pydb [was: The DDD tree ...]

David Relson writes:
 > Perchance, does the latest pydb remember breakpoints after a restart
 > command?  That's the feature I'd most like to see!

It's there. The older pydb didn't have a gdb "run" command, so ddd
*had* to do an "exec" of the program to restart.  The exec breakpoints
and any other debugger state. 

Interestly, the pydb program didn't offer a "restart" as a debugger
command -- ddd did this *outside* of the pydb. (So be grateful that
ddd even gave you that, because outside of ddd there wasn't even
that and still isn't in the stock python debugger pdb.)

When I started working on a Python debugger, the very first thing I
added *was* "run". In fact it was added even before I called the
debugger pydb! (Originally all I was doing was extending the stock
debugger pdb; there's a patch submitted around Jan '06 to the Python
pdb sources gathering dust.)

In current pydb, there is also a "restart" command which does a
re-exec. And sometimes that's really what you want. For example
"imports" occur only once so "restart" would be needed here if you
wanted to debug into an import.

I mention this in the litle demo of some pydb features that I made 
recently:

http://showmedo.com/videos/video?name=pythonBernsteinPydbIntro&fromSeriesID=28

 > 
 > Also, whenever I've tried the pp command, the result is the same as
 > the p command.  

It helps to understand what each does. ;-) 

And you can find the difference described somewhat cryptically here:
   http://bashdb.sourceforge.net/pydb/pydb/lib/subsection-data.html

(In the next pydb release I've expanded that a little to add a link to
the corresponding pprint documentation.)

 > Is there anything special I should do to get
 > different/better output?

If you have Python 2.5 installed that will give different/better
output for dictionaries. :-)

Personally, I don't use "p" all that much. The only thing it provides
is a repr() wrapping which elides output and ensures circular
references don't loop indefinitely. But most of the time I don't want
that as elision, and usually my objects aren't circular.

And I when want more detail about an object, I use examine, which uses
pprint output.