QPaintDevice: Must construct a QApplication before a QPaintDevice

[boris.smirnov@gmail.com]

Hi all,

I have a python script that works without any problem on Windows but
with error:

QPaintDevice: Must construct a QApplication before a QPaintDevice

on Linux.

Where could be the problem?

Thanks.
Boris

[Diez B. Roggisch]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

boris.smirnov@g mail.com wrote:

Hi all,
>
I have a python script that works without any problem on Windows but
with error:
>
QPaintDevice: Must construct a QApplication before a QPaintDevice
>
on Linux.
>
Where could be the problem?

This is the problem:

You Must construct a QApplication before a QPaintDevice

Diez

[boris.smirnov@gmail.com]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Feb 27, 11:46 am, "Diez B. Roggisch" <d...@nospam.we b.dewrote:

boris.smir...@g mail.com wrote:

Hi all,

>

I have a python script that works without any problem on Windows but
with error:

>

QPaintDevice: Must construct a QApplication before a QPaintDevice

>

on Linux.

>

Where could be the problem?

>
This is the problem:
>
You Must construct a QApplication before a QPaintDevice
>
Diez

Yes that I can deduce, but why I'm asking is why it's different on
Windows and Linux. Should it not be platform independent?

[Diez B. Roggisch]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

>

Yes that I can deduce, but why I'm asking is why it's different on
Windows and Linux. Should it not be platform independent?

It should be. I've got no idea why it seems to work on windows but I can say
one thing for sure: that would be an artifact that you shouldn't rely on.
In Qt, you _always_ need a QApplication for anything. So just do as it
requires you to do: first, construct a QApplication. Then things will work.

Diez

[boris.smirnov@gmail.com]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Feb 27, 11:56 am, "Diez B. Roggisch" <d...@nospam.we b.dewrote:

Yes that I can deduce, but why I'm asking is why it's different on
Windows and Linux. Should it not be platform independent?

>
It should be. I've got no idea why it seems to work on windows but I can say
one thing for sure: that would be an artifact that you shouldn't rely on.
In Qt, you _always_ need a QApplication for anything. So just do as it
requires you to do: first, construct a QApplication. Then things will work.
>
Diez

Hmm, as I see the code, I do construct QApplication as first

if __name__ == '__main__':
a = QApplication (sys.argv)
mywidget = Optimizer()
a.setMainWidget (mywidget)
mywidget.show()
Update_StatusLi ne(mywidget)
mywidget.setSta rtconfig()
a.exec_loop ()

[Diez B. Roggisch]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

boris.smirnov@g mail.com wrote:

On Feb 27, 11:56 am, "Diez B. Roggisch" <d...@nospam.we b.dewrote:

Yes that I can deduce, but why I'm asking is why it's different on
Windows and Linux. Should it not be platform independent?

>>
>It should be. I've got no idea why it seems to work on windows but I can
>say one thing for sure: that would be an artifact that you shouldn't rely
>on. In Qt, you _always_ need a QApplication for anything. So just do as
>it requires you to do: first, construct a QApplication. Then things will
>work.
>>
>Diez

>
Hmm, as I see the code, I do construct QApplication as first
>
if __name__ == '__main__':
a = QApplication (sys.argv)
mywidget = Optimizer()
a.setMainWidget (mywidget)
mywidget.show()
Update_StatusLi ne(mywidget)
mywidget.setSta rtconfig()
a.exec_loop ()

I don't see any QPaintDevice here. Where does that come from? You need to
give more information, a stack trace and a reduced example exhibiting the
behaviour.

Diez

[shredwheat]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Feb 27, 3:35 am, "Diez B. Roggisch" <d...@nospam.we b.dewrote:

I don't see any QPaintDevice here. Where does that come from? You need to
give more information, a stack trace and a reduced example exhibiting the
behaviour.

QWidget is derived from QPaintDevice, under Qt, no widgets can be
instantiated before the QApplication.

This source snippet looks like it should be working. If the exception
is happening inside Optimizer() than something really unsusual is
happening. I suspect there is python code somewhere else happening at
import time that is creating a widget. The exception traceback should
show you right where that would be.

[boris.smirnov@gmail.com]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice


shredwheat napísal(a):

On Feb 27, 3:35 am, "Diez B. Roggisch" <d...@nospam.we b.dewrote:

I don't see any QPaintDevice here. Where does that come from? You need to
give more information, a stack trace and a reduced example exhibiting the
behaviour.

>
QWidget is derived from QPaintDevice, under Qt, no widgets can be
instantiated before the QApplication.
>
This source snippet looks like it should be working. If the exception
is happening inside Optimizer() than something really unsusual is
happening. I suspect there is python code somewhere else happening at
import time that is creating a widget. The exception traceback should
show you right where that would be.

Thanks for the reply.
I'm not so familiar with python so could you explain me how to use and
become the exception traceback. I've tried today to read something for
the beginners about traceback and stack trace in the forum, but I
haven't found something really understadable for met.
Some piece of code for my case will help :) and I want to learn it of
course.

Thanks.
Boris

[shredwheat]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

When your programs stops with the error, it should also be printing a
stack trace. This is a list of all the functions that have been called
when Python had the problem.

You shouldn't have to do anything extra to get the stack trace.

[Phil Thompson]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Tuesday 27 February 2007 11:09 pm, shredwheat wrote:

When your programs stops with the error, it should also be printing a
stack trace. This is a list of all the functions that have been called
when Python had the problem.
>
You shouldn't have to do anything extra to get the stack trace.

The error is raised in Qt and aborts immediately. It never gets back to Python
to generate a trace.

He needs to produce a short and complete test which demonstrates the problem,
then we can point out where the QPaintDevice is being created.

Phil

[boris.smirnov@gmail.com]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Feb 28, 8:56 am, Phil Thompson <p...@riverbank computing.co.uk >
wrote:

On Tuesday 27 February 2007 11:09 pm, shredwheat wrote:
>

When your programs stops with the error, it should also be printing a
stack trace. This is a list of all the functions that have been called
when Python had the problem.

>

You shouldn't have to do anything extra to get the stack trace.

>
The error is raised in Qt and aborts immediately. It never gets back to Python
to generate a trace.
>
He needs to produce a short and complete test which demonstrates the problem,
then we can point out where the QPaintDevice is being created.
>
Phil

OK, but before I do a complete test, could anybody tell/explain me why
the same file is working on Windows?
Did anybody already meet with something similar Win vs. Linux?

b.

[boris.smirnov@gmail.com]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Feb 28, 9:07 am, boris.smir...@g mail.com wrote:

On Feb 28, 8:56 am, Phil Thompson <p...@riverbank computing.co.uk >
wrote:
>

On Tuesday 27 February 2007 11:09 pm, shredwheat wrote:

>

When your programs stops with the error, it should also be printing a
stack trace. This is a list of all the functions that have been called
when Python had the problem.

>

You shouldn't have to do anything extra to get the stack trace.

>

The error is raised in Qt and aborts immediately. It never gets back to Python
to generate a trace.

>

He needs to produce a short and complete test which demonstrates the problem,
then we can point out where the QPaintDevice is being created.

>

Phil

>
OK, but before I do a complete test, could anybody tell/explain me why
the same file is working on Windows?
Did anybody already meet with something similar Win vs. Linux?
>
b.

Here is my simple script:

import sys
from qt import *
class Optimizer(QWidg et):
def __init__(self, parent = 0):
QWidget.__init_ _(self)
QGridLayout(sel f)
if __name__ == '__main__':
a = QApplication (sys.argv)
mywidget = Optimizer()
a.exec_loop()

This produces this:

python qt_script_bs_07 0228.py

QPaintDevice: Must construct a QApplication before a QPaintDevice

Any suggestions here?
Thanks



BTW: One question:
when I use "import qt" instead of "from qt import *" I get this error:
Traceback (most recent call last):
File "mscarideidtool _bs_070228.py", line 4, in ?
class Optimizer(QWidg et):
NameError: name 'QWidget' is not defined

What is the difference between "import qt" and "from qt import *" ? I
thought that these are the same.

[Phil Thompson]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Wednesday 28 February 2007 9:00 am, boris.smirnov@g mail.com wrote:

On Feb 28, 9:07 am, boris.smir...@g mail.com wrote:

On Feb 28, 8:56 am, Phil Thompson <p...@riverbank computing.co.uk >

wrote:

On Tuesday 27 February 2007 11:09 pm, shredwheat wrote:
When your programs stops with the error, it should also be printing a
stack trace. This is a list of all the functions that have been
called when Python had the problem.

You shouldn't have to do anything extra to get the stack trace.
>
The error is raised in Qt and aborts immediately. It never gets back to
Python to generate a trace.
>
He needs to produce a short and complete test which demonstrates the
problem, then we can point out where the QPaintDevice is being created.
>
Phil

OK, but before I do a complete test, could anybody tell/explain me why
the same file is working on Windows?
Did anybody already meet with something similar Win vs. Linux?

b.

>
Here is my simple script:
>
import sys
from qt import *
class Optimizer(QWidg et):
def __init__(self, parent = 0):
QWidget.__init_ _(self)
QGridLayout(sel f)
if __name__ == '__main__':
a = QApplication (sys.argv)
mywidget = Optimizer()
a.exec_loop()
>
This produces this:

python qt_script_bs_07 0228.py

>
QPaintDevice: Must construct a QApplication before a QPaintDevice
>
Any suggestions here?

It works fine for me.

Thanks
>
>
>
BTW: One question:
when I use "import qt" instead of "from qt import *" I get this error:
Traceback (most recent call last):
File "mscarideidtool _bs_070228.py", line 4, in ?
class Optimizer(QWidg et):
NameError: name 'QWidget' is not defined
>
What is the difference between "import qt" and "from qt import *" ? I
thought that these are the same.

The first creates a new namespace called "qt" and imports the module's objects
into it. To reference those objects you have to include the namespace name.

The second imports the module's objects into the current namespace.

Phil

[boris.smirnov@gmail.com]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Feb 28, 10:22 am, Phil Thompson <p...@riverbank computing.co.uk >
wrote:

On Wednesday 28 February 2007 9:00 am, boris.smir...@g mail.com wrote:
>
>
>
>
>

On Feb 28, 9:07 am, boris.smir...@g mail.com wrote:

On Feb 28, 8:56 am, Phil Thompson <p...@riverbank computing.co.uk >

>

wrote:
On Tuesday 27 February 2007 11:09 pm, shredwheat wrote:
When your programs stops with the error, it should also be printing a
stack trace. This is a list of all the functions that have been
called when Python had the problem.

>

You shouldn't have to do anything extra to get the stack trace.

>

The error is raised in Qt and aborts immediately. It never gets back to
Python to generate a trace.

>

He needs to produce a short and complete test which demonstrates the
problem, then we can point out where the QPaintDevice is being created.

>

Phil

>

OK, but before I do a complete test, could anybody tell/explain me why
the same file is working on Windows?
Did anybody already meet with something similar Win vs. Linux?

>

b.

>

Here is my simple script:

>

import sys
from qt import *
class Optimizer(QWidg et):
def __init__(self, parent = 0):
QWidget.__init_ _(self)
QGridLayout(sel f)
if __name__ == '__main__':
a = QApplication (sys.argv)
mywidget = Optimizer()
a.exec_loop()

>

This produces this:

python qt_script_bs_07 0228.py

>

QPaintDevice: Must construct a QApplication before a QPaintDevice

>

Any suggestions here?

>
It works fine for me.
>

Thanks

>

BTW: One question:
when I use "import qt" instead of "from qt import *" I get this error:
Traceback (most recent call last):
File "mscarideidtool _bs_070228.py", line 4, in ?
class Optimizer(QWidg et):
NameError: name 'QWidget' is not defined

>

What is the difference between "import qt" and "from qt import *" ? I
thought that these are the same.

>
The first creates a new namespace called "qt" and imports the module's objects
into it. To reference those objects you have to include the namespace name.
>
The second imports the module's objects into the current namespace.
>
Phil- Hide quoted text -
>
- Show quoted text -

OK, I have to apologize because I didn't mention that I use python
version 2.2.1, could it be the problem here? Bugs or something? I have
to use this version since it was delivered with a software that we use
here.

[Phil Thompson]

Re: QPaintDevice: Must construct a QApplication before a QPaintDevice

On Wednesday 28 February 2007 9:26 am, boris.smirnov@g mail.com wrote:

On Feb 28, 10:22 am, Phil Thompson <p...@riverbank computing.co.uk >
>
wrote:

On Wednesday 28 February 2007 9:00 am, boris.smir...@g mail.com wrote:

On Feb 28, 9:07 am, boris.smir...@g mail.com wrote:
On Feb 28, 8:56 am, Phil Thompson <p...@riverbank computing.co.uk >

wrote:
On Tuesday 27 February 2007 11:09 pm, shredwheat wrote:
When your programs stops with the error, it should also be
printing a stack trace. This is a list of all the functions that
have been called when Python had the problem.

You shouldn't have to do anything extra to get the stack trace.
>
The error is raised in Qt and aborts immediately. It never gets
back to Python to generate a trace.
>
He needs to produce a short and complete test which demonstrates
the problem, then we can point out where the QPaintDevice is being
created.
>
Phil

OK, but before I do a complete test, could anybody tell/explain me
why the same file is working on Windows?
Did anybody already meet with something similar Win vs. Linux?

b.
>
Here is my simple script:
>
import sys
from qt import *
class Optimizer(QWidg et):
def __init__(self, parent = 0):
QWidget.__init_ _(self)
QGridLayout(sel f)
if __name__ == '__main__':
a = QApplication (sys.argv)
mywidget = Optimizer()
a.exec_loop()
>
This produces this:
python qt_script_bs_07 0228.py
>
QPaintDevice: Must construct a QApplication before a QPaintDevice
>
Any suggestions here?

It works fine for me.

Thanks
>
BTW: One question:
when I use "import qt" instead of "from qt import *" I get this error:
Traceback (most recent call last):
File "mscarideidtool _bs_070228.py", line 4, in ?
class Optimizer(QWidg et):
NameError: name 'QWidget' is not defined
>
What is the difference between "import qt" and "from qt import *" ? I
thought that these are the same.

The first creates a new namespace called "qt" and imports the module's
objects into it. To reference those objects you have to include the
namespace name.

The second imports the module's objects into the current namespace.

Phil- Hide quoted text -

- Show quoted text -

>
OK, I have to apologize because I didn't mention that I use python
version 2.2.1, could it be the problem here? Bugs or something? I have
to use this version since it was delivered with a software that we use
here.

So what versions of Qt, PyQt and SIP are you using? Were these included with
the software you are using? If so, what is that software?

Phil