Re: open letter about development proposals

[stefan]

> On 2003-08-05 07:10:46 +0000 stefan@wms-network.de wrote:
>
>> Btw: does "Apple's extensive ObjC
>> changes being merged back" include objc++?
>
> Not for the moment; but they are planning to include ObjC++. With a bit
> of luck it will be for gcc 3.4, but that's not sure.

can you estimate how likely it is that these changes are included in 3.4.
any ideas when 3.4 will become available. I can't find a date on the gcc
page.

>
>> However, i think porting gecko
>> can be very hard especially if you take a look at the sources (no, i do
>> not flame c++ but this stuff is really ugly from my point of view).
>

>>> Many consider KHTML/bindings (as used by Apple's Safari, and recently
>>> by
>>> Omni) the better implementation. If/when ObjC++ is available via gcc,
>>> porting KHTML to GNUstep would probably be even easier.
>>
>> Thats what i think too. I already made some (rundimentary)
>> investigations
>> about porting the KWQ-Layer to GNUstep without using objc++ (don't know
>> when this becomes available). I think it's not too difficult to rewrite
>> this layer using pure C to communicate with GNUstep. Is anyone
>> interestet
>> in participating? I think it's too much for one person (or it will take
>> a
>> long time).
>
> Well, it depends in *when* ObjC++ will became available; if that's planned
> for gcc 3.4, I'm not sure it's worth the effort -- just wait for ObjC++
> and port
> WebCore and JavaCore.

I've already done JavaCore (with APPLE_CHANGES) as a sort of 'trace
bullet' to see if it's generally possible. Well, this was quite easy since
JavaCore does not make extensive usage of the Foundation library. And it
uses Carbon which makes it easier to port. For WebCore, i can't say if
it's worth the effort. Maybe i start porting and see how far it goes.

>
>>> Using WebCore (and JavaCore) a browser _completely_ based on GNUstep
>>> could be written within a very short time...
>> and it will probaply give us a nice browser with a complete GNUstep
>> interface. No more need for gtk :-)
>
> yes :-)
> imho KHTML is nicer/cleaner than gecko, structurally speaking. And with
full agree. I spend quite a while trying to figure out how gecko is put
together and i can't say that i've got an understanding of how things
work.
The messy code makes it even more complicated to understand (ok, i stop
here). KHTML seems somewhat more 'straightforeward to me'.


Stefan