This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Problems With egcs/gcc
Martin Kahlert writes:
> Whenever i got things you described it was a memory corruption
> problem or a problem using freed memory or something like that.
Usually, this happens also to me :->
But: the "working" state and "not working" state are perfectly
deterministic; when i say easy to reproduce i don't mean
thru a long but know sequence of steps, but really with
a single operation on the system that can be switched back;
we where also able to make the system switch between the two
state 100 times a second, and we got a completely deterministic
behaviour, and no crash at all.
If you consider that the state switch include quite a lot of
malloc/free, and we never got crashes, this make the memory corruption
case quite difficult (but, of course, not 100% impossible).
Anyway, what about the double to float problem: ever head about
anything like this ? We read in comp.os.linux.developement.system that
some fellow had similar problems in Red Hat 5.1, and solved it
converting the all program to using doubles (i.e. removed float from
all the prototypes); of course, also here, no real prove that the two
things have any relation ...
Maurizio