For the first time ever, Sage has built and pass all the long doctests
on OpenSolaris!
sage -t -long devel/sage/sage/coding/decoder.py
[53.1 s]
sage -t -long devel/sage/sage/schemes/elliptic_curves/ell_rational_field.py
[338.4 s]
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 1615.6 seconds
The hardware/software used as
* Sun Ultra 27
* 3.33 GHz Intel Xeon W3580 (4 cores, 8 threads)
* 12 GB RAM
* OpenSolaris 06/2009 updated to build 134 (March 2010).
* gcc 4.5.0 configured to use the Sun linker and the GNU assembler.
* 32-bit build. (i.e. SAGE64 was *not* exported to yes). 64-bit builds
are unreliable at the minute on this platform.
* ZFS file systems - 2 x 500 MB (in a mirror) and 2 x 2 TB disks (in a mirror).
I'd like to thank Mitesh for merging all the Solaris 10 x86 and
OpenSolaris fixes in this release - many of which I was not expecting
to be merged.
Also to John Palmieri and Jaap Spies, both of whom have made a large
number of changes to get this to work. Also of course to the other
reviewers of patches - too numerous to mention.
It's too early to say whether the changes will cause all doctests to
pass on Solaris x86, as the only machine I have access to (fulvia on
Skynet) is currently unavailable. However, I'm 99% sure they should
pass. I have a VirtualBox virtual machine of Solaris 10 x86, though
I've not set that up yet to build Sage.
At some point in the future, I intend setting up a Solaris zone, and
others can try this Sage, though I'll wait for the final 4.5.3.
I'm not suggesting we officially support Solaris x86 and OpenSolaris
yet, but I think that must be pretty close.
64-bit builds of Sage on any Solaris variant are pretty unstable just now.
I was rather expecting the Cygwin port to be finished before the
32-bit Solaris ports, so I'm somewhat pleased Solaris managed to win.
However, the SYMPOW fix which was merged (#9703) for Solaris should
also resolve the issue of SYMPOW on Cygwin (#9166), though someone
needs to confirm that for sure. (The current version in Sage is *very*
slightly different from the one Mike Hansen tested and said passed all
doctests).
I intend now concentrating efforts on getting 64-bit Solaris versions
working, and doing what I can do add test code.
I think I can treat myself to a pint of beer now - I'm not over keen
on Champagne.
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
