March 2001 - Zope-Dev

Zope not start as win2000 services
by kent sin 12 Mar '01

12 Mar '01

Zope installed from binary can not start as service on NT after modified been made to pythonpath registry. Before that, it runs. The event log says: string object has no attribute upper. Does this means some conflict between the python2 and the 1.52 from zope? Can somebody give me instructions to solve this? Rgs, Kent Sin __________________________________________________ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/

2 1

_getOb or __getattr__?
by Chris Withers 12 Mar '01

12 Mar '01

Hi, I need a product with custom attribute getting code and so looked to the BTreeFolder product for inspiration. It implements both __getattr__ and _getOb which appear to do roughly the same thing. What's the difference? Shane hints that __getattr__ is slower but allows acquisiton. Why is that? TIA, Chris

5 11

Is ZFormulator alive?
by erickson＠kaivo.com 12 Mar '01

12 Mar '01

ZFormulator looks like it will do some pretty neat things, but it also looks like it is "dormant". Does anyone use it, or know if it is compatible with Zope 2.2.x and/or Zope 2.3? Thanks -paul

2 1

SAP DB <--> ZODB ?
by Spicklemire, Jerry 12 Mar '01

12 Mar '01

Kapil said: > looks like it will instantly become the most advanced open source database > out there. my current rankings of the general os rdbms in terms of features > postgresql->interbase->mysql Keep in mind that (even though) it is an effective "general os rdbms", PostgreSQL was originally developed as an experimantal platform for testing concepts in the Object-Relational database realm. As such, it has some unique potential, especially in its ability to define and store Objects as _native_data_types_. Most of the effort and press related to PostgreSQL in the recent past has focused on adding SQL in place PostgreSQL's QEL, and improving stability and performance. Now that most of this important work has been accomplished, my hope is that some of the truly differentiating features can be showcased, e.g. a Zope Storage where an external SQL engine can directly interact with objects as native data. Imagine a SQL statement that updates every instance of a ZClass with new properties, or slipping a super-class in underneath (re-parenting), where it had never been before. also: > the feature list of this database is astonishing > www.sapdb.org > > to name a few > subtransactions > outer joins > sql 92 compliant, possible modes for db2, oracle 7 > very nice prodecural sql language. > lots of statistics information > really rich set of standard functions > scrollable cursors > > this is a great candidate for a zodb storage SAP DB is a fully licensed version of Software AG's (SAG) Adabas D, a descendant of SAG's venerable mainframe class system, Adabas C. Yes, SQL is fully supported, but SQL is after all a query language, not a physical model of the internal structure of an rdbms. This is not to say that Adabas, in any form, is any less of a capable system. Don't be surprised, however, to find something much closer to a network database under the surface, compared to PostgreSQL, which is about two generations newer, according to its "birth date" (1988?). Interestingly, many folks who have been around for several "database revolutions" see the network model as a very good choice for storage of "non-tabular" data types, such as objects, etc. On the other hand, that may be the result of lack of familiarity with the possibilities inherent in a true Object-Relational data model. Later, Jerry S.

1 0

Ensuring 'freshness' of dynamic pages (slightly offtopic)
by Morten W. Petersen 12 Mar '01

12 Mar '01

Hi guys, I've been trying to ensure that documents from a certain website are always fresh, that is, every request for a new page must be validated before the client sees it. I've tried using these HTTP Cache-control directives: no-cache no-store max-age (1) must-revalidate but no such luck. If I for example specify 'Check against server: Never' in IE, the pages are cached, no matter what. Is it possible to ensure that pages are always fresh, even if it is specified on the client side that the client should never try to retrieve a 'fresher' page? I've been contemplating a new solution, which is to append a string like ?ensure_freshness=02349025180.12 (some random, unique id) on each link, but that's not a thing which I wish to do now, at such a late stage in the project. Also, it strikes me that it isn't possible to specify multiple cache-control directives with the current HTTPRequest implementation, am I missing something here, or is this a bug / missing feature? Any ideas? Thanks, Morten

3 3

Re: [Zope-dev] State of ZPatterns
by Itai Tavor 11 Mar '01

11 Mar '01

Phillip J. Eby wrote: >At 10:16 AM 3/11/01 +1100, Itai Tavor wrote: >> >>I'm wondering where TransWarp leaves ZPatterns users. Until a couple >>of weeks ago ZPatterns was the best thing to happen in the Zope world >>since, well, Zope. Now it's described as a 'hack', > >Those things aren't mutually exclusive, you know. :) I realize that... I just got the impression that your own view of the value of ZPatterns suddenly took a nose dive. Guess I was wrong, and I'm happy to know that. > >demoted into >>'maintenance only' mode, and superceded by something that is >>described as being as much better than ZPatterns as ZPatterns was >>better than standard Zope development. > >It hasn't been superceded. TransWarp has several layers of functionality >planned; only one of those layers has been released to date. (By the way, >I also don't recall ever saying even that TransWarp was "better" than >ZPatterns, let alone that it was some giant leap forward. I did say that >it expanded further on the model which was the basis for ZPatterns, however.) Sorry for putting words in your mouth. My understanding was, though, that TransWarp will replace ZPatterns as the best tool for object model based development, which means that the idea of developing ZPatterns code with a view for long term future reuse suddenly isn't that attractive. > >As for "maintenance only", ZPatterns hasn't had any changes in months, >except for patches provided by its users. That's largely because it hasn't >needed any. There's not much you can *add* to ZPatterns, without a major >upheaval. I figure, let the major upheaval be directed at making something >completely different, rather than have major reworking to make something >only marginally better. I fully agree. I just didn't expect something with a 0.4.3b2 version number to remain unchanged. Also, I remember you mentioning that you're working on an SQL DataSkin. If that was really a possibility in the past, and it won't happen, then it's a major loss. Documentation/examples is another area that still needs to be improved (although the community can be considered to be responsible for that as much as yourself). > >So what do we do now? Wrap up current ZPatterns work, writing it off >>as a loss for future reuse? Or can we count on 'maintenance only' >>being sufficient to support continued reused of ZPatterns efforts >>long enough to justify the original development effort? > >Someone asked a similar question of me at the conference. I told them that >if I needed to develop a web-based application today, I would use ZPatterns >with Zope. It works, it's stable, it gets the job done. If you use it for production sites, that's good enough for me :-) > >Also, recall that the RIPP model concept was introduced to the Zope >community last January, and it was many months before ZPatterns' first >release, then many more before it was stable enough to be ready for >production use. Expect the same to be true of TransWarp. The tools >released so far are rock solid, but there isn't anywhere near enough there >to compete with ZPatterns yet. This is ok as long as you're choosing a tool per project. But I'm hate repeating work, and I expect the ZPatterns-based e-commerce app I'm building today to make the e-commerce app I build in 12 months a lot easier to build. So any changes in the basic tools I use bother me. Anyway, you did well calming my concerns. Thanks for that, and for releasing ZPatterns in the first place. -- -- Itai Tavor -- "Je sautille, donc je suis." -- itai(a)optusnet.com.au -- - Kermit the Frog -- -- -- -- "If you haven't got your health, you haven't got anything" --

1 0

ZPatterns: External Attribute Provider question
by Steve Alexander 11 Mar '01

11 Mar '01

Hi Phillip, Is there a good reason that ExternalAttributeProvider sets up a dictionary to put its attributes into, rather than a PersistentMapping? This is the way it seems to store things in the Rack: Rack's BTree | + object_id : PersistentMapping | + ('ZPatterns.Rack', 'Self') : Persistent object, eg ZClass instance + ('ZPatterns.AttributeProviders','ExternalAttributes') : {} # dict filled with external attributes The thing is that when the dict gets updated by the External Attribute Provider, the whole second-level-in PersistentMapping object gets updated. If an External Attribute Provider instead set up a new PersistentMapping in its slot, only that PersistentMapping would change in the ZODB on updates. Do you think this is a reasonable change for the External Attribute Provider class? Or, should I write my own provider, derived from External Attribute Provider, that uses a PersistentMapping instead of a dict? Actually, I can see the use for a Persistent AttributeProvider where you have a configurable slot, so you can group attributes that change at different rates in groups of those that change together. The second half of the slot-key's tuple would be configurable as a property on the attribute provider. -- Steve Alexander Software Engineer Cat-Box limited

2 1

FYI: TransWarp talk slides posted
by Phillip J. Eby 11 Mar '01

11 Mar '01

HTML Version (frames & JavaScript required): http://telecommunity.com/TransWarp/IPC9Talk.htm PowerPoint Version: http://telecommunity.com/TransWarp/IPC9Talk.ppt

1 0

Designing ZPatterns/Python-product-based, reusable applications - take 2
by Itai Tavor 11 Mar '01

11 Mar '01

Hi, I didn't get very far asking this question a little while ago, so I'm going to try again. I started developing ZPatterns applications in the ZODB. Then I figured there would be advantages to moving instead to python product-based development. My goal is to create applications that can be reused easily, but I'm having problems deciding on the best way to structure the application, how the requirement of reusability affects the structure, and how reuse will actually take place. The starting point is an object-model based application, so unlike applications like zCommerce and EMarket, where you start from an API, template, or skeleton and build on that, I'll be trying to reuse parts of an actual, complete, live application. Starting with PD classes: Those would be DataSkin-subclassed python classes. Reuse would involve either using these classes as-is, or subclassing them where changes are needed. No problem here. Then I need Specialists to manage collections of PD classes. I think that these would also benefit from being product-based, so I subclass Specialist to create a manager for each role in the application. Now, since an existing application might be expanded, the Specialists all need factory methods and will all show up in Zope's Add New Object menu. Now consider UI: - Both PD and manager classes will have administrator UI. This would not have to change unless the class is subclassed, so the UI methods can stay in the product and loaded using HTMLFile. - Many of the classes will also have web user UI, and since this has to be designed to fit the design of the web site in which the application is used, the UI methods will have to be located where designers can modify them - in the ZODB. For the Specialists, I can install a default set of UI methods when an instance of the Specialist is installed. But where do I store PD class UI methods? EMarket's solution is to store them in the Specialist, but this not only breaks O-O rules, it can also be very ugly. Say I have a Specialist managing 3 PD classes. Instead of an index_html method for each class, I'll need class1index_html, class2index_html, class3index_html in the Specialist. And each class would have to define: def index_html(self, REQUEST, RESPONSE, **kw): return self.class1index_html(REQUEST, RESPONSE, kw) This is unacceptably ugly. But the only other thing I can think of doing is creating a ZClass for each PD class and store the UI methods there. But that's also pretty unacceptable. Is there a better solution? Next, I need object connections - these are created using SkinScripts. And as far as I can tell, I can't manage SkinScript in the product - they have to be in the ZODB. Which is ok in O-O land as they're not strictly part of the PD classes, but this means that PD classes don't actually work at all until someone comes and fills in the SkinScripts - so the class actually misses a lot of the code it requires to function. Also, it means SkinScript code can't be reused with the rest of the application's components (except by cutting and pasting code pieces from an existing application). I also need storage management - which is done with Rack methods and SkinScripts in the Racks. The SkinScripts present the same problem as the ones used for object connections. Rack methods can be ZODB-based, in which case again they can't be easily reused and need to be manually created before the objects start doing anything, or they can be created in Rack-subclassed classes, but that creates a problem for RDBM storage - unless it's possible to store SQL methods in a python product. Finally, I need to wrap the whole thing up as an application. So I create a Folderish class which installs instances of all Specialists, and add application-wide utility methods. Or do I simply place all Specialists in a Folder? Now, about reuse. O-O reuse, as far as I understand it, takes place at the class level. So how do I start a new application? Do I subclass its main container class and and my own init method that adds any added or subclassed Specialists I use? Or do I create my own application class and import PD classes from the old application? Do I copy the old classes into my new product, or do I import them from the old product, which then would require that the old product is kept on the server, even if the application that product defines isn't used on this server? And, as mentioned above, how do I reuse SkinScripts, Rack methods and SQL methods? Copy and paste? It seems funny that if I create a new instance of the product, I don't get a new copy of a working application... That's it. Sorry it's so long and complex. What would be great is if someone posted a description of the structure of a working application. I'm sure this is something many people are likely to struggle with - it almost could be a HowTo. I hope, now everyone are back from the conference , that I'd have more luck getting some help :-) On the other hand, considering that ZPatterns is now being superceded before it even had a chance to mature, maybe nobody cares to hear about it anymore :-( But more on that in a separate post. TIA Itai -- -- Itai Tavor -- "Je sautille, donc je suis." -- itai(a)optusnet.com.au -- - Kermit the Frog -- -- -- -- "If you haven't got your health, you haven't got anything" --

2 1

State of ZPatterns
by Itai Tavor 11 Mar '01

11 Mar '01

Hi, I'm wondering where TransWarp leaves ZPatterns users. Until a couple of weeks ago ZPatterns was the best thing to happen in the Zope world since, well, Zope. Now it's described as a 'hack', demoted into 'maintenance only' mode, and superceded by something that is described as being as much better than ZPatterns as ZPatterns was better than standard Zope development. So what do we do now? Wrap up current ZPatterns work, writing it off as a loss for future reuse? Or can we count on 'maintenance only' being sufficient to support continued reused of ZPatterns efforts long enough to justify the original development effort? Do any other ZPatterns users share these concerns? Itai -- -- Itai Tavor -- "Je sautille, donc je suis." -- itai(a)optusnet.com.au -- - Kermit the Frog -- -- -- -- "If you haven't got your health, you haven't got anything" --

2 1