Re: [RFA/testsuite] Suppress errors building thread tests

[Michael Elizabeth Chastain <mec@shout.net>]

Recommended for self-approval by Daniel J.

Proofread and tested.  All the pthreads library noise is gone
from gdb.sum.  The results for gcore-thread.exp, print-threads.exp,
and schedlock.exp are fine.

> Note the use of gdb_suppress_entire_file.  I'm just doing that because it's
> what the other tests do; I really dislike gdb_suppress_entire_file, so I
> think I may send along a followup patch to change it to a "return -1". 
> gdb_compile_pthreads already does an "unsupported" if compilation fails. If
> we can't build the threads tests, there's no point in FAILing them, in my
> opinion; this clutters up results when I run simulator testing.  Thoughts?

I'm with you.  If compilation+linking fails, I am fine with one
UNSUPPORTED per test script.  It would be even better if the UNSUPPORTED's
come from a known unsupported list rather than dynamic conditions.

Michael C

===

  2002-12-31  Daniel Jacobowitz  <drow@mvista.com>

	  Fix PR gdb/844
	  * lib/gdb.exp (gdb_compile): Handle "quiet" option.
	  (gdb_compile_pthreads): Pass "quiet" to gdb_compile.

	  * gdb.threads/gcore-thread.exp: Use gdb_compile_pthreads.
	  * gdb.threads/print-threads.exp: Likewise.
	  * gdb.threads/schedlock.exp: Likewise.