pdb/ipdb for python break on editable condition

Question

Say I have code the following code:

for i in range(100):
    print i

In general I can add one line to the code as:

for i in range(100):
    import ipdb;ipdb.set_trace()
    print i

However, now I want to debug it at condition of i == 10, and I don't want to bother by typing c for 10 times in ipdb, how should I do?

In the documentation I found condition bpnumber [condition], but how could I know the bpnumber if there is no list of bpnumber index. The documentation also says b(reak) ([file:]lineno | function) [, condition]. For example, assume the line number of print i is xx. I entered the following in ipdb shell: b xx, i == 10 but nothing as expected happened.

Answer 1

There's a quick dirty way like this:

for i in range(100):
    if i == 10: import ipdb;ipdb.set_trace()
    print i

It works and don't have to busy your mind with any other commands :)

Answer 2

I think you were looking for a more direct solution that did not involve adding lines to the code, and just involved debugger commands.

Your original example of

b xx, i == 10 

doesn't work, because you are setting a breakpoint at the place in your code you inserted the ipdb.set_trace() command. By adding the statement 'b xx, i == 10' in the debugger, you actually have 2 break points (1 conditional and 1 unconditional) defined at the same location (assuming xx is the line were the set_trace() command is).

Alternatively, once you have defined breakpoints in your code using the 'b' command, which apparently works for you. You can add a condition to the breakpoint by

condition bpnumber boolean-expression

for example

condition 1 i == 10

Note: the bpnumber is the number assigned to the breakpoint, not the line in your code. To see a list of breakpoints, just type 'b' with no arguments.

Also, if you want to enter debug mode without using ipdb.set_trace(), you simply run your code with the pdb/ipbd module enabled

python -m pdb foo.py

Answer 3

alright, i did some exploration myself, here is my new understanding of pdb. when you input import ipdb;ipdb.set_trace() you actually add an entry point of ipdb to the line, not really a breakpoint. after you enter ipdb, you can then setup breakpoints

so, to realize what i want for conditional debugging, i should add as follows

import ipdb;ipdb.set_trace()
for i in range(100):
    print i

then after i enter ipdb, i can input b xx, i == 10, and then c or r to run the code, then the code will stop when the condition is met. when I input l, the bpnumber is shown for the line as :

          xx-1                  for i in range(100): 
bpnumber> xx                        print i
          xx+1                      ...

i have to say, the document and all other explanations are so confusing, i wish my answer here clarifies the difference between the "debug entry point" and "debug breakpoint"