A user informed me that my program is stalling when he tries to run it, taking several minutes to show anything on the screen. When he gives up and types Control-C, it ends here:
from wx._core import *
File “/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/_core.py”,
line 4, in
import core
KeyboardInterrupt
At the point in the application where it stalls, I haven’t done anything other than import modules, most of them relatively innocuous. I’ve heard a few other reports of this error, on very diverse systems (a workstation on a company network, and a professor’s laptop unplugged from the network), but all Linux. Also reported by two of the same people is a file dialog taking several minutes to open (after the app finally loads). I’m using Python 2.6.5, wxPython 2.8.10.1, and I think the Linux systems are usually some variant of RedHat/Fedora/CentOS. Has anyone else encountered strange issues like this? We’ve assumed it’s some kind of network weirdness, but the program doesn’t need to be installed on NFS for this to happen.
I've had a file dialog stall or crash when I had a mercurial shell
extension installed, but that's on Windows...not sure if there's
something similar that can happen on Linux or not. Are you doing a
recursive import somewhere where moduleA import moduleB and vice-
versa?
···
On Jul 27, 6:29 pm, Nat Echols <nathaniel.ech...@gmail.com> wrote:
A user informed me that my program is stalling when he tries to run it,
taking several minutes to show anything on the screen. When he gives up and
types Control-C, it ends here:
. . .
File
"/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/__init__.py",
from wx._core import *
File
"/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/_core.py",
line 4, in <module>
import _core_
KeyboardInterrupt
At the point in the application where it stalls, I haven't done anything
other than import modules, most of them relatively innocuous. I've heard a
few other reports of this error, on very diverse systems (a workstation on a
company network, and a professor's laptop unplugged from the network), but
all Linux. Also reported by two of the same people is a file dialog taking
several minutes to open (after the app finally loads). I'm using Python
2.6.5, wxPython 2.8.10.1, and I think the Linux systems are usually some
variant of RedHat/Fedora/CentOS. Has anyone else encountered strange issues
like this? We've assumed it's some kind of network weirdness, but the
program doesn't need to be installed on NFS for this to happen.
A user informed me that my program is stalling when he tries to run it,
taking several minutes to show anything on the screen. When he gives up and
types Control-C, it ends here:
. . .
File
"/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/__init__.py",
from wx._core import *
File
"/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/_core.py",
line 4, in <module>
import _core_
KeyboardInterrupt
At the point in the application where it stalls, I haven't done anything
other than import modules, most of them relatively innocuous. I've heard a
few other reports of this error, on very diverse systems (a workstation on a
company network, and a professor's laptop unplugged from the network), but
all Linux. Also reported by two of the same people is a file dialog taking
several minutes to open (after the app finally loads). I'm using Python
2.6.5, wxPython 2.8.10.1, and I think the Linux systems are usually some
variant of RedHat/Fedora/CentOS. Has anyone else encountered strange issues
like this? We've assumed it's some kind of network weirdness, but the
program doesn't need to be installed on NFS for this to happen.
thanks,
Nat
I've had a file dialog stall or crash when I had a mercurial shell
extension installed, but that's on Windows...not sure if there's
something similar that can happen on Linux or not. Are you doing a
recursive import somewhere where moduleA import moduleB and vice-
versa?
steven@dm-steven:~$ python
Python 2.6.4 (r264:75706, Dec 7 2009, 18:43:55)
[GCC 4.4.1] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Does it make any difference if you use a unicode build of wxPython?
···
On 7/27/10 4:29 PM, Nat Echols wrote:
A user informed me that my program is stalling when he tries to run it,
taking several minutes to show anything on the screen. When he gives up
and types Control-C, it ends here:
. . .
File
"/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/__init__.py",
from wx._core import *
File
"/usr/local/phenix-1.6.3-473/build/intel-linux-2.6-x86_64/base/lib/python2.6/site-packages/wx-2.8-gtk2-ansi/wx/_core.py",
line 4, in <module>
import _core_
KeyboardInterrupt
At the point in the application where it stalls, I haven't done anything
other than import modules, most of them relatively innocuous. I've
heard a few other reports of this error, on very diverse systems (a
workstation on a company network, and a professor's laptop unplugged
from the network), but all Linux. Also reported by two of the same
people is a file dialog taking several minutes to open (after the app
finally loads). I'm using Python 2.6.5, wxPython 2.8.10.1, and I think
the Linux systems are usually some variant of RedHat/Fedora/CentOS. Has
anyone else encountered strange issues like this? We've assumed it's
some kind of network weirdness, but the program doesn't need to be
installed on NFS for this to happen.
It’s possible, but I don’t think so. My understanding was that doing this will usually just fail. I do a lot of importing “inline” in the body of functions, mainly to cut down on the startup time (I use several massive C++ extensions that can take several seconds to load).
Just to clarify, I’ve never been able to independently reproduce this problem - the unnetworked laptop is the only computer that I’ve ever had personal access to, and my attempt to debug it was unsuccessful. I did, however, determine that just running “python -c ‘import wx’” would stall in the same place. So the wxPython C++ backend is doing something that the computers don’t like. Running with PYTHONVERBOSE=1 doesn’t tell me anything I didn’t already know.
Another clarification: I’m distributing a complete Python installation including all required modules, and everything is launched by shell scripts which set PYTHONPATH on the fly. This guarantees compatibility with a wide range of Linux systems, and avoids polluting the user environment with our PYTHONPATH. It does not, however, entirely avoid existing definitions of PYTHONPATH (it will insert our module directories ahead of the user-defined paths), or user modifications to sys.path in ~/.pythonrc - but in the one case I’m currently investigating, neither of these scenarios appears to be happening.
thanks,
Nat
···
On Wed, Jul 28, 2010 at 6:52 AM, Mike Driscoll kyosohma@gmail.com wrote:
I’ve had a file dialog stall or crash when I had a mercurial shell
extension installed, but that’s on Windows…not sure if there’s
something similar that can happen on Linux or not. Are you doing a
recursive import somewhere where moduleA import moduleB and vice-
I haven’t tried this yet. I need to ask a colleague if this is even possible in our environment, because we depend on many Boost.Python extensions and I don’t know how those will interact with the unicode.
-Nat
···
On Wed, Jul 28, 2010 at 9:45 AM, Robin Dunn robin@alldunn.com wrote:
Does it make any difference if you use a unicode build of wxPython?