rawhide report: 20060213 changes 2/2

Jakub Jelinek jakub at redhat.com
Mon Feb 13 08:49:47 UTC 2006


On Mon, Feb 13, 2006 at 09:41:05AM +0100, shrek-m at gmx.de wrote:
> On 13.02.2006 09:18, oldman wrote:
> 
> >Hmm... I didn't see a part 1, But GMane got it!  Thanks Jesse
> 
> thanks,
> http://article.gmane.org/gmane.linux.redhat.fedora.testers/36289
> 
> "bump again for double-long bug on ppc(64)"
> 
> sorry, what is the meaning of "double-long" in this context ?

Please read that as:
"bump again to make sure long double is 128-bit on ppc32"

There has been a bug in gcc <= 4.1.0-0.23, where although
-mlong-double-128 was default for ppc64/s390/s390x/alpha/sparc
and for ppc32 in gcc{-c++} packages, if gcc-ppc32 and/or gcc-c++-ppc32
packages were installed, -mlong-double-64 was the default for 32-bit
code (missing --with-long-double-128 configure option in the configure
command for the gcc-*ppc32 compiler).

	Jakub




More information about the test mailing list