+1 to phillip & grant's comments, i.e. upgrade to 2.5 unless there
are serious problems, e.g. twisted, and don't support multiple versions.
John
On Nov 30, 2006, at 4:33 PM, Phillip J. Eby wrote:
At 04:17 PM 11/30/2006 -0800, Heikki Toivonen wrote:
My main gripe, as usual, is that it
would once again push further the time when we could use the platform
Python to run Chandler (at least on Linux).
I've pretty much come to the conclusion that this isn't going to
happen any time soon anyway, so I wouldn't let that stop us.
Besides, if Chandler is cool and popular perhaps this will push the
vendors to add 2.5 support sooner. ;)
More information about Python 2.5 is here:
http://www.python.org/download/releases/2.5/
So, there are some options available for us:
1) Upgrade and require 2.5
2) Upgrade to 2.5 but let Chandler still run with 2.4
- this would require some version dependent imports for files that
wanted to use certain 2.5 features
3) Don't upgrade
Personally I am slightly in favor of 2) because of the possibility of
using the system Python on Linux. My second choice is 1).
-1 for #2, it just expands our testing and support issues
+1 for #1.
If we decide to upgrade, I think we should aim to do it in alpha5.
Also, I think we should switch to Microsoft Visual C++ 2005 Express
edition as the compiler on Windows.
Does that include an optimizing compiler? I thought that certain
of the "free" compilers didn't include optimization support.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
On Nov 30, 2006, at 4:56 PM, Grant Baillie wrote:
On 30 Nov, 2006, at 16:17, Heikki Toivonen wrote:
Python 2.5 came out a little while ago. It is a fairly significant
release, adding some nice new features in the language. AFAIK at
least
one feature would enable us to fix a repository bug. 2.5 is also
significantly faster than 2.4. Finally, 2.5 contains some security
fixes
that have never been ported to 2.4.
I also believe that 2.5 included build changes which will make it
possible to build it with the free Microsoft Visual C++ 2005 Express
edition. Besides being free, I believe it produces slightly faster
and
better code than our current compiler.
All good reasons for us to upgrade to 2.5 at the first opportunity.
However, there are also downsides.
Have the issues with Twisted & Python 2.5 been resolved? e.g.
<http://glyf.livejournal.com/62308.html>
My main gripe, as usual, is that it
would once again push further the time when we could use the platform
Python to run Chandler (at least on Linux).
More information about Python 2.5 is here:
http://www.python.org/download/releases/2.5/
So, there are some options available for us:
1) Upgrade and require 2.5
2) Upgrade to 2.5 but let Chandler still run with 2.4
- this would require some version dependent imports for files that
wanted to use certain 2.5 features
3) Don't upgrade
Personally I am slightly in favor of 2) because of the possibility of
using the system Python on Linux. My second choice is 1).
If we decide to upgrade, I think we should aim to do it in alpha5.
I'm not so keen on #2 ... it'd be too easy for developers to add
2.5 dependence in their environments. Without 100% code test
coverage, the resulting bugs could be hard to find. Also, it would
require that all the projects we depend on work against both
versions of python.
--Grant
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
