Did it work with a unicode build of 2.5.3.1?
It did.
You don't care about backwards compatibility, do you?
I find it a bit strange that this error occurrs even in PyCrust etc. that are official part of wxPython.
If Stani would fix this issues, would it work also on older wx versions?
(SPE says it should run on 2.5.2.8)
unicode objects to/from wxString objects. Previously the default
Python encoding was always used, but unless the user had tweaked their
sitecustomize.py file it is always "ascii", which would result in
errors if the strings contained character codes >= 128.
SPE writes the user default into sitecustomize.py and relies on that.
Now the startup code is going to set the default encoding
wxPython uses to convert to/from unicode to the value of
locale.getdefaultlocale()[1]
which is certainly not utf8 for you.
Alas, no. I tried to set LOCALE and LANG variables in Windows, but that doesn't help.
Any ideas? (I'm working on WinXP, but my home system is MacOS X...)
It also depends on how SPE and etc. are loading the files. If
I forwarded your hint to the SPE bug tracker.
Anyway, try adding something like this to the SPE code after
it imports
wx and see if that fixes it for you:
wx.SetDefaultPyEncoding("utf8")
In Child.py it helps. Thank you.
That doesn't fix the Notebook sizer problem (I didn't mention yet), but it's a step on the way.
Best regards,
Henning Hraban Ramm
Südkurier Medienhaus / MediaPro
Support/Admin/Development Dept.