[patch] Fix breakpoints in unloaded shared libraries
Tue Dec 22 19:45:00 GMT 2009
On Tue, 22 Dec 2009 19:49:44 +0100, Tom Tromey wrote:
> >>>>> "Jan" == Jan Kratochvil <email@example.com> writes:
> Jan> breakpoints after:
> Jan> warning: Temporarily disabling breakpoints for ...
> Jan> are currently not handled correctly and GDB does even crash (with
> Jan> -lmcheck).
> Jan> 2009-10-25 Jan Kratochvil <firstname.lastname@example.org>
> Jan> gdb/
> Jan> * breakpoint.c (bpstat_stop_status): Check BL->SHLIB_DISABLED.
> Jan> (print_breakpoint_location): New comment. Check LOC->SHLIB_DISABLED.
> Jan> Check LOC validity before printing it. Use LOC instead of B->LOC.
> Jan> Swap the if conditionals order.
> This is ok. Thanks.
The original mail had also the second part:
# 2009-10-25 Jan Kratochvil <email@example.com>
# * gdb.base/unload.c (main): Change the UNLOADSHR parameter to 1.
# Replace the printf call of Y by provided "y-set-1" label. New block
# for the second shared library.
# * gdb.base/unload.exp: Compile also the second library, call
# gdb_load_shlibs also for it. Use now gdb_breakpoint.
# (single pending breakpoint info): Rename to ...
# (pending breakpoint info before run): ... this extended test.
# (libfile2, libname2, libsrcfile2, libsrc2)
# (lib_sl2): New variables.
# (exec_opts): Set also SHLIB_NAME2.
# (pending breakpoint info on first run at shrfunc1)
# (pending breakpoint info on second run at shrfunc1)
# (pending breakpoint info on second run at shrfunc2)
# (print y from libfile, print y from libfile2): New tests.
# * gdb.base/unloadshr.c (shrfunc1): Change the returned value.
# * gdb.base/unloadshr2.c: New.
Do I understand it correctly your approval is for the gdb/ part while the
gdb/testsuite/ part is still left pending for review some other day?
Thanks for the review,
More information about the Gdb-patches