On Jun 28, 4:45 pm, Jason Grout <[EMAIL PROTECTED]> wrote:
> mabshoff wrote:
> > On Jun 28, 2:10 pm, Carl Witty <[EMAIL PROTECTED]> wrote:
<SNIP>
> What is the policy on breaking backward compatibility of pickling at
> major releases (e.g., 4.0)?
I would consider it unacceptable to break backward compatibility for
pickles unless there is a really good reason to do so. It is likely
that as some classes get deprecated pickling support for them will
also have to be removed, but we should be better than say Maple where
upgrades of minor versions regularly seems to massively break code.
Since we now have a deprecation framework in place and are slowly
hashing out the policy associated with it we should also come up with
some agreed upon rules for pickling deprecation.
Since VMWare images should work for the foreseeable future it might be
a good idea to keep at least some of them around so that people can
just download older releases and run them instantly. Compiling older
releases of Sage with current compilers is problematic, i.e. it seem
that no (not even the current release 3.0.3) builds with OpenSuSE 11
since they are shipping a gcc 4.3.1cvs snapshot that identifies itself
as gcc 4.3. So imagine in three years from now when you have to run
Sage 2.10.2 for some reason with you current distribution. You can
build some VMWare image with an older Linux distribution that Sage
2.10.2 supports and then build Sage 2.10.2, but that seems like a
major inconvenience.
> Jason
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
