|
From: Dan K. <da...@ke...> - 2009-09-05 00:46:45
|
So, how easy is it to debug valgrind?
Valgrind is hanging left and right for me on chromium's unit tests today.
It's quite reproducible; would running it under gdb be likely to
find out why?
I tried running valgrind in gdb by editing vg-in-place to read
VALGRIND_LIB="$vgbasedir/.in_place" \
VALGRIND_LIB_INNER="$vgbasedir/.in_place" \
gdb "$vgbasedir/coregrind/valgrind"
and then doing
set args /bin/true
run
That worked fine, but doing it with my real app blew chunks:
Program received signal SIGSEGV, Segmentation fault.
0x0b7e6cf7 in ?? ()
|
|
From: Nicholas N. <n.n...@gm...> - 2009-09-05 20:11:14
|
On Sat, Sep 5, 2009 at 10:17 AM, Dan Kegel<da...@ke...> wrote: > So, how easy is it to debug valgrind? README_DEVELOPERS has much of the accumulated wisdom on the topic. I personally end up using diagnostic printfs more than gdb, but YMMV. N |