[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Is 2.1.2-0pre2 problematic on 386 processor?




Do to a rash of problems on a pure 386 processor, one of which is linked
to glibc strcpy, I started looking at the debian-glibc mailing list.

I see that a change  was made in 2.1.1-13 to debian/sysdeps.mk to add a
cflags variable i386 hack. I do not know the exact change, but I would
guess that the code is commented out in 2.1.2-0pre4. The mailing list
archive implies that this fixed the 386 problems.

In the current 2.1.2-0pre4 the following exists - i.e. commented out.
I do not know what the exact change was to 2.1.1-13, but I suspect that
the code was NOT commented out.

# egcs ends up defining -mcpu=i486 by default, which defines __i486__
# which causes glibc to use the i486-specific inlines. This is a kluge to
# attempt to fix it.
#ifeq ($(DEB_HOST_ARCH),i386)
#cflags += -mcpu=i386
#endif


Should this be commented out - or was there another fix that complensates?
My initial testing implies that the bug may be back....

So, my question is: was the fix backed out?

	Ezra




Reply to: