Robert Bradshaw wrote:
On Mar 17, 2010, at 4:52 PM, Dr. David Kirkby wrote:
William Stein wrote:
On Wed, Mar 17, 2010 at 4:20 PM, Dr. David Kirkby
<david.kir...@onetel.net> wrote:
I started a build of Sage 4.3.4.rc0 on 't2' and one of my own
SPARCs. The
build on 't2' is still going but it completed on on of my SPARCs at
home. I
just started to run the long doctests.
Yeah! All this successful work on Sage+Solaris is great.
Yes.
To
celebrate, I've created a new mailing list:
http://groups.google.com/group/sage-solaris
which is for all discussions related to Solaris and Sage. I would
like to encourage *everybody* reading this with an interest in Solaris
to subscribe, and also like to strongly encourage new threads that are
only relevant to Solaris/Sage to get posted there (rather than
sage-devel).
-- William
Personally I think that was a *very* bad idea. 95% of people won't
read the list, so 95% of people will not read particular problems that
were created by others that have an impact on Solaris.
Are there lists for Debian, Ubuntu, CentOS, Fedora, SUSE, Mandriva or
OS X?
These systems don't cause near the traffic. We have a separate list for
the Windows/cygwin efforts. Most people on sage-devel don't touch
anything that will impact Solaris (e.g. they only mess with Python code
in the Sage library itself), and release managers, etc. will be on that
list.
Now Sage is building on Solaris 10 (SPARC), the volume of traffic about Solaris
10 (on SPARC) should drop. If changes are made that do not affect Solaris 10 on
SPARC, then then there will be no posts about Solaris 10 on SPARC.
If you look at the number of posts by me since 4.3.4.alpha1 was released and
working, there are very few. Some reported things like
* Certain optional packages not installing on Solaris. At least one of which
(ace) was actually broken on every platform.
http://trac.sagemath.org/sage_trac/ticket/8551
* The Mathematica interface not working on Solaris - once again that had been
common to all platforms for 4 months, but a lack of testing had not showed this.
http://trac.sagemath.org/sage_trac/ticket/8495
hopefully fixes that.
This will simply isolate Solaris and make it more difficult to keep
Sage building on Solaris.
That is true to some extent.
I think the word "some" understates this considerably. If updates are made to
packages which break on Solaris, but are only reported on a Solaris-specific
list, then the Sage will soon stop working on Solaris.
I can not think of one single open-source multi-platform project which has a
list dedicated to Solaris.
python - no
gcc - no
perl - no
apache - no
scons - no
ecl - no.
php - no
R - no
Wireshark - no
Firefox - no.
Can you name me one?
The lack of suitable Solaris hardware has hampered the Sage port to Solaris - a
machine I bought from eBay for a US equivalent of $75 is much faster at building
Sage than 't2'. If #6705 was solved, I rekon the build time on 't2' would drop
by about 8 hours, but even then, a $75 machine would be faster.
But now there is (at least in 4.3.4.rc0) a working Solaris build, splitting
Solaris issues off to a Solaris-specific list will probably kill off Sage on
Solaris.
Only time will tell whether I'm right or not.
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
