[IPython-user] ipython.el + pdb rocks!!!

Fernando Perez Fernando.Perez at colorado.edu
Thu Feb 10 19:10:41 CST 2005


Hey Alex,

[cc-ed to the list, in case someone else wants to try this out and have the 
same epiphany I've had today]

I can't thank you enough for ipython.el and the pdb integration.  I know I 
should have tried this long ago, you'd said as much.

Today I've been wrestling a really hard structural bug in a fairly complex 
code, and I was going crazy tracking down the call chain, even with manual 
%pdb usage in ipython.  There were just too jumps I needed to understand, and 
the 'l' command in pdb, while useful, is sometimes just not enough.

I remembered you had mentioned, several times, how your setup involved %pdb 
but within ipython.el.  I tracked (via an old response of yours on the list) 
how to activate pylab with C-c !, and reran my nasty code with %pdb on.  In 
about 30 minutes I've been able to tame this beast, which I knew was going to 
be a major nightmare with the regular tools.  This setup absolutely ROCKS, the 
automatic source jumping with u/d, along with the (still active) ability for 
variable examination, are a dream: you can fly through the source and memory 
state of an entire program in a really incredible way.

Anyway, this is just a bit of fan mail :)

Best,

f

ps. Now that I've suckered up to you, let me ask a question: the one thing 
that's annoying me is keyboard management.  In a terminal, ctrl p/n do 
previous/next _only_ in the history that matches the current line.  Is it 
possible to rebind the keys in Xemacs so that, in an ipython buffer, you get 
also ctrl p/n to do history matching (like ctrl up/down do?).  I'm OK with 
losing ctrl p/n for moving the text caret, the arrow keys still do that.  It's 
just that my muscle memory is tripping me badly, when  I switch from ipython 
in a terminal to ipython in an emacs buffer.




More information about the IPython-user mailing list