On 10 Aug 2013, at 13:03, Luboš Doležel <lu...@dolezel.info> wrote:
> On 08/09/2013 01:36 PM, David Chisnall wrote:
>> On 9 Aug 2013, at 12:22, Luboš Doležel <lu...@dolezel.info> wrote:
>>
>>> Right, sorry, it is @48@0:8{_NSRect={_NSPoint=dd}{_NSSize=dd}}16
>>>
>>> So I assume the right solution is to create a substitutes table to
>>> rewrite CGRect->_NSRect etc.
>>
>> That would certainly work, although the correct solution is probably
>> to rename _NSRect to CGRect and _NSPoint into CGPoint and make the
>> typedefs remain as-is.
>>
>> The runtime is quite conservative about type mapping for structs,
>> because, for example, NSPoint and NSSize have the same
>> representation, but it's probably a bug if something is passed one
>> and expects the other.
>>
>> David
>>
>
> And what about this one?
>
> Calling [NSImage
> -drawInRect:fromRect:operation:fraction:respectFlipped:hints:] with
> incorrect signature. Method has
> v108@0:8{_NSRect={_NSPoint=dd}{_NSSize=dd}}16{_NSRect={_NSPoint=dd}{_NSSize=dd}}48Q80d88C96@100,
> selector has
> v108@0:8{_NSRect={_NSPoint=dd}{_NSSize=dd}}16{_NSRect={_NSPoint=dd}{_NSSize=dd}}48Q80d88C96@100
>
> Note that there are two identical signatures printed in this warning.
That is weird. Do you have a reduced test case that triggers it?
David
_______________________________________________
Gnustep-dev mailing list
Gnustep-dev@gnu.org
https://lists.gnu.org/mailman/listinfo/gnustep-dev
