[Checkins] SVN: Sandbox/faassen/grokcore.site/trunk/TODO.txt Get rid of this too.

Martijn Faassen faassen at infrae.com
Fri Oct 17 13:02:56 EDT 2008


Log message for revision 92329:
  Get rid of this too.
  

Changed:
  D   Sandbox/faassen/grokcore.site/trunk/TODO.txt

-=-
Deleted: Sandbox/faassen/grokcore.site/trunk/TODO.txt
===================================================================
--- Sandbox/faassen/grokcore.site/trunk/TODO.txt	2008-10-17 17:02:25 UTC (rev 92328)
+++ Sandbox/faassen/grokcore.site/trunk/TODO.txt	2008-10-17 17:02:56 UTC (rev 92329)
@@ -1,121 +0,0 @@
-====
-TODO
-====
-
-Meta
-====
-
-- Check demo applications for issues with patterns (grok wiki, ldap address
-  book)
-
-
-Core
-====
-
-- choice fields / sources (theuni)
-
-- testing strategy for the tutorial (faassen)
-
-- make it easier to write tests (wosc, faassen)
-
-- error reporting during grokking (provide file/line-number information
-  on our extrinsically generated errors) (philikon)
-
-- What will happen if we make a utility a old style class and use the
-  MRO stuff. Since we don't support non-persistent local utilities yet this
-  may be a rare case.
-
-- What about local utilities that don't subclass persistent? Perhaps we 
-  can look for IPersistent and give an error if the utility doesn't
-  state it is.
-
-- Change error messages: whenever we report about a callable, add
-  () to the name. Core Python expections do this.
-
-- JSON support similar to XMLRPC.
-
-
-Schema/formlib support
-----------------------
-
-- support nested class 'fields' directly on a view (do we really want this?)
-
-- list form for grok.Container (w/ zc.table?)
-
-- delete action on list form
-
-- make formlib macros available in some form?
-
-- what about subclassing a model that has a 'fields'?
-
-
-Need to discuss
----------------
-
-- Do we want to ship Grok with a javascript library dependency such as
-  MochiKit, to enable out of-the-box AJAX?
-
-- Make it even easier to set up the catalog (intids should be set up
-  automatically if not already present. Perhaps Index grokkers?).
-
-- Testing support. Test grokkers?
-
-- Error pages: make it easy to register application-specific error
-  pages for exceptions.
-
-- Easier queries: integrate hurry.query in some way?
-
-- fall back to a static resource that is defined in a package on a higher
-  level if no static resource directory is defined locally in a package?
-
-- grok.grokkable (to allow grokking of imported things)
-
-- skins
-
-- form redirect
-
-- authentication (pau integration) (faassen)
-
-- sessions (get the session information for something, similar to
-  annotations?)
-
-- menus - define a menu, associate a view with a menu (module-level,
-  class-level)
-
-- making new widgets (faassen, philikon)
-
-- IMPORTANT: different strategies: grok.definefoo() versus n =
-  grok.Foo(), watch out for consistency/symmetry/...
-
-- use ZCML's conflict resolution machinery; actions for Grok.
-
-- do not accept automatic template directory guessing convention for
-  __init__.py, bail out with grok error instead?
-
-- grok.name, grok.template class restrictions (e.g. grok.template
-  should only be usable from grok.View subclasses)
-
-- support grok.template(template) in addition to
-  grok.template('name_of_template')?
-
-- support grok.resource on view class level?
-
-- should grok.context and grok.Model be order-dependent?
-  (so their meaning becomes "below here, this is the context")
-
-- Do we want to initialize model attributes for any schema that the model
-  implements (in addition to the initialization that is taking place for the
-  model-level fields)?
-
-
-Punt
-----
-
-- making new fields
-
-- viewlets / content providers (LATER)
-
-- RDB - via extension: megrok.sqlalchemy for example - make it easy to
-  go between the different schema implementations
-
-- containment constraints (wait for zope 3 to do them right)



More information about the Checkins mailing list