Bug 39571 – Teach the HTML5 parser to lex DOCTYPEs

• Component: New Bugs
• Status: RESOLVED FIXED

[Adam Barth]

Teach the HTML5 parser to lex DOCTYPEs

[Adam Barth]

Created attachment 56828 [details]
Work in progress

[Adam Barth]

Created attachment 56836 [details]
more work in progress

[Adam Barth]

Created attachment 56923 [details]
Patch

[WebKit Review Bot]

Attachment 56923 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style', '--no-squash']" exit_code: 1
WebCore/html/HTML5Lexer.cpp:890:  An else if statement should be written as an if statement when the prior "if" concludes with a return, break, continue or goto statement.  [readability/control_flow] [4]
WebCore/html/HTML5Lexer.cpp:1072:  An else if statement should be written as an if statement when the prior "if" concludes with a return, break, continue or goto statement.  [readability/control_flow] [4]
WebCore/html/HTML5Lexer.cpp:1085:  An else if statement should be written as an if statement when the prior "if" concludes with a return, break, continue or goto statement.  [readability/control_flow] [4]
Total errors found: 3 in 8 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Eric Seidel (no email)]

Comment on attachment 56923 [details]
Patch

WebCore/html/HTML5Lexer.cpp:891
 +                      source.advanceAndASSERTIgnoringCase('d');
Seems it would be simple to write a little helper method (static inline) which did:

static inline void advanceAndASSERTIgnoringCase(SegmentedString& string, char* assertionString)
{
    for (int x = 0;; x++) {
        if (!assertionString[x])
             return;
        string.advanceAndASSERTIngoringCase(assertionString[x]);
     }
}

That would read much cleaner in the callsites.

WebCore/html/HTML5TreeBuilder.cpp:100
 +      if (token.type() == HTML5Token::Uninitialized)
This seems wrong.

WebCore/platform/text/SegmentedString.h:154
 +      void advanceAndASSERTIgnoringCase(UChar expectedCharacter)
We may just want to turn this into a static inline in HTML5Lexer.cpp instead.

I think we need one more round.

[Adam Barth]

Created attachment 56932 [details]
Patch

[WebKit Review Bot]

Attachment 56932 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style', '--no-squash']" exit_code: 1
WebCore/html/HTML5Lexer.cpp:900:  An else if statement should be written as an if statement when the prior "if" concludes with a return, break, continue or goto statement.  [readability/control_flow] [4]
WebCore/html/HTML5Lexer.cpp:1077:  An else if statement should be written as an if statement when the prior "if" concludes with a return, break, continue or goto statement.  [readability/control_flow] [4]
WebCore/html/HTML5Lexer.cpp:1085:  An else if statement should be written as an if statement when the prior "if" concludes with a return, break, continue or goto statement.  [readability/control_flow] [4]
Total errors found: 3 in 8 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Eric Seidel (no email)]

Comment on attachment 56932 [details]
Patch

Would be nice if we had a version of this which took a const String& so we didn't have to write "public" twice:

 1078                         advanceStringAndASSERTIgnoringCase(source, "public");


othewrise looks good.

Seems silly to intentionally violate style just to match the spec wording.

[Adam Barth]

> Would be nice if we had a version of this which took a const String& so we didn't have to write "public" twice:

This might be a premature optimization, but I'm worried about all the string null checks...

> Seems silly to intentionally violate style just to match the spec wording.

My hope is that this problem will be sovled once we convert to using goto.

[Adam Barth]

Committed r60087