208548 - Inconsistency in link determination in HTMLStyleSheetImpl

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect, P4)

Description

[Hixie (not reading bugmail)]

In content/html/style/src/nsHTMLStyleSheet.cpp :

HTMLStyleSheetImpl::RulesMatching(...):
814       if (tag == nsHTMLAtoms::a) {
815         if (mLinkRule || mVisitedRule || mActiveRule) {
816           if (aData->mIsHTMLLink) {
...
HTMLStyleSheetImpl::HasStateDependentStyle(...):
871   *aResult = mActiveRule &&
872              (aData->mStateMask & NS_EVENT_STATE_ACTIVE) &&
873              aData->mStyledContent &&
874              aData->mIsHTMLContent &&
875              aData->mContentTag == nsHTMLAtoms::a &&
876              aData->mStyledContent->HasAttr(kNameSpaceID_None,
877                                             nsHTMLAtoms::href);

It seems that the two pieces of code should use the same algorithm to work out
if the element is a link or not. In fact, it seems that they could both be
simplified:

       if (aData->mIsHTMLLink && (mLinkRule || mVisitedRule || mActiveRule)) {

...and:

   *aResult = mActiveRule &&
              (aData->mStateMask & NS_EVENT_STATE_ACTIVE) &&
              aData->mStyledContent &&
              aData->mIsHTMLContent &&
              aData->mIsHTMLLink;

That would have the side effect of also applying the attributes to <area> and
<link> elements which might even be a good idea.

Anyway. Just something I spotted while browsing the code. It could be
intentional of course.

Comment 1

[Hixie (not reading bugmail)]

Actually this would also require changes to
HTMLStyleSheetImpl::HasAttributeDependentStyle() so maybe it's not such a good idea.

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I think |HasStateDependentStyle| could be fixed (although it's not dangerous if
it returns true more often than needed -- just slower).

Comment 3

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch

Comment 4

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 125102 [details] [diff] [review]
patch

r+sr=me

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Fix checked in to trunk, 2003-06-10 15:10 -0700.