[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
little things ...
Joseph E. Sacco
little things ...
Wed, 23 Feb 1994 22:39:47 -0500 (EST)
I have been doing some alpha testing on glibc-1.07.4[an alpha release :-)]
so I thought I'd see what I ran into building Octave-1.0. I am using gcc-2.58
and libg++-2.5.3 on a SPARC10 running OS4.1.3.
The first "surprise" was that the configuration process opted to use f2c
rather than the unbundled SUN fortran, version 1.4, in stalled on the system.
Is that normal?
The second "surpise" was a multuple definition of "_z_abs"; one from
libf2c.a and the other from libc.a. I got around this by yanking z_abs.o from
libf2c.a. With this change octave was able to link.
Next I ran the tests in the test directory:
Test Run By root on Wed Feb 23 22:16:44 EST 1994
Target is sparc-sun-sunos4.1.3
Host is sparc-sun-sunos4.1.3
=== octave tests ===
Running ./octave.test/fsolve.exp ...
Running ./octave.test/index.exp ...
Running ./octave.test/infnan.exp ...
Running ./octave.test/npsol.exp ...
Running ./octave.test/transpose.exp ...
=== octave Summary ===
# of expected passes 22
# of expected failures 0
# of unexpected successes 0
# of unexpected failures 3
../src/octave version 1.0
I believe that the infnan test requires an "infinity" function. Sun's math
library has one. Glibc does not. Is this an IEEE requirement? If so I will
tell the maintainers of glibc to add it.
I don't know what npsol is? There was a reference to it in the README.NLP
as something you need a license from Stanford in order to use it. So that would
certainly explain the othe two failures. So..., it looks like glibc may be
OK for this application.
I've started looking at some of the examples shown in the info doc. Is
plot (x,t) supposed to work? The parser complains that it does recognize
"plot". I have gnuplot installed. The parser recognizes "gplot x".
What about dynamical linking? The config files seem to indicate that
- little things ...,
Joseph E. Sacco <=