Here are my opinions on 1.x releases, for what they're worth:

1.0 branch:
  I don't see a need for further 1.0.x releases.   The 1.1 branch is
fully backward compatible, so there's no reason people can't upgrade
to it for bugfixes.  Also, releasing another 1.0.x would slow the rate
of adoption of WebTest, which would in turn slow adoption of the 1.5
branch as well as TG 2.0.
  Not having SA support in ModelDesigner & CatWalk is sad, but we
can't continue to let that be a showstopper.
  Finally, having 3 active 1.x branches is a burden on everyone.

1.1 branch:
  This branch is solid & is ready to be released ASAP.   Doing it soon
is important because it's been too long since the Kid->Genshi and
SO->SA default template announcements were made.  More than anything
else, this has contributed to the perception that TurboGears is
stalling and is reason enough to push this release out the door.

Fortunately, there's a lot of other goodness in 1.1, too:

 * In ticket #1786, chrisz migrated TG to PEAK-Rules, and in the
process removed some overly-clever code.  PJE has said that
RuleDispatch won't be ported to Python 2.6, so this is a critical
change for the long-term stability of the framework.

 * (#1592, paj) Multiple database support for SQLAlchemy, is a big win
for large shops.

 * (#1355, paj) You no longer have to call ET() in Genshi templates for widgets.

 * (#1762, me & lmacken)  Make the switch to a framework neutral test
process (WebTest), deprecating testutil.create_request (it will be
removed in the next TG version).   This was a blocker for
re-architecting TG, whether incrementally via CP3, or whole-heartedly
via Pylons.

 * (#1919, chrisz) Fix handling of SCRIPT_NAME in WSGI environments.
No more workarounds for mod_wsgi!

 * (#85, me) Improved SQLObject cache clearing for added safety with
multiprocess web servers.   The mod_wsgi TG integration page bashes us
for this, too.


1.5 Branch:
  The 1.1 & 1.5 branches can't be merged, because doing so would force
our users to re-write their tests at the same time as moving to
CherryPy 3.  That's just too much change all at once.   The branch is
in pretty good shape, however.   I had even hoped that we could do a
release candidate in August, but I think that's a little aggressive
now.   Mainly because of the #1946 multi-root/script_name blocker.
After that's straightened out, I'd like to extend CP3's config checker
to look for config keys that need to be upgraded so that people get
proper warnings (#1949), and I want to re-evaluate the decision to
deprecate config.get (#1948).

  After that, hopefully we can figure out which TG 2.0 features should
be backported.   I'm fond of the RESTful dispatch & use_wsgi_app
stuff.   Also, some people have already shown an interest in porting
TGWebServcies & adding ToscaWidgets support, which is very
encouraging.



Release Naming:

  We should make these releases with the same numbering as their
current branches.   The changes in the 1.1 branch are significant, but
they're not enough to jump half a point on.  Moving to CherryPy 3 is a
better time to jump to 1.5, since jumping numbers gives our users a
warning that there will be backward-incompatible changes (like, say,
filters becoming hooks).

  I went through Trac the other day & made sure that the 1.5 branch
tickets are all in the 1.6 milestone.   If we go with my suggestions,
then the 1.1 milestone should be recreated & tickets batch moved.
Whichever way, it would be nice to have the two sets of numbers in
sync again.


Thanks for reading this far,
-Ken

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"TurboGears Trunk" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/turbogears-trunk?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to