[Zope-dev] DISCUSS: Community checkins for CVS

Paul Everitt paul@zope.com
Sun, 23 Sep 2001 16:35:32 -0400


Joachim Werner wrote:
[snip]

> What I haven't found on the CVS site yet is anything about peer-reviewing
> contributions before they go into the main tree. While I sometimes have the
> feeling that there are fixes from ZC people that should NOT have made it
> into a release, there are many patches from the community that are not
> getting into a release for a long time (this is not a very scientific
> statement, just my personal feeling).


I imagine that the group will decide rules on peer reviewing.  For 
comparison, the Mozilla group has very elaborate rules for checkins, 
while Python has pretty much an innocent until proven guilty culture. 
(That is, you check something in, and if somebody complains, it gets 
removed.)

I don't think it is worthwhile trying to form these rules a priori.


> We need rules like "NO FIXES BETWEEN FINAL BETA AND RELEASE" (Absolutely no
> fixes I mean) -- and those rules should apply to everybody.


Again, we'll let the rules come out of the group.  For instance, what if 
an Emacs #foo.py# accidentally got checked in?  Would you really require 
another beta release for that?  Betas are a cost incurred by hundreds of 
people around the world.

I think the group can do their best to adhere to a policy of doing beta 
cycles for minor changes.


> We maybe also need an improved process for designing new API extensions etc.
> One case for that is the Zope Internationalization Project
> (http://www.eurozope.org/zip/FrontPage), which better sooner than later
> should become a core project. I have the feeling that with the current Wiki
> approach it will take ages to agree on a syntax for internationalization in


Ahh, the "it's the Wiki's fault" argument.  I just checked the zip 
mailing list archive.  9 messages since Aug 1st.  So neither email nor 
Wiki are good choices.  Can you point to an example of a process that 
worked better for designing APIs?

As for internationalization, I'm hoping that EuroZope (or ZIP) will 
recommend a strategy.  I'm on the EuroZope list as well, and from what I 
can tell, there's still a ways to go before consensus is reached.  Let's 
start a discussion over on EuroZope or ZIP and see if consensus can be 
reached.

> Zope. I don't mean that we need a single implementation. But we need an
> agreed-on syntax that is part of the standard Zope package, so that a ZPT or
> DTML Method will not break if it uses translation tags.

Yes, that's needed quite badly.  But I don't think this has to be done 
before we open the CVS to external contributors.

--Paul