Use syn and quote crates for the match_token! macro…
#217
Conversation
| @@ -33,7 +33,7 @@ harness = false | |||
| [features] | |||
| unstable = ["tendril/unstable", "string_cache/unstable"] | |||
| heap_size = ["heapsize", "heapsize_plugin"] | |||
| codegen = ["html5ever_macros"] | |||
| codegen = [] | |||
There was a problem hiding this comment.
This feature is now unused, but removing it would be a breaking change. So let’s keep it until we have some other reason to make a breaking change.
There was a problem hiding this comment.
Can you add a comment with that in the file so we don't accidentally bump the version without removing the feature?
| @@ -7,7 +7,7 @@ | |||
| // option. This file may not be copied, modified, or distributed | |||
| // except according to those terms. | |||
|
|
|||
| //! The tree builder rules, as a single, enormous nested match expression. | |||
There was a problem hiding this comment.
include! doesn’t like top-level attributes.
| @@ -770,7 +770,7 @@ impl<Handle, Sink> TreeBuilderStep | |||
| Done | |||
| } | |||
|
|
|||
| // FIXME: This should be unreachable, but match_token! requires a | |||
| // FIXME: This should be unreachable, but match_token requires a | |||
There was a problem hiding this comment.
Yes, the parser is that stupid. It would find match_token! in a comment and try to parse it as a macro invocation.
There was a problem hiding this comment.
Note: this comment referred to my hand-written parser. syn doesn’t have this problem.
| } | ||
|
|
||
| pub fn expand_match_tokens(from: &Path, to: &Path) { | ||
| // use std::fmt::Write; |
There was a problem hiding this comment.
Do you plan to keep these comments for any specific reasons? Or is it fine to remove them?
There was a problem hiding this comment.
They were temporary, for debugging. I forgot to remove them before committing. Removed now.
Not directly, but there are many tree builder tests that exercise the generated code. |
|
How easy will this be the modify? |
|
syn 0.9.0 is out with many improvements, could the hand-written parser use it? I'm really wary of merging something that can also end up transforming comments. |
|
@nox: I'm a newbie when it comes to procedural macros, but those notes don't show me anything that would make full procedural macros parsable by Did you mean we could parse tags as: match mode {
InitialMode =>
#[derive(tags)]
match token {
CharacterTokens(NotSplit, text) => SplitWhitespace(text),
...
token => {
if !self.opts.iframe_srcdoc {
self.unexpected(&token);
self.set_quirks_mode(Quirks);
}
Reprocess(BeforeHtml, token)
}
}
} |
|
☔ The latest upstream changes (presumably #218) made this pull request unmergeable. Please resolve the merge conflicts. |
SimonSapin
changed the title
match_token! macro…match_token! macro…
|
Alright! I pushed a rewrite that uses the r? @nox I’ll rebase and fix conflicts tomorrow. |
…macro… … instead of the parser and quasi-quoting from Rust’s (unstable) libsyntax. This has significantly worse diagnostics when encountering unexpected syntax (e.g. no indication of which line has the offending code) but this removes all usage of unstable compiler internals that constantly need to be fixed when updating the compiler. Fixes #216.
SimonSapin
force-pushed
the
diy-macros
branch
from
548a0d5 to
63bf17f
Compare
SimonSapin
force-pushed
the
diy-macros
branch
from
63bf17f to
dd9f2cc
Compare
|
@bors-servo r+ |
|
📌 Commit dd9f2cc has been approved by |
Use syn and quote crates for the `match_token!` macro… … instead of the parser and quasi-quoting from Rust’s (unstable) libsyntax. This has significantly worse diagnostics when encountering unexpected syntax (e.g. no indication of which line has the offending code) but this removes all usage of unstable compiler internals that constantly need to be fixed when updating the compiler. Fixes #216. r? @nox <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/html5ever/217) <!-- Reviewable:end -->
|
☀️ Test successful - status-travis |
Token trees are much simpler than an AST. This makes then easier to work with when they are sufficent for the task at hand, such as for example [expanding a procedural macro]( servo/html5ever#217).
Token trees are much simpler than an AST. This makes them easier to work with when they are sufficient for the task at hand, such as for example [expanding a procedural macro]( servo/html5ever#217).
Token trees are much simpler than an AST. This makes them easier to work with when they are sufficient for the task at hand, such as for example [expanding a procedural macro]( servo/html5ever#217).
… instead of the parser and quasi-quoting from Rust’s (unstable) libsyntax.
This has significantly worse diagnostics when encountering unexpected syntax (e.g. no indication of which line has the offending code) but this removes all usage of unstable compiler internals that constantly need to be fixed when updating the compiler.
Fixes #216.
r? @nox
This change is