William Stein wrote:

> On PPC OS X 10.4, Sage doesn't build at all for multiple reasons.
> Either we fix this or deprecate PPC 10.4 support completely.  I think
> deprecation is a fair option, since it's an OS that is 2 versions
> behind:
>      #7107    sage does not build on ppc os x 10.4 anymore, failing with mpfr
>      #7115    building cliquer Cython extension fails on OS X 10.4 PPC
> (with cliquer-1.2.p0.spkg)


In addition to the financial arguments given elsewhere (the cost of 
upgrading hardware to run OS 10.5), I can give another argument for not 
deprecating OS 10.4.

It is quite common for a bug to show up on one platform, but not on 
another. In many cases these are real bugs, which could affect any 
platform, but just happen to show up on one. Then, at a later date, they 
show up on another.

I recall chasing a bug which only happened on a quad processor IBM RS 
6000 running AIX, despite working on numerous other machines including a 
Cray, Suns, Linux PCs, an SGI Octane, HP Visulize C3600, and countless 
other bits of hardware.

When I finally tracked the issue down on AIX, which was due to a locking 
problem in my multi-threaded coded, I realised it have shown up on any 
platform, but just happened to show up on AIX. No doubt, some time in 
the future it would have shown up on other hardware or operating systems.

Note cliquer also has issues on Solaris.
http://sagetrac.org/sage_trac/ticket/6852
which might be an indication the code is not as well written as it could 
be. It uses SCons, which seems to cause problems on lots of code as far 
as I can tell.


As reguards to
http://sagetrac.org/sage_trac/ticket/7107
'sage does not build on ppc os x 10.4 anymore, failing with mpfr' I 
would note that 'internal compiler errors' can be a result of a library 
not being found, which can sometimes be fixed with LD_LIBRARY_PATH.

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