I get a segmentation fault even with the simplest input files
like hello.c. It looks like it's re-using a freed memory resource.
I'm building with the usual ./configure && make routine.
Tested on 3.9/macppc and 3.8/i386. It works fine on Mac OS X.
ktrace -t cns gives me (last few lines on macppc):
18238 cflow CALL mprotect(0x2578f000,0x1000,0)
18238 cflow RET mprotect 0
18238 cflow PSIG SIGSEGV SIG_DFL code 1 addr=0x2578f000 trapno=2
18238 cflow NAMI "cflow.core"
I can step through the program with gdb, but there's some strangeness
when I print the calling arguments and the parameter variable while in
the called function.
I can find where the SEGV occurs (output.c:39 called from line 414),
but where do I find where that last mprotect() is called from?
Is this happening because of the randomized memory allocation?
Any clues appreciated.