On 06/19/10 11:24 PM, William Stein wrote:
On Saturday, June 19, 2010, Dr. David Kirkby<david.kir...@onetel.net> wrote:
I've just done a quick check to see the number of standard packages in Sage
which have a file 'spkg-check' which will execute a test routine after building
in the event the environment variable SAGE_CHECK is set to yes.
The number is depressingly small, at only 19.
./atlas-3.8.3.p12/spkg-check
./cddlib-094f.p6/spkg-check
./eclib-20080310.p10/spkg-check
./ecm-6.2.1.p2/spkg-check
./f2c-20070816.p2/spkg-check
./flint-1.5.0.p4/spkg-check
./givaro-3.2.13rc2.p1/spkg-check
./gsl-1.10.p2/spkg-check
./iconv-1.13.1.p2/spkg-check (Of course, since I added iconv!)
./libfplll-3.0.12.p0/spkg-check
./libgcrypt-1.4.4.p3/spkg-check
./mpfr-2.4.2/spkg-check
./mpir-1.2.2.p1/spkg-check
./ntl-5.4.2.p12/spkg-check
./r-2.10.1.p1/spkg-check
./sagetex-2.2.5/spkg-check
./scipy-0.7.p5/spkg-check
./sqlalchemy-0.5.8/spkg-check
./zn_poly-0.9.p4/spkg-check
That leaves all these which do not run any self-tests, several of which have
self-test routines, but which never get executed in Sage, even if SAGE_CHECK =
yes, since nobody has written a sage-check.
Should spkg-check be added to these to enable self-tests when "SAGE_CHECK=yes"?
Emphatic yes!
I'll do a few until I get bored!
I just noticed that despite the fact flint has a test suite, and an spkg-check
file, the spkg-check suffers from the usual
if [ `uname` = "Darwin" -a "$SAGE64" = "yes" ]; then
echo "64 bit MacIntel"
FLINT_TUNE=" -fPIC -m64 -funroll-loops"
fi
so any attempt to test flint in 64-bit mode on OpenSolaris fails miserably. As
soon as I removed the restriction of building 64-bit on OS X, so all the tests
passed.
That is now
http://trac.sagemath.org/sage_trac/ticket/9277
So those ****ing SAGE64's still haunt me! I thought removing them from
spkg-install would be enough, but no, there are more of them! Luckily, with only
19 spkg-check files, there can't be too many SAGE64 issues and I would expect
many will not need any changing.
Dave
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
