Hi all,
I have run into this exact same issue - has anyone found a fix or is it
best to default to the pre-2.8 DC method?
Cheers :)
Sam
Tim Ansell wrote:
> Hello,
>
> Firstly, sorry about the HTML mail, but wrapping was making the pastes
> unreadable.
>
> I am getting a segfault when using wxFloatCanvas.
>
> The backtrace looks something like this,
>> (gdb) bt
>> #0 _wrap_AlphaPixelData_Accessor_Get (args=0xb0b2670c) at
>> src/gtk/_gdi_wrap.cpp:3191
>> #1 0x0805c9e7 in PyObject_Call (func=0x978e560, arg=0xb0b5f5ec,
>> kw=0xa1888ac) at ../Objects/abstract.c:1860
>> #2 0x080c709f in PyEval_EvalFrameEx (f=0x95d2b9c, throwflag=0) at
>> ../Python/ceval.c:3844
>> #3 0x080ca115 in PyEval_EvalCodeEx (co=0x84a19b0, globals=0x84d2d74,
>> locals=0x0, args=0x95da154, argcount=1, kws=0x95da158, kwcount=0, defs=0x0,
>> defcount=0,
>> closure=0x0) at ../Python/ceval.c:2831
>> #4 0x080c85d9 in PyEval_EvalFrameEx (f=0x95da00c, throwflag=0) at
>> ../Python/ceval.c:3660
>> ...
>>
>
> If I go up a couple of frames and use the python gdb macros I can see
> the following,
>
>> (gdb) up
>> #1 0x0805c9e7 in PyObject_Call (func=0x978e560, arg=0xb0b5f5ec,
>> kw=0xa1888ac) at ../Objects/abstract.c:1860
>> 1860 ../Objects/abstract.c: No such file or directory.
>> in ../Objects/abstract.c
>> Current language: auto; currently c
>> (gdb) up
>> #2 0x080c709f in PyEval_EvalFrameEx (f=0x95d2b9c, throwflag=0) at
>> ../Python/ceval.c:3844
>> 3844 ../Python/ceval.c: No such file or directory.
>> in ../Python/ceval.c
>> (gdb) pyframe
>> *//usr/lib/python2.5/site-packages/wx-2.8-gtk2-unicode/wx/_gdi.py (1082):
>> Get/*
>>
>> (gdb) up
>> #3 0x080ca115 in PyEval_EvalCodeEx (co=0x84a19b0, globals=0x84d2d74,
>> locals=0x0, args=0x95da154, argcount=1, kws=0x95da158, kwcount=0, defs=0x0,
>> defcount=0,
>> closure=0x0) at ../Python/ceval.c:2831
>> 2831 in ../Python/ceval.c
>> (gdb) up
>> #4 0x080c85d9 in PyEval_EvalFrameEx (f=0x95da00c, throwflag=0) at
>> ../Python/ceval.c:3660
>> 3660 in ../Python/ceval.c
>> (gdb) pyframe
>> */./extra/wxFloatCanvas/FloatCanvas.py (2342): GetHitTestColor/*
>>
>> (gdb) up
>> #5 0x080ca115 in PyEval_EvalCodeEx (co=0x918c338, globals=0x917f604,
>> locals=0x0, args=0x960d200, argcount=2, kws=0x960d208, kwcount=0, defs=0x0,
>> defcount=0,
>> closure=0x0) at ../Python/ceval.c:2831
>> 2831 in ../Python/ceval.c
>> (gdb) up
>> #6 0x080c85d9 in PyEval_EvalFrameEx (f=0x960d0ac, throwflag=0) at
>> ../Python/ceval.c:3660
>> 3660 in ../Python/ceval.c
>> (gdb) pyframe
>> */./extra/wxFloatCanvas/FloatCanvas.py (2379): MouseOverTest/*
>>
>
> It appears that the code which is causing the segfault is in the
> HitTest code (which I have included below),
>
>> def GetHitTestColor(self, xy):
>> if self._ForegroundHTBitmap:
>> pdata = wx.AlphaPixelData(self._ForegroundHTBitmap)
>> else:
>> pdata = wx.AlphaPixelData(self._HTBitmap)
>> if not pdata:
>> raise RuntimeError("Trouble Accessing Hit Test bitmap")
>> pacc = pdata.GetPixels()
>> pacc.MoveTo(pdata, xy[0], xy[1])
>> return pacc.Get()[:3]
>>
> The Get line is the one which fails (line number 2342 in
> FloatCanvas.py). I think it may be related to something changing the
> _HTBitmap while in the HitTest code (however I tried putting locks
> around the _HTBitmap accesses and was still able to get a segfault).
>
> If I change it to use the older hit test code which uses a DC,
> everything works fine.
>
> Any ideas how to work around this bug?
>
> Tim 'Mithro' Ansell
> ------------------------------------------------------------------------
>
> _______________________________________________
> FloatCanvas mailing list
> [email protected]
> http://mail.mithis.com/cgi-bin/mailman/listinfo/floatcanvas
>
_______________________________________________
FloatCanvas mailing list
[email protected]
http://paulmcnett.com/cgi-bin/mailman/listinfo/floatcanvas
