[Zope-CMF] Re: CMF roadmap update

Tres Seaver tseaver at palladion.com
Tue Apr 25 12:45:26 EDT 2006

Hash: SHA1

Florent Guillaume wrote:
> Jens Vagelpohl wrote:
>> Hash: SHA1
>> On 25 Apr 2006, at 15:08, Chris Withers wrote:
>>> Jens Vagelpohl wrote:
>>>>> I sent an email about this a couple days ago. Basically, I'm down
>>>>> to one failing unit test in CMFUid and have discovered that CMFUid
>>>>> is basically broken in 2.0/trunk. I also sent an email on this
>>>>> issue to the list, hoping that the developers who wrote CMFUid and
>>>>> lobbied hard to get it included would take notice and do something
>>>>> about it (or at least acknowledge it), but all I see is silence so
>>>>> far.
>>>> At least I have found what the problem with CMFUid is, and it is
>>>> mostly a misunderstanding how it is supposed to work. Mea culpa. I'm
>>>> still left with the one failing unit test on the events branch, though.
>>> Having such an unknown and unmaintained piece of code in the core of
>>> the cmf scares me.
>>> How would people feel about deprecating it for 2.1 and removing it in
>>> 2.3 if no-one steps up who wants it?
>> Well, I have my own opinion about that, but the course of action
>> depends mostly on those people who are using it. No one seems to,
>> judged by the complete silence.
> Grégoire Weber is the one that coded it and included it.
>> As I have found, it is only used in one specific situations: If you
>> create a Favorite pointing to a piece of content, then that piece gets
>> tagged with a UID, and the UID identifies the content piece for the
>> Favorite. So you can copy/paste/whatever the content and the Favorite
>> still knows how to find it.
> Given that it's unmaintained, that Plone has its own UID tool, that CPS
> does it differently, I'm for deprecating it quickly and slating it for
> removal earlier than the usual 1 year.
> I've also alreay pointed out the overengineering of having 3 tools for a
> simple UID management.

The intent was to allow replacement of one bit of policy (e.g., the
generation of a UID / UUID for a given object) without requiring
replacement of the other bits.  Another, similarly-pluggable
implementaiton would be to have a single tool containing a plugin
registry (as PAS does), with interfaces for each of the plugins.

- -0 on deprecating it yet;  let's see what the folks who *do* use it have
to say about their future intent.  For instance, the Plone
implementation might want to fold into what we are doing in the CMF.

- --
Tres Seaver          +1 202-558-7113          tseaver at palladion.com
Palladion Software   "Excellence by Design"    http://palladion.com
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org


More information about the Zope-CMF mailing list