| Age | Commit message (Collapse) | Author | Lines |
|---|
|
this is because the mobile sidebar cannot be resized,
unlike on desktop.
|
|
|
|
rustdoc: add a handle that makes sidebar resizing more obvious
This aims to make the resizable sidebars more obvious
Preview: <https://notriddle.com/rustdoc-html-demo-12/sidebar-resize-handle/std/index.html>
This change is based on some discussion on [lolbinarycat's idea], but with a more "traditional" design. Specifically, while very few systems use exactly this design, most of them use [a skeumorph](https://ux.stackexchange.com/questions/80463/what-do-the-3-close-horizontal-bars-not-hamburger-menu-represent-and-what-is-t/80591#80591) of a grip texture:
- This design is similar to the one used in the Rust Playground, and almost identical to UX StackExchange:
<details><img src="https://github.com/user-attachments/assets/39a6bb69-4895-4fd0-87da-b87913bc7309"></details>
<details><img src="https://github.com/user-attachments/assets/a41942e1-651b-410b-b855-2aafe8fe54f4"></details>
- In Jira, resizable sidebars have a stack of four dots, but only in one row.
<details><img src="https://github.com/user-attachments/assets/13047998-02bf-47e6-b796-16f393f870b0"></details>
- In The GIMP, resizable sidebars have a stack of three dots.
<details><img src="https://github.com/user-attachments/assets/138f5c21-3069-4bbe-b306-0bb9a4bf0318"></details>
- In [old Windows], "panes" are defined to have the same border style as a window, which has a raised appearance. To evoke this, the PR adds a lightweight "shadow" border, darker than the sidebar itself
<details><img src="https://github.com/user-attachments/assets/301da4b8-6c48-4131-b741-1689af84670a"></details>
- In [NeXT], a drag point usually had an innie, whether the line in a slider or the circle in a scroller; I can also hide and show the favorites bar in Workspace by dragging on a circular "grip spot"
<details><img src="https://github.com/user-attachments/assets/b13c2d30-a3a8-4672-90fa-58c1fdf19f42"></details>
- In [old Mac], drag handles for things usually had a "grip track" of parallel lines.
<details><img src="https://github.com/user-attachments/assets/1fbecc67-ffbc-4ed6-a8c5-a9ff085638db"></details>
*This design is far closer to old Mac than anything else*, though they've put it in the bottom corner instead of the middle.
- [OSX] kept that, but the "Source List" part of the Finder still had the circle grip for a time the same way Workspace did (resulting in an odd mishmash, if you compare the source list sidebar with the other grip tracks embedded in the scrollbars).
<details><img src="https://github.com/user-attachments/assets/551b8f9d-2dd8-4291-917f-dc88741a2b97"></details>
[lolbinarycat's idea]: https://github.com/rust-lang/rust/pull/139420
[old Windows]: https://archive.org/details/windowsinterface00micr/page/n9/mode/2up
[old Mac]: https://archive.org/details/apple-hig/1996_Human_Interface_Guidelines_for_Mac_OS_8_%28WWDC_Release%29/page/16/mode/2up
[NeXT]: https://archive.org/details/apple-hig/1993%20NeXTSTEP%20User%20Interface%20Guidelines%20-%20Release%203/page/145/mode/2up
[OSX]: https://dn721903.ca.archive.org/0/items/apple-hig/MacOSX_HIG_2005_09_08.pdf#page=267
|
|
|
|
also export metrics from librustdoc
Addresses the issue mentioned here: [#t-docs-rs > metrics intitiative @ 💬](https://rust-lang.zulipchat.com/#narrow/channel/356853-t-docs-rs/topic/metrics.20intitiative/near/515714331)
The previous implementation only emitted metrics from rustc, but it turns out running `cargo doc` only calls `rustc` for dependencies, and not for the root crate being documented. We are planning to gather a sample dataset from docs.rs ci via `cargo doc` so as things stood this would not emit any metrics for any of the crates themselves that were published.
This change adds the same logic from `rustc_driver_impl` to `librustdoc` to also dump metrics at the end of its execution if they are enabled.
Note: The hash's generated by librustdoc will likely be completely different from the ones generated by rustc. This is because rustc is actually doing the various passes needed to fully calculate the stable version hash. My understanding of how rustdoc works is that the hashes generated will be working with partial information due to it only doing the work required to generate docs. The hashes will still be unique per crate and will work for the purposes of the metrics proof of concept, it would not be possible to correlate metrics generated by rustdoc with those generated by rustc for the same crate. This is fine for the purposes of the PoC but a future full implementation of metrics may want to address this issue.
|
|
r=GuillaumeGomez
rustdoc: Replace unstable flag `--doctest-compilation-args` with a simpler one: `--doctest-build-arg`
Tracking issue: https://github.com/rust-lang/rust/issues/134172.
Context: https://github.com/rust-lang/rust/pull/137096#issuecomment-2776318800
Yeets the ad hoc shell-like lexer for 'nested' program arguments.
No FCP necessary since the flag is unstable.
I've chosen to replace `compilation` with `build` because it's shorter (you now need to pass it multiple times in order to pass many arguments to the doctest compiler, so it matters a bit) and since I prefer it esthetically.
**Issue**: Even though we don't process the argument passed to `--doctest-build-arg`, we end up passing it via an argument file (`rustc `@argfile`)` which delimits arguments by line break (LF or CRLF, [via](https://doc.rust-lang.org/rustc/command-line-arguments.html#path-load-command-line-flags-from-a-path)) meaning ultimately the arguments still get split which is unfortunate. Still, I think this change is an improvement over the status quo.
I'll update the tracking issue if/once this PR merges. I'll also add the (CR)LF issue to 'unresolved question'.
r? GuillaumeGomez
r? notriddle
|
|
|
|
This change is based on some discussion on [lolbinarycat's idea],
but with a more "traditional" design. Specifically, this is the
closest thing I could find to a consensus across many systems I
looked at for inspiration:
- In Jira, resizable sidebars have a stack of four dots.
- In The GIMP, resizable sidebars have a stack of three dots.
- In [old Windows], "panes" are defined to have the same border
style as a window, which has a raised appearance.
- In [NeXT], a drag point usually had an innie, whether the line in a
slider or the circle in a scroller; I can also hide and show the
favorites bar in Workspace by dragging on a circular "grip spot"
- In [old Mac], drag handles for things usually had a "grip track"
of parallel lines.
- [OSX] kept that, but the "Source List" part of the Finder still had
the circle grip for a time the same way Workspace did
[lolbinarycat's idea]: https://github.com/rust-lang/rust/pull/139420
[old Windows]: https://archive.org/details/windowsinterface00micr/page/n9/mode/2up
[old Mac]: https://archive.org/details/apple-hig/1996_Human_Interface_Guidelines_for_Mac_OS_8_%28WWDC_Release%29/page/16/mode/2up
[NeXT]: https://archive.org/details/apple-hig/1993%20NeXTSTEP%20User%20Interface%20Guidelines%20-%20Release%203/page/145/mode/2up
[OSX]: https://dn721903.ca.archive.org/0/items/apple-hig/MacOSX_HIG_2005_09_08.pdf#page=267
|
|
All uses have been removed. And it's nonsensical: an identifier by
definition has at least one char.
The commits adds an is-non-empty assertion to `Ident::new` to enforce
this, and converts some `Ident` constructions to use `Ident::new`.
Adding the assertion requires making `Ident::new` and
`Ident::with_dummy_span` non-const, which is no great loss.
The commit amends a couple of places that do path splitting to ensure no
empty identifiers are created.
|
|
|
|
rustdoc-json: Remove newlines from attributes
Fixes #140689
Not sure if this needs to bump `FORMAT_VERSION` or not.
r? ``@GuillaumeGomez``
cc ``@obi1kenobi``
|
|
|
|
r=notriddle
[rustdoc] Ensure that temporary doctest folder is correctly removed even if doctests failed
Fixes #139899.
The bug was due to the fact that if any doctest fails for any reason, we call `exit` (or it's called inside `libtest` if not edition 2024), meaning that `TempDir`'s destructor isn't called, and therefore the temporary folder isn't cleaned up.
Took me a while to figure out how to reproduce but finally I was able to reproduce the bug with:
`````rust
#![doc(test(attr(deny(warnings))))]
//! ```
//! let a = 12;
//! ```
`````
And then I ensured that panicking doctests were cleaned up as well:
`````rust
//! ```
//! panic!();
//! ```
`````
And finally I checked if it was fixed for merged doctests too (`--edition 2024`).
To make this work, I needed to add a new public function in `libtest` too which would call a function once all tests have been run.
So only issue is: I have absolutely no idea how we can add a regression test for this fix. If anyone has an idea...
r? `@notriddle`
|
|
Follow up https://github.com/rust-lang/rust/pull/127127
|
|
failed
|
|
Unify sidebar buttons to use the same image
Part of https://github.com/rust-lang/rust/issues/139832.
The source sidebar looks like this with the new image:
You can test it [here](https://rustdoc.crud.net/imperio/sidebar-images/src/foo/foo.rs.html).
r? `@notriddle`
|
|
r=jieyouxu,wesleywiser
Implement RFC 3503: frontmatters
Tracking issue: #136889
Supercedes #137193. This implements [RFC 3503](https://github.com/rust-lang/rfcs/blob/master/text/3503-frontmatter.md).
This might break rust-analyzer. Will look into how to fix that.
Suggestions welcome for how to improve diagnostics.
|
|
Supercedes #137193
|
|
By consecutively initializing `tracing` and `rustc_log`, Rustdoc assumes
that these involve 2 different tracing crates.
I would like to be able to build rustdoc against the same tracing crate
that rustc_log is also built against. Previously this arrangement would
crash rustdoc:
thread 'main' panicked at rust/compiler/rustc_log/src/lib.rs:142:65:
called `Result::unwrap()` on an `Err` value: SetGlobalDefaultError("a global default trace dispatcher has already been set")
stack backtrace:
0: rust_begin_unwind
1: core::panicking::panic_fmt
2: core::result::unwrap_failed
3: rustc_log::init_logger
4: rustc_driver_impl::init_logger
5: rustdoc::main
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
error: the compiler unexpectedly panicked. this is a bug.
note: we would appreciate a bug report: https://github.com/rust-lang/rust/issues/new?labels=C-bug%2C+I-ICE%2C+T-rustdoc&template=ice.md
note: please make sure that you have updated to the latest nightly
query stack during panic:
end of query stack
|
|
|
|
|
|
|
|
Introduce `BoxMarker` to improve pretty-printing correctness
Box opening/closing is really easy to get wrong in the pretty-printers. This PR makes it much harder to get wrong.
r? `@Urgau`
|
|
Remove `weak` alias terminology
I find the "weak" alias terminology to be quite confusing. It implies the existence of "strong" aliases (which do not exist) and I'm not really sure what about weak aliases is "weak". I much prefer "free alias" as the term. I think it's much more obvious what it means as "free function" is a well defined term that already exists in rust.
It's also a little confusing given "weak alias" is already a term in linker/codegen spaces which are part of the compiler too. Though I'm not particularly worried about that as it's usually very obvious if you're talking about the type system or not lol. I'm also currently trying to write documentation about aliases and it's somewhat awkward/confusing to be talking about *weak* aliases, when I'm not really sure what the basis for that as the term actually *is*.
I would also be happy to just find out there's a nice meaning behind calling them "weak" aliases :-)
r? `@oli-obk`
maybe we want a types MCP to decide on a specific naming here? or maybe we think its just too late to go back on this naming decision ^^'
|
|
Fix detection of main function if there are expressions around it
Fixes #140162.
Fixes #139651.
Once this is merged, we can backport and I'll send a follow-up to emit a warning in case a `main` function is about to be "wrapped" (and therefore not run).
r? `@fmease`
try-job: x86_64-mingw-1
|
|
The pretty-printers open and close "boxes" of text a lot. The open and
close operations must be matched. The matching is currently all implicit
and very easy to get wrong. (#140280 and #140246 are two recent
pretty-printing fixes that both involved unclosed boxes.)
This commit introduces `BoxMarker`, a marker type that represents an
open box. It makes box opening/closing explicit, which makes it much
easier to understand and harder to get wrong.
The commit also removes many comments are on `end` calls saying things
like "end outer head-block", "Close the outer-box". These demonstrate
how confusing the implicit approach was, but aren't necessary any more.
|
|
Correctly display stdout and stderr in case a doctest is failing
Fixes https://github.com/rust-lang/rust/issues/140289.
Since the doctest is actually running itself, we need to handle the output directly inside it.
cc `@fmease`
r? `@notriddle`
|
|
|
|
a function
|
|
|
|
|
|
r=notriddle
Fix impl block items indent
Fixes #139771.
Now, all impl block "before impl block items" indent are the same (ie, item info and documentation).
With this fix, it looks like this:
You can test it [here](https://rustdoc.crud.net/imperio/fix-impl-block-items-indent/foo/struct.Context.html).
r? `@notriddle`
|
|
Stabilize flags for doctest cross compilation
This makes the following changes in preparation for supporting doctest cross-compiling in cargo:
- Renames `--runtool` and `--runtool-arg` to `--test-runtool` and `--test-runtool-arg` to maintain consistency with other `--test-*` arguments.
- Stabilizes the `--test-runtool` and `--test-runtool-arg`. These are needed in order to support cargo's `target.runner` option which specifies a runner to execute a cross-compiled doctest (for example, qemu).
- Stabilizes the `--enable-per-target-ignores` flag by removing it and making it unconditionally enabled. This makes it possible to disable a doctest on a per-target basis, which I think will be helpful for rolling out this feature.
These changes were suggested in https://rust-lang.zulipchat.com/#narrow/channel/266220-t-rustdoc/topic/stabilizing.20doctest.20xcompile/near/409281127
The intent is to stabilize the doctest-xcompile feature in cargo. This will help ensure that for projects that do cross-compile testing that their doctests are also covered. Currently there is a somewhat surprising behavior that they are ignored.
Closes https://github.com/rust-lang/rust/issues/64245
try-job: x86_64-msvc-1
|
|
|
|
|
|
|
|
|
|
|
|
Rollup of 5 pull requests
Successful merges:
- #139981 (Don't compute name of associated item if it's an RPITIT)
- #140077 (Construct OutputType using macro and print [=FILENAME] help info)
- #140081 (Update `libc` to 0.2.172)
- #140094 (Improve diagnostics for pointer arithmetic += and -= (fixes #137391))
- #140128 (Use correct annotation for CSS pseudo elements)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Use correct annotation for CSS pseudo elements
The list of CSS pseudo elements is pretty short so it was easy to go through. Even though the `:` is accepted, it's incorrect.
For a description of CSS pseudo elements: https://developer.mozilla.org/en-US/docs/Web/CSS/Pseudo-elements
r? ``@notriddle``
|
|
Remove `token::{Open,Close}Delim`
By replacing them with `{Open,Close}{Param,Brace,Bracket,Invisible}`.
PR #137902 made `ast::TokenKind` more like `lexer::TokenKind` by
replacing the compound `BinOp{,Eq}(BinOpToken)` variants with fieldless
variants `Plus`, `Minus`, `Star`, etc. This commit does a similar thing
with delimiters. It also makes `ast::TokenKind` more similar to
`parser::TokenType`.
This requires a few new methods:
- `TokenKind::is_{,open_,close_}delim()` replace various kinds of
pattern matches.
- `Delimiter::as_{open,close}_token_kind` are used to convert
`Delimiter` values to `TokenKind`.
Despite these additions, it's a net reduction in lines of code. This is
because e.g. `token::OpenParen` is so much shorter than
`token::OpenDelim(Delimiter::Parenthesis)` that many multi-line forms
reduce to single line forms. And many places where the number of lines
doesn't change are still easier to read, just because the names are
shorter, e.g.:
```
- } else if self.token != token::CloseDelim(Delimiter::Brace) {
+ } else if self.token != token::CloseBrace {
```
r? `@petrochenkov`
|
|
|
|
Fix error when an intra doc link is trying to resolve an empty associated item
Fixes https://github.com/rust-lang/rust/issues/140026.
Assigning ```@nnethercote``` since they're the one who wrote the initial change.
I updated rustdoc code instead of compiler's because I think it makes more sense that the caller ensures on their side that the name they're looking for isn't empty.
r? ```@nnethercote```
|
|
By replacing them with `{Open,Close}{Param,Brace,Bracket,Invisible}`.
PR #137902 made `ast::TokenKind` more like `lexer::TokenKind` by
replacing the compound `BinOp{,Eq}(BinOpToken)` variants with fieldless
variants `Plus`, `Minus`, `Star`, etc. This commit does a similar thing
with delimiters. It also makes `ast::TokenKind` more similar to
`parser::TokenType`.
This requires a few new methods:
- `TokenKind::is_{,open_,close_}delim()` replace various kinds of
pattern matches.
- `Delimiter::as_{open,close}_token_kind` are used to convert
`Delimiter` values to `TokenKind`.
Despite these additions, it's a net reduction in lines of code. This is
because e.g. `token::OpenParen` is so much shorter than
`token::OpenDelim(Delimiter::Parenthesis)` that many multi-line forms
reduce to single line forms. And many places where the number of lines
doesn't change are still easier to read, just because the names are
shorter, e.g.:
```
- } else if self.token != token::CloseDelim(Delimiter::Brace) {
+ } else if self.token != token::CloseBrace {
```
|
|
r=nnethercote
Improve `clean_maybe_renamed_item` function code a bit
Follow-up of #139846.
This is what I tried to say in there: the `name` variable can be unwrapped in most cases so better do it directly once and for all if possible and move the cases where it's not possible above.
r? `@nnethercote`
|
|
|
|
|
|
|
|
rustdoc/clean: Fix lowering of fn params (fixes correctness & HIR vs. middle parity regressions)
**(0)** PR #136411 aimed to stop rendering unnamed params of fn ptr types as underscores in the common case (e.g., `fn(_: i32)` → `fn(i32)`) to make the rendered output stylistically more conventional.
**(0.a)** However, since the cleaning fn that the PR modified is also used for lowering the HIR params of foreign fns and required assoc fns in traits, it accidentally butchered the rendering of the latter two:
```rs
pub trait Trait { fn assoc_fn(_: i32); } // as well as (Rust 2015 only): fn assoc_fn(i32);
unsafe extern "C" { pub fn foreign_fn(_: i32); }
// Since 1.86 the fns above gets mis-rendered as:
pub fn assoc_fn(: i32) // <-- BUTCHERED
pub unsafe extern "C" fn foreign_fn(: i32) // <-- BUTCHERED
```
**(0.b)** Furthermore, it broke parity with middle cleaning (which includes inlined cross-crate re-exports) re-regressing parts of #44306 I once fixed in PR #103885.
**(1)** Lastly, PR #139035 introduced an ICE triggered by the following input file:
```rs
trait Trait { fn anon(()) {} } // internal error: entered unreachable code
```
---
This PR fixes all of these regressions and in the first commit renames several types and fns to be more ~~correct~~ descriptive and legible.
~~It also refactors `Param.name` to be of type `Option<Symbol>` instead `Symbol` (where `None` ~ `kw::Empty`), so rendering mistakes like that can no longer creep in like that (ignoring tests). CC #137978.~~ Independently done in PR #139846 a day prior.
|
|
Remove `name_or_empty`
Another step towards #137978.
r? ``@jdonszelmann``
|
