[Checkins] SVN: zc.comment/trunk/src/zc/comment/test.zcml Remove layer="default" from resourceDirectory and resourceLibrary directives in

[Marius Gedminas]

On Wed, Sep 13, 2006 at 03:20:54PM -0400, Benji York wrote:
> Marius Gedminas wrote:
> >Log message for revision 70156:
> >  Remove layer="default" from resourceDirectory and resourceLibrary 
> >  directives in
> >  ZCML.  This is a deprecated spelling of IDefaultBrowserLayer, which is 
> >  used by
> >  default anyway if you do not explicitly specify a layer attribute.
> 
> I wonder (and don't want to take the time to check, if someone else 
> knows off-hand) if this was properly deprecated and has now finally been 
> removed, or if this checkin was a bit too aggressive (and should 
> therefore be reverted).

As far as I could tell my checkin was a no-op: the same layer is used
for resource definitions now as was used before.

Zope 3 rev 70047 was the one that deprecated <browser:layer>; Philipp
von Weitershausen questioned that checkin on zope3-dev on Sep 8, but I
don't see any answers to his email.

As a result of that checkin[1] functional tests for various zc.* packages
(such as zc.classifieds or zc.task) started failing with a

  ZopeXMLConfigurationError: ...
  ConfigurationError: ('Invalid value for', 'layer', "ImportError: Couldn't import default, No module named default")

  [1] I'm not really sure it was that checkin's fault---it was
      comitted a week ago, but I only started seeing the test failures
      in zc.checkins yesterday

There were three ways to fix the problem:

  (1) revert the change in Zope 3

  (2) edit all zcml files that <include package="zope.app" /> and add
      <include package="zope.app.layers" file="meta.zcml" /> above it

  (3) change resourceDirectory and resourceLibrary directives to refer
      to the default layer in a different way (either by spelling it as
      zope.publisher.browser.interfaces.IDefaultBrowserLayer, or by
      omitting the attribute altogether, because IDefaultBrowserLayer is
      the default value for it).

I chose option (3) because it seemed to be the most obviously correct
and the least controversial.

Marius Gedminas
-- 
Where do you think you're going today?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://mail.zope.org/pipermail/checkins/attachments/20060914/7504df7b/attachment.bin