Re: check with SGML::Parser::OpenSP (and branches)

On Aug 15, 2005, at 7:24, Bjoern Hoehrmann wrote:
>>  I would suggest committing it to CVS
>> as soon as you can, after a bit of cleaning up if you like.
>
> Well, if I commit to HEAD stuff like automated builds on qa-dev will
> stop working and releasing a new version not based on S::P::O from
> HEAD would be more difficult. I'm happy to do this, but it might
> not be the best option.

There was a lot of discussion on IRC about this, and (not surprisingly) 
Terje was suggesting that the only way for us to release versions not 
based on SPO (which we're about to do anyway, there are already a few 
bugfixes in the pipe, plus some HTML::Templates related caching tweaks) 
was to create a 0_7 branch, and do modularized development on HEAD.

More surprisingly perhaps, I am started to be convinced that this may 
indeed be our best bet. I am far from being fond of branches, but I'd 
rather see cvs commits for a version based on S:P:O than waiting even 
longer, and branching seems to be the least awful way to manage both 
m12n'd development and post-release bugfix and optimization.

In any case, if we don't want to get back to the branching mess we've 
experienced with 0_6, the rules should be (at least):
- No feature work on the 0_7 branch,
- merge into HEAD at each 0_7 release,
- and release 0.7.x versions often

As an application of the 3rd rule, I think we should target for a 
bugfix release for bugs 1839, 1845 and HTL:Template cache within a 
week.

Therefore, I suggest:
- that you should commit your changes to CVS. If there is something to 
do to qa-dev to make it work, we'll do it. openSP and SPO should 
realatively easily install from CVS on that system.
- that Terje should advise on (or proceed with) making a 0_7 branch 
based on tag validator-0_7_0-release
- that Ville should commit his HTML::Template patches.

Any objection to such a plan?
-- 
olivier

Received on Monday, 15 August 2005 08:21:51 UTC