[AdamMagued]

Fixes #446.

Problem and Root Cause

The protobuf example in _examples/protobuf did not define a custom lexer and instead relied on Participle default lexer (TextScannerLexer). Go underlying text/scanner.Scanner defaults to GoTokens, which sets scanner.SkipComments. This discards all // and /* ... */ comments at the scanner level before tokens reach Participle. Consequently, attempts to preserve, capture, or match comments fail because no comment tokens exist in the token stream.

Solution

1. Define protoLexer using lexer.MustSimple to tokenize protobuf source including line comments (//[^\r\n]), block comments (/ ... */), whitespace, identifiers, strings, floats, ints, and punctuation.
2. Configure the protobuf parser with participle.Lexer(protoLexer) and participle.Elide("Comment", "Whitespace"). This enables standard Participle comment handling: comments throughout the file are cleanly elided by default, while allowing explicit capture on AST nodes (such as Entry).
3. Add Comment string @Comment to Entry so entry-level comments can be preserved in the AST when desired.
4. Add regression test TestComments in _examples/protobuf/main_test.go verifying that line comments, block comments, and inline comments parse properly without syntax errors and entry comments are captured into the AST.

Verification

• Run go test -v ./... in _examples/protobuf (both TestExe and TestComments pass).
• Run go test ./... and cd _examples && go test ./... (all 20 example packages pass).
• Verified gofmt -l is clean on modified files.

[alecthomas]

Thanks for the PR, but that issue should have been closed - the current example is working as intended.