pgdb: Debugging Python extensions made easier
Hi guys, http://luaforge.net/frs/?group_id=327 pgdb.zip is an addition to scite-debug, which adds source debugging to the popular SciTE programmer's editor. gdbpy.zip is a standalone version which can be run from Emacs. These allow you to single-step from Python to C code in a debugger session. To use pgdb, install scite-debug (either on top of an existing SciTE installation or with one of the packaged all-in-one versions) and unzip into the same directory. gdbpy requires Lua 5.1, but there are no other dependencies. In both cases, read the release notes. I know the FAQ says that this is not possible, but the FAQ is somewhat outdated. GDB is quite happy with pending breakpoints to unresolved libraries, but you have to reassure it. This technique is a generalization of one that I used to solve the equivalent problem with debugging Lua extensions. steve d. -- http://mail.python.org/mailman/listinfo/python-list
Re: pgdb: Debugging Python extensions made easier
They are not insurmontable problems. But you will still see things like this: http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/66510 The point being that with a newish GDB (the current mingw GDB is still 5.x) there's no need for such tedious tricks, since it handles pending breakpoints fine. The additional trick I use for debugging plain non-debug python is to use symbol-file to load some arbitrary symbols, so GDB will not complain about breakpoints. steve d. On Apr 15, 2:41 am, Paul Rubin <http://[EMAIL PROTECTED]> wrote: > SteveD <[EMAIL PROTECTED]> writes: > > pgdb.zip is an addition to scite-debug, which adds source debugging to > > the popular SciTE programmer's editor. ... > > I know the FAQ says that this is not possible, but the FAQ is somewhat > > outdated. GDB is quite happy with pending breakpoints to unresolved > > libraries, but you have to reassure it. > > I'm not sure what FAQ or what is supposed to be impossible but > I've used gdb in the past to debug python extensions without > running into insurmountable problems that I remember. -- http://mail.python.org/mailman/listinfo/python-list
