Bug 47735 – Remove some functions from class Node and make a few others non-virtual

• Component: DOM
• Status: RESOLVED FIXED

[Darin Adler]

Remove some functions from class Node and make a few others non-virtual

[Darin Adler]

Created attachment 70919 [details]
Patch

[Darin Adler]

Reported bug 47755 about the false positive style violation found in this patch.

[WebKit Review Bot]

Attachment 70919 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style']" exit_code: 1
WebCore/dom/Node.h:38:  Code inside a namespace should not be indented.  [whitespace/indent] [4]
Total errors found: 1 in 9 files


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

[WebKit Review Bot]

Attachment 70919 [details] did not build on chromium:
Build output: http://queues.webkit.org/results/4478010

[Alexey Proskuryakov]

Comment on attachment 70919 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=70919&action=review

r=me

> WebCore/ChangeLog:62
> +        (WebCore::enclosingInline): Moved function here from the Node class. Also
> +        fixed an incorrect cast this function did, which assumed the function
> +        result was always an element, which the function does not guarantee.

Does this change observable behavior?

> WebCore/dom/Node.h:370
> +    // see @ref traversePreviousNode()

Is there a reason to keep a part of this comment in Javadoc style?

[Dimitri Glazkov (Google)]

Note the cr-linux redness. I think it's because of the FloatPoint.h include removed.

[Darin Adler]

(In reply to comment #6)
> Note the cr-linux redness. I think it's because of the FloatPoint.h include removed.

I’ll add an include of MathExtras.h to HistoryItem.cpp to pick up the ceil function.

[Darin Adler]

(In reply to comment #5)
> > WebCore/ChangeLog:62
> > +        (WebCore::enclosingInline): Moved function here from the Node class. Also
> > +        fixed an incorrect cast this function did, which assumed the function
> > +        result was always an element, which the function does not guarantee.
> 
> Does this change observable behavior?

I don’t think the incorrect cast had yet led to anything that would create incorrect code. I thought of it as a “ticking time bomb”, though.

> > WebCore/dom/Node.h:370
> > +    // see @ref traversePreviousNode()
> 
> Is there a reason to keep a part of this comment in Javadoc style?

I guess I could have gone further and tried to wipe out the half-baked attempt to use Javadoc style. I wasn’t sure whether I should.

[Darin Adler]

Created attachment 71063 [details]
Already-reviewed patch: Posted because it’s the only way to get the Early Warning System to test it

[WebKit Review Bot]

Attachment 71063 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style']" exit_code: 1
WebCore/dom/Node.h:38:  Code inside a namespace should not be indented.  [whitespace/indent] [4]
Total errors found: 1 in 10 files


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

[Alexey Proskuryakov]

> I guess I could have gone further and tried to wipe out the half-baked attempt to use Javadoc style. I wasn’t sure whether I should.

It seems strange to keep them in comments that were modified (this may be the only one, in fact).

[Darin Adler]

(In reply to comment #11)
> > I guess I could have gone further and tried to wipe out the half-baked attempt to use Javadoc style. I wasn’t sure whether I should.
> 
> It seems strange to keep them in comments that were modified (this may be the only one, in fact).

Fixed now locally (although not in the patch I uploaded for the early warning system).

[Darin Adler]

Committed r69987: <http://trac.webkit.org/changeset/69987>

[Lucas Forschler]

Mass moving XML DOM bugs to the "DOM" Component.