I can wait because it is automated. So after you check in, I could easily pull
a new tree and apply the changes.
Also, for some reason I was having problems with the DOM interfaces, and now I
am not. I now have those fixed in my tree.
Eric P. found some other items I was skipping and I have fixed those also.
(nsIFrame and <class name>::GetIID())
So let me know when you have checked in and it is clear for pulling a new tree
and fixing them.
Rod
Johnny Stenback wrote:
> Rod Spears wrote:
>
> > I did some analysis on WinNT and discovered no increase in size when
> > using the macro NS_GET_IID (which uses templates) over statically
> > defining the IID using the macro NS_DEFINE_IID. And as you might expect,
> > getting rid of statically defined IIDs does save some space.
> >
> > The size analysis was done with a non-debug build.
> >
> > I have a small app that has scrubbed all the source in layout and
> > removed all NS_DEFINE_IID macros where they could be replaced by
> > directly inserting NS_GET_IID(<class name>).
> >
>
> If the app you wrote is fairly easy to run over the whole layout
> directory then I'd like to ask you to not check this in yet since I have
> a huge amount of changes (content model related) in my tree that are
> getting close to being complete and doing this IID change would most
> likely cause conflicts in my tree.
>
> I'm all for doing this change but if it's easy enough to do later (maybe
> in a couple of weeks) then please hold off untill my changes are in.
>
> Hmm, I had a look at the patch and it seems like your app doesn't get
> rid of static IID's for DOM interfaces (and some other random interfaces
> too), why is that?
>
> > This is good for these reasons:
> >
> > 1) More code now uses a more standard calling convention
> > 2) Unneeded NS_DEFINE_IID have been removed
> > 3) Some size reduction
> >
> > I have filed a bug:
> > http://bugzilla.mozilla.org/show_bug.cgi?id=61274
> >
> > The diffs can be viewed with NS 6 at:
> > http://www.cableone.net/rnspears/getiid/index.html
> >
> >
> > Rod
>
> --
> jst
S/MIME Cryptographic Signature
