We Keep Coding

pdb bypass error/Jump failed: can only jump from a 'line' trace event

I'm trying to debug a Python program using pdb. The program could be like this:
def main():
a = 1
print(b)
c = 2
d = 3
Apparently, print(b) is a typo which should be print(a) but it is not important and I can fix it with the text editor but I want to bypass this error and continue debugging.
I tried jump, like jump 4(assuming "c=2" is line 4) but I was given error "Jump failed: f_lineno can only be set by a line trace function", which means I need to give a line trace function when I'm programming.
So, is there a way to deal with this problem, or is there some other way to bypass the error line when using pdb?

TLDR: this is pdb's post-mortem mode in which jumping is not supposed to work. But it's still very useful.
Painting by Rembrandt (public domain)
I reproduce it with python 3.8.2 as *** Jump failed: can only jump from a 'line' trace event by running the script "under pdb" like so: python3 -m pdb -c c script.py and trying to jump to another line in pdb prompt which then appears.
What's happened: an unhandled exception, in this case NameError: name 'b' is not defined caused python to stop interpreting the script; pdb intercepted this situation and entered its post-mortem mode.
As Almar Klein nicely put it in his blog post,
Post-mortem debugging refers to the concept of entering debug mode after something has broken. There is no setting of breakpoints involved, so it's very quick and you can inspect the full stack trace, making it an effective way of tracing errors.
Although jump, next, return won't work in post-mortem, bt, up, down, ll and pp, along with the possibility to import modules and run arbitrary python code directly in the pdb's interactive shell can be very effective ways to get the root cause. In our simple example the root cause of the NameError is shown immediately upon a quick ll: pdb prefixes the offending line of code with >>.
Had we not passed -c c (meaning continue), pdb would have shown its prompt and paused before the first line of your program is interpreted, so you'd have a chance to step through the whole program or set a breakpoint before or at the offending line, and jump over it, never entering the post-mortem.
Even in post-mortem, you can prepare a breakpoint anywhere in the program, e.g. break 2 for line 2, and say c or continue so pdb will finish post-mortem, reload the file, and restart the program with the updated set of breakpoints.
Another way to deal with it is to import pdb and pdb.set_trace() in suspicious code - or since python 3.7, simply breakpoint() - and run the python program normally (not "under" pdb anymore) which allows then to jump, next, return etc, as well as everything else - when the breakpoint is hit.
If your Python program is started through behave:
prefer to run behave with --no-capture whenever using pdb or similar debuggers (whether post-mortem mode or not), to avoid problems with behave's stdin/stdout capturing making pdb unresponsive and/or its prompt invisible.
best of all, if you want to end up in the pdb post-mortem mode automatically while potentially still supporting capturing, set the post_mortem environment variable (can also name it differently) to any value (but only on dev machine, not for automated CI or production!) and commit the following permanently into environment.py:
def after_step(context, step):
import os
if 'post_mortem' in os.environ and step.status == 'failed':
import pdb
# Similar to "behave --no-capture" calling stop_capture() ensures visibility of pdb's prompts,
# while still supporting capture until an uncaught error occurs.
# Warning: this does rely on behave's internals which might change
context._runner.stop_capture() # pylint: disable=protected-access
pdb.post_mortem(step.exc_traceback)

I'm not sure, but this may be a bug that was fixed in Mar 2018, so you may need to (patch, upgrade, reinstall?) your Python.

how to disable pdb.set_trace() without stopping python program and edit the code

I suspect that I have issue in one of my loops, so I setup a break points with pdb.set_trace()
import pdb
for i in range(100):
print("a")
pdb.set_trace()
print("b")
after check variable in this loop for a few times, I decide continue this programming without further breaks. So I try to get the break number with b command, no breaks listed. I guess this line of code don't setup a break point. but How Do I get ride of this "break points" without stopping the program and change the code?

to my knowledge, you could not bypass set_trace, but you could neutralize it, once debugger stopped, type:
pdb.set_trace = lambda: 1
then continue, it wont break again.

Setting a breakpoint (requires Python 3.7):
breakpoint()
Disabling breakpoints set with the breakpoint() function:
import os
os.environ["PYTHONBREAKPOINT"] = "0"
Long story:
In the 3.7 version of Python, the breakpoint() built-in function for setting breakpoints was introduced. By default, it will call pdb.set_trace(). Also, since the 3.7 version of Python the PYTHONBREAKPOINT environment variable is available. It is considered when the breakpoint() function is used.
So, in order to disable these breakpoints (set with the breakpoint() function), one can just set the PYTHONBREAKPOINT environment variable like this:
import os
os.environ["PYTHONBREAKPOINT"] = "0"
It may be useful to mention here sys.breakpointhook() which was also added in the 3.7 version of Python and allows to customize breakpoints behavior.

Unfortunately pdb is missing a bunch of functionality (even basic stuff like display lists), and you've found another example of that here. The good news is that pdb++ is a great drop-in replacement for pdb, and one of the things it solves is exactly the problem of disabling set_trace. So you can simply do:
pip install pdbpp
and then at the (Pdb++) prompt, type
pdb.disable()
Easy! And you will get lots of other useful goodies on top of that.

It is possible to start a Python script without PDB control but then hit a stray set_trace() left there. To prevent breaking into debugger every time set_trace() is encountered, a similar trick as above (changing the symbol's reference to point to a harmless function) can be applied.
However, the namespace of the debuggee has to be modified, not one of the debugger itself. Simply overwriting pdb.set_trace = lambda:1 or set_trace = lambda:1 did not work for me.
The following trick worked from the pdb prompt:
pdb> globals()['set_trace'] = lambda:1
This line first calls globals() to get access to a dict of the program under debugging, and then modifies the reference of set_trace there.

One way around this is to not write the breakpoints in the script itself, but rather set breakpoints when you start python with python -m pdb my_script.py
You then get into a prompt first, before execution of the script starts, and you can write for example
(Pdb) b 321
to set a breakpoint on line 321 (you can also specify file and specify conditions b 321, i == 50)
Then
(Pdb) c
(for continue) to start the execution of the actual script. The breakpoints you set in this way, you can clear when you're done with them with:
(Pdb) cl
(for clear)

Step-by-step debugging with IPython

From what I have read, there are two ways to debug code in Python:
With a traditional debugger such as pdb or ipdb. This supports commands such as c for continue, n for step-over, s for step-into etc.), but you don't have direct access to an IPython shell which can be extremely useful for object inspection.
Using IPython by embedding an IPython shell in your code. You can do from IPython import embed, and then use embed() in your code. When your program/script hits an embed() statement, you are dropped into an IPython shell. This allows the full inspection of objects and testing of Python code using all the IPython goodies. However, when using embed() you can't step-by-step through the code anymore with handy keyboard shortcuts.
Is there any way to combine the best of both worlds? I.e.
Be able to step-by-step through your code with handy pdb/ipdb keyboard shortcuts.
At any such step (e.g. on a given statement), have access to a full-fledged IPython shell.
IPython debugging as in MATLAB:
An example of this type of "enhanced debugging" can be found in MATLAB, where the user always has full access to the MATLAB engine/shell, and she can still step-by-step through her code, define conditional breakpoints, etc. From what I have discussed with other users, this is the debugging feature that people miss the most when moving from MATLAB to IPython.
IPython debugging in Emacs and other editors:
I don't want to make the question too specific, but I work mostly in Emacs, so I wonder if there is any way to bring this functionality into it. Ideally, Emacs (or the editor) would allow the programmer to set breakpoints anywhere on the code and communicate with the interpreter or debugger to have it stop in the location of your choice, and bring to a full IPython interpreter on that location.

What about ipdb.set_trace() ? In your code :
import ipdb; ipdb.set_trace()
update: now in Python 3.7, we can write breakpoint(). It works the same, but it also obeys to the PYTHONBREAKPOINT environment variable. This feature comes from this PEP.
This allows for full inspection of your code, and you have access to commands such as c (continue), n (execute next line), s (step into the method at point) and so on.
See the ipdb repo and a list of commands. IPython is now called (edit: part of) Jupyter.
ps: note that an ipdb command takes precedence over python code. So in order to write list(foo) you'd need print(list(foo)), or !list(foo) .
Also, if you like the ipython prompt (its emacs and vim modes, history, completions,…) it's easy to get the same for your project since it's based on the python prompt toolkit.

You can use IPython's %pdb magic. Just call %pdb in IPython and when an error occurs, you're automatically dropped to ipdb. While you don't have the stepping immediately, you're in ipdb afterwards.
This makes debugging individual functions easy, as you can just load a file with %load and then run a function. You could force an error with an assert at the right position.
%pdb is a line magic. Call it as %pdb on, %pdb 1, %pdb off or %pdb 0. If called without argument it works as a toggle.

(Update on May 28, 2016) Using RealGUD in Emacs
For anyone in Emacs, this thread shows how to accomplish everything described in the OP (and more) using
a new important debugger in Emacs called RealGUD which can operate with any debugger (including ipdb).
The Emacs package isend-mode.
The combination of these two packages is extremely powerful and allows one to recreate exactly the behavior described in the OP and do even more.
More info on the wiki article of RealGUD for ipdb.
Original answer:
After having tried many different methods for debugging Python, including everything mentioned in this thread, one of my preferred ways of debugging Python with IPython is with embedded shells.
Defining a custom embedded IPython shell:
Add the following on a script to your PYTHONPATH, so that the method ipsh() becomes available.
import inspect
# First import the embed function
from IPython.terminal.embed import InteractiveShellEmbed
from IPython.config.loader import Config
# Configure the prompt so that I know I am in a nested (embedded) shell
cfg = Config()
prompt_config = cfg.PromptManager
prompt_config.in_template = 'N.In <\\#>: '
prompt_config.in2_template = ' .\\D.: '
prompt_config.out_template = 'N.Out<\\#>: '
# Messages displayed when I drop into and exit the shell.
banner_msg = ("\n**Nested Interpreter:\n"
"Hit Ctrl-D to exit interpreter and continue program.\n"
"Note that if you use %kill_embedded, you can fully deactivate\n"
"This embedded instance so it will never turn on again")
exit_msg = '**Leaving Nested interpreter'
# Wrap it in a function that gives me more context:
def ipsh():
ipshell = InteractiveShellEmbed(config=cfg, banner1=banner_msg, exit_msg=exit_msg)
frame = inspect.currentframe().f_back
msg = 'Stopped at {0.f_code.co_filename} at line {0.f_lineno}'.format(frame)
# Go back one level!
# This is needed because the call to ipshell is inside the function ipsh()
ipshell(msg,stack_depth=2)
Then, whenever I want to debug something in my code, I place ipsh() right at the location where I need to do object inspection, etc. For example, say I want to debug my_function below
Using it:
def my_function(b):
a = b
ipsh() # <- This will embed a full-fledged IPython interpreter
a = 4
and then I invoke my_function(2) in one of the following ways:
Either by running a Python program that invokes this function from a Unix shell
Or by invoking it directly from IPython
Regardless of how I invoke it, the interpreter stops at the line that says ipsh(). Once you are done, you can do Ctrl-D and Python will resume execution (with any variable updates that you made). Note that, if you run the code from a regular IPython the IPython shell (case 2 above), the new IPython shell will be nested inside the one from which you invoked it, which is perfectly fine, but it's good to be aware of. Eitherway, once the interpreter stops on the location of ipsh, I can inspect the value of a (which be 2), see what functions and objects are defined, etc.
The problem:
The solution above can be used to have Python stop anywhere you want in your code, and then drop you into a fully-fledged IPython interpreter. Unfortunately it does not let you add or remove breakpoints once you invoke the script, which is highly frustrating. In my opinion, this is the only thing that is preventing IPython from becoming a great debugging tool for Python.
The best you can do for now:
A workaround is to place ipsh() a priori at the different locations where you want the Python interpreter to launch an IPython shell (i.e. a breakpoint). You can then "jump" between different pre-defined, hard-coded "breakpoints" with Ctrl-D, which would exit the current embedded IPython shell and stop again whenever the interpreter hits the next call to ipsh().
If you go this route, one way to exit "debugging mode" and ignore all subsequent breakpoints, is to use ipshell.dummy_mode = True which will make Python ignore any subsequent instantiations of the ipshell object that we created above.

You can start IPython session from pudb and go back to the debugging session as you like.
BTW, ipdb is using IPython behind the scenes and you can actually use IPython functionality such as TAB completion and magic commands (the one starts with %). If you are OK with ipdb you can start it from IPython using commands such as %run and %debug. ipdb session is actually better than plain IPython one in the sense you can go up and down in the stack trace etc. What is missing in ipdb for "object inspection"?
Also, python.el bundled with Emacs >= 24.3 has nice ipdb support.

Looks like the approach in #gaborous's answer is deprecated.
The new approach seems to be:
from IPython.core import debugger
debug = debugger.Pdb().set_trace
def buggy_method():
debug()

Prefixing an "!" symbol to commands you type in pdb seems to have the same effect as doing something in an IPython shell. This works for accessing help for a certain function, or even variable names. Maybe this will help you to some extent. For example,
ipdb> help(numpy.transpose)
*** No help on (numpy.transpose)
But !help(numpy.transpose) will give you the expected help page on numpy.transpose. Similarly for variable names, say you have a variable l, typing "l" in pdb lists the code, but !l prints the value of l.

You can start IPython from within ipdb.
Induce the ipdb debugger1:
import idpb; ipdb.set_trace()
Enter IPython from within in the ipdb> console2:
from IPython import embed; embed()
Return to the ipdb> console from within IPython:
exit
If you're lucky enough to be using Emacs, things can be made even more convenient.
This requires using M-x shell. Using yasnippet and bm, define the following snippet. This will replace the text ipdb in the editor with the set-trace line. After inserting the snippet, the line will be highlighted so that it is easily noticeable and navigable. Use M-x bm-next to navigate.
# -*- mode: snippet -*-
# name: ipdb
# key: ipdb
# expand-env: ((yas-after-exit-snippet-hook #'bm-toggle))
# --
import ipdb; ipdb.set_trace()
1 All on one line for easy deletion. Since imports only happen once, this form ensures ipdb will be imported when you need it with no extra overhead.
2 You can save yourself some typing by importing IPython within your .pdbrc file:
try:
from IPython import embed
except:
pass
This allows you to simply call embed() from within ipdb (of course, only when IPython is installed).

Did you try this tip?
Or better still, use ipython, and call:
from IPython.Debugger import Tracer; debug_here = Tracer()
then you can just use
debug_here()
whenever you want to set a breakpoint

the right, easy, cool, exact answer for the question is to use %run macro with -d flag.
In [4]: run -d myscript.py
NOTE: Enter 'c' at the ipdb> prompt to continue execution.
> /cygdrive/c/Users/mycodefolder/myscript.py(4)<module>()
2
3
----> 4 a=1
5 b=2

One option is to use an IDE like Spyder which should allow you to interact with your code while debugging (using an IPython console, in fact). In fact, Spyder is very MATLAB-like, which I presume was intentional. That includes variable inspectors, variable editing, built-in access to documentation, etc.

If you type exit() in embed() console the code continue and go to the next embed() line.

The Pyzo IDE has similar capabilities as the OP asked for. You don't have to start in debug mode. Similarly to MATLAB, the commands are executed in the shell. When you set up a break-point in some source code line, the IDE stops the execution there and you can debug and issue regular IPython commands as well.
It does seem however that step-into doesn't (yet?) work well (i.e. stopping in one line and then stepping into another function) unless you set up another break-point.
Still, coming from MATLAB, this seems the best solution I've found.

From python 3.2, you have the interact command, which gives you access to the full python/ipython command space.

Running from inside Emacs' IPython-shell and breakpoint set via pdb.set_trace() should work.
Checked with python-mode.el, M-x ipython RET etc.

Developing New Code
Debugging inside IPython
Use Jupyter/IPython cell execution to speed up experiment iterations
Use %%debug for step through
Cell Example:
%%debug
...: for n in range(4):
...: n>2
Debugging Existing Code
IPython inside debugging
Debugging a broken unit test: pytest ... --pdbcls=IPython.terminal.debugger:TerminalPdb --pdb
Debugging outside of test case: breakpoint(), python -m ipdb, etc.
IPython.embed() for full IPython functionality where needed while in the debugger
Thoughts on Python
I agree with the OP that many things MATLAB does nicely Python still does not have and really should since just about everything in the language favors development speed over production speed. Maybe someday I will contribute more than trivial bug fixes to CPython.
https://github.com/ipython/ipython/commit/f042f3fea7560afcb518a1940daa46a72fbcfa68
See also Is it possible to run commands in IPython with debugging?

If put import ipdb; ipdb.set_trace() at cell outside function, it will occur error.
Using %pdb or %debug, you can only see the filnal error result. You cannot see the code doing step by step.
I use following skill:
%%writefile temp.py
.....cell code.....
save the code of cell to file temp.py.
and then
%run -i -d temp.py, it will run the cell code by pdb .
-i: run the file in IPython’s namespace instead of an empty one.
-d: run your program under the control of pdb, the Python debugger.

Running a python debug session from a program, not from the console

I'm writing a little python IDE, and I want to add simple debugging. I don't need all the features of winpdb.
How do I launch a python program (by file name) with a breakpoint set at a line number so that it runs until that line number and halts?
Note that I don't want to do this from the command-line, and I don't want to edit the source (by inserting set_trace, for example). And I don't want it to stop at the first line so I have to run the debugger from there. I've tried all the obvious ways with pdb and bdb, but I must be missing something.

Pretty much the only viable way to do it (as far as I know) is to run Python as a subprocess from within your IDE. This avoids "pollution" from the current Python interpreter, which makes it fairly likely that the program will run in the same way as if you had started it independently. (If you have issues with this, check the subprocess environment.) In this manner, you can run a script in "debug mode" using
p = subprocess.Popen(args=[sys.executable, '-m', 'pdb', 'scriptname.py', 'arg1'],
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
This will start up Python at the debugger prompt. You'll need to run some debugger commands to set breakpoints, which you can do like so:
o,e = p.communicate('break scriptname.py:lineno')
If this works, o should be the normal output of the Python interpreter after it sets a breakpoint, and e should be empty. I'd suggest you play around with this and add some checks in your code to ensure whether the breakpoints were properly set.
After that, you can start the program running with
p.communicate('continue')
At this point you'd probably want to hook the input, output, and error streams up to the console that you're embedding in your IDE. You would probably need to do this with an event loop, roughly like so:
while p.returncode is None:
o,e = p.communicate(console.read())
console.write(o)
console.write(e)
You should consider that snippet to be effectively pseudocode, since depending on how exactly your console works, it'll probably take some tinkering to get it right.
If this seems excessively messy, you can probably simplify the process a bit using the features of Python's pdb and bdb modules (I'm guessing "Python debugger" and basic debugger" respectively). The best reference on how to do this is the source code of the pdb module itself. Basically, the way the responsibilities of the modules are split is that bdb handles "under the hood" debugger functionality, like setting breakpoints, or stopping and restarting execution; pdb is a wrapper around this that handles user interaction, i.e. reading commands and displaying output.
For your IDE-integrated debugger, it would make sense to adjust the behavior of the pdb module in two ways that I can think of:
have it automatically set breakpoints during initialization, without you having to explicity send the textual commands to do so
make it take input from and send output to your IDE's console
Just these two changes should be easy to implement by subclassing pdb.Pdb. You can create a subclass whose initializer takes a list of breakpoints as an additional argument:
class MyPDB(pdb.Pdb):
def __init__(self, breakpoints, completekey='tab',
stdin=None, stdout=None, skip=None):
pdb.Pdb.__init__(self, completekey, stdin, stdout, skip)
self._breakpoints = breakpoints
The logical place to actually set up the breakpoints is just after the debugger reads its .pdbrc file, which occurs in the pdb.Pdb.setup method. To perform the actual setup, use the set_break method inherited from bdb.Bdb:
def setInitialBreakpoints(self):
_breakpoints = self._breakpoints
self._breakpoints = None # to avoid setting breaks twice
for bp in _breakpoints:
self.set_break(filename=bp.filename, line=bp.line,
temporary=bp.temporary, conditional=bp.conditional,
funcname=bp.funcname)
def setup(self, f, t):
pdb.Pdb.setup(self, f, t)
self.setInitialBreakpoints()
This piece of code would work for each breakpoint being passed as e.g. a named tuple. You could also experiment with just constructing bdb.Breakpoint instances directly, but I'm not sure if that would work properly, since bdb.Bdb maintains its own information about breakpoints.
Next, you'll need to create a new main method for your module which runs it the same way pdb runs. To some extent, you can copy the main method from pdb (and the if __name__ == '__main__' statement of course), but you'll need to augment it with some way to pass in the information about your additional breakpoints. What I'd suggest is writing the breakpoints to a temporary file from your IDE, and passing the name of that file as a second argument:
tmpfilename = ...
# write breakpoint info
p = subprocess.Popen(args=[sys.executable, '-m', 'mypdb', tmpfilename, ...], ...)
# delete the temporary file
Then in mypdb.main(), you would add something like this:
def main():
# code excerpted from pdb.main()
...
del sys.argv[0]
# add this
bpfilename = sys.argv[0]
with open(bpfilename) as f:
# read breakpoint info
breakpoints = ...
del sys.argv[0]
# back to excerpt from pdb.main()
sys.path[0] = os.path.dirname(mainpyfile)
pdb = Pdb(breakpoints) # modified
Now you can use your new debugger module just like you would use pdb, except that you don't have to explicitly send break commands before the process starts. This has the advantage that you can directly hook the standard input and output of the Python subprocess to your console, if it allows you to do that.

Can I put a breakpoint in a running Python program that drops to the interactive terminal?

I'm not sure if what I'm asking is possible at all, but since python is an interpreter it might be. I'm trying to make changes in an open-source project but because there are no types in python it's difficult to know what the variables have as data and what they do. You can't just look up the documentation on the var's type since you can't be sure what type it is. I want to drop to the terminal so I can quickly examine the types of the variables and what they do by typing help(var) or print(var). I could do this by changing the code and then re-running the program each time but that would be much slower.
Let's say I have a program:
def foo():
a = 5
my_debug_shell()
print a
foo()
my_debug_shell is the function I'm asking about. It would drop me to the '>>>' shell of the python interpreter where I can type help(a), and it would tell me that a is an integer. Then I type 'a=7', and some 'continue' command, and the program goes on to print 7, not 5, because I changed it.

http://docs.python.org/library/pdb.html
import pdb
pdb.set_trace()

Here is a solution that doesn't require code changes:
python -m pdb prog.py <prog_args>
(pdb) b 3
Breakpoint 1 at prog.py:3
(pdb) c
...
(pdb) p a
5
(pdb) a=7
(pdb) ...
In short:
start your program under debugger control
set a break point at a given line of code
let the program run up to that point
you get an interactive prompt that let's you do what you want (type 'help' for all options)

Python 3.7 has a new builtin way of setting breakpoints.
breakpoint()
The implementation of breakpoint() will import pdb and call pdb.set_trace().
Remember to include the braces (), since breakpoint is a function, not a keyword.

A one-line partial solution is simply to put 1/0 where you want the breakpoint: this will raise an exception, which will be caught by the debugger. Two advantages of this approach are:
Breakpoints set this way are robust against code modification (no dependence on a particular line number);
One does not need to import pdb in every program to be debugged; one can instead directly insert "breakpoints" where needed.
In order to catch the exception automatically, you can simply do python -m pdb prog.py… and then type c(ontinue) in order to start the program. When the 1/0 is reached, the program exits, but variables can be inspected as usual with the pdb debugger (p my_var). Now, this does not allow you to fix things and keep running the program. Instead you can try to fix the bug and run the program again.
If you want to use the powerful IPython shell, ipython -pdb prog.py… does the same thing, but leads to IPython's better debugger interface. Alternatively, you can do everything from within the IPython shell:
In IPython, set up the "debug on exception" mode of IPython (%pdb).
Run the program from IPython with %run prog.py…. When an exception occurs, the debugger is automatically activated and you can inspect variables, etc.
The advantage of this latter approach is that (1) the IPython shell is almost a must; and (2) once it is installed, debugging can easily be done through it (instead of directly through the pdb module). The full documentation is available on the IPython pages.

You can run the program using pdb, and add breakpoints before starting execution.
In reality though, it's usually just as fast to edit the code and put in the set_trace() call, as another user stated.

Not sure what the real question is. Python gives you the 'pdb' debugger (google yourself) and in addition you can add logging and debug output as needed.