Jason Grout wrote: > Dr. David Kirkby wrote: >> I'd be interested what you get if you build this program, which was >> written by one of the gcc guys to try to get to the bottom of this issue >> with mpfr not building. >> >> On the Sun T5240 ('t2') donated by Sun to the Sage project, it dumps core: >> >> kir...@t2:[~] $ ./a.out >> n=0 >> n=1 >> Abort (core dumped) >> >> On my own Blade 2000 it runs like this: >> >> drkir...@kestrel:[~] $ ./a.out >> n=0 >> n=1 >> n=2 >> n=3 >> n=4 >> n=5 >> n=6 >> n=7 >> n=8 >> n=9 >> n=10 >> >> What does it do on your machine? > > > It runs like on your Blade 2000. > > > CPU: Mobile Intel(R) Pentium(R) 4 - M CPU 2.00GHz > > OS: Ubuntu 9.04 > GCC: gcc (Ubuntu 4.3.3-5ubuntu4) 4.3.3 > > $ ./test_program > n=0 > n=1 > n=2 > n=3 > n=4 > n=5 > n=6 > n=7 > n=8 > n=9 > n=10 > > > Jason
Thank you. This looks like it is a bug in the Sun implementation of memset() of Solaris in the library used on for the Sun T2+ proessors. i and j are both 2^31, so their sum is 2^32, which should be 0 in 32-bit code (which this is). So the program should replace 0 bytes of memory with the number 0 (in other words, do nothing.) But clearly it is doing something as it corrupts the value of the buffer. This should be a relief to the developers of gcc, mpir and mpfr!! Sun was the last party I expected to be guilty, but that is where I think the fault lies now. Dave --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---