RE: Grief with Dejagnu; Testing gcc 2.95.3.test4 on SCO OS5.0.4

[David Korn <dkorn at pixelpower dot com>]

>-----Original Message-----
>From: David Gressett [mailto:gressett@iglobal.net]
>Sent: 26 February 2001 08:25
>To: gcc@gcc.gnu.org


>There are problems with  gcc 2.95.3.test4 on SCO Openserver 5.0.4:
>when doing a make bootstrap, these failures occur:
>
>1.
>make[5]: Entering directory `/u7/objdir2/gcc'
>./xgcc -B/usr/local/i386-pc-sco3.2v5.0.4/bin/ -B./ 
>-I/usr/local/i386-pc-sco3.2v5.0.4/include -O2   -DIN_GCC     
>-O2 -g -O2 
>-I./include   -g1  -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED    -g -O2 -I. 
>-I../../gcc-2.95.3.test4/gcc -I../../gcc-2.95.3.test4/gcc/config 
>-I../../gcc-2.95.3.test4/gcc/../include \
>   -c -fexceptions ../../gcc-2.95.3.test4/gcc/cp/exception.cc
>xgcc: Internal compiler error: program as got fatal signal 11
>
>make bootstrap can then be restarted and will finish properly.

  Yeah, that is the legendary 'erratic signal 11 thing'.  Some people think
it's caused by hardware problems: bad RAM or trouble with the memory timing
parameters in your bios.  It crops up under Linux and Cygwin as well, so
that would suggest it isn't a simple software problem.  OTOH you can run a
memory checker continuously for days on end (I've tried) without spotting
any errors; it only seems to happen in complex real world situations such
as a gcc or linux kernal build, when the system is heavily loaded with
lots of forks and subshells all over the place.  At times I wonder if there
is some really obscure, as-yet-unnoticed race condition in the canonical
*nix forking/process handling code, that has been preserved across these
different ports because of the way in which Unices have descended from each
other.  Anyway, it probably doesn't represent a real failure unless you
can repeat it on a second machine.

>The latest Dejagnu does not work with OS 5.0.4 any better than the
>previous version. I still get errors which make it impossible to run
>the full suite of tests:
>
>WARNING: couldn't find the global config file
>WARNING: couldn't find tool init file
>ERROR: couldn't find tool config file for unix
>
>TCL_LIBRARY and DEJAGNULIBS were set properly.

  Hmm.  Did you try cd into your gcc top level build dir and

export $DEJAGNU=`pwd`/gcc/testsuite/site.exp

because that fixed the ERROR: for me under Cygwin.

>Final question: My SCO 5.0.4 machine cannot access the outside world; how 
>do I get the contrib/test_summary shell script to produce an output which
>I can move to a Windows machine which does have the mail access 
>needed to 
>report the test results? (Assuming that I actually have some to report)

Pipe the output to a file:

 test_summary >summary.txt

You'll want to snip the first and last lines, which look something like

mail -S "subject line" gcc-testresults@gcc-gnu.org <<EOF
   [body gets mailed to list]
EOF

      DaveK
-- 
"Never give succor to the mentally ill; it is a bottomless pit."  
  -- William S. Burroughs


**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

This footnote also confirms that this email message has been swept by
MIMEsweeper for the presence of computer viruses.

www.mimesweeper.com
**********************************************************************