libg++-2.2 test suite

Article View: gnu.g++.lib.bug

Article #298

#298

From: Neil McKellar
Date: Wed, 14 Oct 1992 09:26

28 lines
1445 bytes

Hi.  I am installing libg++-2.2 on an SGI running IRIX 4.0.4 using gcc 2.2.2.
I ran configure.

  % configure mips-sgi-irix4.0.4
  % make

It compiled fine, so I decided to run the test suite.

  % make check

Things went fine until we got to tComplex.cc in libg++/tests.
tComplex.cc compiles fine, but fails on

  a / b = (0.5609760976, 0.0487804878)

  where a = (2.0, 3.0)
        b = (4.0, 5.0)

At this point, close_enough() fails to assert() and dumps core.
Since the division by (1,0) earlier in the test succeeded, I can't understand
why this calculation fails.  Everything has been done using doubles for the
real values, so a loss of accuracy should not be the problem.

It seems strange that anything in the test suite should fail.  I imagine this
stuff is fairly well proven.  Have I done something silly during setup?
Any help or ideas about this would be appreciated.  Thanks in advance.

Neil McKellar (mckellar@cs.ualberta.ca)

Technical Details

Message-ID: <mckellar.719097989@spirit-riv>

Path: arc1-ca.pugleaf.net!sparky!uunet!convex!linac!pacific.mps.ohio-state.edu!cis.ohio-state.edu!cs.ualberta.ca!mckellar

Show Full Headers

Complete Article Headers

Newsgroups: gnu.g++.lib.bug
Path: arc1-ca.pugleaf.net!sparky!uunet!convex!linac!pacific.mps.ohio-state.edu!cis.ohio-state.edu!cs.ualberta.ca!mckellar
From: mckellar@cs.ualberta.ca (Neil McKellar)
Subject: libg++-2.2 test suite
Message-ID: <mckellar.719097989@spirit-riv>
Keywords: gnu,g++,libg++
Sender: gnulists@ai.mit.edu
Organization: University of Alberta, Edmonton, Canada
Distribution: gnu
Date: Wed, 14 Oct 1992 09:26:29 GMT
Approved: bug-lib-g++@prep.ai.mit.edu
Lines: 30