August 1999 - Zope - Zope lists

RE: [Zope] Idea: ZTest: Integrated Use-case based web site testin g.
by Rob Page 27 Aug '99

27 Aug '99

> Once I was thinking about a scheduler object for Zope and actually > made the first steps and designed it around Tim O Malleys(?) eventloop > module. However I gave up when I considered autorization > details. Do you > have any good ideas on how to deal with authorization? What about giving the scheduler object Proxy Roles like you can for other documents..?? --Rob

1 0

RE: [Zope] Idea: ZTest: Integrated Use-case based web site testin g.
by Alexander Staubo 27 Aug '99

27 Aug '99

I agree that it would be nice -- I've been thinking of it earlier, then submitted to chron, because I'm lazy -- I'm just saying that we do have the tools, and running ZClient externally really isn't that different. Ok, I'm sold. I disagree slightly about your event object design though. Imho the event object should be similar to a DTML Method -- so the event "action" would in fact be DTML. The immediate benefit is that you can do anything -- you're not restricted to launching URLs, and you get the Zope context in the box. You could allow passing the calling event object to other parts of the site. If event objects descended from DTMLMethod, you could also use event objects just like DTML Methods and call them up -- manual execution if you will. The scheduler should support concurrency, not just queueing. In fact, all we need is a simple "max concurrent tasks" semaphore to regulate the thread count from the master thread. A scheduler also needs a logging/reporting system. A history API would be cool -- you could so <dtml-in "MyScheduler.pastEvents()"> and build your own schedule reports. Maybe all output from the event DTML should be attached to the history entry. And so on. Who writes this thing? This being an open source initiative(TM), I'm all for chipping in. Anyone want to join the Scheduler Task Force? -- Alexander Staubo http://www.mop.no/~alex/ "Give me an underground laboratory, half a dozen atom smashers and a beautiful girl in a diaphanous veil waiting to be turned into a chimpanzee, and I care not who writes the nation's laws." --S. J. Perelman > -----Original Message----- > From: Martijn Pieters [mailto:mj@antraciet.nl] > Sent: 27. august 1999 09:47 > To: Alexander Staubo > Subject: RE: [Zope] Idea: ZTest: Integrated Use-case based web site > testin g. > > > At 08:53 27/08/99 , you wrote: > >Aside: Do you really need this within Zope if you can run a ZClient > >script with chron? While a scheduler within Zope isn't a bad > idea, the > >infrastructure -- at least on Unix and NT -- is already there. > > I do see the use of a web-administrable scheduler for Zope. > This way users > without telnet to the server can schedule calls to Zope > objects as well. It > would also make cross-platform, easily transportable > scheduled applications > possible. Just imagine: You develop on an NT server, then > just copy over > Data.fs to a Sun server at the ISP, and all scheduled calls > are set up as > well.. No fuss, no hassle. > > A scheduler object would need: > * Permissions, so a site maintainer can determine who can > schedule calls. > * A separate thread to manage it's queue. > > Scheduled events would need: > * Interval settings (one time only, repeat, number of > repeats, repeat interval) > * URL to call > * Security settings > * Call parameters (including call type, GET or POST) > * Maybe a file field to upload? > > -- > Martijn Pieters, Web Developer > | Antraciet http://www.antraciet.nl > | Tel: +31-35-7502100 Fax: +31-35-7502111 > | mailto:mj@antraciet.nl http://www.antraciet.nl/~mj > | PGP: http://wwwkeys.nl.pgp.net:11371/pks/lookup?op=get&search=0xA8A32149 ------------------------------------------

3 2

RE: [Zope] Semi-persistent objects, volatile attributes
by Alexander Staubo 27 Aug '99

27 Aug '99

> Alexander Staubo wrote: > > > > Is it possible to have persistent objects with attributes > that are only > > persistent in memory? > > Yes, but it takes some work and the mechanism need to be formalized. > Would you want to specify this on a per-class or per-instance basis? That would be great. Personally? Per class. But wouldn't it be possible to do something like self._p_nodeactivate = 1 or was that being too naive? :) [snip] > > > > def my_method(self): > > self._v_foo = (hasattr(self, '_v_foo') and self._v_foo) or > > '<default>' > > This is one way. > > > ...ad tedium if you have lots of methods. > > Yup. > > > Does the persistence engine > > support an "onload" mechanism which you can override to > know when your > > instance has been loaded and needs to initialize itself? > > Yup, it's called __setstate__. __setstate__ is called when > the object state is loaded. You can overload __setstate__ > and initialize any volitile variables when you load your state: > > > class C: > def __setstate__(self, state): > C.inheritedAttribute('__setstate__')(self, state) > self._v_spam=..... Just what I was looking for. Yum, thanks. For future reference and posterity's sake, is this stuff documented anywhere except in the UML documentation? > Jim > > -- > Jim Fulton mailto:jim@digicool.com Python Powered! > Technical Director (888) 344-4332 http://www.python.org > Digital Creations http://www.digicool.com http://www.zope.org -- Alexander Staubo http://www.mop.no/~alex/ "Give me an underground laboratory, half a dozen atom smashers and a beautiful girl in a diaphanous veil waiting to be turned into a chimpanzee, and I care not who writes the nation's laws." --S. J. Perelman

1 0

managing properties dynamically
by Heiko Stoermer 27 Aug '99

27 Aug '99

another one from me... :-) I have a form that contains input fields for some values (say 'header' and 'text'). After submitting the form I have to manage_changeProperties() of an object that has the above properties, but also many more. Of course only the property values entered in the form are supposed to be changed. now comes the tricky bit: the method that is supposed to change properties does know the target object, but it does not yet know which properties are to be changed. These are defined dynamically in the preceding form (and there exist various of these). In theory the solution is pretty simple: <with target object> <in propertyItems() of the target object> <if there is a corresponding form field in REQUEST> --> manage_changeProperties(current property = REQUEST[current]) </if> </in> </with> The arrow indicates the "problem area": I have the name of the property (and thus the name of the input field) in a variable called "current". if if put  nothing happens. Zope simply runs over the code and does _nothing_. What goes wrong here, and how do I get it running? Thanks so far, Heiko -- Heiko Stoermer MIG Augsburg

1 0

Zope 2.0.0b5 released...
by Jim Fulton 27 Aug '99

27 Aug '99

Zope 2.0.0 beta 6 is available at: http://www.zope.org/Download/Releases/Zope-2.0.0b6 This release has a number of minor bug fixes. -- Jim Fulton mailto:jim@digicool.com Python Powered! Technical Director (888) 344-4332 http://www.python.org Digital Creations http://www.digicool.com http://www.zope.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 2

Zope 2.0.0b6 released...
by Jim Fulton 27 Aug '99

27 Aug '99

Zope 2.0.0 beta 6 is available at: http://www.zope.org/Download/Releases/Zope-2.0.0b6 This release has a number of minor bug fixes. Jim -- Jim Fulton mailto:jim@digicool.com Python Powered! Technical Director (888) 344-4332 http://www.python.org Digital Creations http://www.digicool.com http://www.zope.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.

1 0

Semi-persistent objects, volatile attributes
by Alexander Staubo 27 Aug '99

27 Aug '99

Is it possible to have persistent objects with attributes that are only persistent in memory? Effectively, an object that is always cached, for volatile attributes to be preserved during the Zope process' execution. Is there a way for an object to load itself permanently into the cache, say? I figure I could store the attributes in the product module as part of a dictionary keyed by the object instance, but this feels very dirty. Speaking of volatile attributes, is there an idiom for handling these? Since they're not stored, and __init__() is only ever called when the object is initially created (not persistently loaded), it seems you have to verify their existence, something like: def my_method(self): self._v_foo = (hasattr(self, '_v_foo') and self._v_foo) or '<default>' ...ad tedium if you have lots of methods. Does the persistence engine support an "onload" mechanism which you can override to know when your instance has been loaded and needs to initialize itself? Pointers to appropriate documentation would be nice. I've been rummaging through the ZODB UML stuff and while it's enlightening, the documentation is also quite fragmented. -- Alexander Staubo http://www.mop.no/~alex/ "Give me an underground laboratory, half a dozen atom smashers and a beautiful girl in a diaphanous veil waiting to be turned into a chimpanzee, and I care not who writes the nation's laws." --S. J. Perelman

2 1

RE: [Zope] Re: Saving Pipe3 object
by Alexander Staubo 27 Aug '99

27 Aug '99

The underscore prefix should be enough: Your variable can be named _long_R_Pipe and it won't be pickled. DC uses the _v_ prefix convention for various stuff, possibly meaning "value"... Who knows. -- Alexander Staubo http://www.mop.no/~alex/ "QED?" said Russell. "It's Latin," said Morgan. "It means, 'So there you bastard'." --Robert Rankin, _Nostramadus Ate My Hamster_ > -----Original Message----- > From: Li Dongfeng [mailto:mavip5@inet.polyu.edu.hk] > Sent: 26. august 1999 15:13 > To: ZOPE List > Subject: [Zope] Re: Saving Pipe3 object > > > I have solved this problem. Giving an non-persistent attribute > a name starting with "_v_" can avoid pickling it. So I used > setattr(self, "_v_long_R_Pipe", long_R) > and the attribute can be access outside of the defining method. > > Li Dongfeng wrote: > > > I'm using ZOPE to develope a statistics server using R(a GPLed > > statistical software). For this, I started the R executable with > > popen2.Popen in an external method, but I need to exit the method > > without shutting down the pipe, so I used > > setattr(self, "some_name", the_pipe) to save the Pipe3 object > > as an attribute, > > and got an error message: Object unpicklable. > > > > I understand a open pipe cannot be restored from a file, > > but can I save the Pipe3 object reference somewhere > > in the running ZOPE's name space without pickling? > > Then I can use it in other method to control the > > running R process. > > > > This is a general problem if you want to use Zope > > to control other long-running processes. > > > _______________________________________________ > Zope maillist - Zope(a)zope.org > http://www.zope.org/mailman/listinfo/zope > > (To receive general Zope announcements, see: > http://www.zope.org/mailman/listinfo/zope-announce > > For developer-specific issues, zope-dev(a)zope.org - > http://www.zope.org/mailman/listinfo/zope-dev ) >

4 3

Re: [ZCommerce] Worthy group agenda: ZOPE dynamic pages and search engines... engines...
by Joerg Herbst 27 Aug '99

27 Aug '99

> CodeIt and other zCommerce folks, > > Thanks for starting this important list. > > The biggest concern we have about moving our sites to Z2/PTK is the > 'invisibility' factor which can haunt sites based largely on dynamic pages > rather than static HTML files. > > We have heard anecdotal comments by folks on the general list saying, "No > sweat, I look at my logs and search engines crawlers and robots index my > site all the time." Crawling and actually indexing are two different things. > And which crawlers? There are no blanket easy answers here. > > Search engine indexing can be the lifeblood of certain types of sites. One > of our sites, http://popculturestore.com, started out as a raunchy > Fusion-generated collection of template-based pages for the hundreds of > collectible items in the store's inventory. > > With little effort we started getting orders from all over the world. Each > hard-coded page was indexed on its own right. > > Then we made The Pop Culture Store MUCH better for human visitors with some > Perl scripts. Within weeks we dropped off the face of the earth! Search > engines were not indexing our 'bread and butter' pages. > > Of course the situation should be different with the URL-addressing of ZOPE > pages. But how different? > > We need an 'ad hoc' study group to create or identify a 'guinea pig' > e-commerce site which we can use to test and confirm the 'indexability' of > ZOPE pages with EACH of the MAJOR search engines. (We use First Place > Software's WebPosition Gold for SE submissions and indexing validation. It > would be an excellent tool for running such formal tests.) > > Saying that SE indexing of ZOPE sites 'should work' is not going to cut it > for REAL business folks with REAL sites that depend on REAL visibility. > > We need to find out exactly what our 'search engine friendliness' factor is > and then have real information with which to respond to this and similar > questions. > > If we find that major search engines have trouble with ZOPE dynamic pages, > then we have identified a development requirement to allow ZOPE site > designers to include a 'shell of static doorway pages' generated and updated > as part of their template-based site design. > > We believe overcoming customers' objections to dynamic page designs will be > one of the BIGGEST hurdles in getting ZOPE used for e-commerce. > > We are very interested in collaborating with others to address this issue. > Due to entrepreneurial commitments, however, we cannot lead such an effort. > > Thanks again for starting this list, > --Jim and Timlynn-- > Hi, I guess this is not just a ZCommerce Problem, it's a general Problem of dynamically websites, built for a shop or just for information use. Any project with a lot of database stored content seems to have this problem. Maybe we can reach some effect with providing META Information on dynamic pages and with the intensive use of links. But for a general solution we need to know how robots work when meeting Zope webpages, what they do and what not. I'm not that great with such Zope details, but maybe somebody else know, how we can handle search engines for Zope. Joerg Herbst

2 1

RE: [Zope] Idea: ZTest: Integrated Use-case based web site testin g.
by Alexander Staubo 27 Aug '99

27 Aug '99

> -----Original Message----- > From: Martijn Pieters [mailto:mj@antraciet.nl] > Sent: 27. august 1999 10:18 > To: Alexander Staubo; Zope Mailing List (E-mail) > Subject: RE: [Zope] Idea: ZTest: Integrated Use-case based web site > testin g. [snip] > >Who writes this thing? This being an open source > initiative(TM), I'm all > >for chipping in. Anyone want to join the Scheduler Task Force? > > Well, we are getting a nice list of requirements already =). > Anyone having > some list space? Or shall we just move over to Zope-Dev? I have list space to spare, but I don't think this is big enough for a long-winded discussion -- let's implement rather than jerk off. ;) If you're interested, Martijn, we two can kick off this one. Or, we can set up a list and start jerking off. ;) > > -- > Martijn Pieters, Web Developer > | Antraciet http://www.antraciet.nl > | Tel: +31-35-7502100 Fax: +31-35-7502111 > | mailto:mj@antraciet.nl http://www.antraciet.nl/~mj > | PGP: http://wwwkeys.nl.pgp.net:11371/pks/lookup?op=get&search=0xA8A32149 ------------------------------------------ -- Alexander Staubo http://www.mop.no/~alex/ "Give me an underground laboratory, half a dozen atom smashers and a beautiful girl in a diaphanous veil waiting to be turned into a chimpanzee, and I care not who writes the nation's laws." --S. J. Perelman

1 0