Re: Is GNUstep really cross platform?

[John Davidorff Pell]

I plan on contributing. :-)

Also, I never said that I needed the apple runtime, I would be happy 
with a GNU runtime installed into GNUstep's dir tree. :-)

JP

P.S. I would be happy to test it, if you like. :-)

On Dec 5, 2003, at 5:24 AM, David Ayers wrote:

> John Davidorff Pell wrote:
>
> > We're going no where fast.
>
> Very true.  Maybe you would like to contribute?
>
> Isn't it a bit inconsistent insisting on using the Apple Runtime but 
> GNUstep core?.  The Runtimes are incompatible. GNUstep and Cocoa are 
> incompatible.  Though you can write code to work with both *if* you're 
> prepared to make sure that you're using a common subset and/or 
> adequate mappings.
>
> It will take proficient Cocoa/Darwin developers to deal with 
> incompatibilities.  Andrew has already offered to help on getting the 
> GNU runtime to compile on Darwin (and I think his approach is aimed at 
> being more user friendly than my approach (which tries not to hide any 
> of the complications)).  So this is a matter of time.
>
> If it is important for you that GNUstep core works with the Apple 
> runtime, then please contribute or request for a proposal if you're 
> actually prepared to invest money in it.  I'm sure there are enough 
> folks who will respond.  Note that you will probably end up with a 
> patched Apple Runtime, but maybe you could even submit those patches 
> to Apple, and with some luck... you never know...
>
> Cheers,
> David
>
> PS: Having said that, it would be nice if OS X users could pinpoint 
> the real issues of why GNUstep-base doesn't work with the Apple 
> Runtime and track them on savannah (once she's back online), even if 
> the issues can't be solved in GNUstep alone and are there for no bugs 
> of GNUstep per se.
>
> PPS: FYI I'm locally reworking GSObjCAddClassBehavior as it was 
> believed to cause a problem with the Apple Runtime.  I'll have someone 
> with OS X test it once it's ready, but I haven't found any real issues 
> in the old implementation.  It was a bit ackward when applied to the 
> Apple runtime, but that's all.  My local implementation is more 
> "symetric" (i.e. should be easier to maintain) but it is tiny bit less 
> efficient for the GNU runtime than the current implementation.  But 
> I'm sure it's not issue here.



--
Every time you share on a P2P network, God kills a kitten.
Please think of the kittens.

smime.p7s
Description: S/MIME cryptographic signature