Bug 174469 – -Wimplicit-fallthrough warning in ComputedStyleExtractor::propertyValue

• Component: CSS
• Status: RESOLVED FIXED

[Michael Catanzaro]

This new warning from GCC 7 has found a problem in ComputedStyleExtractor::propertyValue:

[2701/5861] Building CXX object Source.../css/CSSComputedStyleDeclaration.cpp.o
../../Source/WebCore/css/CSSComputedStyleDeclaration.cpp: In member function ‘WTF::RefPtr<WebCore::CSSValue> WebCore::ComputedStyleExtractor::propertyValue(WebCore::CSSPropertyID, WebCore::EUpdateLayout)’:
../../Source/WebCore/css/CSSComputedStyleDeclaration.cpp:3344:13: warning: this statement may fall through [-Wimplicit-fallthrough=]
             }
             ^
../../Source/WebCore/css/CSSComputedStyleDeclaration.cpp:3345:9: note: here
         case CSSPropertyTextIndent: {
         ^~~~

Looking at the code, the fallthrough there appears to almost surely be undesired. But I have no clue how to fix it.

[Radar WebKit Bug Importer]

<rdar://problem/33311638>

[Michael Catanzaro]

Anyone know how to fix this one?

[Simon Fraser (smfr)]

I don't think this is harmful. The switch in the previous block returns under every condition.

[Michael Catanzaro]

Um, so it does. I'm surprised I missed that. Let's just silence the warning, then.

[Michael Catanzaro]

Created attachment 317452 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 317452 [details]
Patch

Or maybe just return nullptr?

[Michael Catanzaro]

(In reply to Simon Fraser (smfr) from comment #6)
> Comment on attachment 317452 [details]
> Patch
> 
> Or maybe just return nullptr?

I assume that if we don't want that code to ever be hit, an assert is most appropriate?

[Simon Fraser (smfr)]

Comment on attachment 317452 [details]
Patch

Sure, i guess.

[WebKit Commit Bot]

Comment on attachment 317452 [details]
Patch

Clearing flags on attachment: 317452

Committed r220354: <http://trac.webkit.org/changeset/220354>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.