On Fri, 2008-08-22 at 12:14 -0700, Garrett D'Amore wrote: > 1) Does Python versioning "make sense" here? I.e. is Python 2.5 largely > compatible with 2.4, and is it the case that developers should start > using the new version? (Put another way, is there any reason that > developers should be building new programs against 2.4 after 2.5 > integrates?)
Yes. Python itself comes with a large number of modules, however there are a lot of extra modules and bindings that we add on top. For example libxml2-python, wrappers for most the GNOME libraries, etc. It will take time to make all these modules available for both 2.4 and 2.5. At that point existing python programs can all move to 2.5 and 2.4 can be phased out. > 2) Assuming that there are no surprises in the question above, wouldn't > it make more sense to upgrade 2.4 to "Uncommitted Obsolete" instead of > "Uncommitted"? (I.e. discourage new users from using 2.4 and direct > them to 2.5 instead.) It should be made Obsolete when all extra Python are delivered for Python 2.5. I expect that to happen around the time the Python 2.6 case is submitted (2.6 is currently in beta). > 3) Is the project planning to undertake an effort to upgrade any Python > 2.4 dependents delivered by OpenSolaris to Python 2.5? (Either as part > of this project, or as part of a follow on effort?) Yes, as a follow-on effort. > Thanks. (FWIW, I think its unlikely that any answers to any of the > questions are going to negatively affect any decision to deliver Python > 2.5. I'm more interested in what happens to Python 2.4, and what > precedent we might be setting, if any, when Python 2.6, 2.7... etc. > comes out.) We actually have a precedent: we moved from 2.3 to 2.4 in the past. S10 GA had 2.3. It now delivers both 2.3 and 2.4, 2.3 was declared Obsolete when 2.4 integrated and then 2.3 was removed from Nevada. We are going to follow a similar pattern. Thanks, Laca
