Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
On Jun 21, 2009, at 3:38 PM, Laurence Rowe wrote: > Jim Fulton wrote: >> I don't agree. The semantics are different. For example, you often >> want to traverse to things in a template that you don't want to >> expose >> via URL. We currently (or last time I checked) expose ++resource+ >> +name in URLs and this is a bug. > > What use is a resource without being URL accessible? It's used fairly > often in Plone products to expose static css / js / images. The way to access a resource in a URL is /@@/name, as Hanno mentioned. The ++resource++name form is intended for use in ZPT paths. It can be used anywhere in a path, which defeats effective caching of resources. Jim -- Jim Fulton Zope Corporation ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
On Sun, Jun 21, 2009 at 9:38 PM, Laurence Rowe wrote: > Jim Fulton wrote: >> I don't agree. The semantics are different. For example, you often >> want to traverse to things in a template that you don't want to expose >> via URL. We currently (or last time I checked) expose ++resource+ >> +name in URLs and this is a bug. > > What use is a resource without being URL accessible? It's used fairly > often in Plone products to expose static css / js / images. You are a victim of Five here and the different semantics of it compared to zope.app.publisher. In zope.app.publisher the url for a resource is something like: /@@/ where is basically IAbsoluteURL for zope.site.getSite(). In Zope2 / Five we have: /++resource++ instead. The expression used in page templates to access resources is the same, though. The reason for this semantic difference is entirely historical. Five didn't support local component registries / sites at the time when support for resources have been added. This all happened before those local registries where simplified in one of the early Zope 3.x releases. Hanno ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
Jim Fulton wrote: > I don't agree. The semantics are different. For example, you often > want to traverse to things in a template that you don't want to expose > via URL. We currently (or last time I checked) expose ++resource+ > +name in URLs and this is a bug. What use is a resource without being URL accessible? It's used fairly often in Plone products to expose static css / js / images. Laurence ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
Gaaa. As I did deeper, it's even more muddled that I feared. I'll start a separate thread. Jim On Jun 21, 2009, at 8:48 AM, Jim Fulton wrote: > > On Jun 21, 2009, at 5:36 AM, Christian Theune wrote: > >> On Sat, 2009-06-20 at 13:04 -0600, Shane Hathaway wrote: >>> Jim Fulton wrote: Why? traverseName is part of zope.app.publication's implementation. Now it's oddly split off in a very separate package. >>> >>> The publisher traversal code is very similar to the code in >>> zope.traversing, so I thought the best thing to do is put it in the >>> same >>> package as zope.traversing, so that traversal would be maintained >>> in one >>> place. >> >> I think there's a valid goal around here. Publisher traversal and >> traversal as used in templates have two different implementations >> that >> partly overlap at some points and its hard to explain why there are >> two >> of them. >> >> I'm not sure whether we can get a single unified implementation, but >> for >> me that would be nice. > > > I don't agree. The semantics are different. For example, you often > want to traverse to things in a template that you don't want to expose > via URL. We currently (or last time I checked) expose ++resource+ > +name in URLs and this is a bug. > > The only place this traversal code is used outside of publication is > by the browser menu code that tries to traverse to a menu item to see > if it is accessible to the user. This was to avoid having to put > security declarations on menu items (a worthy goal) but we've found > that this entailed far too much expense and complexity. I'm not sure > anyone is even using this menu code any more and if they are, I bet > they are or should be making explicit security declarations. > > Jim > > -- > Jim Fulton > Zope Corporation > > > ___ > Zope-Dev maillist - Zope-Dev@zope.org > http://mail.zope.org/mailman/listinfo/zope-dev > ** No cross posts or HTML encoding! ** > (Related lists - > http://mail.zope.org/mailman/listinfo/zope-announce > http://mail.zope.org/mailman/listinfo/zope ) -- Jim Fulton Zope Corporation ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
On Jun 21, 2009, at 5:36 AM, Christian Theune wrote: > On Sat, 2009-06-20 at 13:04 -0600, Shane Hathaway wrote: >> Jim Fulton wrote: >>> Why? traverseName is part of zope.app.publication's implementation. >>> Now it's oddly split off in a very separate package. >> >> The publisher traversal code is very similar to the code in >> zope.traversing, so I thought the best thing to do is put it in the >> same >> package as zope.traversing, so that traversal would be maintained >> in one >> place. > > I think there's a valid goal around here. Publisher traversal and > traversal as used in templates have two different implementations that > partly overlap at some points and its hard to explain why there are > two > of them. > > I'm not sure whether we can get a single unified implementation, but > for > me that would be nice. I don't agree. The semantics are different. For example, you often want to traverse to things in a template that you don't want to expose via URL. We currently (or last time I checked) expose ++resource+ +name in URLs and this is a bug. The only place this traversal code is used outside of publication is by the browser menu code that tries to traverse to a menu item to see if it is accessible to the user. This was to avoid having to put security declarations on menu items (a worthy goal) but we've found that this entailed far too much expense and complexity. I'm not sure anyone is even using this menu code any more and if they are, I bet they are or should be making explicit security declarations. Jim -- Jim Fulton Zope Corporation ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
On Jun 20, 2009, at 1:10 PM, Jim Fulton wrote: > Why? traverseName is part of zope.app.publication's > implementation. Now it's oddly split off in a very separate > package. This makes customizing publication behavior more difficult. > I recently made proxying overridable and missed traverseName. > > This should be moved back to zope.app.publication. The only other > thing that uses this is zope.app.publsher.browser.menu. That can and > should get to these methods via request.publication. > > Or, better yet, traverseRelativeURL and traversePath should be moved > to the browser module and should get traverseName from > request.publication. > > I'll go ahead and do this. Unless something other than zope.app.publication.ZopePublication is using PublicationTraverser (or PublicationTraverse) as a base class. I'm guessing that this isn't likely. The tests for PublicationTraverser either need it to provide traverseName or they need the request to have a publication that has traverseName. But the tests don't want to depend on zope.app.publication. Also, traverseRelativeURL really wants to use the publication getDefaultTraversal method. I'm going to wager that the intent of zope.traverser.publicationtraverse is *not* to provide a publication base class. I'm going to refactor the tests so that they have a test publication with minimal traverseName and getDefaultTraversal. Jim -- Jim Fulton Zope Corporation ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
On Sat, 2009-06-20 at 13:04 -0600, Shane Hathaway wrote: > Jim Fulton wrote: > > Why? traverseName is part of zope.app.publication's implementation. > > Now it's oddly split off in a very separate package. > > The publisher traversal code is very similar to the code in > zope.traversing, so I thought the best thing to do is put it in the same > package as zope.traversing, so that traversal would be maintained in one > place. I think there's a valid goal around here. Publisher traversal and traversal as used in templates have two different implementations that partly overlap at some points and its hard to explain why there are two of them. I'm not sure whether we can get a single unified implementation, but for me that would be nice. Christian -- Christian Theune · c...@gocept.com gocept gmbh & co. kg · forsterstraße 29 · 06112 halle (saale) · germany http://gocept.com · tel +49 345 1229889 7 · fax +49 345 1229889 1 Zope and Plone consulting and development signature.asc Description: This is a digitally signed message part ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
Jim Fulton wrote: > Why? traverseName is part of zope.app.publication's implementation. > Now it's oddly split off in a very separate package. The publisher traversal code is very similar to the code in zope.traversing, so I thought the best thing to do is put it in the same package as zope.traversing, so that traversal would be maintained in one place. > Or, better yet, traverseRelativeURL and traversePath should be moved > to the browser module and should get traverseName from > request.publication. > > I'll go ahead and do this. That's fine. Shane ___ Zope-Dev maillist - Zope-Dev@zope.org http://mail.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope )
Re: [Zope-dev] [Checkins] SVN: zope.traversing/trunk/src/zope/traversing/ Moved the publicationtraverse module from zope.app.publication and added tests.
Why? traverseName is part of zope.app.publication's implementation. Now it's oddly split off in a very separate package. This makes customizing publication behavior more difficult. I recently made proxying overridable and missed traverseName. This should be moved back to zope.app.publication. The only other thing that uses this is zope.app.publsher.browser.menu. That can and should get to these methods via request.publication. Or, better yet, traverseRelativeURL and traversePath should be moved to the browser module and should get traverseName from request.publication. I'll go ahead and do this. Jim On May 22, 2009, at 8:35 PM, Shane Hathaway wrote: > Log message for revision 100262: > Moved the publicationtraverse module from zope.app.publication and > added tests. > > > Changed: > A zope.traversing/trunk/src/zope/traversing/publicationtraverse.py > A zope.traversing/trunk/src/zope/traversing/tests/ > test_publicationtraverse.py > > -=- > Added: zope.traversing/trunk/src/zope/traversing/ > publicationtraverse.py > === > --- zope.traversing/trunk/src/zope/traversing/ > publicationtraverse.py(rev 0) > +++ zope.traversing/trunk/src/zope/traversing/publicationtraverse.py > 2009-05-23 00:35:46 UTC (rev 100262) > @@ -0,0 +1,129 @@ > + > ## > +# > +# Copyright (c) 2001, 2002 Zope Corporation and Contributors. > +# All Rights Reserved. > +# > +# This software is subject to the provisions of the Zope Public > License, > +# Version 2.1 (ZPL). A copy of the ZPL should accompany this > distribution. > +# THIS SOFTWARE IS PROVIDED "AS IS" AND ANY AND ALL EXPRESS OR > IMPLIED > +# WARRANTIES ARE DISCLAIMED, INCLUDING, BUT NOT LIMITED TO, THE > IMPLIED > +# WARRANTIES OF TITLE, MERCHANTABILITY, AGAINST INFRINGEMENT, AND > FITNESS > +# FOR A PARTICULAR PURPOSE. > +# > + > ## > +"""Publication Traverser > + > +$Id: publicationtraverse.py 67630 2006-04-27 00:54:03Z jim $ > +""" > +__docformat__ = 'restructuredtext' > +from types import StringTypes > + > +from zope.component import queryMultiAdapter > +from zope.publisher.interfaces import NotFound > +from zope.security.checker import ProxyFactory > +from zope.traversing.namespace import namespaceLookup > +from zope.traversing.namespace import nsParse > +from zope.traversing.interfaces import TraversalError > +from zope.publisher.interfaces import IPublishTraverse > +from zope.publisher.interfaces.browser import IBrowserPublisher > + > +class DuplicateNamespaces(Exception): > +"""More than one namespace was specified in a request""" > + > +class UnknownNamespace(Exception): > +"""A parameter specified an unknown namespace""" > + > +class PublicationTraverser(object): > +"""Traversal used for publication. > + > +The significant differences from > +zope.traversing.adapters.traversePathElement() are: > + > +- Instead of adapting each traversed object to ITraversable, this > + version multi-adapts (ob, request) to IPublishTraverse. > + > +- This version wraps a security proxy around each traversed > object. > + > +- This version raises NotFound rather than LocationError. > + > +- This version has a method, traverseRelativeURL(), that > + supports "browserDefault" traversal. > +""" > + > +def traverseName(self, request, ob, name): > +nm = name # the name to look up the object with > + > +if name and name[:1] in '@+': > +# Process URI segment parameters. > +ns, nm = nsParse(name) > +if ns: > +try: > +ob2 = namespaceLookup(ns, nm, ob, request) > +except TraversalError: > +raise NotFound(ob, name) > + > +return ProxyFactory(ob2) > + > +if nm == '.': > +return ob > + > +if IPublishTraverse.providedBy(ob): > +ob2 = ob.publishTraverse(request, nm) > +else: > +# self is marker > +adapter = queryMultiAdapter((ob, request), > IPublishTraverse, > +default=self) > +if adapter is not self: > +ob2 = adapter.publishTraverse(request, nm) > +else: > +raise NotFound(ob, name, request) > + > +return ProxyFactory(ob2) > + > +def traversePath(self, request, ob, path): > + > +if isinstance(path, StringTypes): > +path = path.split('/') > +if len(path) > 1 and not path[-1]: > +# Remove trailing slash > +path.pop() > +else: > +path = list(path) > + > +# Remove single dots > +path = [x for x in path if x != '.'] > + > +path.reverse()