements to error reporting (still in the works)
r? @arielb1 or @pnkfelix
In #46382 the logic around linkage preservation with ThinLTO ws tweaked but the
loop that registered all otherwise exported GUID values as "don't internalize
me please" was erroneously too conservative and only asking "external" linkage
items to not be internalized. Instead we actually want the inversion of that
condition, everything *without* "local" linkage to be internalized.
This commit updates the condition there, adds a test, and...
Closes#46543
It's just not useful. It also makes it hard to have tests that probe
internal state, since the interning number is very sensitive.
Dumping the number in the case of gensym is not ideal but will do for
now.
It's inefficient, and the substitution there doesn't account for the
extra regions used by NLL inference, so it's a bad thing to encourage.
As it happens all callers already know if they have a closure or not,
from what I can tell.
ui tests are the future, especially since the
recent improvement where we have gained
checking and requiring of //~ERROR comments.
The tidy feature-gate test check is intended
to be 50% an actual insurance that there is
a check, and 50% to be a teacher that such
checks are required.
With this commit applied, newbies might
interpret stuff wrongly and create tests
that don't fail but succeed instead.
This is not what feature gate tests are
for though. Therefore, in a later step,
when only ui tests are allowed to be feature
gate tests, we will add checking to ensure
that a file marked as gate test is actually
required to be a compilation failure.
Right now implementing such a check is a
bit annoying as one needs to only do it
when the compile-fail test is in the
ui test suite :/.
compiletest: account for `ui` reference files when deciding to skip
The stamp files for compiletest were ignoring `.stderr` and `.stdout` files. This was driving me crazy.
r? @alexcrichton
Stabilize abi_sysv64
Closes#36167, stabilizing the use of the "sysv64" ABI on x64 platforms where it is not the default ABI.
FCP on this is complete in the tracking issue.
rustc_mir: promote references of statics from other statics.
Fixes#46522 by also allowing `STATIC_REF` in MIR const-qualification, not just AST rvalue promotion.
Display `\t` in diagnostics code as four spaces
Follow up to #44386 using the unicode variable width machinery from #45711 to replace tabs in the source code when displaying a diagnostic error with four spaces (instead of only one), while properly accounting for this when calculating underlines.
Partly addresses #44618.
MIR borrowck: implement union-and-array-compatible semantics
Fixes#44831.
Fixes#44834.
Fixes#45537.
Fixes#45696 (by implementing DerefPure semantics, which is what we want going forward).
r? @nikomatsakis
Give compile_error macro examples
I cannot get Rust to build at all with it complaining about GCC not being a valid C compiler or something, so letting TravisCI be my tester...
Fixes#46171
