Skip to content

Fix grammar for block comments#2191

Merged
traviscross merged 2 commits intorust-lang:masterfrom
ehuss:fix-comment-nested
Mar 2, 2026
Merged

Fix grammar for block comments#2191
traviscross merged 2 commits intorust-lang:masterfrom
ehuss:fix-comment-nested

Conversation

@ehuss
Copy link
Contributor

@ehuss ehuss commented Feb 25, 2026

This fixes an issue with the block comment grammar with nested block comments. For example, this would fail to match /*/* test */*/ because of the following:

The open slash of the inner block comment matches ~[* !] which then misinterprets the rest of the inner block comment as being normal characters, and prevents the nesting from working correctly.

The original intent with this formulation was to prevent it from matching an inner or outer block doc comment. This is no longer needed because we are now defining the order with the COMMENT production, and the doc comments come before block comments.

This also changes the order of BLOCK_COMMENT_OR_DOC, but that is not strictly necessary because once inside a block comment, everything is comment text. rustc does not expose nested comments as being distinct. I mainly ordered them to be consistent with COMMENT (and in case I forgot anything).

The implementation for this is at Cursor::block_comment.

@rustbot rustbot added the S-waiting-on-review Status: The marked PR is awaiting review from a maintainer label Feb 25, 2026
src/comments.md Outdated
Comment on lines +20 to +23
`/**/`
| `/***/`
| `/*`
^
( ~[`*` `!`] | `**` | BLOCK_COMMENT_OR_DOC )
( BLOCK_COMMENT_OR_DOC | ~`*/` )*
^ ( BLOCK_COMMENT_OR_DOC | ~`*/` )*
Copy link
Contributor

@traviscross traviscross Mar 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

With the changes, the first two branches aren't needed, right?

Separately, we might as well avoid using ~ here on this multiple character lookahead.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Indeed, it is no longer needed, I have removed it and the use of ~.

This fixes an issue with the block comment grammar with nested block
comments. For example, this would fail to match `/*/* test */*/` because
of the following:

The open slash of the inner block comment matches `~[`*` `!`]` which
then misinterprets the rest of the inner block comment as being normal
characters, and prevents the nesting from working correctly.

The original intent with this formulation was to prevent it from
matching an inner or outer block doc comment. This is no longer needed
because we are now defining the order with the COMMENT production, and
the doc comments come before block comments.

This also changes the order of BLOCK_COMMENT_OR_DOC, but that is not
strictly necessary because once inside a block comment, everything is
comment text. rustc does not expose nested comments as being distinct. I
mainly ordered them to be consistent with COMMENT (and in case I forgot
anything).

The implementation for this is at [`Cursor::block_comment`](https://github.com/rust-lang/rust/blob/859951e3c7c9d0322c39bad49221937455bdffcd/compiler/rustc_lexer/src/lib.rs#L782-L817).
@ehuss ehuss force-pushed the fix-comment-nested branch from 92d669c to 37ed14b Compare March 1, 2026 21:56
In the grammar rule for `BLOCK_COMMENT`, we indent the rule expression
for the content of the block comment.  We had put the cut ahead of
this rule, but the resulting indentation makes it seem as though the
cut is scoped to this subpart when it in fact applies to all remaining
parts of the rule.  The potential confusion here is particularly
subtle because the rule for the content of a block comment is followed
by a Kleene star -- if the cut only applied to this subpart, it
would be superfluous.

Let's move the cut to after the opening `/*` to avoid this.
@traviscross
Copy link
Contributor

Looks good; thanks.

@traviscross traviscross enabled auto-merge March 2, 2026 01:22
@traviscross traviscross added this pull request to the merge queue Mar 2, 2026
Merged via the queue into rust-lang:master with commit 2dfed35 Mar 2, 2026
6 checks passed
@rustbot rustbot removed the S-waiting-on-review Status: The marked PR is awaiting review from a maintainer label Mar 2, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants