Re: [Pyobjc-dev] Re: NSAnyEventMask as unsigned?

[Bob I. <bo...@re...>]

On Oct 29, 2003, at 2:03 PM, Ronald Oussoren wrote:

>
> On 29 okt 2003, at 19:50, Bob Ippolito wrote:
>
>>
>> On Oct 29, 2003, at 1:43 PM, Ronald Oussoren wrote:
>>
>>>
>>> On 29 okt 2003, at 19:09, Bob Ippolito wrote:
>>>>
>>>> In any case, because this is Python, any sort of bit flag or mask=20=

>>>> really needs to treated as if it were unsigned or else bad things=20=

>>>> start happening and warnings get thrown all over the place (or=20
>>>> probably break in 2.4).
>>>
>>> There won't be warnings because we create the int/long objects from=20=

>>> C, but things might still break in 2.4.
>>
>> Before I committed the fix, I got an exception when I used=20
>> NSAnyEventMask
>
> That's probably because you used NSAnyEventMask as an unsigned, and=20
> PyObjC doesn't like that. That has nothing to do with the int/long=20
> integration process (e.g. the warning you get when typing 0xFFFFFFFF=20=

> in a python script).

No, I used NSAnyEventMask as-is from AppKit, which was -1 (signed). =20
The problem is that the signature of =96[NSApplication=20
nextEventMatchingMask:untilDate:inMode:dequeue:] says that=20
nextEventMatchingMask is an unsigned (as would be expected for a set of=20=

bit flags).  I meant to mention that in my last email, but I hit send=20
before I remembered.

> I'm currently updating the code generator script to detect 'Mask'=20
> values and convert those as 'unsigned int' values. That should take=20
> care of this problem.

Excellent.

-bob