rustdoc: avoid many `Symbol` to `String` conversions.
Particularly when constructing file paths and fully qualified paths.
This avoids a lot of allocations, speeding things up on almost all
examples.
r? `@GuillaumeGomez`
I seem to recall that in general, it's best to request an allocation
with a size that's a power of 2. The low estimate of 5 was probably a
little too low as well.
Rollup of 9 pull requests
Successful merges:
- #92045 (Don't fall back to crate-level opaque type definitions.)
- #92381 (Suggest `return`ing tail expressions in async functions)
- #92768 (Partially stabilize `maybe_uninit_extra`)
- #92810 (Deduplicate box deref and regular deref suggestions)
- #92818 (Update documentation for doc_cfg feature)
- #92840 (Fix some lints documentation)
- #92849 (Clippyup)
- #92854 (Use the updated Rust logo in rustdoc)
- #92864 (Fix a missing dot in the main item heading)
Failed merges:
- #92838 (Clean up some links in RELEASES)
r? `@ghost`
`@rustbot` modify labels: rollup
Deduplicate box deref and regular deref suggestions
Remove the suggestion code special-cased for Box deref.
r? ```@camelid```
since you introduced the code in #90627
Suggest `return`ing tail expressions in async functions
This PR fixes#92308.
Previously, the suggestion to `return` tail expressions (introduced in #81769) did not apply to `async` functions, as the suggestion checked whether the types were equal disregarding `impl Future<Output = T>` syntax sugar for `async` functions. This PR changes that in order to fix a potential papercut.
I'm not sure if this is the "right" way to do this, so if there is a better way then please let me know.
I amended an existing test introduced in #81769 to add a regression test for this, if you think I should make a separate test I will.
Swift has specific syntax that desugars to `Option<T>` similar to our
`?` operator, which means that people might try to use it in Rust. Parse
it and gracefully recover.
Currently all generators are named with a `generator$N` suffix,
regardless of where they come from. This means an `async fn` shows up as
a generator in stack traces, which can be surprising to async
programmers since they should not need to know that async functions are
implementated using generators.
This change generators a different name depending on the generator kind,
allowing us to tell whether the generator is the result of an async
block, an async closure, an async fn, or a plain generator.
Closure capture cleanup & refactor
Follow up of #89648
Each commit is self-contained and the rationale/changes are documented in the commit message, so it's advisable to review commit by commit.
The code is significantly cleaner (at least IMO), but that could have some perf implication, so I'd suggest a perf run.
r? `@wesleywiser`
cc `@arora-aman`
Update cargo
6 commits in 358e79fe56fe374649275ca7aebaafd57ade0e8d..06b9d31743210b788b130c8a484c2838afa6fc27
2022-01-04 18:39:45 +0000 to 2022-01-11 23:47:29 +0000
- Port cargo to clap3 (rust-lang/cargo#10265)
- feat: support rustflags per profile (rust-lang/cargo#10217)
- Make bors ignore the PR template so it doesn't end up in merge messages (rust-lang/cargo#10267)
- Be resilient to most IO error and filesystem loop while walking dirs (rust-lang/cargo#10214)
- Remove the option to disable pipelining (rust-lang/cargo#10258)
- Always ask rustc for messages about artifacts, and always process them (rust-lang/cargo#10255)
rustdoc: Preserve rendering of macro_rules matchers when possible
Fixes#92331. This approach restores the behavior prior to #86282 **if** the matcher token held by the compiler **and** the matcher token found in the source code are identical TokenTrees. Thus #86208 remains fixed, but without regressing formatting for the vast majority of macros which are not macro-generated.
[code coverage] Fix missing dead code in modules that are never called
The issue here is that the logic used to determine which CGU to put the dead function stubs in doesn't handle cases where a module is never assigned to a CGU (which is what happens when all of the code in the module is dead).
The partitioning logic also caused issues in #85461 where inline functions were duplicated into multiple CGUs resulting in duplicate symbols.
This commit fixes the issue by removing the complex logic used to assign dead code stubs to CGUs and replaces it with a much simpler model: we pick one CGU to hold all the dead code stubs. We pick a CGU which has exported items which increases the likelihood the linker won't throw away our dead functions and we pick the smallest to minimize the impact on compilation times for crates with very large CGUs.
Fixes#91661Fixes#86177Fixes#85718Fixes#79622
r? ```@tmandry```
cc ```@richkadel```
This PR is not urgent so please don't let it interrupt your holidays! 🎄🎁
Welcome opaque types into the fold
r? ```@nikomatsakis``` because idk who else to bug on the type_op changes
The commits have explanations in them. The TLDR is that
* 5c46002273 stops the "recurse and replace" scheme that replaces opaque types with their canonical inference var by just doing that ahead of time
* bdeeb07bf6 does not affect anything on master afaict, but since opaque types generate obligations when instantiated, and lazy TAIT instantiates opaque types *everywhere*, we need to properly handle obligations here instead of just hoping no problematic obligations ever come up.
Make rlib metadata strip works with MIPSr6 architecture
Because MIPSr6 has many differences with previous MIPSr2 arch, the previous rlib metadata stripping code in `rustc_codegen_ssa` is only for MIPSr2/r3/r5 (which share the same elf e_flags).
This commit fixed this problem. It makes `rustc_codegen_ssa` happy when compiling rustc for MIPSr6 target or hosts.
e_flags REF: e356027016/llvm/include/llvm/BinaryFormat/ELF.h (L562)
Migrate rustdoc from Tera to Askama
See #84419.
Should probably get a benchmarking run to verify if it has the intended effect on rustdoc performance.
cc `@jsha` `@jyn514.`
Add note about upstream commit musl-patch-configure.diff is derived from
The patch landed in upstream over the weekend so I'm adding a note so we can more easily track when our patch file can be removed.
rustdoc: Display "private fields" instead of "fields omitted"
Also:
* Always use `/* */` block comments
* Use the same message everywhere, rather than sometimes prefixing
with "some"
When I first read rustdoc docs, I was confused why the fields were being
omitted. It was only later that I realized it was because they were
private. It's also always bothered me that rustdoc sometimes uses `//`
and sometimes uses `/*` comments for these messages, so this change
makes them all use `/*`.
Technically, I think fields can be omitted if they are public but
`doc(hidden)` too, but `doc(hidden)` is analogous to privacy. It's
really just used to emulate "doc privacy" when -- because of technical
limitations -- an item has to be public. So I think it's fine to include
this under the category of "private fields".
r? ```@jsha```
Error when selected impl is not const in constck
Catches bad things when checking a `default_method_body_is_const` body, such as:
```rust
self.map(/* .. */).is_sorted();
```
When `Map` does not yet have a `const` `impl` for `Iterator`.
r? ```@oli-obk```
Store a `Symbol` instead of an `Ident` in `VariantDef`/`FieldDef`
The field is also renamed from `ident` to `name`. In most cases,
we don't actually need the `Span`. A new `ident` method is added
to `VariantDef` and `FieldDef`, which constructs the full `Ident`
using `tcx.def_ident_span()`. This method is used in the cases
where we actually need an `Ident`.
This makes incremental compilation properly track changes
to the `Span`, without all of the invalidations caused by storing
a `Span` directly via an `Ident`.
Actually instantiate the opaque type when checking bounds
Before this change, `instantiate_opaque_types` was a no-op, because it only works relative to the defined opaque type inference anchor. If it is a no-op, the for loop will not actually have anything to iterate over, and thus nothing is checked at all.
The field is also renamed from `ident` to `name. In most cases,
we don't actually need the `Span`. A new `ident` method is added
to `VariantDef` and `FieldDef`, which constructs the full `Ident`
using `tcx.def_ident_span()`. This method is used in the cases
where we actually need an `Ident`.
This makes incremental compilation properly track changes
to the `Span`, without all of the invalidations caused by storing
a `Span` directly via an `Ident`.
These closures were quite complex and part of a quite complex function.
The fact that they are closures makes mistakes likely when refactoring.
For example, earlier, I meant to use `resolved`, an argument of the
closure, but I instead typed `res`, which captured a local variable and
caused a subtle bug that led to a confusing test failure.
Extracting them as functions makes the code easier to understand and
refactor.
This currently calls `std` a "crate" in one part of the message and a
"module" in another part. The next commits fix this so it says "crate"
in both places.
This hack was added in 6ab1f05697.
I don't know what change allowed removing the hack, but that commit
added a test (which I presume covered the hack's behavior), and all
tests are passing with this change. So, I think it should be good.
rustdoc: Remove the intra-doc links side channel
The side channel made the code much more complex and harder to
understand. It was added as a temporary workaround in
0c99d806ea, but it's no longer necessary.
The addition of `UrlFragment` in #92088 was the key to getting rid of
the side channel. The semantic information (rather than the strings that
used to be used for fragments) that is now captured by `UrlFragment` is
enough to obviate the side channel. An additional change had to be made
to `UrlFragment` in this PR to make this possible: it now records
`DefId`s rather than item names.
This PR also consolidates the checks for anchor conflicts into one place.
r? `@Manishearth`
As discussed in
https://github.com/rust-lang/rust/pull/92142#issuecomment-1008239473,
tests that contain multiple files order their results differently on
Windows and Linux which the test runner currently can't handle
correctly. For now, ignore the "bin" part of the test and only include
the unused library dependency which is what the test really cares about
anyway.
Normalize generator-local types with unevaluated constants
Normalize generator-interior types in addition to (i.e. instead of just) erasing regions, since sometimes we collect types with unevaluated const exprs.
Fixes#84737Fixes#88171Fixes#92091Fixes#92634
Probably also fixes#73114, but that one has no code I could test. It looks like it's the same issue, though.
Normalize struct tail type when checking Pointee trait
Let's go ahead and implement the FIXMEs by properly normalizing the struct-tail type when satisfying a Pointee obligation. This should fix the ICE when we try to calculate a layout depending on `<Ty as Pointee>::Metadata` later.
Fixes#92128Fixes#92577
Additionally, mark the obligation as ambiguous if there are any infer types in that struct-tail type. This has the effect of causing `<_ as Pointee>::Metadata` to be properly replaced with an infer variable ([here](https://github.com/rust-lang/rust/blob/master/compiler/rustc_trait_selection/src/traits/project.rs#L813)) and registered as an obligation... this turns out to be very important in unifying function parameters with formals that are assoc types.
Fixes#91446
Mak DefId to AccessLevel map in resolve for export
hir_id to accesslevel in resolve and applied in privacy
using local def id
removing tracing probes
making function not recursive and adding comments
Move most of Exported/Public res to rustc_resolve
moving public/export res to resolve
fix missing stability attributes in core, std and alloc
move code to access_levels.rs
return for some kinds instead of going through them
Export correctness, macro changes, comments
add comment for import binding
add comment for import binding
renmae to access level visitor, remove comments, move fn as closure, remove new_key
fmt
fix rebase
fix rebase
fmt
fmt
fix: move macro def to rustc_resolve
fix: reachable AccessLevel for enum variants
fmt
fix: missing stability attributes for other architectures
allow unreachable pub in rustfmt
fix: missing impl access level + renaming export to reexport
Missing impl access level was found thanks to a test in clippy
Also:
* Always use `/* */` block comments
* Use the same message everywhere, rather than sometimes prefixing
with "some"
When I first read rustdoc docs, I was confused why the fields were being
omitted. It was only later that I realized it was because they were
private. It's also always bothered me that rustdoc sometimes uses `//`
and sometimes uses `/*` comments for these messages, so this change
makes them all use `/*`.
Technically, I think fields can be omitted if they are public but
`doc(hidden)` too, but `doc(hidden)` is analogous to privacy. It's
really just used to emulate "doc privacy" when -- because of technical
limitations -- an item has to be public. So I think it's fine to include
this under the category of "private fields".
Hash `Ident` spans in all HIR structures
This PR removes all of the `#[stable_hasher(project(name))]`
attributes used in HIR structs. While these attributes are not known
to be causing any issues in practice, we need to hash these in
order for the incremental system to work correctly -
a query could be otherwise be incorrectly marked green
when a change occures in one of the `Span`s that it uses.
rustdoc: Introduce a resolver cache for sharing data between early doc link resolution and later passes
The refactoring parts of https://github.com/rust-lang/rust/pull/88679, shouldn't cause any slowdowns.
r? `@jyn514`
Don't resolve blocks in foreign functions
Although it is an error for a foreign function to have a block, it is still possible at the level of the AST. #74204 made AST lowering skip over blocks belonging to foreign functions, since they're invalid. However, resolve still treated these blocks normally, resulting in a mismatch between the HIR and resolve, which could cause an ICE under certain circumstances. This PR changes resolve to skip over blocks belonging to foreign functions, as AST lowering does.
Fixes#91370.
r? ``@cjgillot``
Add a query for resolving an impl item from the trait item
This makes finding the item in an impl that implements a given trait item a query. This is for a few reasons:
- To slightly improve performance
- To avoid having to do name resolution during monomorphisation
- To make it easier to implement potential future features that create anonymous associated items
rustdoc: Remove apparently unnecessary conditional in `doc_value`
I need to remove this conditional for #91072, but while it seems
unnecessary, we are not certain. So, the plan is to first remove the
conditional and see if any regressions pop up before doing the refactor.
This way, it will be easier to revert if there are subtle regressions.
r? `@jyn514`
Min capture computation can already handle the same place appearing twice,
and previous commits made CaptureInfo construction very cheap, so just
delegate all work to min capture and let InferBorrowKind and
process_collected_capture_information handle everything linearly.
Region info is completely unnecessary for upvar capture kind computation
and is only needed to create the final upvar tuple ty. Doing so makes
creation of UpvarCapture very cheap and expose further cleanup opportunity.
Break the loop
A missing break statement lead to an infinite loop in bootstrap.py.
I also added a short sleep so it's not constantly running at 100%. But I can remove that if it's not wanted.
Fixes#76661
Fix spacing and ordering of words in pretty printed Impl
Follow-up to #92238 fixing one of the FIXMEs.
```rust
macro_rules! repro {
($item:item) => {
stringify!($item)
};
}
fn main() {
println!("{}", repro!(impl<T> Struct<T> {}));
println!("{}", repro!(impl<T> const Trait for T {}));
}
```
Before: `impl <T> Struct<T> {}`
After: `impl<T> Struct<T> {}`
Before: `impl const <T> Trait for T {}` 😿
After: `impl<T> const Trait for T {}`
Delay remaining `span_bug`s in drop elaboration
This follows changes from #67967 and converts remaining `span_bug`s into
delayed bugs, since for const items drop elaboration might be executed
on a MIR which failed borrowck.
Fixes#81708.
Fixes#91816.
I had the epiphany that now that fragments are "semantic" -- rather than
just strings -- they fill the role that used to be handled by the side
channel. I think I may be able to get rid of the other uses of the side
channel using this technique too.
Rollup of 7 pull requests
Successful merges:
- #92058 (Make Run button visible on hover)
- #92288 (Fix a pair of mistyped test cases in `std::net::ip`)
- #92349 (Fix rustdoc::private_doc_tests lint for public re-exported items)
- #92360 (Some cleanups around check_argument_types)
- #92389 (Regression test for borrowck ICE #92015)
- #92404 (Fix font size for [src] links in headers)
- #92443 (Rustdoc: resolve associated traits for non-generic primitive types)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Rustdoc: resolve associated traits for non-generic primitive types
Fixes#90703
This seems to work:
<img width="457" alt="image" src="https://user-images.githubusercontent.com/2807772/147774059-9556ff96-4519-409e-8ed0-c33ecc436171.png">
I'm just afraid I might have missed some cases / broken previous functionality.
I also have not written tests yet, I will have to take a look to see where tests are and how they are structured, but any help there is also appreciated.
Some cleanups around check_argument_types
Split out in ways from my rebase/continuation of #71827
Commits are mostly self-explanatory and these changes should be fairly straightforward
Fix rustdoc::private_doc_tests lint for public re-exported items
Closes#72081
This involves changing the lint to check the access level is exported, rather than public.
The [exported access level](e91ad5fc62/compiler/rustc_middle/src/middle/privacy.rs (L24)) accounts for public items and items accessible to other crates with the help of `pub use` re-exports.
The pattern of re-exporting public items from a private module is usage seen in a number of popular crates.
`thorin` is a Rust implementation of a DWARF packaging utility that
supports reading DWARF objects from archive files (i.e. rlibs) and
therefore is better suited for integration into rustc.
Signed-off-by: David Wood <david.wood@huawei.com>
In #79570, `-Z split-dwarf-kind={none,single,split}` was replaced by `-C
split-debuginfo={off,packed,unpacked}`. `-C split-debuginfo`'s packed
and unpacked aren't exact parallels to single and split, respectively.
On Unix, `-C split-debuginfo=packed` will put debuginfo into object
files and package debuginfo into a DWARF package file (`.dwp`) and
`-C split-debuginfo=unpacked` will put debuginfo into dwarf object files
and won't package it.
In the initial implementation of Split DWARF, split mode wrote sections
which did not require relocation into a DWARF object (`.dwo`) file which
was ignored by the linker and then packaged those DWARF objects into
DWARF packages (`.dwp`). In single mode, sections which did not require
relocation were written into object files but ignored by the linker and
were not packaged. However, both split and single modes could be
packaged or not, the primary difference in behaviour was where the
debuginfo sections that did not require link-time relocation were
written (in a DWARF object or the object file).
This commit re-introduces a `-Z split-dwarf-kind` flag, which can be
used to pick between split and single modes when `-C split-debuginfo` is
used to enable Split DWARF (either packed or unpacked).
Signed-off-by: David Wood <david.wood@huawei.com>
Rollup of 7 pull requests
Successful merges:
- #92092 (Drop guards in slice sorting derive src pointers from &mut T, which is invalidated by interior mutation in comparison)
- #92388 (Fix a minor mistake in `String::try_reserve_exact` examples)
- #92442 (Add negative `impl` for `Ord`, `PartialOrd` on `LocalDefId`)
- #92483 (Stabilize `result_cloned` and `result_copied`)
- #92574 (Add RISC-V detection macro and more architecture instructions)
- #92575 (ast: Always keep a `NodeId` in `ast::Crate`)
- #92583 (⬆️ rust-analyzer)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Set font size proportional to user's font size
According to MDN (https://developer.mozilla.org/en-US/docs/Web/CSS/font-size),
> To maximize accessibility, it is generally best to use values that are relative to the user's default font size.
> Defining font sizes in px is not accessible, because the user cannot change the font size in some browsers.
Note that changing font size (in browser or OS settings) is distinct from the zoom functionality triggered with Ctrl/Cmd-+. Zoom
functionality increases the size of everything on the page, effectively applying a multiplier to all pixel sizes. Font size changes apply to just text.
For relative font sizes, we could use `em`, as we do in several places already. However that has a problem of "compounding" (see MDN article for details). The compounding problem is nicely solved by `rem`, which make font sizes relative to the root element, not the parent element.
Since we were using a hodge-podge of pixel sizes, em, rem, and percentage sizes before, this change switches everything to rem, while keeping the same size relative to our old default of 16px.
16px is still the default on most browsers, for users that haven't set a larger or smaller font size.
Part of #59845. Note: this will conflict with #92404. We should merge that first (once it's done) and I'll resolve the merge conflicts.
r? `@GuillaumeGomez`
Demo: https://rustdoc.crud.net/jsha/font-size-access/std/string/struct.String.html
