The current codegen tests only compare IR line counts between similar
rust and C programs, the latter getting compiled with clang. That looked
like a good idea back then, but actually things like lifetime intrinsics
mean that less IR isn't always better, so the metric isn't really
helpful.
Instead, we can start doing tests that check specific aspects of the
generated IR, like attributes or metadata. To do that, we can use LLVM's
FileCheck tool which has a number of useful features for such tests.
To start off, I created some tests for a few things that were recently
added and/or broken.
Currently, for `use` declarations with multiple paths, only the `use` item itself is saved in the AST map, not the individual path nodes. This can lead to a problem when a span of a specific path node is needed.
For example, #24818 caused an ICE because of this, in `ImportResolver::check_for_conflicting_import()`.
Fixes#25763.
Needed to support:
```rust
match X {
pattern if Y ...
}
for pattern in Y {}
```
IMO, this shouldn't require an RFC because it can't interfere with any future language changes (because `pattern if` and `pattern in` are already legal in rust) and can't cause any ambiguity.
- Adds explanations for E0055, E0089, E0192, E0261-E0263, E0318
- Improves explanations for E0250, E0368, E0372.
- Converts 15 diagnostics to have error codes (E0380-E0394). Adds an explanation for E0380.
- The E0087-E0090 messages currently look like "expected {} parameter(s) found {} parameter(s)". This changes them to either use "parameter" or "parameters", based on the number.
This is, in part, more progress towards #24407
rustdoc: Associated type fixes
The first commit fixes a bug with "dud" items in the search index from
misrepresented `type` items in trait impl blocks.
For a trait *implementation* there are typedefs which are the types for
that particular trait and implementor. Skip these in the search index.
There were lots of dud items in the search index due to this (search for
Item, Iterator's associated type).
Add a boolean to clean::TypedefItem so that it tracks whether the it is
a type alias on its own, or if it's a `type` item in a trait impl.
The second commit fixes a bug that made signatures and where bounds
using associated types (if they were not on `Self`) incorrect.
The third commit fixes so that where clauses in type alias definititons
are shown.
Fixes#22442Fixes#24417Fixes#25769
Changes:
- adds explanations for E0185, E0186, E0202, E0326
- fixes the explanation for E0053. The previous description was too narrow; there are other error cases.
- changes the error message for E0202 to be specific for associated types, since it seems inherent associated constants are implemented.
Part of #24407
Closes#25046 (by rejecting the code that causes the ICE) and #24946. I haven't been able to deal with the array size or recursion issues yet for associated consts, though my hope was that the change I made for range match patterns might help with array sizes, too.
This PR is pretty much orthogonal to #25065.
This flag indicates that when files are being replaced or added to archives (the
`r` flag) that the new file should not be inserted if it is not newer than the
file that already exists in the archive. The compiler never actually has a use
case of *not* wanting to insert a file because it already exists, and this
causes rlibs to not be updated in some cases when the compiler was re-run too
quickly.
Closes#18913
A regression was introduced by commit 7b1916d253#25612. Negative signed integer literals less than -9223372036854775808i64 were no longer properly reported as #[warn(overflowing_literals)].
Also adding missing test cases to test/compile-fail/lint-type-overflow.rs which could have detected the regression.
Further explanation:
The expression `(negative && v > max as u64 + 1)` relies on the fact that algebraically speaking `-min == max + 1` to avoid negation and removing the need for `min` completely.
If i128 or i256 are ever added, it should also work for these types without requiring a change to `min != i64::MIN &&` also simplifying maintenance.
r? @pnkfelix
Currently, for `use` declarations with multiple paths, only the `use`
item itself is saved in the AST map, not the individual path nodes. This
can lead to a problem when a span of a specific path node is needed.
For example, #24818 caused an ICE because of this, in
`ImportResolver::check_for_conflicting_import()`.
Fixes#25763.
Functions such as `fn foo<I: Iterator>(x: I::Item)` would not
render correctly and displayed `I` instead of `I::Item`. Same thing
with `I::Item` appearing in where bounds.
This fixes the bug by using paths for generics.
Fixes#24417
This is a port of @eddyb's `const-fn` branch. I rebased it, tweaked a few things, and added tests as well as a feature gate. The set of tests is still pretty rudimentary, I'd appreciate suggestions on new tests to write. Also, a double-check that the feature-gate covers all necessary cases.
One question: currently, the feature-gate allows the *use* of const functions from stable code, just not the definition. This seems to fit our usual strategy, and implies that we might (perhaps) allow some constant functions in libstd someday, even before stabilizing const-fn, if we were willing to commit to the existence of const fns but found some details of their impl unsatisfactory.
r? @pnkfelix
fixes#25683
I have a very nonscientific measurement of the data via valgrind/massif [here](https://gist.github.com/Manishearth/4c47f15f6835cb3957c4)
I measured the memory usage for both --pretty=expanded and -Z no-trans
It *seems* like there's a 20-25MB decrease during expansion on stage2 librustc; but I'm not quite sure.
r? @eddyb
(have not yet run tests, but it compiles fine, might want to wait before giving r+)
cc @nrc @huon
The `debug_builders` feature is up for 1.1 stabilization in #24028. This commit stabilizes the API as-is with no changes.
Some nits that @alexcrichton mentioned that may be worth discussing now if anyone cares:
* Should `debug_tuple_struct` and `DebugTupleStruct` be used instead of `debug_tuple` and `DebugTuple`? It's more typing but is a technically more correct name.
* `DebugStruct` and `DebugTuple` have `field` methods while `DebugSet`, `DebugMap` and `DebugList` have `entry` methods. Should we switch those to something else for consistency?
cc @alexcrichton @aturon
- add feature gate
- add basic tests
- adjust parser to eliminate conflict between `const fn` and associated
constants
- allow `const fn` in traits/trait-impls, but forbid later in type check
- correct some merge conflicts
For a trait *implementation* there are typedefs which are the types for
that particular trait and implementor. Skip these in the search index.
There were lots of dud items in the search index due to this (search for
Item, Iterator's associated type).
Add a boolean to clean::TypedefItem so that it tracks whether the it is
a type alias on its own, or if it's a `type` item in a trait impl.
Fixes#22442
When taking the address of an unsized field we generate a rvalue datum
for the field and then convert it to an lvalue datum. At that point,
cleanup is scheduled for the field, leading to multiple drop calls.
The problem is that we generate an rvalue datum for the field, since the
pointer does not own the data and there's already cleanup scheduled
elsewhere by the true owner. Instead, an lvalue datum must be created.
Thanks to @eddyb for identifying the underlying cause and suggesting the
correct fix.
Fixes#25549.
I think I didn't run tests properly - my second call to
select_all_obligations_or_error has made 3 tests fail. However, this is
just an error message change - integer fallback never worked with casts.
