Open GlL issue

> No joy here.  I gunzipped the package in the build directory.  Used  
> the
> changeset tool, and followed the instructions here:
> http://wiki.squeak.org/squeak/946
>
> To access the world menu, I had entered Cobalt.sh, gotten the base
> window, left clicked
> to get the world menu, open, file.  Obtained the file and then
> leftclicked to get the install into new changeset option.  Did that.
> Then leftclicked in the world menu to and selected save.  Exited  
> Cobalt.
>
> Restarted Cobalt.  same error message.  I can't do much once the error
> message comes up.  For example if I middle click, I get the halo menu,
> but any button freezes the window(s) in Cobalt.  I can still move the
> mouse, but no buttons can be accessed except the outer window frame
> which lets me close cobalt.  I cannot copy or select any of the  
> text. I
> have attached a screenshot.  In the screenshot, the window is purple,
> but when it typically happens it is black but somehow the border is  
> just
> visible.
>
> The world menu shows that this is Croquet1.0beta, latest update #2,
> Current Changeset CurrentWork
>
> So, it appears that maybe I patched the software incorrectly (but it
> seemed simple), the patch did not take effect, or it failed to fix the
> problem.
>
> Regards,
> Les H
> On Sat, 2008-07-05 at 10:23 -0700, Joshua Gargus wrote:
>> Bert provided a fix for this in 2006.  I've packaged it into a
>> changeset because the OpenGL packages have changed enough to make
>> Monticello merging a pain.
>>
>> Les, can you let us know if this fixes your problem?  If so, let's
>> make sure that the fix makes it into the repository.
>>
>>
>>
>> Cheers,
>> Josh
>>
>>
>>
>> On Jul 5, 2008, at 9:31 AM, Les wrote:
>>
>>> I simply do not wish to give up on Croquet.  For some reason no one
>>> else
>>> seems to be having the issue I am seeing.  I am unsure why, so I
>>> looked
>>> at the code to see where it is calling the problem
>>> gloadTransposeMatrixf
>>> function.
>>>
>>> The module is RenderSpace: ogl port: portal depth: depth ghostFrame:
>>> ghost
>>> The call appears 31 lines into the function.
>>>
>>> The call states:
>>>
>>> ogl gLoadTransposeMatrixf: currentCameraInv
>>>
>>> currentCameraInv contains:
>>> 1.0 0.0 0.0 0.0
>>> 0.0 1.0 0.0 -0.699999988079071
>>> 0.0 0.0 1.0 -4.0
>>> 0.0 0.0 0.0 1.0
>>>
>>> The location in currentCameraInv looks like a legal positioin, but  
>>> the
>>> function doesn't make sense to me.  I suspect the
>>> gloadTransposeMatrixf
>>> call is in error.  Is this a correct?
>>>
>>> I hate debugging someone else's code, so I have put off digging into
>>> this thinking that someone else might be seeing the error, or would
>>> tell
>>> me that there was no function like gloadTransposeMatrixf.  I hope  
>>> that
>>> you all do realize that I can contribute to the process, but that
>>> what I
>>> wanted to do was develop an actual application, not fix the  
>>> underlying
>>> code.
>>>
>>> Regards,
>>> Les H
>>>
>>>
>>
> <Screenshot--home-lesh-Software-Cobalt-cobalt-base-current-build-
> cobalt.1.image.png>

> Sorry, I do know better, but I was a bit wrapped up in what I was
> seeing.
> Here is the log
> Regards,
> Les H
> On Sun, 2008-07-06 at 12:31 +0200, Bert Freudenberg wrote:
>> Am 06.07.2008 um 10:00 schrieb Joshua Gargus:
>>
>>> If the bug shows up, can you provide as much information as
>>> possible from the debugger (Squeak will probably be more responsive
>>> than when the error occurs while running Cobalt)?  In particular,
>>> I'm interested in the receiver and argument of/to #unsignedCharAt:
>>
>>
>> Attaching a debug log is much more informative than a screenshot.
>>
>> - Bert -
>>
>>
> <SqueakDebug.log>

> I suspected this might be the problem... glGetString() is returning a  
> null pointer.
>
> Could you please go to the OpenGLConstants class, highlight the class  
> variable GLExtensions, and 'print it'?  I'm wondering if maybe the  
> constants haven't been initialized (when I evaluate 'anOpenGL  
> glGetString: nil', the result is also a null pointer).  The proper  
> value for that constant is 7939.
>
> If it turns out that 'GLExtensions' is nil, try evaluating  
> 'OpenGLConstants initialize. GLExtContstants initialize'.
>
> Keep us posted,
> Josh
>
>
> On Jul 6, 2008, at 9:14 AM, Les wrote:
>
> > Sorry, I do know better, but I was a bit wrapped up in what I was
> > seeing.
> > Here is the log
> > Regards,
> > Les H
> > On Sun, 2008-07-06 at 12:31 +0200, Bert Freudenberg wrote:
> >> Am 06.07.2008 um 10:00 schrieb Joshua Gargus:
> >>
> >>> If the bug shows up, can you provide as much information as
> >>> possible from the debugger (Squeak will probably be more responsive
> >>> than when the error occurs while running Cobalt)?  In particular,
> >>> I'm interested in the receiver and argument of/to #unsignedCharAt:
> >>
> >>
> >> Attaching a debug log is much more informative than a screenshot.
> >>
> >> - Bert -
> >>
> >>
> > <SqueakDebug.log>
>

> Regards,
> Les H
>
> On Sun, 2008-07-06 at 09:26 -0700, Joshua Gargus wrote:
>> I suspected this might be the problem... glGetString() is returning a
>> null pointer.
>>
>> Could you please go to the OpenGLConstants class, highlight the class
>> variable GLExtensions, and 'print it'?  I'm wondering if maybe the
>> constants haven't been initialized (when I evaluate 'anOpenGL
>> glGetString: nil', the result is also a null pointer).  The proper
>> value for that constant is 7939.
>>
>> If it turns out that 'GLExtensions' is nil, try evaluating
>> 'OpenGLConstants initialize. GLExtContstants initialize'.
>>
>> Keep us posted,
>> Josh
>>
>>
>> On Jul 6, 2008, at 9:14 AM, Les wrote:
>>
>>> Sorry, I do know better, but I was a bit wrapped up in what I was
>>> seeing.
>>> Here is the log
>>> Regards,
>>> Les H
>>> On Sun, 2008-07-06 at 12:31 +0200, Bert Freudenberg wrote:
>>>> Am 06.07.2008 um 10:00 schrieb Joshua Gargus:
>>>>
>>>>> If the bug shows up, can you provide as much information as
>>>>> possible from the debugger (Squeak will probably be more  
>>>>> responsive
>>>>> than when the error occurs while running Cobalt)?  In particular,
>>>>> I'm interested in the receiver and argument of/to #unsignedCharAt:
>>>>
>>>>
>>>> Attaching a debug log is much more informative than a screenshot.
>>>>
>>>> - Bert -
>>>>
>>>>
>>> <SqueakDebug.log>
>>
>