It's not obvious why this was ever a separate class name, since even
in c4219a4783 when it was first added,
it was styled identically to regular `fn` links.
`@!has` (and `@!matches`) with two arguments used to treat the second
argument as a literal string of HTML code. Now, that feature has been
renamed into `@!hasraw` (and `@!matchesraw`), and the arity-2 `@!has`
version is an error.
These uses thought the second argument was being treated as an XPath, as
with the arity-3 version, but in fact was being treated as literal HTML.
Because these were checking for the *absence* of the string, the tests
silently did nothing -- an XPath string won't ever be showing up in the
test's generated HTML!
Previously, tidy-html5 (`tidy`) would complain about a few things in our
HTML. The main thing is that `<summary>` tags can't contain `<div>`s.
That's easily fixed by changing out the `<div>`s for `<span>`s with
`display: block`.
However, there's also a rule that `<span>`s can't contain heading
elements. `<span>` permits only "phrasing content"
https://developer.mozilla.org/en-US/docs/Web/HTML/Element/span, and
`<h3>` (and friends) are "Flow content, heading content, palpable
content".
https://developer.mozilla.org/en-US/docs/Web/HTML/Element/Heading_Elements
We have a wrapping `<div>` that goes around each `<h3>`/`<h4>`,
etc. We turn that into a `<section>` rather than a `<span>` because
`<section>` permits "flow content".
https://developer.mozilla.org/en-US/docs/Web/HTML/Element/section
After this change we get only three warnings from tidy, run on
struct.String.html:
line 6 column 10790 - Warning: trimming empty <span>
line 1 column 1118 - Warning: <link> proprietary attribute "disabled"
line 1 column 1193 - Warning: <link> proprietary attribute "disabled"
The empty `<span>` is a known issue - there's a span in front of the
search box to work around a strange Safari issue.
The `<link>` attributes are the non-default stylesheets. We can probably
refactor theme application to avoid using this proprietary "disabled"
attribute.
Instead of using a depth counter and adding "../" to get to the top,
this commit makes rustdoc actually compare the path of what it's
linking from to the path that it's linking to. This makes the resulting
HTML shorter.
Here's a comparison of one of the largest (non-source) files in the
Rust standard library docs (about 4% improvement before gzipping).
$ wc -c struct.Wrapping.old.html struct.Wrapping.new.html
2387389 struct.Wrapping.old.html
2298538 struct.Wrapping.new.html
Most if it can be efficiently gzipped away.
$ wc -c struct.Wrapping.old.html.gz struct.Wrapping.new.html.gz
70679 struct.Wrapping.old.html.gz
70050 struct.Wrapping.new.html.gz
But it also makes a difference in the final DOM size, reducing it from 91MiB to 82MiB.
