Re: __NSCF** classes and libobjc2 dependency on C++ runtime

[David Chisnall]

On 4 Jun 2013, at 15:45, Chan Maxthon <address@hidden> wrote:

> Indeed I am not that sure about how GNUstep currently handles exceptions but 
> I am giving you ideas from some crash reports with symbols from an Apple 
> system.

You are not giving me ideas.  We already have an implementation of this that 
works better than the Apple implementation.  Why would you:

1) Think I didn't know how the Apple implementation works?  I talk fairly often 
to the developers of Apple's compiler and runtime?
2) Think that I would need you to explain how to implement something that I 
already have working?

> And the conclusion on toll-free bridging is about the latest Apple 
> implementation. You should really get a Mac and poke around. Just for your 
> information, what I poked around with is OS X 10.8.3 and iOS 6.1.3, both.

I have a Mac, but reverse engineering at this level of detail is of dubious 
legality.  By posting this kind of thing, you potentially open GNUstep 
developers up to legal problems.  And, for this, we should thank you?

> With non-fragile ABI it is almost impossible to construct a C struct that can 
> perfectly mirror Objective-C objects' memory layout, especially under some 
> corner circumstances.

That's what the non-fragile ABI means and why @defs() is not permitted with the 
non-fragile ABI.  

If you want to poke around in the internals of OS X, then that's fine.  Have 
fun, learn things.  But please be advised that:

1) You are most likely breaking Apple's T&Cs.
2) As a result, we most likely can't accept any code from you.
3) If you are going to post the results of anything other than the input-output 
semantics of Apple's functions, then we will need to ban you from the lists to 
avoid legal complications.

David