Re: i386/sid results, updates to summaries, i386/ia64 diff

[Jeff Licquia <licquia@progeny.com>]

On Fri, 2003-10-17 at 04:53, Matt Taggart wrote:
> Fail on ia64 but not i386
> =========================
> > /tset/LI18NUX2K.L1/base/fwide/T.fwide 4 FAIL
> > /tset/LI18NUX2K.L1/base/vfprintf/T.vfprintf 5 FAIL
> 
> Unanalysed. These are glibc functions.
> 
> > /tset/LSB.os/files/dev_tty/T.dev_tty 1 UNRESOLVED
> 
> Unanalysed.

I've been seeing this on all my test runs as well, and on i386.  Weird.

> Fail on i386 but not ia64
> =========================
> > /tset/POSIX.os/procenv/times/T.times 4 FAIL
> > /tset/POSIX.os/procenv/times/T.times 6 UNINITIATED
> > /tset/POSIX.os/procprim/fork/T.fork 5 UNRESOLVED
> 
> Unanalysed.

I have not seen these failures in my test runs, at least not recently.

> Fail on both(that we don't have waivers for)
> ============
> > /tset/POSIX.os/devclass/c_iflag/T.c_iflag 6 FAIL
> > /tset/POSIX.os/devclass/c_iflag/T.c_iflag 7 FAIL
> > /tset/POSIX.os/devclass/c_iflag/T.c_iflag 8 FAIL
> > /tset/POSIX.os/devclass/c_iflag/T.c_iflag 9 FAIL
> 
> Unanalysed.

I get this on sarge, but not woody.

> > /tset/POSIX.os/devclass/tcgetattr/T.tcgetattr 1 FAIL
> > /tset/POSIX.os/devclass/tcgetattr/T.tcgetattr 2 FAIL
> 
> Unanalysed. This is a glibc function. Did not fail in Jeff's
> i386/woody and sarge testing. Jeff did you patch for this?

This does fail on sarge, but not woody.  I do use a patched glibc on
woody; the single patch actually comes from AJ, and appears to be a port
of some SuSE fixes to glibc 2.2.

The patch is here:

  http://hackers.progeny.com/~licquia/lsb/patches/woody/glibc_2.2.5-11.5/various-lsb-fixes-suse.dpatch.txt

It's supposedly aimed at the following LSB tests:

    LSB.os/devclass/grantpt_L/T.grantpt_L-2       FAIL
    LSB.os/devclass/grantpt_L/T.grantpt_L-3       FAIL
    LSB.os/genuts/ftw/T.ftw-4                     FAIL
    LSB.os/genuts/nftw/T.nftw-10                  FAIL
    POSIX.os/procenv/sysconf/T.sysconf-10         FAIL

> > /tset/POSIX.os/procenv/getlogin/T.getlogin 1 FAIL
> 
> This one is supposedly dependent on how the tests are run. In this case I
> ssh'd into the system as vsx0 which is the recommended way to do it. I've
> also tried running tests from a linux virtual console(should work) and
> with `su - vsx0`(known to fail). This test needs some investigation.

I haven't been seeing this failure on either sarge or woody.