On 12-02-14 01:53 PM, Joel Brobecker wrote: >> ChangeLog: > > I haven't looked at the validity of the patch (Pedro has a better > understanding of this area, for instance), but I still noticed some > trivial deviations from the GNU Coding Standards. > > Your ChangeLog entry, for instance. Lines should be folded at 70 chars > (hard limit is 80 chars). Other violations highlighted below: Ok. > >> + /* Set pspace with frame's pspace */ > > End the sentence with a period (and two spaces before the '*/'). I removed the comment. It is stating what code unambiguously states already. > >> + if (valid&& pspace == NULL) { >> + warning(_("Trying to set NULL pspace.")); >> + } > > Wrong formatting for first and second line. Yes, thank you. Moved 'warning' and check down. > > Why are you adding an empty line here? > >> # Copyright 2012 Free >> # Software Foundation, Inc. > > Missing (C), and please join the two lines. If you copied some of > the testcase from another testcase, then you need to preserve the > original copyright years, I think. Added '(C)'. I copied only the copyright notice, the tests are new. >> #include > > Is there a way to trigger the same problem without the dependency > on stdio.h? Many systems do not provide it (bareboard). I would > think that all you need is to define a function that has the same > profile as printf, no? > Yes, I removed printf and stdio, they are not important and the issue is still reproduced (the only important thing is that foo actually gets inlined which -O2 optimization level should do). New patch, fixed tests attached. Change logs: 2012-02-14 Aleksandar Ristovski * frame.c (find_frame_sal): Initialise sal->pspace field from frame data. * stack.c (set_last_displayed_sal): Perform sanity check of the data passed in, in particular, validate that PSPACE is not NULL if requesting valid last_displayed_* data. Testsuite: 2012-02-14 Aleksandar Ristovski * gdb.base/break-inline.exp: New test. * gdb.base/break-inline.c: New test. Thanks, Aleksandar