http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53718
--- Comment #9 from janus at gcc dot gnu.org 2012-06-20 09:19:28 UTC --- (In reply to comment #8) > It would be nice if someone could compile gfortran.dg/select_type_12.f03 (and > the tests) through valgrind (I cannot do it under darwin). While valgrind 3.6.1 gives me some kind of internal error (cf. https://bugs.kde.org/show_bug.cgi?id=277045), valgrind 3.7.0 seems to work well and produces the following output: > valgrind `gfortran-4.8 -v select_type_12.f03 2>&1 | grep f951` ==32473== Memcheck, a memory error detector ==32473== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al. ==32473== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info ==32473== Command: /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/f951 select_type_12.f03 -quiet -dumpbase select_type_12.f03 -mtune=generic -march=x86-64 -auxbase select_type_12 -version -fintrinsic-modules-path /usr/lib64/gcc/x86_64-unknown-linux-gnu/4.8.0/finclude -o /tmp/ccOu2tOb.s ==32473== GNU Fortran (GCC) version 4.8.0 20120620 (experimental) [trunk revision 188819] (x86_64-unknown-linux-gnu) compiled by GNU C version 4.6.3 20120531 [gcc-4_6-branch revision 188067], GMP version 5.0.5, MPFR version 3.1.0-p1, MPC version 0.8.2 GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096 GNU Fortran (GCC) version 4.8.0 20120620 (experimental) [trunk revision 188819] (x86_64-unknown-linux-gnu) compiled by GNU C version 4.6.3 20120531 [gcc-4_6-branch revision 188067], GMP version 5.0.5, MPFR version 3.1.0-p1, MPC version 0.8.2 GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096 ==32473== ==32473== HEAP SUMMARY: ==32473== in use at exit: 425,388 bytes in 1,978 blocks ==32473== total heap usage: 9,348 allocs, 7,370 frees, 4,853,954 bytes allocated ==32473== ==32473== LEAK SUMMARY: ==32473== definitely lost: 15,827 bytes in 90 blocks ==32473== indirectly lost: 18,024 bytes in 62 blocks ==32473== possibly lost: 192 bytes in 1 blocks ==32473== still reachable: 391,345 bytes in 1,825 blocks ==32473== suppressed: 0 bytes in 0 blocks So, apparently there are no errors. Moreover, the patch seems to survive a regtest on x86_64-unknown-linux-gnu without any failures in the gfortran testsuite.