December 1999 - Zope-Dev

Re: [Zope-dev] Python Methods commited
by Anthony Baxter 18 Dec '99

18 Dec '99

>>> Evan Simpson wrote > > * Object traversal, and traversal_subpath > > In retrospect, these aren't necessary if people are willing to use > SiteAccess. If they prove to be popular, you should probably remove their > "magical" quality and put an "Allow traversal?" checkbox on the main edit > form. I'd _really_ prefer it to stay. SiteAccess rules really do tend to be an utter bastard to debug, particularly compared to pythonmethods. And I don't want to _think_ about trying to put a siteaccess rule into a ZClass. > That reminds me: Do you know why tuple, list, map, reduce, and filter aren't > surfaced? Is it the "risk" of turning strings into lists? Given how PythonMethods mean you're actually writing python, I'm finding that you pretty quickly notice the lack of some of these. (For a quick example, start writing ZCatalog based things in a PM - things like uniqueValuesFor return tuples, but searchResults will only accept a list.. bah). Anthony -- Anthony Baxter <anthony(a)interlink.com.au> It's never too late to have a happy childhood.

1 0

Adding External Methods
by Alexander Limi 17 Dec '99

17 Dec '99

When I try to add an external method, and supply the wrong function name, Zope instead complains that it can't find the module file. This is a bit misleading, and causes confusion from the newbie standpoint. It should be able to tell whether the file is missing, or just the function. Alexander Limi. http://mp3.no

1 0

RDF Use cases
by Martijn Pieters 17 Dec '99

17 Dec '99

Hello Zopatistas, Digital Creations is in the process of extending the Zope architecture towards new emerging standards, and a new interface based on Mozilla. See: http://www.digicool.com/News/mozilla http://www.zope.org/Resources/Mozilla One of these standards is RDF, the Resource Description Format. This is a foundation for processing metadata and for interoperability. With it, you can describe content and content relationships, rate content, define the structure of a website, and much, much more. All this in a machine readable format. More information about RDF can be found at the W3 Consortium website: http://www.w3c.com/Metadata/RDF Before we dive in and start writing code, however, DC will first try and define what we are up against. One thing we do, is define Use Cases. Use Cases describe ways in which people would use RDF. A few examples: - RDF Sitemaps. The browser of visitor of a Zope generated website can retrieve a RDF Sitemap of the website, for presentation to the visitor. For a preliminary specification of RDF Sitemaps, see: http://rudolf.opensource.ac.uk/about/specs/sitemap.html - Content syndication. A website administrator makes some of its content available via a RDF datasource. - Metadata embedded in HTML. HTML pages returned from Zope have embedded in their header an RDF description of the document, including information on author, language, publisher, etc. This could be based on the Dublin Core metadata proposal. - Advanced management of Zope objects. When describing the contents of a ZODB database in RDF, RDF clients like Mozilla can query and manipulate these contents with alternate interfaces. This is your opportunity to help in this process. We want to know from you all what your ideas are for using RDF in Zope. Nothing is too crazy or inventive! Let me know what you can come up with. -- Best regards, Martijn Pieters mailto:mj@digicool.com

1 0

ZopeTalk (an invitation)
by Eron Lloyd 17 Dec '99

17 Dec '99

Greetings, fellow Zopers! Wanna talk Zope? Have a handful of newbie questions? Get together with others 24 hours a day in the EFNet Zope IRC channel. The irc.ais.net server is a good one to try - just type "/join #zope" to get in on the discussion. We'll send an invitation once a month to remind everybody about it (this could cut down on mailing list traffic...). See you there! Aquarius :-)

1 0

Final GenericUserFolder beta (GUF)
by Stuart 'Zen' Bishop 17 Dec '99

17 Dec '99

This should have been out a while ago - I blame my playstation... Added optional userExists hook to improve performance for sites with large user populations. Implemented basic user authentication properly. I believe all outstanding issues have been dealt with - let me know if this is not the case. -- ___ // Zen (alias Stuart Bishop) Work: zen(a)cs.rmit.edu.au // E N Senior Systems Alchemist Play: zen(a)shangri-la.dropbear.id.au //__ Computer Science, RMIT WWW: http://www.cs.rmit.edu.au/~zen

1 0

Re: [Zope-dev] __cmp__ in python classes?
by Jim Fulton 16 Dec '99

16 Dec '99

Benno Rice wrote: > > On Tue, Dec 14, 1999 at 01:52:22PM +0000, Jim Fulton wrote: > > > > > > I thought this looked familliar. Unfortuantely, I don't have that > > > patch anymore. I squeezed some old brain cells REALLY hard, and I think I > > > recall trick. It's only a few lines, let's see if I can recreate it > > > today. > > > > Let me search my old mail first. > > > > > BTW, AFAIK I don't have the ability to commit to the CVS tree. > > > > Hm. OK, assuming I find your mail, I'll check it in. This probably > > won't happen till tomorrow. > > Were you able to find the patch? No. I just checked in a more ambitious patch that: - Implements the numeric protocol. This means that wrapped numeric objects should work. - Fixed comparison betwee wrapped and unwrapped objects. - Pass wrapped objects to comparison operators, so that acquired data can be used in comparisons. :) I also fixed a bug in some numeric coersion code in ExtensionClass. So two C files were updated, ExtensionClass.c and Acquisition.c. You'll need to rebuild the corresponding extensions (by running make in lib/python) after checking these out. Jim -- Jim Fulton mailto:jim@digicool.com Technical Director (888) 344-4332 Python Powered! Digital Creations http://www.digicool.com http://www.python.org Under US Code Title 47, Sec.227(b)(1)(C), Sec.227(a)(2)(B) This email address may not be added to any commercial mail list with out my permission. Violation of my privacy with advertising or SPAM will result in a suit for a MINIMUM of $500 damages/incident, $1500 for repeats.

2 1

Inheritance problems with ZClasses with Python base classes - workaround
by Itamar Shtull-Trauring 16 Dec '99

16 Dec '99

In short, and I haven't verified it in a test case, if your ZClass inherits from both ZClasses and from regular python classes, the python classes must be first, or you won't be able to access the python classes methods. Instead you might get a "missing doc string" error, or perhaps some other error. I need to try this with a simpler test case, but simply changing the order of inheritance in my ZClass so python classes were first solved my problem. And yes, all my functions had doc strings and yes, the methods that didn't work originally were only defined in the python classes. What I tried, proof of my claim: CatalogAware python class -> Workflow python class -> ZWorkflow ZClass ObjectManager python class -> OrderedObjectManager python class (mine) -> Foodbase pytohn class and then Food inherited from ZWorkflow and FoodBase. This didn't work. ****** But inheriting directly from just Workflow and FoodBase did work - ******* ******* wrapping Workflow in a ZClass is what caused my problem ******* Inheritng directly from Workflow, then from a new plain-vanilla ZClass I created and then from FoodBase didn't work. So I change the order: Inheriting from Workflow, then FoodBase (both python classes), then from the ZClass solved the problem! -- Itamar S.T. itamars(a)ibm.net

1 0

Re: [Zope-dev] Why is TemplateDict so opaque? -and- Have you seen this problem?
by Dave Parker 16 Dec '99

16 Dec '99

Anthony Baxter wrote: > I'm pretty sure that a patch would get added if it was written and submitted. > :-) If I were not forehead deep in a "user level" project I'd probably give it a try. But at the moment I'm struggling to get this thing done. Anyway, it's really the namespace-going-away issue that kills me more than anything else (see original message) - I think if that one got fixed 98% of my "problems with zope" would be gone. As it stands I basically can'd do more than the most basic abstractions in dtml space because I'm limited to a depth of two expression-evaluated method calls - ARGHGH! > (don't look at me, I'm still fighting my way through the ZODB machinery. > The horror, the horror.) Oof. ;-)

1 0

Re: [Zope-dev] Why is TemplateDict so opaque? -and- Have you seen this problem?
by Anthony Baxter 16 Dec '99

16 Dec '99

>>> Dave Parker wrote > [sorry folks, I'm tired, so this is halfway to a rant :-( ] it's ok. I was wandering around the office kicking things and looking dazed a couple of weeks ago when I finally delved into this stuff. > I'm not happy that I now know what TemplateDict is - I really wish I > hadn't had to go there, but now that I'm there I have a much better > understanding of why it is impossible to get information out of it; it > has "do you have this key" and "get the value of this key" and > absolutely nothing else. The problem is that it's implemented as a multimapping. To do keys(), items(), &c you'd have to walk the multimapping and search each object. I guess you could simply do something like this (untested, written from just looking at the MultiMapping.c code. Probably could be implemented far more easily...) t = templatedictobj l = [] while 1: try: l.append(t.pop()) except IndexError: break # l now has the dictionaries from the namespace stack l.reverse() d = {} for i in l: d.update(i) t.push(i) print d.items(), d.keys(), d.values() this code just pops all the namespace objects off the multimapping, stores them in a list, merges them together and pushes them back onto the multimapping. it's horribly inefficient. and probably buggy. :) Anthony

2 1

ugh. bug(?) in ZCatalog.ZopeFindAndApply
by Anthony Baxter 16 Dec '99

16 Dec '99

ooo, this has been an annoying one to track down. When ZCatalog's ZopeFindAndApply is called, it looks at container.aq_base.objectItems() to list the objects in a container, then indexes them. This means that if you're indexing an attribute that's acquired from higher up the chain than the immediate enclosing folder, it won't get picked up. It _will_ when you call manage_catalogReIndex, as it does a resolve_url of the object. This caused much confusion for me, as the Catalog bug I was seeing would be there sometimes and not others, depending on whether a reindex had happened. Two possible fixes here: First is to force a resolve_url before indexing. Second is to index the object in the initial acquisition context (and pass the correct object to the recursive call) Not sure which is the better fix - both work for me. The second is probably cheaper, and "more correct" Note that there's another fix needed for this to work that's in the collector - there's a call to 'find' that should be 'string.find' (line numbers may be off - there's other magic stuff in my local copy.) Can anyone else confirm that this is an appropriate approach? Or is ZopeFindAndApply meant to act this way? ta, Anthony RCS file: /cvs-repository/Zope2/lib/python/Products/ZCatalog/ZCatalog.py,v retrieving revision 1.42 diff -u -r1.42 ZCatalog.py --- ZCatalog.py 1999/11/05 18:58:54 1.42 +++ ZCatalog.py 1999/12/16 07:39:38 @@ -533,7 +559,14 @@ ) ): if apply_func: - apply_func(ob, (apply_path+'/'+p)) + # needed because earlier we retrieved + # container.aq_base.objectItems, otherwise + # we don't see anything acquired from more + # than one acquisition level higher + if REQUEST: + ptmp = apply_path+'/'+p + ob = self.resolve_url(ptmp, REQUEST) + apply_func(ob, (ptmp)) else: add_result((p, ob)) dflag=0 RCS file: /cvs-repository/Zope2/lib/python/Products/ZCatalog/ZCatalog.py,v retrieving revision 1.42 diff -u -r1.42 ZCatalog.py --- ZCatalog.py 1999/11/05 18:58:54 1.42 +++ ZCatalog.py 1999/12/16 07:39:38 @@ -532,14 +558,16 @@ role_match(ob, obj_permission, obj_roles) ) ): + ob_real = getattr(obj, id) if apply_func: - apply_func(ob, (apply_path+'/'+p)) + apply_func(ob_real, (apply_path+'/'+p)) else: - add_result((p, ob)) + add_result((p, ob_real)) dflag=0 if search_sub and hasattr(bs, 'objectItems'): - self.ZopeFindAndApply(ob, obj_ids, obj_metatypes, + ob_real = getattr(obj, id) + self.ZopeFindAndApply(ob_real, obj_ids, obj_metatypes, obj_searchterm, obj_expr, obj_mtime, obj_mspec, obj_permission, obj_roles,

1 0