------- Comment #2 from Jean-pierre dot vial at wanadoo dot fr 2006-11-19 08:55 ------- (In reply to comment #1) > This is a bug in your instation of GMP/MPFR. It is causing gfortran to crash. > No, I found the explanation, it has nothing to do with gmp: it is a parallelism problem in the makefile. I have a multiprocessor workstation (2 dual-core opterons = 4 processors) and if I run the makefile with the option make -j4 a I usually do, the sub-make for libgfortran is started when the required new gfortran does not yet exist, so the make-kinds-h.sh shell-script fails without meaningful error message. running make in single-thread mode (without the -j4 option) solves this problem. I then fall back on bug 29867 which is completely different.
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29866