Ok 
* launching winpython command prompt
* spyder --reset
* spyder --show-console
* asking for a IPython qt console inside Ipython ...
==> it goes a little further, opening a qt console sub-window + displaying 
the initial blabla "to connect to another client to this kerrnel,..."
but then it fails before proposing the first "[1]:"  prompt.

Sorry I can't do further test tonight, my eyes are closing.

Sheers,

Le lundi 19 août 2013 23:08:47 UTC+2, Pierre Raybaut a écrit :
>
> Ok.
>
> Now, let's say you go back to the original IPython file contents: the bug 
> should be triggered each time you try to open an IPython console.
> Is removing all Spyder settings has any effect on this situation? (doing a 
> reset: `spyder --reset`)
> (BTW, to test all this, you may open the WinPython command prompt and type 
> `spyder --show-console`)
>  
> -Pierre
>
>
> 2013/8/19 <[email protected] <javascript:>>
>
>> Yes , with the previous move, I couldn't trigger the bug.
>>
>> Excuse me for this perhaps stupid remark : if the bug is triggered by a 
>> too close succession of zoom_in + zoom_out on a slow machine, like a VM or 
>> an old Vista beast, could a 'delay' or a 'wait of completion' do the trick 
>> also ?
>>
>> Le lundi 19 août 2013 22:37:58 UTC+2, Pierre Raybaut a écrit :
>>>
>>> That's a nasty one, for sure. I was sure that it was the revision which 
>>> caused this issue... but shortly after I wrote my last message, I realized 
>>> that it was still crashing (I just had the "chance" to run it three times 
>>> without a single crash...).
>>>  
>>> However, do you confirm that commenting lines 354-376 in 
>>> IPython\qt\console\console_**widget.py is really working (I mean, 100%)?
>>>  
>>> -Pierre
>>>
>>>
>>> 2013/8/19 <[email protected]>
>>>
>>> ok, 
>>>>
>>>> so un-commenting previous change + trying spyder +ipython qt console
>>>> ==> fail (expected, nearly hoped).
>>>>
>>>> but, commenting the 4 lines indicated and trying again fails also just 
>>>> after 
>>>> (and yes, i killed the python process in the background before)
>>>> ==> unless I did something wrong, i cannot confirm the last move as a 
>>>> success.
>>>>
>>>> Le lundi 19 août 2013 21:52:24 UTC+2, Pierre Raybaut a écrit :
>>>>>
>>>>> Ok, I think I've narrowed it down.
>>>>> This bug may be a shortcut conflict in Qt introduced with the 
>>>>> following revision:
>>>>> https://code.google.com/p/**spyd**erlib/source/detail?r=**d703c902**
>>>>> 66fa<https://code.google.com/p/spyderlib/source/detail?r=d703c90266fa>
>>>>>  
>>>>> This is very surprising but it's apparently causing a hard crash.
>>>>>  
>>>>> So, to confirm that this is it, could you please try and comment lines 
>>>>> 631-634 in spyderlib\widgets\editor.py:
>>>>>  
>>>>>         self.zoominsc = newsc(QKeySequence.ZoomIn,
>>>>>                               lambda: self.emit(SIGNAL('zoom_in()'))**
>>>>> **)
>>>>>         self.zoomoutsc = newsc(QKeySequence.ZoomOut,
>>>>>                                lambda: self.emit(SIGNAL('zoom_out()')*
>>>>> ***))
>>>>> (and uncomment the lines in IPython\qt\console\console_**wid**get.py, 
>>>>> of course)
>>>>>  
>>>>> Thanks,
>>>>> -Pierre
>>>>>
>>>>>
>>>>> 2013/8/19 Pierre Raybaut <[email protected]>
>>>>>
>>>>>> You're right, being able to reproduce the bug is a good thing but in 
>>>>>> that case, the bug is not 100% reproducible. Sometimes it crashes, 
>>>>>> sometimes not. And I don't understand why.
>>>>>>  
>>>>>> Anyway, I was able to find a way to prevent the application from 
>>>>>> crashing when an IPython console is opened by commenting out lines 
>>>>>> 354-376 
>>>>>> in IPython\qt\console\console_**wid**get.py:
>>>>>>  
>>>>>>         self.increase_font_size = QtGui.QAction("Bigger Font", 
>>>>>>
>>>>>>                 self,
>>>>>>
>>>>>>                 shortcut=QtGui.QKeySequence.**Zo**omIn,
>>>>>>
>>>>>>                 shortcutContext=QtCore.Qt.**Widg**
>>>>>> etWithChildrenShortcut,
>>>>>>
>>>>>>                 statusTip="Increase the font size by one point",
>>>>>>
>>>>>>                 triggered=self._increase_font_****size)
>>>>>>
>>>>>>         self.addAction(self.**in**crease_font_size)
>>>>>>
>>>>>>
>>>>>>          self.decrease_font_size = QtGui.QAction("Smaller Font",
>>>>>>
>>>>>>                 self,
>>>>>>
>>>>>>                 shortcut=QtGui.QKeySequence.**Zo**omOut,
>>>>>>
>>>>>>                 shortcutContext=QtCore.Qt.**Widg**
>>>>>> etWithChildrenShortcut,
>>>>>>
>>>>>>                 statusTip="Decrease the font size by one point",
>>>>>>
>>>>>>                 triggered=self._decrease_font_****size)
>>>>>>
>>>>>>         self.addAction(self.decrease_**f**ont_size)
>>>>>>
>>>>>>
>>>>>>          self.reset_font_size = QtGui.QAction("Normal Font",
>>>>>>
>>>>>>                 self,
>>>>>>
>>>>>>                 shortcut="Ctrl+0",
>>>>>>
>>>>>>                 shortcutContext=QtCore.Qt.**Widg**
>>>>>> etWithChildrenShortcut,
>>>>>>
>>>>>>                 statusTip="Restore the Normal font size",
>>>>>>
>>>>>>                 triggered=self.reset_font)
>>>>>>
>>>>>>         self.addAction(self.reset_**font**_size)
>>>>>>
>>>>>>
>>>>>> Could you please try and comment those lines on your side to see if 
>>>>>> it has the same effect?
>>>>>>
>>>>>>  
>>>>>>
>>>>>> -Pierre
>>>>>>
>>>>>>
>>>>>> 2013/8/19 <[email protected]>
>>>>>>
>>>>>> I would rather consider a perfectly reproducible bug very good news. 
>>>>>>> Wouldn't it be possible to escaladate the bug to ipython team ?
>>>>>>>
>>>>>>> They are already doing a change in "qt.console.rich_ipython_**widge*
>>>>>>> *t", related to old "qt console" / "py3k" bugs.
>>>>>>> (https://github.com/ipython/**ip**ython/pull/4054/files<https://github.com/ipython/ipython/pull/4054/files>
>>>>>>> )
>>>>>>>
>>>>>>>
>>>>>>> Le lundi 19 août 2013 19:22:14 UTC+2, Pierre Raybaut a écrit :
>>>>>>>>
>>>>>>>> I'm now able to reproduce this issue with a WinPython 3.3.2.2 
>>>>>>>> distribution installed on a fresh Win7 Pro 64bit VM.
>>>>>>>>  
>>>>>>>> So far my conclusion is that python.exe crashes when calling 
>>>>>>>> IPythonWidget's contructor (when object is instantiated) which is not 
>>>>>>>> good 
>>>>>>>> news as this widget is from 
>>>>>>>> IPython.qt.console.rich_**ipytho****n_widget. 
>>>>>>>> In other words, I'm not close to a solution...
>>>>>>>>  
>>>>>>>> Have you tried upgrading Spyder to 2.3.0dev5 in your "old" 
>>>>>>>> WinPython 3.3 install? (the one that works)
>>>>>>>>  
>>>>>>>> -Pierre
>>>>>>>>
>>>>>>>>
>>>>>>>> 2013/8/19 big stone <[email protected]>
>>>>>>>>
>>>>>>>>>  Hello Pierre,
>>>>>>>>>
>>>>>>>>> I just tried and it changed nothing, (and the debug vs 2010 asked 
>>>>>>>>> me a service pack, that I'm currently installing).
>>>>>>>>> But first :
>>>>>>>>> - I had 2 success, by starting a ipython qt console, then closing 
>>>>>>>>> spyder (and not the ipython console)
>>>>>>>>> - then it failed when I , if I did note well the chain of events, 
>>>>>>>>> when I :
>>>>>>>>>   * started again a ipython qt console, 
>>>>>>>>>   * then closed the ipython qt console but not spyder,
>>>>>>>>>   * then retried
>>>>>>>>> ==> May it be a 'wrong' chain of actions that you didn't 
>>>>>>>>> anticipate users could do ?
>>>>>>>>> ==> is there a magic spyder --reset or something like that that I 
>>>>>>>>> should do or look ?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>  
>>>>>>>>>
>>>>>>>>> 2013/8/19 Pierre Raybaut <[email protected]>
>>>>>>>>>
>>>>>>>>>>  After some thinking, I might have a solution to show a 
>>>>>>>>>> traceback for this crash.
>>>>>>>>>> Just modify the code of spyderlib\widgets\**internalshel****l.py 
>>>>>>>>>> in your WinPython installation and replace DEBUG by True at line 168:
>>>>>>>>>> http://code.google.com/p/**spyde****rlib/source/browse/**
>>>>>>>>>> spyderlib/**w**idgets/**internalshell.py#168<http://code.google.com/p/spyderlib/source/browse/spyderlib/widgets/internalshell.py#168>
>>>>>>>>>>   
>>>>>>>>>> This should enable the DEBUG mode only for the internal shell 
>>>>>>>>>> which could be sufficient here.
>>>>>>>>>>  
>>>>>>>>>> -Pierre
>>>>>>>>>>
>>>>>>>>>> Le lundi 19 août 2013 10:27:38 UTC+2, Pierre Raybaut a écrit :
>>>>>>>>>>
>>>>>>>>>>> Unfortunately, I can't reproduce this bug.
>>>>>>>>>>>  
>>>>>>>>>>> So, considering the fact that there is no traceback (hard 
>>>>>>>>>>> crash), the only way to find out where this comes from is to run 
>>>>>>>>>>> Spyder 
>>>>>>>>>>> from source checkout and update to older revisions until the bug 
>>>>>>>>>>> disappears 
>>>>>>>>>>> (or the opposite: update to the 2.3.0dev3 revision if it's working 
>>>>>>>>>>> and 
>>>>>>>>>>> update to newer revisions until the bug shows up).
>>>>>>>>>>> If you need assistance, please do not hesitate to ask.
>>>>>>>>>>>  
>>>>>>>>>>> This may also be related to this bug report:
>>>>>>>>>>> http://code.google.com/p/**winpy******thon/issues/detail?id=77<http://code.google.com/p/winpython/issues/detail?id=77>
>>>>>>>>>>>  
>>>>>>>>>>> -Pierre
>>>>>>>>>>>
>>>>>>>>>>> Le samedi 17 août 2013 03:12:10 UTC+2, Jim Passmore a écrit :
>>>>>>>>>>>
>>>>>>>>>>>> Confirming similar problem running dev5, in a freshly unzipped 
>>>>>>>>>>>> 32-bit WinPython 3.3.2.2, on 32-bit Win7.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> On Monday, August 12, 2013 12:10:28 PM UTC-4, 
>>>>>>>>>>>> [email protected] wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Hello,
>>>>>>>>>>>>>
>>>>>>>>>>>>> I'm experimenting on windows vista 32bit the new spyder 
>>>>>>>>>>>>> 2.3.0.dev5 (as a non-manager of my pc). 
>>>>>>>>>>>>> (via winpython 3.3.2.2, so python3.3).
>>>>>>>>>>>>>
>>>>>>>>>>>>> Spyder fails to launch reliabily an IPython (1.0.0) console. 
>>>>>>>>>>>>> Procedure =
>>>>>>>>>>>>> - click on "Interpreters-> open an IPython Console",
>>>>>>>>>>>>> - I get an opening IPython console the trick about the way to 
>>>>>>>>>>>>> connect to an existing IPython Kernel,
>>>>>>>>>>>>> - then after about 5 seconds python.exe stops with a debug 
>>>>>>>>>>>>> message :  "Exception non gérée à 0x00000000 dans python.exe : 
>>>>>>>>>>>>> 0xC0000005: 
>>>>>>>>>>>>> Access violation."
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Does anyone experience the same unfortunate problem ?
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Nota :
>>>>>>>>>>>>> - among the fails, I had two success,
>>>>>>>>>>>>> - a "python.exe" process remains in the background (and seems 
>>>>>>>>>>>>> to be the launched ipython kernel)
>>>>>>>>>>>>> - I tried with mathjax installed, without more success,
>>>>>>>>>>>>> - I removed Pylab activation when opening IPython console, no 
>>>>>>>>>>>>> more success,
>>>>>>>>>>>>> - on the exact same PC, I have not problems with previous 
>>>>>>>>>>>>> spyder (2.3.0.dev3) and python1.0.0, except a kind 
>>>>>>>>>>>>> remindermessage 
>>>>>>>>>>>>>    "The top-level `frontend` package has been deprecated. "
>>>>>>>>>>>>>
>>>>>>>>>>>>>  -- 
>>>>>>>>>> You received this message because you are subscribed to a topic 
>>>>>>>>>> in the Google Groups "spyder" group.
>>>>>>>>>> To unsubscribe from this topic, visit 
>>>>>>>>>> https://groups.google.com/d/**to****pic/spyderlib/E4uKC3SQ3iU/**
>>>>>>>>>> unsu****bscribe<https://groups.google.com/d/topic/spyderlib/E4uKC3SQ3iU/unsubscribe>
>>>>>>>>>> .
>>>>>>>>>>  To unsubscribe from this group and all its topics, send an email 
>>>>>>>>>> to spyderlib+...@**googlegroups.com****.
>>>>>>>>>> To post to this group, send email to [email protected].
>>>>>>>>>>
>>>>>>>>>> Visit this group at http://groups.google.com/**group****
>>>>>>>>>> /spyderlib <http://groups.google.com/group/spyderlib>.
>>>>>>>>>> For more options, visit https://groups.google.com/**grou****
>>>>>>>>>> ps/opt_out <https://groups.google.com/groups/opt_out>.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>  -- 
>>>>>>>>> You received this message because you are subscribed to the Google 
>>>>>>>>> Groups "spyder" group.
>>>>>>>>> To unsubscribe from this group and stop receiving emails from it, 
>>>>>>>>> send an email to spyderlib+...@**googlegroups.com****.
>>>>>>>>> To post to this group, send email to [email protected].
>>>>>>>>>
>>>>>>>>> Visit this group at 
>>>>>>>>> http://groups.google.com/**group****/spyderlib<http://groups.google.com/group/spyderlib>
>>>>>>>>> .
>>>>>>>>> For more options, visit https://groups.google.com/**grou****
>>>>>>>>> ps/opt_out <https://groups.google.com/groups/opt_out>.
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>  -- 
>>>>>>> You received this message because you are subscribed to the Google 
>>>>>>> Groups "spyder" group.
>>>>>>> To unsubscribe from this group and stop receiving emails from it, 
>>>>>>> send an email to spyderlib+...@**googlegroups.com**.
>>>>>>> To post to this group, send email to [email protected].
>>>>>>> Visit this group at 
>>>>>>> http://groups.google.com/**group**/spyderlib<http://groups.google.com/group/spyderlib>
>>>>>>> .
>>>>>>> For more options, visit 
>>>>>>> https://groups.google.com/**grou**ps/opt_out<https://groups.google.com/groups/opt_out>
>>>>>>> .
>>>>>>>
>>>>>>
>>>>>>
>>>>>  -- 
>>>> You received this message because you are subscribed to the Google 
>>>> Groups "spyder" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send 
>>>> an email to spyderlib+...@**googlegroups.com.
>>>> To post to this group, send email to [email protected].
>>>> Visit this group at 
>>>> http://groups.google.com/**group/spyderlib<http://groups.google.com/group/spyderlib>
>>>> .
>>>> For more options, visit 
>>>> https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>
>>>> .
>>>>
>>>
>>>  -- 
>> You received this message because you are subscribed to the Google Groups 
>> "spyder" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to [email protected] <javascript:>.
>> To post to this group, send email to [email protected]<javascript:>
>> .
>> Visit this group at http://groups.google.com/group/spyderlib.
>> For more options, visit https://groups.google.com/groups/opt_out.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"spyder" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/spyderlib.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to