[Stackless] Still getting crashes with PySide

Christian Tismer tismer at stackless.com
Tue Nov 20 15:09:44 CET 2012


Howdy,

this seems to be an important one, because of PySide's popularity.
While I'm not convinced that PySide is really bug-free... I have
to ask the Wing guys about how stable it is meanwhile.

What is your impression about the games that stackless plays
with the heaptype - do you see a way to get rid of this potential
source of hidden bugs?

How much of soft-switching would we loose if we stop using
this extra table at all. What alternatives do we have?

Since I'm working with pyside myself in a project, I get into the weird 
situation
of not using Stackless, which is kind of ironic ;-)


cc'ed to Wingware - hi to John & Stefan

cheers - chris

On 15.11.12 12:51, Kristján Valur Jónsson wrote:
> I was unaware of Anselm's patch until now.
> The thing is, stackless has always been able to cope with PyTypeObjects coming from external, because they do not have the Py_TPFLAGS_HAVE_STACKLESS_EXTENSION flag on them and can thus be recognized.  I think Anselm's patch just guarded us against another of those issues.
> But is that the issue that was causing the crashes with PySide?  I thought PySide were doing something more sinister, like copying static types and not realizing it had to copy the trailing stackless extensions too.  But I could be wrong.
> However, the remaining errors could be just another case where an external type object's size is incorrectly used.
>
>> -----Original Message-----
>> From: stackless-bounces at stackless.com [mailto:stackless-
>> bounces at stackless.com] On Behalf Of Richard Tew
>> Sent: 13. nóvember 2012 18:26
>> To: The Stackless Python Mailing List
>> Subject: Re: [Stackless] Still getting crashes with PySide
>>
>> I thought we fixed that with Anselm's patch?
>>
>> On Wed, Nov 14, 2012 at 2:43 AM, Kristján Valur Jónsson
>> <kristjan at ccpgames.com> wrote:
>>> So, previously the issue with PySide was how it had a conflicting
>> redefinition of the PyTypeObject structure from stackless.  I think.
>>> Has it stopped doing that?
>>>
>>>> -----Original Message-----
>>>> From: stackless-bounces at stackless.com [mailto:stackless-
>>>> bounces at stackless.com] On Behalf Of Richard Tew
>>>> Sent: 13. nóvember 2012 07:33
>>>> To: stackless at stackless.com
>>>> Subject: Re: [Stackless] Still getting crashes with PySide
>>>>
>>>> On Mon, Nov 12, 2012 at 5:24 PM, Richard Tew
>>>> <richard.m.tew at gmail.com>
>>>> wrote:
>>>>> Hi,
>>>>>
>>>>> Is anyone using PySide with Stackless and getting crashes?
>>>> Hi,
>>>>
>>>> Okay, other than the crash on exit I emailed about above, I'm now
>>>> getting something more serious.  I've tried reproducing it with
>>>> mainline Python 2.7.3 and failed.  There's something broken with
>>>> Stackless, and it is directly related to when it is used with PySide.
>>>>
>>>> In the case of the following stacktrace, the VM appears to get
>>>> confused or corrupt state.  The sequence of activity is basically to
>>>> open a PySide window, open a file dialog, select a file, use the
>>>> given path to do file operations.  Why it has suddenly started
>>>> repeatedly happening, I do not know.  I've been doing what is now the
>>>> reproduction case for the last hour or so without a sign of it.
>>>>
>>>> $ /c/Python27/python qtui.py
>>>> XXX lineno: 90, opcode: 0
>>>> Traceback (most recent call last):
>>>>    File "qtui.py", line 192, in on_file_open_menu
>>>>      line_count = run.UI_display_file(file_path)
>>>>    ...
>>>>      hunk_id = file_info.uint32(f.read(4))
>>>> SystemError: unknown opcode
>>>>
>>>> I'll try and find time to look into it, but this isn't how I want to
>>>> spend my time
>>>> :-)
>>>>
>>>> Cheers,
>>>> Richard.
>>>>
>>>> _______________________________________________
>>>> Stackless mailing list
>>>> Stackless at stackless.com
>>>> http://www.stackless.com/mailman/listinfo/stackless
>>>
>>>
>>> _______________________________________________
>>> Stackless mailing list
>>> Stackless at stackless.com
>>> http://www.stackless.com/mailman/listinfo/stackless
>> _______________________________________________
>> Stackless mailing list
>> Stackless at stackless.com
>> http://www.stackless.com/mailman/listinfo/stackless
>
>
> _______________________________________________
> Stackless mailing list
> Stackless at stackless.com
> http://www.stackless.com/mailman/listinfo/stackless


-- 
Christian Tismer             :^)   <mailto:tismer at stackless.com>
Software Consulting          :     Have a break! Take a ride on Python's
Karl-Liebknecht-Str. 121     :    *Starship* http://starship.python.net/
14482 Potsdam                :     PGP key -> http://pgp.uni-mainz.de
phone +49 173 24 18 776  fax +49 (30) 700143-0023
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
       whom do you want to sponsor today?   http://www.stackless.com/




More information about the Stackless mailing list