I've been looking at the gcc checks. The only failures I see that I
can't explain are all associated with libmudflap:pass41-frag.cxx. This
test fails if any optimization is used.
It is really very simple:
int
main (int argc, char *argv[])
{
std:string myStr = "Hello, World!";
std:cout << myStr << std:endl;
return 0;
}
but it shows a segment violation. I tested it with a number instead of
a string and also with a constant string instead of a variable and both
those variations pass.
libmudflap is a debugging library. It has to be specifically built into
a program with options like -fmudflap and output turned on with an
environment variable MUDFLAP_OPTIONS.
http://gcc.gnu.org/wiki/Mudflap_Pointer_Debugging
We always seem to get some failure in gcc when testing libmudflap. The
failing routines seem to change with each release.
I am wondering if we ought to disable building libmudflap using the
obscure option --disable-libmudflap (It's not in configure's --help).
This would eliminate a lot of questions about check failures. I doubt
that it is used very often and it does seem to give some false positives.
What do you think?
-- Bruce
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
