[IPython-dev] [patch] IPython.set_trace() implemented

[Gökhan SEVER]

>
> Great catch! You are absolutely right, ipython seems to be still
> active. Yes, so this approach has it's limitations, that it will alter
> the environment.
>
> Never the less, it actually doesn't hurt me at all, because I use
> set_trace() to debug some particular part of the code and once I am
> done, I don't care if the program crashes or what, since I want to
> exit it anyway.
>

You are welcome :) To tell the truth I don't care too much about this
either, only for the sake of the integrity of the work.


>
> So this should be mentioned in the docstring too. I'll wait for
> ipython developers if they agree with the patch and if so, I'll update
> the docstring with this information.
>
> Ondrej
>

I might have devised a way to launch a pylab session from Python following
your footsteps. What I did was in the Shell.py
replaced the below line as is shown (My matplotlibrc file uses QT4 backend
currently but it works with WX too, so Mayavi will not complain :)

    #user_opts = set([s.replace('-','') for s in argv[:3]])
    user_opts = set(['pylab'])

Then replaced __init.py__ set_trace()

    #Shell.IPShell(user_ns=dict(globals, **locals)).mainloop()
    Shell.start(user_ns=dict(globals, **locals)).mainloop()

now IPython.set_trace() works fine with pylab too :), however this also
shows the similar strange behaviours as Shell.IPShell case...

Hope someone from the lead IPy dev team response this topic and the patch
will be neatly added into the code-base.

Gökhan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ipython-dev/attachments/20090623/059b018b/attachment.html>