Since versions 4.2  Sage would build on Solaris 10 SPARC using gcc.
The last remaining issue to prevent this:

http://sagetrac.org/sage_trac/ticket/6579

which was solved 3 months ago. However, it would not build if the Sun
compilers were installed - since the Sage library chose the Sun C++
rather than g++ (for reasons unknown, it chose gcc rather than the Sun
C compiler).

The bug that prevented this happening was solved a few weeks ago

http://trac.sagemath.org/sage_trac/ticket/6595

and integrated into sage-4.3.1.alpha0. However, something else broke
the build on Solaris, so 4.3.1, so now there is no Solaris build at
all.

I thought there was a policy that code had to work on Solaris SPARC
before being committed into Sage - clearly if there is such a policy,
the code is not being tested, and at least one release was made,
despite I made the problem a 'blocker'.

People often CC me onto tickets to test on Solaris. Unfortunately, I
don't have the time to test every patch on Solaris. Yet 't2' is
available for anyone to test their changes. That machine was supplied
free by Sun, who obviously want to see Sage running on Solaris.

I feel there needs to be a documented policy that code gets tested on
Solaris and not put in a new release if it fails, or the support for
Solaris is dropped.

I was wondering if there was any logic in creating a 4.3.0.1 release,
which is just 4.3 with

http://trac.sagemath.org/sage_trac/ticket/6595

integrated. That would allow Sage to build on Solaris 10 (SPARC),
irrespective of whether the Sun compilers were present.

To say I was pissed off about the 4.3.1 release breaking the Solaris
port would be an understatement.

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

Reply via email to