[IPython-dev] Re: [SciPy-dev] Generic gui_thread + IPython: solution already exists!

[Prabhu Ramachandran]

>>>>> "JH" == John Hunter <jdhunter at ace.bsd.uchicago.edu> writes:

>>>>> "Prabhu" == Prabhu Ramachandran <prabhu_r at users.sf.net> writes:
    Prabhu> OK, so I assume that you do have a wx._core.PyApp_MainLoop
    Prabhu> attribute, right?

    JH> Yes

Thanks.

    JH> I don't have _tkinter.so.  What does that mean?  Perhaps my tk
    JH> is compiled statically?

ldd on the Python executable and try.

[...]
    JH> This may be my problem - I don't seem to have pthread compiled
    JH> in
[...]
    JH> peds-pc311:~> ldd /usr/lib/libtcl8.4.so
    JH>         libdl.so.2 => /lib/libdl.so.2 (0xb7543000) libm.so.6
    JH>         => /lib/tls/libm.so.6 (0xb7521000) libc.so.6 =>
    JH>         /lib/tls/libc.so.6 (0xb73e9000) /lib/ld-linux.so.2 =>
    JH>         /lib/ld-linux.so.2 (0x80000000)

Yes, that is a possibility.  Actually, since the error is random it
might be a little subtler.  update_tk() is run outside of runcode.  So
this might be breaking thread safety somehow.  I've attached a
modified Shell.py that tries to do the right thing by calling
update_tk from inside runcode.  Please let me know if it gets any
better and if there are no freezes anymore.  This *might* just do the
trick.

Thanks!

cheers,
prabhu

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: Shell.py
URL: <http://mail.python.org/pipermail/ipython-dev/attachments/20041116/6a6303c8/attachment.ksh>