- Zope-Dev - Zope lists

Re: [Zope-dev] Properties Introspection
by Tres Seaver 29 Feb '00

29 Feb '00

Josh Zeidner <jmz_phylogenic(a)hotmail.com> asked: > > I am trying to develop a way to dynamically introspect an objects > properties( specifically, a ZClass ). How can I determine which sub-objects > of a Zope object are properties? Can I do this in DTML or do I need to use > python? Umm, does the 'hasProperty()' method do what you want? <dtml-if "hasProperty( 'Foo' )"> <p>My Foo property is <dtml-var Foo>.</p> </dtml-if> -- ========================================================= Tres Seaver tseaver(a)palladion.com 713-523-6582 Palladion Software http://www.palladion.com

1 0

Re: [Zope-dev] Zope unit testing framework?
by Tres Seaver 29 Feb '00

29 Feb '00

Martijn Faassen <faassen(a)vet.uu.nl> > Hi there, > > Recently I've been reading some books [1] which put emphasis on the idea > of unit testing. At the Python conference Joel Shprentz held a short talk > about doing unit tests in Python [2] > > Basically a unit tests are some simple test functions that are associated > with a unit of code (a python module). The tests exercise the code in > the module, and either succeed or fail (there may also be an error due > to an uncaught exception). Each time you change something in the module > you run the associated tests. The idea is that this makes debugging easier, > catching bugs more quickly and giving better diagnostics. Doing unit tests > in Python isn't hard to do. > > I'm having more trouble thinking of a good way to do unit tests in Zope, > though. Partially this is because Zope has various parts (the management > interface for folders, ZClasses, product development), which each may need > a different testing strategy. Partially this is because Zope currently mixes > model and view together (though there are thoughts to do something about > this [3]) -- unit tests are about testing code (methods and such), not > user interfaces. Another problem is that much Zope code is not easy to test > in isolation from Zope. Too much of the code depends on various Zope > facilities (base classes, object database integration). > > Has anybody any ideas on how to do unit testing within the Zope framework? I'd > like to try adopting somekind of unit test strategy for my Zope > development. I got addicted to unit testing after reading John Lakos' excellent book, _Large Scale C++ Software Design_. He emphasises not only unit tests, but a more elaborate strategy, "design for testability:" not only does each module come with its own tests, but the system knows about dependencies between modules, and guarantees to test the depended-on modules before their dependents (so as to avoid spurious errors). In the Zope world, I think this has to translate to somthing like: 1. Each Python module should have a set of associated tests, which would by preference live inside the "if __name__ == '__main__'" block at the bottom of the module (an alternative would be to use Tim Peters' DocTest stuff, which embeds the unit tests in docstrings). One would then have a script, like the one which compiles all the Python modules during installation, to run these modules (perhaps using some fancy dependency analysis based on the import statements?) 2. For testing ZClasses, I am afraid that some kind of ZClient-based test will be required. Each test would: * Generate a suitably-unique folder name. * Construct a folder of that name off the root of the ZODB. * Build all scaffolding objects inside the new folder. * Run the tests, capturing the results and comparing them against expectations. * Finally, blow away the sandbox folder. Perhaps, as you point out, some of the effort to disentangle the core Python model from the HTML view will make such testing easier to lay down in the Zope core itself. Given a suitable framework, product developers could even allow third parties to contribute tests highlighting bugs they have found. > > [1] > Martin Fowler, ed., Refactoring: Improving the Design of Existing Code, > Addison Wesley Longman, 1999 This is the best book I have read on the actual practice of "Extreme Programming" -- I recommend it highly. > > Kent Beck, Extreme Programming Explained, Addison Wesley Longman, 1999 This book is a somewhat higher-level view of XP, and an "apologetic" for it (in the old, technical sense). Still quite useful to read, but not as much meat for the working programmer to digest. > > [2] > http://www.python.org/workshops/2000-01/proceedings/posters/shprentz/shpren… > http://c2.com/cgi/wiki?PythonUnit I have put up the code I currently use for unit testing Python modules at: http://www.zope.org/Members/tseaver/PyUnit > > [3] > http://www.zope.org/Resources/Mozilla/ZWiki/ZopeMVC > http://www.zope.org/Resources/Mozilla/ZWiki/ModelViewController -- ========================================================= Tres Seaver tseaver(a)palladion.com 713-523-6582 Palladion Software http://www.palladion.com

3 3

Re: [Zope-dev] The Debugger Is Your Friend
by Tres Seaver 28 Feb '00

28 Feb '00

As Zsa-Zsa would say:: I am just *sure* that you are putting this lovely document on zope.org as a How-To. -- ========================================================= Tres Seaver tseaver(a)palladion.com 713-523-6582 Palladion Software http://www.palladion.com

2 1

RE: [Zope-dev] Compiling Zope with cygwin
by Miller, John M. 28 Feb '00

28 Feb '00

Anthony Pfrunder did some neat work on this a ways back, you might want to try hunting him down. Once upon a time, long ago (about eight months, pre 2.0) I had Zope compiling from sources on my NT box. Of course I've almost completely forgotten how I got it to work, but here's the jist of what I did: I started off with the Python 1.5.2 binary install for windows. Haven't tried compiling Python using mingw32 however. Check out http://starship.python.net/crew/kernr/mingw32/Notes.html and follow the instructions there. After that it was a direct path to building Zope, though I may have had to do a little trickery to get the Extensions to compile sanely. If I remember correctly I simply deleted the code that compiled the extensions in the build script for Zope and instead used compmgw.py to build the Extensions. Not much help, I know, but hopefully it will put you a little further down the path, John > -----Original Message----- > From: Michel Pelletier [SMTP:michel@digicool.com] > Sent: Monday, February 28, 2000 1:40 PM > To: zope-dev(a)zope.org > Subject: [Zope-dev] Compiling Zope with cygwin > > > Has anyone gotten this to work? I can get Python to compile, and I can > get Zope to compile, but I can't get Python to compile w/ threads, and > thus Zope complains and bombs. Does anyone know if it's possible to > compile a python extension against one python but run it with another? > I'll keep playing if no one has a better clue. > > -Michel > > _______________________________________________ > Zope-Dev maillist - Zope-Dev(a)zope.org > http://lists.zope.org/mailman/listinfo/zope-dev > ** No cross posts or HTML encoding! ** > (Related lists - > http://lists.zope.org/mailman/listinfo/zope-announce > http://lists.zope.org/mailman/listinfo/zope )

1 0

Properties Introspection
by Josh Zeidner 28 Feb '00

28 Feb '00

Hello, I am trying to develop a way to dynamically introspect an objects properties( specifically, a ZClass ). How can I determine which sub-objects of a Zope object are properties? Can I do this in DTML or do I need to use python? -thanks Josh Zeidner ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com

1 0

Re: [Zope-dev] Compiling Zope with cygwin
by Tres Seaver 28 Feb '00

28 Feb '00

Michel Pelletier <michel(a)digicool.com> > To: zope-dev(a)zope.org > Subject: > > Has anyone gotten this to work? I can get Python to compile, and I can > get Zope to compile, but I can't get Python to compile w/ threads, and > thus Zope complains and bombs. Does anyone know if it's possible to > compile a python extension against one python but run it with another? > I'll keep playing if no one has a better clue. I think you want to be using the mingwin32 stuff to get threading right (the plain ole cygwin libc is flaky about threads -- mingwin32 uses the MS RTL). Robert Kern has a set of directions up about it on the starship, under: http://starship.python.net/crew/kernr/ -- ========================================================= Tres Seaver tseaver(a)palladion.com 713-523-6582 Palladion Software http://www.palladion.com

1 0

zClasses with zClasses - can they act in a folderish way?
by morton＠dennisinter.com 28 Feb '00

28 Feb '00

Ive been experimenting with zClasses nested within zClasses. I can get all kinds of behaviours happening and its extremely usefull. What I cant seem to achieve is to get a nested zClass to display in a folderish manner in the managment tree. I can get zClasses at the top level of a product to display in a folderish way (i.e. having a '+' control to expand its subtrees), but I cant get a zClass embedded inside a zClass to display this way. Ive tried subclassing from Folder, ObjectManager, nothing, etc etc, but nothing seems to work. Any suggestions anyone?

1 0

Compiling Zope with cygwin
by Michel Pelletier 28 Feb '00

28 Feb '00

Has anyone gotten this to work? I can get Python to compile, and I can get Zope to compile, but I can't get Python to compile w/ threads, and thus Zope complains and bombs. Does anyone know if it's possible to compile a python extension against one python but run it with another? I'll keep playing if no one has a better clue. -Michel

1 0

Re: [Zope-dev] Zopifying an FTP'd/POST'ed HTML document?
by Graham Fawcett 28 Feb '00

28 Feb '00

Chris Withers wrote: > Hi, > > Could one of you two please forward the response to the question to the zope > list, 'cos it didn't make it and I'd really like to know what it was :-) > > cheers, > > Chris Here it is, Chris. Itamar Shtull-Trauring wrote: > Graham Fawcett wrote: > > > > Has anywone worked on a mechanism for taking HTML documents that are FTP'd > > or HTTP-POST'ed by end-users and manipulating them into Zopish documents > > -- inserting and <..._footer> tags, for instance? > > Check out load_site.py in Zope's utilities folder, expecially the -D > option. Or my version of this code: > http://www.zope.org/Members/itamar/load_sitepossible only between equals > > Itamar S.T. itamars(a)ibm.net A good approach to the problem--it keeps all the hacked-off bits of the original message. When I get the time, I think I'll still hook this into a custom FTPd to Zopify/unZopify HTML documents published to the server. (The usefulness of an unZopify method might be questionable in some cases, but I'd like to be able to "give back what I got" if that's what the client asks for.) And BTW, Itamar, sorry for the (twice! oops..) misplaced thanks--whether you're the author or not, I'm equally grateful for your help. -- Graham

1 0

RE: Why missing connection._storage.load?
by Martijn Pieters 28 Feb '00

28 Feb '00

From: Loren Stafford [mailto:lstaffor@dynalogic.com] > > My product's initialize(context) procedure creates a ZCatalog if that > catalog doesn't already exist. I've defined an additional method > armed_event() for the Catalog object. Later, the same product > _init_.py > module calls armed_event() (OK -- it's called from a thread spawned by > initialize(context)). Do you create a new DB connection? > > When the catalog is created in the same product load process, > the call to > the catalog method armed_event() yields this output from pdb: > > > > C:/Zope214\lib\python\Products\ZScheduler\ZSchedule.py(47)arme > d_event() > > ...which is what I would expect. > > However, when the catalog already exists (from a previous > product load), > the call to the catalog method yields this output: > > > C:\Zope214\lib\python\ZODB\Connection.py(386)setstate() > > I didn't write any setstate methods, so I presume this is > extension class or > persistence voodoo. If so, the fact that pdb doesn't show how > it got to > setstate must be due to the fact that that logic is written > in C. Somewhere, > somehow, (perhaps via an inherited __setstate__ method) Zope > has figured out > that the catalog object has to be loaded from the ZOSB before it can > proceed, and calls on Connection.py\setstate to help do so. > Right, so far? > Indeed, via the Persistence class, the ZODB is asked to retrieve the object. How do you access the Catalog? You should obtain a new Root object for your new thread, and traverse to the Catalog via that reference. When you use the ZServer Monitor, you have to do the same thing: import Zope app=Zope.app() You then have your own connection to the database. When you are done with it (basically every time you go to sleep in the thread), you close it again: app._p_jar.close() del app This way you keep conflicts down, performance up. > OK, tracing along in setstate(), we get to this statement and > exception: > > > C:\Zope214\lib\python\ZODB\Connection.py(391)setstate() > -> p, serial = self._storage.load(oid, self._version) > (Pdb) s > AttributeError: "'None' objec...ribute 'load'" > > C:\Zope214\lib\python\ZODB\Connection.py(391)setstate() > -> p, serial = self._storage.load(oid, self._version) > (Pdb) > > OOps! This connection's storage object has no load method. > How can that > happen? Because pdb isn't showing how it got here, it's hard > to see what the > cause is. Can someone who's been here before help me out? > > Was the Catalog object not defined properly or not > initialized properly? If > not, how is the Catalog storage object supposed to be initialized? The storage is not defined. Probably because the Conection has been closed in the meantime, which is what happens as soon as initialisation of products is done. See lib/python/Zope/__init__.py HTH -- Martijn Pieters, Software Engineer | Digital Creations http://www.digicool.com | Creators of Zope http://www.zope.org | mailto:mj@digicool.com ICQ: 4532236 | PGP: http://wwwkeys.nl.pgp.net:11371/pks/lookup?op=get&search=0xA8A32149 -------------------------------------------

2 1