pyobjc-dev — CVS commit notices go here

[Pyobjc-dev] PyObjC on the iPhone

[Jay F. \(saurik\) <sa...@sa...>]

So, a few weeks ago I ported PyObjC to the iPhone, and I'm realizing that I should be doing a better job of telling people how to get it ;P. I don't know if anyone is even interested in this, but I have written a little article about it on my website (http://www.saurik.com/id/5). Most of it is for iPhone developers who want to get started with PyObjC, but it also explains how to get it and provides a link to some example iPhone application code.

(The short version of getting it is to add http://apptapp.saurik.com/ as an Installer source, install Cydia Packger, and then install iPhone/Python from the Python category.)

I had to bypass the build environment (as I didn't see how to make it do cross compilation in any easy way) and I made a bunch of modifications to the iPhone assembler to support libffi almost entirely out of the box (I also ported Java a few months ago, which needed this for a setup I built that's similar to PyObjC called JocStrap), but otherwise PyObjC pretty much worked. The changes and build scripts can be found at the Telesphoreo project:

http://svn.telesphoreo.org/trunk/data/pyobjc/port.diff
http://svn.telesphoreo.org/trunk/data/pyobjc/make.sh

(Sorry about the massive single-file .diff, when I was working on that I was sufficiently in a rush as to not spend the time to break it into a couple patches for the couple things I changed.)

Sincerely,
Jay Freeman (saurik)
sa...@sa...
http://www.saurik.com/

Re: [Pyobjc-dev] PyCon 2008

[Noah G. <noa...@gm...>]

On Mar 4, 2008, at 5:14 PM, Zac Bir wrote:

> On Tue, Mar 4, 2008 at 1:23 PM, Phil Frost  
> <ph...@ma...> wrote:
>> I'll probably be at PyCon, at least for the sprints. I might do an
>> impromptu intro to PyObjC if there is interest, and I might do a
>> sprint if inspiration finds me, but otherwise I'll just be mingling.
>

+1  I would love to attend an impromput intro to PyObjC.


> I'd love to get together and go briefly over the steps involved in
> migrating from a py2app-based build strategy to an Xcode-based build
> strategy.
>
> Zac
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Pyobjc-dev mailing list
> Pyobjc-de

Noah Gift / http://noahgift.com

Re: [Pyobjc-dev] PyCon 2008

[Zac B. <zb...@ur...>]

On Tue, Mar 4, 2008 at 1:23 PM, Phil Frost <ph...@ma...> wrote:
> I'll probably be at PyCon, at least for the sprints. I might do an
>  impromptu intro to PyObjC if there is interest, and I might do a
>  sprint if inspiration finds me, but otherwise I'll just be mingling.

I'd love to get together and go briefly over the steps involved in
migrating from a py2app-based build strategy to an Xcode-based build
strategy.

Zac

Re: [Pyobjc-dev] PyCon 2008

[Phil F. <ph...@ma...>]

I'll probably be at PyCon, at least for the sprints. I might do an  
impromptu intro to PyObjC if there is interest, and I might do a  
sprint if inspiration finds me, but otherwise I'll just be mingling.

On Feb 28, 2008, at 13:03 , Robert Vadnais wrote:

> Anybody else heading to PyCon in a couple weeks?  I assume it's too
> late to organize a formal PyObjC sprint, but perhaps we could pull off
> an underground version.
>
> Bob
>
>
> ---------------------------------------------------------------------- 
> ---
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Pyobjc-dev mailing list
> Pyo...@li...
> https://lists.sourceforge.net/lists/listinfo/pyobjc-dev
>

[Pyobjc-dev] PyObjc at PyCon

[Noah G. <noa...@gm...>]

I haven't heard much about PyObjc at PyCon 08.  I would be interested  
in attending an informal talk, etc.

Noah Gift

Re: [Pyobjc-dev] PyCon 2008

[James K. <JK...@ne...>]

On 2/28/08 1:03 PM, "Robert Vadnais" <pyo...@bo...> wrote:
> Anybody else heading to PyCon in a couple weeks?  I assume it's too
> late to organize a formal PyObjC sprint, but perhaps we could pull off
> an underground version.

Hey Bob,

I'll be at PyCon, and it appears they have the ability to handle unscheduled
talks:

http://us.pycon.org/2008/conference/openspace/

I'd certainly be interested in attending.

Jim

Re: [Pyobjc-dev] PyCon 2008

[Ronald O. <ron...@ma...>]

On 28 Feb, 2008, at 19:03, Robert Vadnais wrote:

> Anybody else heading to PyCon in a couple weeks?  I assume it's too
> late to organize a formal PyObjC sprint, but perhaps we could pull off
> an underground version.

I won't be there. Heck, I won't even be on the right continent :-)

Ronald

Re: [Pyobjc-dev] Learning Cocoa with pyobjc

[Greg E. <gre...@ca...>]

Ronald Oussoren wrote:

> Objective-C isn't too bad, the object model is close to that of  Python. 
> For GUI programming you mostly don't have to worry about the  pointy 
> bits in C, which makes it pretty easy to get to know enough of  
> Objective-C to be able to understand sample code.

The main thing you'll need to get used to is remembering to
put underscores in the right places in the Python versions
of Objective-C method names. It feels quite screwy at first...

-- 
Greg

[Pyobjc-dev] PyCon 2008

[Robert V. <pyo...@bo...>]

Anybody else heading to PyCon in a couple weeks?  I assume it's too  
late to organize a formal PyObjC sprint, but perhaps we could pull off  
an underground version.

Bob

Re: [Pyobjc-dev] Learning Cocoa with pyobjc

[Ronald O. <ron...@ma...>]

On 28 Feb, 2008, at 17:16, Anders Olme wrote:

> Ok. Guess we all have to learn objective c sometime :).

Objective-C isn't too bad, the object model is close to that of  
Python. For GUI programming you mostly don't have to worry about the  
pointy bits in C, which makes it pretty easy to get to know enough of  
Objective-C to be able to understand sample code.

Ronald

Re: [Pyobjc-dev] Learning Cocoa with pyobjc

[Anders O. <and...@gm...>]

Ok. Guess we all have to learn objective c sometime :).
Best Regards Anders Olme

Re: [Pyobjc-dev] Learning Cocoa with pyobjc

[Ronald O. <ron...@ma...>]

On 28 Feb, 2008, at 11:54, Anders Olme wrote:

> Hello,
>
> Do you have to have knowledge in objective-c/cocoa to be able to use  
> pyobjc or can you start learning coca in python directly?

You can start learning Cocoa in Python, but it is helpful to have some  
knowlegde of Objective-C because almost all documentation and code  
snippets on the web are written in Objective-C.

Ronald

[Pyobjc-dev] Learning Cocoa with pyobjc

[Anders O. <and...@gm...>]

Hello,

Do you have to have knowledge in objective-c/cocoa to be able to use pyobjc
or can you start learning coca in python directly?

Best Regards Anders Olme

Re: [Pyobjc-dev] endian bug in PyDocURLProtocol example

[Ronald O. <ron...@ma...>]

 Tom,

Thanks for the fix. I've applied it in revision 1992.

Ronald

On Wednesday, February 27, 2008, at 08:37AM, "Tom Insam" <to...@je...> wrote:
>The function 'fourCharToInt' in PyDocEvents.py
>
>http://svn.red-bean.com/pyobjc/trunk/pyobjc/pyobjc-framework-WebKit/Examples/PyDocURLProtocol/PyDocEvents.py
>
>uses struct.unpack without specifying the endianness of the unpack -  
>it wants to be
>
>  struct.unpack('>l', code)[0]
>
>or it won't work on intel machines.
>
>I haven't tested the example otherwise - I was merely cribbing it for  
>writing my own URL handler. The way it mysteriously fails to work on  
>intel machines while producing no error messages whatsoever is  
>especially entertaining.
>
>Tom Insam
>to...@je...
>
>
>-------------------------------------------------------------------------
>This SF.net email is sponsored by: Microsoft
>Defy all challenges. Microsoft(R) Visual Studio 2008.
>http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>_______________________________________________
>Pyobjc-dev mailing list
>Pyo...@li...
>https://lists.sourceforge.net/lists/listinfo/pyobjc-dev
>
>

[Pyobjc-dev] endian bug in PyDocURLProtocol example

[Tom I. <to...@je...>]

The function 'fourCharToInt' in PyDocEvents.py

http://svn.red-bean.com/pyobjc/trunk/pyobjc/pyobjc-framework-WebKit/Examples/PyDocURLProtocol/PyDocEvents.py

uses struct.unpack without specifying the endianness of the unpack -  
it wants to be

  struct.unpack('>l', code)[0]

or it won't work on intel machines.

I haven't tested the example otherwise - I was merely cribbing it for  
writing my own URL handler. The way it mysteriously fails to work on  
intel machines while producing no error messages whatsoever is  
especially entertaining.

Tom Insam
to...@je...

Re: [Pyobjc-dev] weakref-like behavior on ObjC subclasses

[James R E. <ea...@ba...>]

On Feb 26, 2008, at 10:58 , Ronald Oussoren wrote:

> That's right, the Python proxy to an Objective-C object retains the  
> Objective-C object, all references from Python share a single Python  
> proxy object (and use Python's refcounting to decide when that proxy  
> object should die).


That's what I figured and what I was afraid of.  Would that also mean  
that a Leopard-only approach of using an NSHashTable with  
NSPointerFunctionsZeroingWeakMemory would also fail, because the  
Python proxy to the object in the hash table would maintain a strong  
reference?

As for my particular use-case, I can achieve good-enough behaviour by  
letting windowWillClose: trigger the unregistering (deregistration?)  
of all the controller's listeners, but it's not ideal.

Thanks!
James

--
It's dead, Jim.

Re: [Pyobjc-dev] weakref-like behavior on ObjC subclasses

[Ronald O. <ron...@ma...>]

On 26 Feb, 2008, at 0:26, James R Eagan wrote:

> I have a class that implements an observer pattern in python.   
> Observables maintain weakrefs to listeners that have been registered  
> with them such that, when a listener goes out of scope, the  
> observable doesn't keep it alive.  Of course, this mechanism fails  
> as soon as a listener inherits from an ObjC class because weakref  
> doesn't support ObjC classes.
>
> At this point, my workaround is to store a strong reference if the  
> listener can't be weakref'd.  I'm not happy about this solution  
> because it now creates different behaviors depending on whether the  
> object can be weakref'd.  If it can, then the listener is  
> automatically unregistered from the observable when its lifespan is  
> up.  If it can't, however, then the object must be manually  
> unregistered before deletion will ever occur.
>
> This is where the problem arises in PyObjC.  Is there any mechanism  
> through which I can approximate a weak reference to an ObjC object  
> in Python?

Not that I know of. PyObjC explicitly disables weakref support for  
Objective-C classes because it is AFAIK impossible to get the  
semantics right. This is due to (lack of) interaction between the  
Python and Objective-C memory managers, which makes it impossible to  
keep the weakref valid until the last strong reference to an object  
goes away, even when that last reference is in Objective-C.

There's currently some talk on python-dev about making 'weakref.ref' a  
generic function (like "len()"), which would allow us to implement a  
custom weakref type for Objective-C objects. I haven't put much  
thought into that yet, Cocoa might not have enough hooks to be able to  
implement such a weakref type.

>
>
> I tried doing some magic by checking the retainCount() of my  
> objects, but that seems to always be the same, no matter how many  
> Python objects hold a ref to the ObjC instance.

That's right, the Python proxy to an Objective-C object retains the  
Objective-C object, all references from Python share a single Python  
proxy object (and use Python's refcounting to decide when that proxy  
object should die).

Ronald

[Pyobjc-dev] weakref-like behavior on ObjC subclasses

[James R E. <ea...@ba...>]

I have a class that implements an observer pattern in python.   
Observables maintain weakrefs to listeners that have been registered  
with them such that, when a listener goes out of scope, the observable  
doesn't keep it alive.  Of course, this mechanism fails as soon as a  
listener inherits from an ObjC class because weakref doesn't support  
ObjC classes.

At this point, my workaround is to store a strong reference if the  
listener can't be weakref'd.  I'm not happy about this solution  
because it now creates different behaviors depending on whether the  
object can be weakref'd.  If it can, then the listener is  
automatically unregistered from the observable when its lifespan is  
up.  If it can't, however, then the object must be manually  
unregistered before deletion will ever occur.

This is where the problem arises in PyObjC.  Is there any mechanism  
through which I can approximate a weak reference to an ObjC object in  
Python?

I tried doing some magic by checking the retainCount() of my objects,  
but that seems to always be the same, no matter how many Python  
objects hold a ref to the ObjC instance.

Thanks in advance!
James

PS: I'd love to be able to use an NSHashTable with  
NSHashTableZeroingWeakMemory, but I need to maintain 10.4 support on  
this project. :-(

--
Beauty is in the eye of the beer holder...

Re: [Pyobjc-dev] More wiimote wrapping fun!

[John H. <joh...@gm...>]

On Feb 21, 2008, at 2:41 PM, Ronald Oussoren wrote:

>
> On 21 Feb, 2008, at 18:57, John Harrison wrote:
>
>> I can't get a signature.  Attempting to do so I get:
>> AttributeError: 'function' object has no attribute 'signature'
>>
>>
>> calling dir(rawIRData_) gives:
>> ['__call__', '__class__', '__delattr__', '__dict__', '__doc__',  
>> '__get__', '__getattribute__', '__hash__', '__init__',  
>> '__module__', '__name__', '__new__', '__reduce__',  
>> '__reduce_ex__', '__repr__', '__setattr__', '__str__',  
>> 'func_closure', 'func_code', 'func_defaults', 'func_dict',  
>> 'func_doc', 'func_globals', 'func_name']
>>
>> What connects the objc.informal_protocol definition with the class  
>> wii_remote_delegate?  I can't see anything that seems to use the  
>> same name for the thing in two places.  Is that where the problem is?
>
> Is the delegate a subclass of NSObject (either directly or  
> indirectly)?
>
> Ronald

Here's the start of the code for the delegate:

class wii_remote_delegate(NSObject):
     def irPointMovedX_Y_(self, px, py):
         pass

     #This works
     def rawIRDataHackX1_Y1_S1_X2_Y2_S2_X3_Y3_S3_X4_Y4_S4_(self,  
*posargs, **kwdargs):
         ircontrol.set_points(posargs[0],posargs[1],posargs[3],posargs 
[4])

     #This does not work
     def rawIRData_(self, irData):
         print 'rawIRData: %s' % (irData,)

>
>>
>>
>> thanks,
>> John
>>
>> On Feb 21, 2008, at 3:20 AM, Ronald Oussoren wrote:
>>
>>>
>>> On 20 Feb, 2008, at 1:02, John Harrison wrote:
>>>
>>>> I've been following the discussions concerning the wiimote wrapper
>>>> with interest as I have this insane dream of integrating head
>>>> tracking into my laser/pySight controlled video game in order to  
>>>> make
>>>> it even sillier than it currently is.  Alas, I have encountered
>>>> problems along the way.
>>>>
>>>> Working with some people that I've found on this list and the
>>>> darwiinremote project I've got something that works somewhat for
>>>> button presses and the irPointMovedX_Y_ method.  My problem is  
>>>> that I
>>>> want to capture the raw IR data, and this portion of the code  
>>>> seem to
>>>> not work.
>>>>
>>>> Here is the relevant information from the most recent  
>>>> WiiRemote.h file:
>>>>
>>>>
>>>> typedef struct {
>>>> 	int x, y, s;
>>>> } IRData;
>>>>
>>>> @interface NSObject (WiiRemoteDelegate)
>>>> - (void) rawIRData: (IRData[4]) irData;
>>>>
>>>>
>>>>
>>>>
>>>> in python I have the following:
>>>>
>>>> WiiRemoteDelegate = objc.informal_protocol(
>>>>    "WiiRemoteDelegate",
>>>>    [
>>>>        objc.selector(None,
>>>>            selector="irPointMovedX:Y:",
>>>>            signature="v@:ff", isRequired=False),
>>>>        objc.selector(None,
>>>>            selector="rawIRData:",
>>>>            signature="v@:[4{IRData=iii}]", isRequired=False)
>>>> ...
>>>>
>>>>    def rawIRData_(self, *posargs, **kwdargs):
>>>>    #def rawIRData_(self, irData):
>>>>        #print 'rawIRData: %s' % (irData,)
>>>>        for x in posargs:
>>>>            print '%x' % (x)
>>>>
>>>>
>>>> No matter what I do I get the same data in my delegate,  
>>>> formatted as
>>>> hex here:
>>>> 1322f2c
>>>>      13
>>>> bfffe350
>>>> 1313b20
>>>> 1313b00
>>>> bfffe3a0
>>>>       0
>>>> 154c396
>>>> 94aa0ca4
>>>> 142b9c8
>>>> 1322ed0
>>>>      13
>>>>
>>>> This looks to me like I'm not getting the ints that I would expect
>>>> but that I'm getting references to the IRData struct or something.
>>>
>>> That's odd, it looks like the informal protocol isn't actually  
>>> used. What's the value of rawIRData_.signature? This should be v@: 
>>> [4{IRData=iii}]. Please use the commented-out version of the  
>>> method prototype for this, not the one with '*posargs'.
>>>
>>> BTW. I've downloaded the WiiRemote framework a while back and my  
>>> parsing script cannot deal with some features in the header files  
>>> (I haven't even tried fixing that yet). Furthermore, you won't be  
>>> able to use the Mii structure from python because that contains  
>>> bitfields and those aren't supported yet.
>>>
>>> Ronald
>>>>
>>>>
>>>> Is there something wrong with selector and/or signature?  Do I need
>>>> to add method encodings to the signature?  What would a properly
>>>> encoded signature look like?  If these are pointers is there a  
>>>> way to
>>>> dereference them in python?
>>>>
>>>> I've been banging my head against this for more time than I'd  
>>>> like to
>>>> admit and keep getting either bus errors or junk data.
>>>>
>>>> Thanks!
>>>> John
>>>>
>>>> ------------------------------------------------------------------- 
>>>> ------
>>>> This SF.net email is sponsored by: Microsoft
>>>> Defy all challenges. Microsoft(R) Visual Studio 2008.
>>>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>>>> _______________________________________________
>>>> Pyobjc-dev mailing list
>>>> Pyo...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/pyobjc-dev
>>>
>>
>

Re: [Pyobjc-dev] More wiimote wrapping fun!

[Ronald O. <ron...@ma...>]

On 21 Feb, 2008, at 18:57, John Harrison wrote:

> I can't get a signature.  Attempting to do so I get:
> AttributeError: 'function' object has no attribute 'signature'
>
>
> calling dir(rawIRData_) gives:
> ['__call__', '__class__', '__delattr__', '__dict__', '__doc__',  
> '__get__', '__getattribute__', '__hash__', '__init__', '__module__',  
> '__name__', '__new__', '__reduce__', '__reduce_ex__', '__repr__',  
> '__setattr__', '__str__', 'func_closure', 'func_code',  
> 'func_defaults', 'func_dict', 'func_doc', 'func_globals', 'func_name']
>
> What connects the objc.informal_protocol definition with the class  
> wii_remote_delegate?  I can't see anything that seems to use the  
> same name for the thing in two places.  Is that where the problem is?

Is the delegate a subclass of NSObject (either directly or indirectly)?

Ronald

>
>
> thanks,
> John
>
> On Feb 21, 2008, at 3:20 AM, Ronald Oussoren wrote:
>
>>
>> On 20 Feb, 2008, at 1:02, John Harrison wrote:
>>
>>> I've been following the discussions concerning the wiimote wrapper
>>> with interest as I have this insane dream of integrating head
>>> tracking into my laser/pySight controlled video game in order to  
>>> make
>>> it even sillier than it currently is.  Alas, I have encountered
>>> problems along the way.
>>>
>>> Working with some people that I've found on this list and the
>>> darwiinremote project I've got something that works somewhat for
>>> button presses and the irPointMovedX_Y_ method.  My problem is  
>>> that I
>>> want to capture the raw IR data, and this portion of the code seem  
>>> to
>>> not work.
>>>
>>> Here is the relevant information from the most recent WiiRemote.h  
>>> file:
>>>
>>>
>>> typedef struct {
>>> 	int x, y, s;
>>> } IRData;
>>>
>>> @interface NSObject (WiiRemoteDelegate)
>>> - (void) rawIRData: (IRData[4]) irData;
>>>
>>>
>>>
>>>
>>> in python I have the following:
>>>
>>> WiiRemoteDelegate = objc.informal_protocol(
>>>    "WiiRemoteDelegate",
>>>    [
>>>        objc.selector(None,
>>>            selector="irPointMovedX:Y:",
>>>            signature="v@:ff", isRequired=False),
>>>        objc.selector(None,
>>>            selector="rawIRData:",
>>>            signature="v@:[4{IRData=iii}]", isRequired=False)
>>> ...
>>>
>>>    def rawIRData_(self, *posargs, **kwdargs):
>>>    #def rawIRData_(self, irData):
>>>        #print 'rawIRData: %s' % (irData,)
>>>        for x in posargs:
>>>            print '%x' % (x)
>>>
>>>
>>> No matter what I do I get the same data in my delegate, formatted as
>>> hex here:
>>> 1322f2c
>>>      13
>>> bfffe350
>>> 1313b20
>>> 1313b00
>>> bfffe3a0
>>>       0
>>> 154c396
>>> 94aa0ca4
>>> 142b9c8
>>> 1322ed0
>>>      13
>>>
>>> This looks to me like I'm not getting the ints that I would expect
>>> but that I'm getting references to the IRData struct or something.
>>
>> That's odd, it looks like the informal protocol isn't actually  
>> used. What's the value of rawIRData_.signature? This should be v@: 
>> [4{IRData=iii}]. Please use the commented-out version of the method  
>> prototype for this, not the one with '*posargs'.
>>
>> BTW. I've downloaded the WiiRemote framework a while back and my  
>> parsing script cannot deal with some features in the header files  
>> (I haven't even tried fixing that yet). Furthermore, you won't be  
>> able to use the Mii structure from python because that contains  
>> bitfields and those aren't supported yet.
>>
>> Ronald
>>>
>>>
>>> Is there something wrong with selector and/or signature?  Do I need
>>> to add method encodings to the signature?  What would a properly
>>> encoded signature look like?  If these are pointers is there a way  
>>> to
>>> dereference them in python?
>>>
>>> I've been banging my head against this for more time than I'd like  
>>> to
>>> admit and keep getting either bus errors or junk data.
>>>
>>> Thanks!
>>> John
>>>
>>> -------------------------------------------------------------------------
>>> This SF.net email is sponsored by: Microsoft
>>> Defy all challenges. Microsoft(R) Visual Studio 2008.
>>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>>> _______________________________________________
>>> Pyobjc-dev mailing list
>>> Pyo...@li...
>>> https://lists.sourceforge.net/lists/listinfo/pyobjc-dev
>>
>

Re: [Pyobjc-dev] DeprecationWarning: Not all Objective-C arguments are present in the Python argument-list

[James R E. <ea...@cc...>]

On Feb 21, 2008, at 04:40 , Ronald Oussoren wrote:

> The DeprecationWarning is bogus here. What this is trying to  
> accomplish is warn about mismatches between the Python signature and  
> Objective-C signature of a method, but this is giving false  
> positives on some occassions. In your case you get the warning  
> because the decorated function has a variable number of arguments.

Okay, that makes me feel a lot better. :-)  Is there a way I can  
suppress the DeprecationWarning?  It's not the end of the world if I  
can't, but I'd rather not dump those to the console when my app runs.

The closest I've found is to use Michele Simionato's decorator  
module[1], which ensures that the wrapper generated by the decorator  
actually matches the wrapped function's signature (à la  
inspect.getargspec).  I don't like this approach because it uses eval  
to generate the wrapper.  I also haven't tested it to see if it  
successfully convinces PyObjC, but I assume it would.

[1]: http://www.phyast.pitt.edu/~micheles/python/documentation.html

Cheers!
James

--
"I'm using LaTeX math mode.  How much more sexy can I be?"  -- Je77

Re: [Pyobjc-dev] More wiimote wrapping fun!

[John H. <joh...@gm...>]

I can't get a signature.  Attempting to do so I get:
AttributeError: 'function' object has no attribute 'signature'


calling dir(rawIRData_) gives:
['__call__', '__class__', '__delattr__', '__dict__', '__doc__',  
'__get__', '__getattribute__', '__hash__', '__init__', '__module__',  
'__name__', '__new__', '__reduce__', '__reduce_ex__', '__repr__',  
'__setattr__', '__str__', 'func_closure', 'func_code',  
'func_defaults', 'func_dict', 'func_doc', 'func_globals', 'func_name']

What connects the objc.informal_protocol definition with the class  
wii_remote_delegate?  I can't see anything that seems to use the same  
name for the thing in two places.  Is that where the problem is?

thanks,
John

On Feb 21, 2008, at 3:20 AM, Ronald Oussoren wrote:

>
> On 20 Feb, 2008, at 1:02, John Harrison wrote:
>
>> I've been following the discussions concerning the wiimote wrapper
>> with interest as I have this insane dream of integrating head
>> tracking into my laser/pySight controlled video game in order to make
>> it even sillier than it currently is.  Alas, I have encountered
>> problems along the way.
>>
>> Working with some people that I've found on this list and the
>> darwiinremote project I've got something that works somewhat for
>> button presses and the irPointMovedX_Y_ method.  My problem is that I
>> want to capture the raw IR data, and this portion of the code seem to
>> not work.
>>
>> Here is the relevant information from the most recent WiiRemote.h  
>> file:
>>
>>
>> typedef struct {
>> 	int x, y, s;
>> } IRData;
>>
>> @interface NSObject (WiiRemoteDelegate)
>> - (void) rawIRData: (IRData[4]) irData;
>>
>>
>>
>>
>> in python I have the following:
>>
>> WiiRemoteDelegate = objc.informal_protocol(
>>     "WiiRemoteDelegate",
>>     [
>>         objc.selector(None,
>>             selector="irPointMovedX:Y:",
>>             signature="v@:ff", isRequired=False),
>>         objc.selector(None,
>>             selector="rawIRData:",
>>             signature="v@:[4{IRData=iii}]", isRequired=False)
>> ...
>>
>>     def rawIRData_(self, *posargs, **kwdargs):
>>     #def rawIRData_(self, irData):
>>         #print 'rawIRData: %s' % (irData,)
>>         for x in posargs:
>>             print '%x' % (x)
>>
>>
>> No matter what I do I get the same data in my delegate, formatted as
>> hex here:
>>  1322f2c
>>       13
>> bfffe350
>>  1313b20
>>  1313b00
>> bfffe3a0
>>        0
>>  154c396
>> 94aa0ca4
>>  142b9c8
>>  1322ed0
>>       13
>>
>> This looks to me like I'm not getting the ints that I would expect
>> but that I'm getting references to the IRData struct or something.
>
> That's odd, it looks like the informal protocol isn't actually  
> used. What's the value of rawIRData_.signature? This should be v@:[4 
> {IRData=iii}]. Please use the commented-out version of the method  
> prototype for this, not the one with '*posargs'.
>
> BTW. I've downloaded the WiiRemote framework a while back and my  
> parsing script cannot deal with some features in the header files  
> (I haven't even tried fixing that yet). Furthermore, you won't be  
> able to use the Mii structure from python because that contains  
> bitfields and those aren't supported yet.
>
> Ronald
>>
>>
>> Is there something wrong with selector and/or signature?  Do I need
>> to add method encodings to the signature?  What would a properly
>> encoded signature look like?  If these are pointers is there a way to
>> dereference them in python?
>>
>> I've been banging my head against this for more time than I'd like to
>> admit and keep getting either bus errors or junk data.
>>
>> Thanks!
>> John
>>
>> --------------------------------------------------------------------- 
>> ----
>> This SF.net email is sponsored by: Microsoft
>> Defy all challenges. Microsoft(R) Visual Studio 2008.
>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> _______________________________________________
>> Pyobjc-dev mailing list
>> Pyo...@li...
>> https://lists.sourceforge.net/lists/listinfo/pyobjc-dev
>

Re: [Pyobjc-dev] More wiimote wrapping fun!

[Ronald O. <ron...@ma...>]

On 20 Feb, 2008, at 1:02, John Harrison wrote:

> I've been following the discussions concerning the wiimote wrapper
> with interest as I have this insane dream of integrating head
> tracking into my laser/pySight controlled video game in order to make
> it even sillier than it currently is.  Alas, I have encountered
> problems along the way.
>
> Working with some people that I've found on this list and the
> darwiinremote project I've got something that works somewhat for
> button presses and the irPointMovedX_Y_ method.  My problem is that I
> want to capture the raw IR data, and this portion of the code seem to
> not work.
>
> Here is the relevant information from the most recent WiiRemote.h  
> file:
>
>
> typedef struct {
> 	int x, y, s;
> } IRData;
>
> @interface NSObject (WiiRemoteDelegate)
> - (void) rawIRData: (IRData[4]) irData;
>
>
>
>
> in python I have the following:
>
> WiiRemoteDelegate = objc.informal_protocol(
>     "WiiRemoteDelegate",
>     [
>         objc.selector(None,
>             selector="irPointMovedX:Y:",
>             signature="v@:ff", isRequired=False),
>         objc.selector(None,
>             selector="rawIRData:",
>             signature="v@:[4{IRData=iii}]", isRequired=False)
> ...
>
>     def rawIRData_(self, *posargs, **kwdargs):
>     #def rawIRData_(self, irData):
>         #print 'rawIRData: %s' % (irData,)
>         for x in posargs:
>             print '%x' % (x)
>
>
> No matter what I do I get the same data in my delegate, formatted as
> hex here:
>  1322f2c
>       13
> bfffe350
>  1313b20
>  1313b00
> bfffe3a0
>        0
>  154c396
> 94aa0ca4
>  142b9c8
>  1322ed0
>       13
>
> This looks to me like I'm not getting the ints that I would expect
> but that I'm getting references to the IRData struct or something.

That's odd, it looks like the informal protocol isn't actually used.  
What's the value of rawIRData_.signature? This should be v@: 
[4{IRData=iii}]. Please use the commented-out version of the method  
prototype for this, not the one with '*posargs'.

BTW. I've downloaded the WiiRemote framework a while back and my  
parsing script cannot deal with some features in the header files (I  
haven't even tried fixing that yet). Furthermore, you won't be able to  
use the Mii structure from python because that contains bitfields and  
those aren't supported yet.

Ronald
>
>
> Is there something wrong with selector and/or signature?  Do I need
> to add method encodings to the signature?  What would a properly
> encoded signature look like?  If these are pointers is there a way to
> dereference them in python?
>
> I've been banging my head against this for more time than I'd like to
> admit and keep getting either bus errors or junk data.
>
> Thanks!
> John
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Pyobjc-dev mailing list
> Pyo...@li...
> https://lists.sourceforge.net/lists/listinfo/pyobjc-dev

Re: [Pyobjc-dev] DeprecationWarning: Not all Objective-C arguments are present in the Python argument-list

[Ronald O. <ron...@ma...>]

On 20 Feb, 2008, at 7:00, James R Eagan wrote:

> I've been using the following decorator in PyObjC 1.4 without  
> problems, but now PyObjC 2.0 gives me a nice DeprecationWarning.  Am  
> I doing something naughty and/or dangerous?  If so, is there a safer  
> way to accomplish the same thing in PyObjC2?

The DeprecationWarning is bogus here. What this is trying to  
accomplish is warn about mismatches between the Python signature and  
Objective-C signature of a method, but this is giving false positives  
on some occassions. In your case you get the warning because the  
decorated function has a variable number of arguments.

I guess it is time to start working on a set of unittests for this  
feature to find all types of false positives.

Ronald

>
>
> First, the decorator:
>
> def trap_exceptions(func):
>   ''' Trap exceptions and prevent them from being passed along to  
> the ObjC runtime. '''
>   def trapper(self, *args, **kwargs):
>       try:
>           return func(self, *args, **kwargs)
>       except:
>           sys.stderr.write('Exception in %s: ' % (func.__name__))
>           traceback.print_exception(*sys.exc_info())
>   return trapper
>
> Typical usage is on an Objective-C entrypoint.  e.g.:
>
>   @trap_exceptions
>   def performDragOperation_(self, sender):
>       if self.delegate is not None \
>               and  
> self.delegate.respondsToSelector_('performDragOperation:'):
>           return self.delegate.performDragOperation_(sender)
>       return False
>
> This worked fine pre-Leopard.  When an exception was thrown  
> somewhere in (e.g.) performDragOperation_, I'd get a stack trace on  
> the console and everything would carry along as if nothing  
> happened.  In PyObjC2, the decorator appears to behave as expected,  
> but I also get this warning for every @trap_exceptions'ed method (at  
> NIB load time):
>
> /System/Library/Frameworks/Python.framework/Versions/2.5/Extras/lib/ 
> python/PyObjC/PyObjCTools/NibClassBuilder.py:219:  
> DeprecationWarning: Not all Objective-C arguments are present in the  
> Python argument-list of <unbound selector  
> observeValueForKeyPath:ofObject:change:context: of CBIconView at  
> 0x4f1fef0>
> return metaClass(name, bases, methods)
>
> Thanks!
> James
>
> PS:  Thanks for all the work you've done on PyObjC 2.0.  There are  
> lots of wonderful improvements here!  I'm very much looking forward  
> to the Tiger backport. :-)
>
> --
> Doctors are sadists who like to play God and watch lesser people  
> scream
> 			-- Bren MacGuff in "Juno"
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/_______________________________________________
> Pyobjc-dev mailing list
> Pyo...@li...
> https://lists.sourceforge.net/lists/listinfo/pyobjc-dev

Re: [Pyobjc-dev] More wiimote wrapping fun!

[John H. <joh...@gm...>]

For those that are interested in getting the data from rawIRData in  
the wiimote framework, I've developed a hack that works, but is  
ugly.  I've rolled my own framework by added a method to the delegate  
in WiiRemote.h:

- (void) rawIRDataHackX1:(int) x1 Y1:(int) y1 S1:(int)s1 X2:(int) x2  
Y2:(int) y2 S2:(int)s2 X3:(int) x3 Y3:(int) y3 S3:(int)s3 X4:(int) x4  
Y4:(int) y4 S4:(int)s4;


	
I also added this to the end of the handleIRData method in WiiRemote.m:

	
	if ([_delegate respondsToSelector:@selector  
(rawIRDataHackX1:Y1:S1:X2:Y2:S2:X3:Y3:S3:X4:Y4:S4:)])
		[_delegate rawIRDataHackX1:irData[0].x Y1:irData[0].y S1:irData 
[0].s X2:irData[1].x Y2:irData[1].y S2:irData[1].s X3:irData[2].x  
Y3:irData[2].y S3:irData[2].s X4:irData[3].x Y4:irData[3].y S4:irData 
[3].s];
	

and then in my python code:

WiiRemoteDelegate = objc.informal_protocol(
     "WiiRemoteDelegate",
     [
         objc.selector(None,
             selector="irPointMovedX:Y:",
             signature="v@:ff", isRequired=False),
         objc.selector(None,
             selector="rawIRData:",
             signature="v@:[4{?=iii}]", isRequired=False),
         objc.selector(None,
              
selector="rawIRDataHackX1:Y1:S1:X2:Y2:S2:X3:Y3:S3:X4:Y4:S4:",
             signature="v@:iiiiiiiiiiii", isRequired=False),

and

    def rawIRDataHackX1_Y1_S1_X2_Y2_S2_X3_Y3_S3_X4_Y4_S4_(self,  
*posargs, **kwdargs):
         for x in posargs:
              print '%8x' % (x)

This works and I get data for the IR points.  If anyone can tell me  
how to properly wrap the normal rawIRData method I'd appreciate it,  
but at least I can continue development with this solution.