279806 - kNameSpaceID_None casts to pointer

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Boris Zbarsky [:bzbarsky]]

nodeinfo->Equals(kNameSpaceID_None, nsHTMLAtoms::foo) compiles, becase the
namespace id casts to nsIAtom* and we call the (localname, prefix) version of
the function.

sicking and I poked at this and found a change that keeps at least gcc 3.2 from
casting the namespace to nsIAtom*.  Thoughts on checking this in?

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: Like this, say...

Comment 2

[Johnny Stenback (:jst)]

Comment on attachment 172385 [details] [diff] [review]
Like this, say...

Sure, but I wonder if we should just bump the namespace ID's by one and just
not use 0 at all in stead? Unless we think we have code that checks
!namespaceID...

Comment 3

[Boris Zbarsky [:bzbarsky]]

That's also an option (and I'm willing to bet that we don't have code like
that)....  Either way.

Comment 4

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

I sort of prefer the approach in the patch, static consts are supposed to be the
c++ replacement for this type of macro anyway (I'm told). Also, there might be
plugins/embedders that are using nsIContent, but I guess they would break on a
regular basis anyway.

Comment 5

[Boris Zbarsky [:bzbarsky]]

I think I prefer not changing the constant values.

Should I change the others to static const PRInt32 as well, or just check this
in as-is?

Comment 6

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Either way is fine with me :)

It might be a good idea to include a pointer to this bug in the comment btw.
Though i guess you can always get that info from bonsai.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Checked in that patch as-is.

This bug has no more info than the comment in the patch really, so....