[wxPython] Problems with wxPython 2.2.5 under Linux.

hi,

I have been using wxPython-2.2.5 (sparingly) for a while under Debian
GNU/Linux 2.2 (stable).

I have the following installed:

Python 1.5.2 (standard with the distro)
Python 2.1 (compiled and installed)
libgtk1.2-1.2.8-1 (i compiled the debian source package from testing
                   and installed it)
libwxgtk2.2-python_2.2.5.1 (similar to the above).
XFree86 4.0.2-7 (using C.P. Botha's debs for potato)

I also compiled wxPython2.2.5 separately for Python2.1 and installed
it appropriately.

I had posted part of these errors a while back (sometime in March). I
recently saw that some folks were using Debian so thought I could
verify if these problems were common to us all.

All of these problems are seen with Python1.5.2 and Python2.1 and are
demonstrated with demo.py.

     (1) The tree widget has problems and clicking on the +/- thing
     doesnt work all the time. You have to click on another node and
     then click back on the node you want. For instance I want to
     look at the "controls" so I click on the '+' then I dont want it
     and click on the '-' but it wont do anything if i click on the
     '-'. I have to click on another node and then try again.

     (2) The windows do not seem to be repainting properly. For
     instance as soon as I start the demo.py I get this "Tip of the
     day" and when I close it part of the background notebook widget
     is not repainted to be white and becomes a grey colour. The rest
     of the background (just behind the tip of the day window) is
     white. I hope I am clear.

     (3) Click on Controls->wxToolBar. On the toolbar frame that pops
     up just check a setting on the combobox and close the frame. I
     get a segfault. Here is the backtrace.

Program received signal SIGSEGV, Segmentation fault.
0x407f0223 in wxTextCtrl::DoGetBestSize () from /usr/lib/libwx_gtk-2.2.so.0
(gdb) back
#0 0x407f0223 in wxTextCtrl::DoGetBestSize () from /usr/lib/libwx_gtk-2.2.so.0
#1 0x40c031fa in g_main_set_poll_func () from /usr/lib/libglib-1.2.so.0
#2 0x40c022d9 in g_get_current_time () from /usr/lib/libglib-1.2.so.0
#3 0x40c028e3 in g_get_current_time () from /usr/lib/libglib-1.2.so.0
#4 0x40c02a7c in g_main_run () from /usr/lib/libglib-1.2.so.0
#5 0x40b281e7 in gtk_main () from /usr/lib/libgtk-1.2.so.0
#6 0x407bbe28 in wxApp::MainLoop () from /usr/lib/libwx_gtk-2.2.so.0
#7 0x404fdafe in wxPyApp::MainLoop (this=0x83d32d0) at src/helpers.cpp:89
<...>

My questions are:

(1) Are these problems linux specific? If so are they debian
specific?

(2) Where could the problems be? In gtk or wxGtk or wxPython?

(3) Are these issues solved with the 2.3.0 release?

Any help would be appreciated.

thanks,
prabhu

All of these problems are seen with Python1.5.2 and Python2.1 and are
demonstrated with demo.py.

     (1) The tree widget has problems and clicking on the +/- thing
     doesnt work all the time. You have to click on another node and
     then click back on the node you want. For instance I want to
     look at the "controls" so I click on the '+' then I dont want it
     and click on the '-' but it wont do anything if i click on the
     '-'. I have to click on another node and then try again.

     (2) The windows do not seem to be repainting properly. For
     instance as soon as I start the demo.py I get this "Tip of the
     day" and when I close it part of the background notebook widget
     is not repainted to be white and becomes a grey colour. The rest
     of the background (just behind the tip of the day window) is
     white. I hope I am clear.

Hmm... I was sure that both of these were solved in 2.2.5 or before. At
least #1 was just a problem in the demo, (#2 may have been too, I don't
recall.) Do you have the 2.2.5 version of the demo too? If not you can get
it here:

http://prdownloads.sourceforge.net/wxpython/wxPython-demo-2.2.5.tar.gz

     (3) Click on Controls->wxToolBar. On the toolbar frame that pops
     up just check a setting on the combobox and close the frame. I
     get a segfault. Here is the backtrace.

Program received signal SIGSEGV, Segmentation fault.
0x407f0223 in wxTextCtrl::DoGetBestSize () from

/usr/lib/libwx_gtk-2.2.so.0

(gdb) back
#0 0x407f0223 in wxTextCtrl::DoGetBestSize () from

/usr/lib/libwx_gtk-2.2.so.0

#1 0x40c031fa in g_main_set_poll_func () from /usr/lib/libglib-1.2.so.0
#2 0x40c022d9 in g_get_current_time () from /usr/lib/libglib-1.2.so.0
#3 0x40c028e3 in g_get_current_time () from /usr/lib/libglib-1.2.so.0
#4 0x40c02a7c in g_main_run () from /usr/lib/libglib-1.2.so.0
#5 0x40b281e7 in gtk_main () from /usr/lib/libgtk-1.2.so.0
#6 0x407bbe28 in wxApp::MainLoop () from /usr/lib/libwx_gtk-2.2.so.0
#7 0x404fdafe in wxPyApp::MainLoop (this=0x83d32d0) at src/helpers.cpp:89
<...>

This is one I havn't heard of. It may be some issue of the build as I don't
think g_main_set_poll_func should be calling wxTextCtrl::DoGetBestSize, but
since it is a virtual something like mismatched headers or compilation flags
could be causing a seg fault like this.

···

--
Robin Dunn
Software Craftsman
robin@AllDunn.com Java give you jitters?
http://wxPython.org Relax with wxPython!

hi,

    > Hmm... I was sure that both of these were solved in 2.2.5 or
    > before. At least #1 was just a problem in the demo, (#2 may
    > have been too, I don't recall.) Do you have the 2.2.5 version
    > of the demo too? If not you can get it here:

Yes, the demo that i am using is the one that ships with
wxPython2.2.5.

<snip>

    >> Program received signal SIGSEGV, Segmentation fault.
    >> 0x407f0223 in wxTextCtrl::DoGetBestSize () from
    > /usr/lib/libwx_gtk-2.2.so.0
    >> (gdb) back #0 0x407f0223 in wxTextCtrl::DoGetBestSize () from
    > /usr/lib/libwx_gtk-2.2.so.0
    >> #1 0x40c031fa in g_main_set_poll_func () from
    >> /usr/lib/libglib-1.2.so.0 #2 0x40c022d9 in g_get_current_time
    >> () from /usr/lib/libglib-1.2.so.0 #3 0x40c028e3 in

<snip>

    > This is one I havn't heard of. It may be some issue of the
    > build as I don't think g_main_set_poll_func should be calling
    > wxTextCtrl::DoGetBestSize, but since it is a virtual something
    > like mismatched headers or compilation flags could be causing
    > a seg fault like this.

Yes, it is vague. I was hoping other linux+wxPython users could
confirm such behaviour.

thanks,
prabhu