You can clone with
Using latest code on ubuntu 10.10 and pdbtrack doesn't work... Are others seeing this too?
for me it's working. I'm on Arch GNU/Linux.
Tested with python 3 and 2. Standard pdb and ipdb.
I'm closing this until you provide more details.
Here's how I reproduce on Gentoo and Ubuntu, python 2.x.
Create a test file (/tmp/test.py)
import pdb; pdb.set_trace()
Type M-x shell and run python /tmp/test.py from the shell. pdbtrack used to split the buffer at this point and trace through the code. This no longer works.
I see what's going on. Pdbtrack is not hooked to the standard shell as it might be in other python modes. In this python.el, pdbtrack is only activated for python comint processes (the ones this mode starts), so what you want is to open the file and do C-c C-c (python-shell-send-buffer). That would do the trick.
BTW it should be possible to add the pdbtrack filter function globally, but haven't tested and I don't think I would activate such thing by default.
Hmmm, I would consider this a regression. It used to work. Also, python-mode.el has this feature, which people are using. If you want to be a good candidate for drop-in replacement for it, I would enable this by default.
An example of where I use it frequently is debugging django. In that case I'm not wanting the execute the buffer, but run a command (python manage.py runserver) that will eventually hit a breakpoint in the buffer. Can you suggest another way to do this?