Remove the first rule if it’s @charset in "parse a stylesheet"

Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 
 name = "cssparser"
-version = "0.5.7"
+version = "0.5.8"
 authors = [ "Simon Sapin <[email protected]>" ]
 
 description = "Rust implementation of CSS Syntax Level 3"

src/css-parsing-tests/rule_list.json

@@ -6,6 +6,11 @@
 
 "@foo", [["at-rule", "foo", [], null]],
 
+"@charset; @foo", [
+    ["at-rule", "charset", [], null],
+    ["at-rule", "foo", [], null]
+],
+
 "@foo bar; \t/* comment */", [["at-rule", "foo", [" ", ["ident", "bar"]], null]],
 
 " /**/ @foo bar{[(4", [["at-rule", "foo",

src/css-parsing-tests/stylesheet.json

@@ -6,6 +6,13 @@
 
 "@foo", [["at-rule", "foo", [], null]],
 
+"@charset 4 {} @foo", [["at-rule", "foo", [], null]],
+
+"@foo; @charset 4 {}", [
+    ["at-rule", "foo", [], null],
+    ["at-rule", "charset", [" ", ["number", "4", 4, "integer"], " "], []]
+],
+
 "@foo bar; \t/* comment */", [["at-rule", "foo", [" ", ["ident", "bar"]], null]],
 
 " /**/ @foo bar{[(4", [["at-rule", "foo",

src/css-parsing-tests/stylesheet_bytes.json

@@ -41,32 +41,27 @@
 
 
 {"css_bytes": "@charset \"ISO-8859-5\"; @\u00E9"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "щ", [], null]],
+[[["at-rule", "щ", [], null]],
  "iso-8859-5"],
 
 {"css_bytes": "@Charset \"ISO-8859-5\"; @\u00E9",
  "comment": "@charset has to match an exact byte pattern"},
-[[["at-rule", "Charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "�", [], null]],
+[[["at-rule", "�", [], null]],
  "utf-8"],
 
 {"css_bytes": "@charset  \"ISO-8859-5\"; @\u00E9",
  "comment": "@charset has to match an exact byte pattern"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "�", [], null]],
+[[["at-rule", "�", [], null]],
  "utf-8"],
 
 {"css_bytes": "@charset 'ISO-8859-5'; @\u00E9",
  "comment": "@charset has to match an exact byte pattern"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "�", [], null]],
+[[["at-rule", "�", [], null]],
  "utf-8"],
 
 {"css_bytes": "@charset \"ISO-8859-5\" ; @\u00E9",
  "comment": "@charset has to match an exact byte pattern"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"], " "], null],
-  ["at-rule", "�", [], null]],
+[[["at-rule", "�", [], null]],
  "utf-8"],
 
 
@@ -79,35 +74,30 @@
 
 {"css_bytes": "@charset \"UTF-16LE\"; @\u00C3\u00A9",
  "comment": "@charset can only specify ASCII-compatible encodings"},
-[[["at-rule", "charset", [" ", ["string", "UTF-16LE"]], null],
-  ["at-rule", "é", [], null]],
+[[["at-rule", "é", [], null]],
  "utf-8"],
 
 
 {"css_bytes": "\u00EF\u00BB\u00BF @charset \"ISO-8859-5\"; @\u00E9",
  "comment": "BOM takes precedence over @charset"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "�", [], null]],
+[[["at-rule", "�", [], null]],
  "utf-8"],
 
 {"css_bytes": "\u00EF\u00BB\u00BF @charset \"ISO-8859-5\"; @\u00C3\u00A9",
  "comment": "BOM takes precedence over @charset"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "é", [], null]],
+[[["at-rule", "é", [], null]],
  "utf-8"],
 
 {"css_bytes": "@charset \"ISO-8859-5\"; @\u00E9",
  "protocol_encoding": " Iso-8859-2",
  "comment": "Protocol takes precedence over @charset"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "é", [], null]],
+[[["at-rule", "é", [], null]],
  "iso-8859-2"],
 
 {"css_bytes": "@charset \"ISO-8859-5\"; @\u00E9",
  "protocol_encoding": "kamoulox",
  "comment": "Unknow protocol encoding falls back to @charset"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "щ", [], null]],
+[[["at-rule", "щ", [], null]],
  "iso-8859-5"],
 
 
@@ -120,15 +110,13 @@
 {"css_bytes": "@charset \"ISO-8859-5\"; @\u00E9",
  "environment_encoding": "ISO-8859-2",
  "comment": "@character takes precedence over environment"},
-[[["at-rule", "charset", [" ", ["string", "ISO-8859-5"]], null],
-  ["at-rule", "щ", [], null]],
+[[["at-rule", "щ", [], null]],
  "iso-8859-5"],
 
 {"css_bytes": "@charset \"kamoulox\"; @\u00E9",
  "environment_encoding": "ISO-8859-2",
  "comment": "@character with unknown encoding falls back to environment encoding"},
-[[["at-rule", "charset", [" ", ["string", "kamoulox"]], null],
-  ["at-rule", "é", [], null]],
+[[["at-rule", "é", [], null]],
  "iso-8859-2"],
 
 {"css_bytes": "@\u00E9",

src/parser.rs

@@ -471,6 +471,9 @@ impl<'i, 't> Parser<'i, 't> {
         if next_byte.is_some() && !self.stop_before.contains(Delimiters::from_byte(next_byte)) {
             debug_assert!(delimiters.contains(Delimiters::from_byte(next_byte)));
             self.tokenizer.advance(1);
+            if next_byte == Some(b'{') {
+                consume_until_end_of_block(BlockType::CurlyBracket, &mut *self.tokenizer);
+            }
         }
         result
     }

src/rules_and_declarations.rs

@@ -4,6 +4,7 @@
 
 // https://drafts.csswg.org/css-syntax/#parsing
 
+use std::ascii::AsciiExt;
 use std::ops::Range;
 use std::borrow::Cow;
 use super::{Token, Parser, Delimiter, SourcePosition};
@@ -264,6 +265,7 @@ where P: QualifiedRuleParser<QualifiedRule = R> + AtRuleParser<AtRule = R> {
     pub parser: P,
 
     is_stylesheet: bool,
+    any_rule_so_far: bool,
 }
 
 
@@ -285,6 +287,7 @@ where P: QualifiedRuleParser<QualifiedRule = R> + AtRuleParser<AtRule = R> {
             input: input,
             parser: parser,
             is_stylesheet: true,
+            any_rule_so_far: false,
         }
     }
 
@@ -300,6 +303,7 @@ where P: QualifiedRuleParser<QualifiedRule = R> + AtRuleParser<AtRule = R> {
             input: input,
             parser: parser,
             is_stylesheet: false,
+            any_rule_so_far: false,
         }
     }
 }
@@ -318,9 +322,17 @@ where P: QualifiedRuleParser<QualifiedRule = R> + AtRuleParser<AtRule = R> {
                 Ok(Token::WhiteSpace(_)) | Ok(Token::Comment(_)) => {}
                 Ok(Token::CDO) | Ok(Token::CDC) if self.is_stylesheet => {}
                 Ok(Token::AtKeyword(name)) => {
-                    return Some(parse_at_rule(start_position, name, self.input, &mut self.parser))
+                    let first_stylesheet_rule = self.is_stylesheet && !self.any_rule_so_far;
+                    self.any_rule_so_far = true;
+                    if first_stylesheet_rule && name.eq_ignore_ascii_case("charset") {
+                        let delimiters = Delimiter::Semicolon | Delimiter::CurlyBracketBlock;
+                        let _ = self.input.parse_until_after(delimiters, |_input| Ok(()));
+                    } else {
+                        return Some(parse_at_rule(start_position, name, self.input, &mut self.parser))
+                    }
                 }
                 Ok(_) => {
+                    self.any_rule_so_far = true;
                     self.input.reset(start_position);
                     return Some(parse_qualified_rule(self.input, &mut self.parser)
                                 .map_err(|()| start_position..self.input.position()))