[Grok-dev] Re: gocept.zope3instance
Philipp von Weitershausen
philipp at weitershausen.de
Thu Jan 11 06:22:07 EST 2007
On 11 Jan 2007, at 12:10 , Christian Theune wrote:
>> I have two bug reports to make:
>> 1.) The advertised "user = user:passwd" doesn't work. Instead, it
>> "admin-user = user" and "admin-password = passwd". This isn't so
>> much of
>> a bug as a documentation issue. Btw, it doesn't look like it supports
>> password managers for encrypted passwords. I wonder if
>> zc.recipe.zope3instance does...
> Thanks. Updated in SVN.
Thanks. You might also want to "python setup.py register" again to
update the Cheeseshop page.
>> 2.) When the recipe fails, e.g. due to a missing configuration
>> option in
>> buildout, it exits gracefully with a nice error message (I assume the
>> error message comes from zc.buildout). When you fix the problem in
>> buildout.cfg and want to run the buildout again, gocept.zope3recipe
>> complains that it can't create the instance directory again. The
>> is this line:
>> # XXX In theory we could just delete the parts directory
>> if it
>> # exists already. Or not?
> This is not so clear unfortunately, at least to me. I've gotta ask Jim
> about what this means in buildout terminology ...
Hmmm, I guess. zc.recipe.zope3recipe at least deletes the whole
instance and recreates it every time you call bin/buildout. That's
not really an option, of course, when you're using buildout to update
a production environment, or you need to keep your Data.fs around
even for development.
I certainly wouldn't want bin/buildout to delete Data.fs in a
>> -recipe = zc.recipe.zope3instance
>> +recipe = gocept.zope3instance
>> database = testdata
>> -user = grok:grok
>> +admin-user = grok
>> +admin-password = grok
>> +address = 8080
>> eggs = setuptools
>> zcml = zope.annotation
>> @@ -40,9 +42,11 @@
> Please notice, that the zcml option does not exist anymore.
Huh? The Cheeseshop page says something else.
> We have created an approach where you can created a derived version
> of the ZCML
> and add your pieces there.
> The exact structuring needs more discussion
> though. Eventually we want to provide multiple profiles of
> skeletons in
> the recipe.
Aha. Use case?
More information about the Grok-dev