`rustdoc`: compute correct line number for indented rust code blocks.
This PR fixes a bug in `rustdoc` where it computes the wrong line number for indented rust code blocks (and subsequent blocks) it finds in markdown strings. To fix this issue, we decrement the line number if we find characters between the code block and the preceding line ending. I noticed this issue as I was trying to use `rustdoc` to extract examples from The Rust Reference and run them through the [Rust Model Checker](https://github.com/model-checking/rmc).
I noticed that `Pin::new()`'s search summary looked off, and I realized
that the reason is that it has inline code containing `Pin<P>`, which is
not escaped and thus renders as a paragraph tag!
I don't think the boxing helped performance, in fact I think it
potentially made it worse. The data was still being copied, but now it
was through a pointer. Thinking about it more, I think boxing might only
help when you're passing a big object around by value all the time,
rather than the slowdown being that you're cloning it.
There was no need to clone `id_map` because it was reset before each
item was rendered. `deref_id_map` was not reset, but it was keyed by
`DefId` and thus was unlikely to have collisions (at least for now).
Now we just clone the fields that need to be cloned, and instead create
fresh versions of the others.
[librustdoc] Only split lang string on `,`, ` `, and `\t`
Split markdown lang strings into tokens on `,`.
The previous behavior was to split lang strings into tokens on any
character that wasn't a `_`, `_`, or alphanumeric.
This is a potentially breaking change, so please scrutinize! See discussion in #78344.
I noticed some test cases that made me wonder if there might have been some reason for the original behavior:
```
t("{.no_run .example}", false, true, Ignore::None, true, false, false, false, v(), None);
t("{.sh .should_panic}", true, false, Ignore::None, false, false, false, false, v(), None);
t("{.example .rust}", false, false, Ignore::None, true, false, false, false, v(), None);
t("{.test_harness .rust}", false, false, Ignore::None, true, true, false, false, v(), None);
```
It seemed pretty peculiar to specifically test lang strings in braces, with all the tokens prefixed by `.`.
I did some digging, and it looks like the test cases were added way back in [this commit from 2014](3fef7a74ca) by `@skade.`
It looks like they were added just to make sure that the splitting was permissive, and aren't testing that those strings in particular are accepted.
Closes https://github.com/rust-lang/rust/issues/78344.
Before:
The Rust Standard LibraryThe Rust Standard Library is the foundation of portable Rust software, a set of minimal and battle-tested shared abstractions for the broader Rust ecosystem. It offers core types, like `Vec<T>` and `Option<T>`, library-defined operations on language primitives, standard macros, I/O and multithreading, among many other things.
After:
The Rust Standard Library
Only split doctest lang strings on `,`, ` `, and `\t`. Additionally, to
preserve backwards compatibility with pandoc-style langstrings, strip a
surrounding `{}`, and remove leading `.`s from each token.
Prior to this change, doctest lang strings were split on all
non-alphanumeric characters except `-` or `_`, which limited future
extensions to doctest lang string tokens, for example using `=` for
key-value tokens.
This is a breaking change, although it is not expected to be disruptive,
because lang strings using separators other than `,` and ` ` are not
very common
Previously Markdown documentation was not rendered to HTML for search results,
which led to the output not being very readable, particularly for inline code.
This PR fixes that by rendering Markdown to HTML with the help of pulldown-cmark
(the library rustdoc uses to parse Markdown for the main text of documentation).
However, the text for the title attribute (the text shown when you hover over an
element) still uses the plain-text rendering since it is displayed in browsers
as plain-text.
Only these styles will be rendered; everything else is stripped away:
* *italics*
* **bold**
* `inline code`
rustdoc: do not use plain summary for trait impls
Fixes#38386.
Fixes#48332.
Fixes#49430.
Fixes#62741.
Fixes#73474.
Unfortunately this is not quite ready to go because the newly-working links trigger a bunch of linkcheck failures. The failures are tough to fix because the links are resolved relative to the implementor, which could be anywhere in the module hierarchy.
(In the current docs, these links end up rendering as uninterpreted markdown syntax, so I don't think these failures are any worse than the status quo. It might be acceptable to just add them to the linkchecker whitelist.)
Ideally this could be fixed with intra-doc links ~~but it isn't working for me: I am currently investigating if it's possible to solve it this way.~~ Opened #73829.
EDIT: This is now ready!
