This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
Re: egcs 1.1b make check problems under 2.0.98
- To: howarth@bromo.med.uc.edu
- Subject: Re: egcs 1.1b make check problems under 2.0.98
- From: Geoff Keating <geoffk@ozemail.com.au>
- Date: Thu, 22 Oct 1998 16:31:44 +1000
- CC: libc-alpha@cygnus.com
- References: <9810202301.ZM8823@bromo.med.uc.edu>
> From: "Jack Howarth" <howarth@bromo.med.uc.edu>
> Date: Tue, 20 Oct 1998 23:01:37 -0400
> On linuxppc running Linux 2.1.125 built with egcs 1.1b
> and glibc 2.0.98 we are seeing failures now in the make
> check of egcs 1.1b. In particular we find that the tests
> in libio are failing. For example, when we execute...
>
> g++ hounddog.cc -O2 -o hounddog
> ./hounddog -b0 < hounddog.inp
>
> ...we should see...
>
> Code: -1 followed by 6 chars
> Code: 1 followed by 3 chars
> Code: 2 followed by 0 chars
> Code: -1 followed by 3 chars
> Code: 3 followed by 4 chars
> Code: 1 followed by 0 chars
> Code: -1 followed by 3 chars
>
> ...instead we see...
>
> Code: -1 followed by 6 chars
> Code: 1 followed by 3 chars
> Code: 2 followed by 0 chars
> Code: -1 followed by 3 chars
> Code: 2 followed by 6 chars
> Code: 1 followed by 0 chars
> Code: -1 followed by 3 chars
>
> There is also a report that this is seen on Intel linux as well.
> Could this be due to the recent changes in bits/string2.h?
It seems unlikely, since I compile my libio with
-D__NO_STRING_INLINES. However, I am now seeing this, so I will debug
it and tell everyone what I find...
--
Geoffrey Keating <geoffk@ozemail.com.au>