I was looking through this huge LLDB python reference manual, and can't find still how to handle breakpoint hit with my code. I mean, I want to run my foo() when some breakpoint was hit.
Please help me with some code snippets!
Or point me the needed class in Python reference manual.
Thanx!
On the page:
http://lldb.llvm.org/python-reference.html
there is a section titled:
RUNNING A PYTHON SCRIPT WHEN A BREAKPOINT GETS HIT
which has some useful info.
What you are doing is "adding a command to your breakpoint". The lldb command for this is "breakpoint command add". All the basic lldb commands are in the form "noun [sub-noun [...]] verb options"; knowing that might help translate from the thing you want to do to where in the command set the command for that operation lives. Anyway, the help on "breakpoint command add" has other examples that might be useful.
Then condensed version is, make a python module containing a function with this signature:
breakpoint_function (frame, bp_loc, dict)
Bring that module into lldb with the command:
(lldb) command script import "path to your .py file"
You can also use the module name in "command script import" if you've set up the PYTHONPATH to point to it, but unlike Python's "import" you don't have to, the command will take care of that for you.
Then make a breakpoint, and use "br com a" to add your function to it:
(lldb) br com a -F MyModule.breakpoint_function
Now when a breakpoint gets hit, your function will get called with the following arguments:
The "frame" argument is the frame that hit the breakpoint. You can get the thread from the frame & thus the complete stack if you need it.
The "bploc" argument is the "Breakpoint Location" that hit the breakpoint. In lldb one "breakpoint specification" (which is what you are setting with "break set") can resolve to many locations. For instance, a "source pattern" breakpoint might match many source patterns in your code. So you might want to know which one was actually hit.
The "dict" option is so we can squirrel some stuff away and pass it to Python, it should be left alone.
One other thing to keep in mind is that though the script interpreter (accessible with the "script" command) defines lldb.thread, lldb.frame etc. convenience variables, these variables are NOT set up when your breakpoint command is running. So if you've used these variables in the script interpreter while prototyping your command, you'll have to find them from the frame you were passed in if you need them in the breakpoint command.
Note, Python breakpoint commands don't currently work in Xcode 6, though that should be fixed by the time it is done.