[pacman-dev] strange valgrind result

[Xavier]

When running valgrind with pactest, I always get the following leak :

> ./pactest/pactest.py -p  ./src/pacman/.libs/lt-pacman -t  pactest/tests/remove011.py --valgrind
==3131==
==3131== 20 bytes in 1 blocks are still reachable in loss record 1 of 1
==3131==    at 0x402314C: calloc (in
/usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
==3131==    by 0x439E0E5: _dlerror_run (in /lib/libdl-2.10.1.so)
==3131==    by 0x439DD2B: dlsym (in /lib/libdl-2.10.1.so)
==3131==    by 0x402EC78: load_library_symbols (in
/usr/lib/libfakeroot/libfakeroot-0.so)
==3131==    by 0x402F676: tmp___xstat64 (in
/usr/lib/libfakeroot/libfakeroot-0.so)
==3131==    by 0x402D082: __xstat64 (in /usr/lib/libfakeroot/libfakeroot-0.so)
==3131==    by 0x4047D73: alpm_option_set_root (stat.h:455)
==3131==    by 0x804E57F: main (pacman.c:864)

It happens on both maint and master.
I dont remember seeing this error last time I tried valgrind with
pactest. Ok it was a long time ago, but I don't think this code have
changed.
if I comment out the stat in set_root :   if(stat(dbpath, &st) == -1
|| !S_ISDIR(st.st_mode)) {
Then I get the same error for set_dbpath, and etc...

Also when I run directly valgrind on pacman (without pactest), it
works fine. So might be only a new fakeroot related problem.

Anyway, it's not important, but if anyone has any clue, just tell :)