Bug 152983 – Picture element needs to work properly with the preload scanner

• Component: Images
• Status: RESOLVED FIXED

[Dave Hyatt]

The preloader needs to do source selection properly on picture elements.

[Dave Hyatt]

Created attachment 268703 [details]
Patch

[WebKit Commit Bot]

Attachment 268703 [details] did not pass style-queue:


ERROR: Source/WebCore/html/parser/HTMLPreloadScanner.cpp:124:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
ERROR: Source/WebCore/html/parser/HTMLPreloadScanner.cpp:177:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
Total errors found: 2 in 14 files


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

[Dave Hyatt]

Created attachment 268704 [details]
Patch

[WebKit Commit Bot]

Attachment 268704 [details] did not pass style-queue:


ERROR: Source/WebCore/html/parser/HTMLPreloadScanner.cpp:124:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
ERROR: Source/WebCore/html/parser/HTMLPreloadScanner.cpp:177:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
Total errors found: 2 in 14 files


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

[Dean Jackson]

Comment on attachment 268704 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=268704&action=review

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:124
> +        if (m_tagId == TagId::Source && !pictureState.isEmpty() && pictureState.last() == false && m_mediaMatched && !m_srcSetAttribute.isEmpty()) {

Why not !pictureState.last() ?

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:177
> +        bool alreadyMatchedSource = inPicture && pictureState.last() == true;

Same here.

[Chris Dumez]

Comment on attachment 268704 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=268704&action=review

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:128
> +                pictureState.removeLast();

I believe we could replace these 2 lines with: pictureState.last() = true;
and avoid any vector resizing.

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:174
> +    void processAttribute(const AtomicString& attributeName, const String& attributeValue, Document& document, Vector<bool> pictureState)

const Vector<bool>& ?

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:210
> +                RefPtr<MediaQuerySet> mediaSet = MediaQuerySet::createAllowingDescriptionSyntax(attributeValue);

Ref<>

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:213
> +                m_mediaMatched = evaluator.evalCheckingViewportDependentResults(mediaSet.get(), viewportDependentMediaQueryResults);

.ptr() instead of .get() if you switch to Ref<>.

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:326
> +    bool m_mediaMatched { true };

Would be clearer with a 'is' / 'was' prefix, as per coding style.

> Source/WebCore/html/parser/HTMLPreloadScanner.cpp:364
>              m_inStyle = false;

We probably want to add an early return after this to avoid running your new check.

[Dave Hyatt]

Fixed in r194865.