Bug#385580: gcc generates incorrect code on s390

[Matthias Klose <doko@cs.tu-berlin.de>]

reassign 385580 gcc-4.1
clone 385580 -1
reassign -1 coreutils
severity 385580 important
forwarded 385580 http://gcc.gnu.org/PR28146
retitle 385580 [4.1 regression, fixed in 4.2] gcc generates incorrect code on s390
tags 385580 + upstream
tags 385580 + fixed-upstream
thanks

according to PR 28789 this cannot be fixed in 4.1.  Please either
build using gcc-3.4 on s390 or use the workaround mentioned in the
redhat report.

Michael Stone writes:
> Severity: serious
> Package: gcc
> 
> gcc has been building parts of coreutils on s390 (sha384 & sha512) in a 
> way that silently generates incorrect results. (Caught by tests during 
> the coreutils build process.) This has been traced to a bug in gcc:
> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=196369
> 
> Info about the fix is here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28146 
> 
> severity set to serious because it causes ftbfs (and because other 
> packages without test scripts may be simply outputting incorrect answers 
> rather than ftbfs)