rustdoc: Render [src] links for trait implementors

[vi]

Should close #43893.

No tests [yet].

r? @QuietMisdreavus

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @QuietMisdreavus (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[shepmaster]

No tests [yet].

Do you expect there to be tests in this PR? I'm going to assume "yes" and mark this as effectively a WIP.

[QuietMisdreavus]

The last time i added [src] links i didn't add tests for them, but i have discovered the test src/test/rustdoc/src-links.rs that ensures these are present for full-file links, or other tests like issue-34274.rs that checks this for specific place. If you want to set up a test for this you can use that latter file as a base, and i can walk you through it. (Personally i wouldn't gate the PR on them, but since other [src] links are tested it would be good to cover that.)

The PR as-is seems to put the links in a place where the CSS isn't expecting to style them. This leaves them looking weird when adjacent to other src links, like here on Iterator:

I'd like to get with @GuillaumeGomez to double-check what the best CSS change for this is, but I'm pretty sure that's all it will take to get these right.

[vi]

No tests [yet].

Do you expect there to be tests in this PR? I'm going to assume "yes" and mark this as effectively a WIP.

If this is required. [yet] means "if needed per some rues there will also be a test, but if the change is too trivial and/or the test environment is too tricky to setup and/or there is some impending change devaluing existing tests then not".

[QuietMisdreavus]

Thanks for writing that test! The last thing i wanted to make sure this had was the CSS change we discussed on Gitter yesterday. @GuillaumeGomez suggested adding the following block to src/librustdoc/html/static/rustdoc.css:

ul.item-list > li > .out-of-band {
    font-size: 19px;
}

And then changing the part that sets the font-family for all the sans-serif text to add ul.item-list > li > .out-of-band to its list of targets. These changes will make sure the [src] links look like all the others, since the existing CSS missed these new ones.

[GuillaumeGomez]

I don't even see any css update. Did you forgot to push the file?

[vi]

Saw the messages, implemented the CSS change.

[vi]

Means "why not spaces"? Because of other indenting whitespace in this file is tabs.

Continued list of targets which was split into mutiple lines is additionally indented to avoid blending with the rules part which is singly indented.

[vi]

Just put it on the next line and it's fine, no?

Mainly because of I didn't notice a precedent of multi-line target line nearby and just did as I usually do.

Fixed and re-pushed.

the need of the alignment

To allow assumption that each non-indented line is some section (oneline or multiline), not a continuation of something previous. But since non-indented style is already used elsewhere in the file, it's better to just follow it.

[GuillaumeGomez]

Please remove all unneeded empty lines in here (like in the impl blocks).

[vi]

Which empty lines? Before and after the // ignore-corss-compile? They are just retained from issue-34274.rs, which served as a template. I don't think they should be removed (especially the first empty line).

like in the impl blocks

Means "and likewise remove them in impl blocks" or "like you did in impl blocks"?

[GuillaumeGomez]

Like this:

// Copyright 2016 The Rust Project Developers. See the COPYRIGHT
// file at the top-level directory of this distribution and at
// http://rust-lang.org/COPYRIGHT.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.

// ignore-cross-compile

#![crate_name = "foo"]

pub trait SomeTrait {}
pub struct SomeStruct;

// @has foo/trait.SomeTrait.html '//a/@href' '../src/foo/issue-43893.rs.html#23-26'
impl SomeTrait for usize {}

// @has foo/trait.SomeTrait.html '//a/@href' '../src/foo/issue-43893.rs.html#29-32'
impl SomeTrait for SomeStruct {}

[vi]

impl SomeTrait for usize {}

{}

Other [src] link test (issue-34274.rs) tests for single line link. I thought it would be useful to also provide multiple line implementation, so that the link contains line range, not a single line.

[QuietMisdreavus]

Having multi-line tests is useful to make sure that the link points to the full span.

[vi]

That's why I contracted one impl to one line, but left the other impl as three lines in the update.

[vi]

@GuillaumeGomez, Done, but one impl is intentionally left as 3 lines instead of just {}.

[GuillaumeGomez]

Ok, then it's all good now. Thanks!

@bors: r+ rollup

[bors]

📌 Commit acef039 has been approved by GuillaumeGomez

[bors]

⌛ Testing commit acef039 with merge 61bad30...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: GuillaumeGomez
Pushing 61bad30 to master...