This more concrete wrapper type can write a prefix the very first time something
is written to it. This allows removing plenty of useless monomorphisations caused
by the former W/SequenceWriter<W> pair of types.
Fixes warnings from rust-lang/rust#44229 when `--enable-commonmark` is
passed to rustdoc.
This is mostly a global find-and-replace for bare URIs on lines by
themselves in doc comments.
… rather than the start location of the current construct.
This likely places the error just *after* of the unexpected token
whereas before would be best, but that’s likely a much bigger change.
See https://bugzilla.mozilla.org/show_bug.cgi?id=1378861
Devirtualize CSS error reporting.
This removes a trait object from the path of reporting a CSS error.
---
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] There are tests for these changes
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/18209)
<!-- Reviewable:end -->
Passing SourceLocation into constructor instead of assigning
immediately after construction cleans up the code and helps to
prevent leaving an invalid SourceLocation in the future.
A future patch series has some values that should be separated by spaces. This
allows us to re-use the code for serialization, but the types do get a little
clunky. The separator is now indicated with an associated type.
Fix#16165.
Also, it turns out that the CSSFontFaceRule IDL specified in the
css-fonts spec is not web-compatible.
Instead browsers implement a .style attribute like in CSSStyleRule:
https://github.com/w3c/csswg-drafts/issues/825
This in turn requires preserving data about which descriptors
were set or not (distinguishing unset from set to a value that happens
to be the initial value),
so this commit also makes every field `Option<_>`.
