pkgsrc/math/scilab/files
frueauf 7b74a65cd4 Update scilab to 2.4.1. This is based on the work provided in pr 6889
by Dan McMahill. It fixes several serious problems for NetBSD/alpha
and some for NetBSD/sparc. Thanx for doing the work, Dan!
1999-03-05 11:22:34 +00:00
..
md5 Update scilab to 2.4.1. This is based on the work provided in pr 6889 1999-03-05 11:22:34 +00:00
README.next_rev Update scilab to 2.4.1. This is based on the work provided in pr 6889 1999-03-05 11:22:34 +00:00

$NetBSD: README.next_rev,v 1.1 1999/03/05 11:22:34 frueauf Exp $

NOTE (mcmahill@alum.mit.edu, March 3, 1999):

There is a bug in NetBSD/alpha-1.3I which causes some problems with scilab.

The relevent alpha bugs are:
    
port-alpha/6805, with  port-alpha/6634, and port-alpha/4424 looking related.
Once those are fixed I'll update scilab to make it fully alpha functional.

The effect is that any scilab operation which causes a divide by zero generates
a floating point exception which is not trapped.  This causes a core dump.

Once the alpha bugs listed above are fixed, the following changes should be made
to the scilab package source:

Once fpsetmask(0) on the alpha is fixed to no longer cause a core dump, the
following patch files should be updated:
patch-ai
patch-an
patch-ao
patch-ap
patch-aq

The lines 

+#if (defined(netbsd) && !defined(__alpha__))  || defined(freebsd)
+  fpsetmask(0);
+#endif

should be changed to

+#if defined(netbsd) || defined(freebsd)
+  fpsetmask(0);
+#endif



In addition, once the alpha bug is fixed, patch-bt should be removed and the
pkgsrc Makefile should have the following lines removed:

post-patch:
      @if [ ${MACHINE_ARCH} != "alpha" ]; then \
      ${MV} ${WRKSRC}/scilab.star.orig ${WRKSRC}/scilab.star