[Zope-dev] Re: Possible performance problem in Page Template engine?
Ok, little correction here: I used ploneout on plone-trunk and it then uses Zope trunk and thus 2.11. This problem seems to be a new problem coming from refactoring: r78767 | andreasjung | 2007-08-12 12:55:45 +0200 (So, 12 Aug 2007) | 4 lines - Collector #2339: ZPT: fixed unicode issue when using the 'structure' directive - ZopeContext: added evaluateStructure() and refactored _handleText() in order to deal with unicode issues when using the 'structure' directive -- christian ___ 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 )
[Zope-dev] Possible performance problem in Page Template engine?
Hi everybody! When I was trying to fix portlets in Plone I was wondering why the render() method of my portlet actually gets called multiple times. After a bit of investigation I ended up in the Five product in browser/providerexpressions.py where provider expressions also seem to be called twice for each provider statement. The problem seems to be in Products/PageTemplates/Expressions.py in line 199/200 (in Zope 2.10.4): def evaluateStructure(self, expr): """ customized version in order to get rid of unicode errors for all and ever """ text = super(ZopeContext, self).evaluateStructure(expr) return self._handleText(text, expr) This construct seems to evaluate expr twice. One time in evaluateStructure() of the super class which is basically a call to self.evaluate(expr) (to be found in zope/tales/tales.py around line 713: def evaluateStructure(self, expr): return self.evaluate(expr) and then again in _handleText(): def _handleText(self, text, expr): text = self.evaluate(expr) ... I now tried to remove the evaluation of expr in handleText() and in a very simple test using ab on my site I got twice the speed! I might be missing something here of course but if every expression is evaluated twice then this is of course huge performance drain. And of course this fix was just a quick idea without much understanding of the underlying engine (e.g. not knowing if _handleText() is also called from somewhere else etc.) So any Page Template experts here who can confirm this? cheers, Christian -- Christian Scholz video blog: http://comlounge.tv COM.lounge blog: http://mrtopf.de/blog Luetticher Strasse 10Skype: HerrTopf 52064 Aachen Homepage: http://comlounge.net Tel: +49 241 400 730 0 E-Mail [EMAIL PROTECTED] Fax: +49 241 979 00 850 IRC: MrTopf, Tao_T connect with me: http://mrtopf.de/connect ___ 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] serialization with Ape
Hi again!
> Christian Scholz wrote:
> > root_mapper, conns = createMapper(fspath)
> > ei=ExportImport(root_mapper,conns)
> > ei.exportObject(object)
>
> Ah-ha, I just realized what went wrong. You need to tell exportObject()
> where to export. PathKeychainGenerator refused to guess. Try changing
> the last line to:
>
> ei.exportObject(object, keychain=('some/path/under/fspath',))
I now tested it and had the following results:
- the last item of the path needs to be the same id as the object has
- only objects derived from Objectmanager in the end can be serialized.
- they are not written to disk as long as I don't do
conns['fs'].vote()
conns['fs'].finish()
After that it worked quite well :)
I now did the import the other way round with
obj=ei.importObject(keychain)
and then using _setObject with that object in order to store it again
in the usual ZODB (Data.fs). Is this ok? Or do I again need to do some copy?
> It's still obtuse. I'll think about how to make that cleaner.
Well, for me the whole thing actually is magic. But maybe it helps
if there is an example of usage in the docstring.
But it seems working now, so fine :-)
Thanks again!
Christian
--
COM.lounge http://comlounge.net/
communication & design [EMAIL PROTECTED]
___
Zope-Dev maillist - [EMAIL PROTECTED]
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] serialization with Ape
Hi! I actually now have a (as it seems) working version of my first approach, just using _setObject() etc. which even works recursively. Has this some drawbacks except it looks a bit like a hack due to the call of commit()? I also have the deserialization working and basically it's just a copying between two ZODBs I'd say. On Wed, Jun 11, 2003 at 10:12:46AM -0400, Shane Hathaway wrote: > Christian Scholz wrote: > > I then detected the fascades in io.py which I tried to use (dunno if they're > > thought to be used for such a purpose actually ;-): > > > > root_mapper, conns = createMapper(fspath) > > ei=ExportImport(root_mapper,conns) > > ei.exportObject(object) > > This is *exactly* the purpose ExportImport is meant for, but it's quite > new and unfinished. I'm glad you found it, although it's only in CVS, > not in the released version of Ape. Well, I am looking a bit on the repository and was reading log messages ;-) > > Unfortunately this results in an exception > > > > Type: MappingError > > Error Value: Path keychains require a name > > > > So can maybe somebody point me to the easiest way to serialize an object > > using the fsmapper (Shane? ;-) ? > > You're on the right track. I wonder what went wrong. I looked into that a bit and in core/io.py line 316 you do return kgen.makeKeychain(event, None, 1) so name is None here as it seems. I just dunno if it's a special case because I was doing something wrong or not ;-) > > I am also interested in reading a serialized object from the fs again > > and writing it to the fs. Also I wonder if I need to delete objects which > > have been modified before another export or if there's a possibility to > > just change it). > > Ape will assign new OIDs on import, so you don't have to delete objects > first. Well, at least with my first approach I need to do that because I just do a folder._setObject() and this fails then of course. > > (and the best would be of course some simple methods for reading/writing > > a whole directory recursively. I don't know if something like this is already > > in there somehow or how it works internally.. seems all a bit magic to me ;-) > > Again, ExportImport is designed to do this. Will it also do it recursively? I found out that when I do it via folder._setObject() it will also serialize all content which is quite nice :) What is actually the result of an export? Where in the fs is it stored then? Is the whole path to the object in ZODB is taken ? But thanks for the answer and thanks for Ape! :) -- christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] serialization with Ape
Hi list! :) I am just playing around with Ape (very cool, btw :) and wondering how it can be used from the outside, e.g. not controlled by transactions. Main idea is to manually copy parts of the ZODB to the filesystem (from a product). I first tried the transaction based process, as I found it in some tests: # init (createMapper imported from fsmapper) root_mapper, conns = createMapper(fspath) conn=conns['fs'] resource = StaticResource(root_mapper) storage = ApeStorage(resource, conns) db = ApeDB(storage, resource) get_transaction().begin() # test object to store object = getattr(self,'standard_template.pt') c=db.open() app=c.root()['Application'] id=object.getId() app._setObject(id,object.aq_base,set_owner=0) get_transaction().commit() db.close() which actually stores it but does not seem to me that clean because of all the transaction stuff. I would here then recursively run through all the objects and do that inside one transaction. I then detected the fascades in io.py which I tried to use (dunno if they're thought to be used for such a purpose actually ;-): root_mapper, conns = createMapper(fspath) ei=ExportImport(root_mapper,conns) ei.exportObject(object) Unfortunately this results in an exception Type: MappingError Error Value: Path keychains require a name So can maybe somebody point me to the easiest way to serialize an object using the fsmapper (Shane? ;-) ? I am also interested in reading a serialized object from the fs again and writing it to the fs. Also I wonder if I need to delete objects which have been modified before another export or if there's a possibility to just change it). (and the best would be of course some simple methods for reading/writing a whole directory recursively. I don't know if something like this is already in there somehow or how it works internally.. seems all a bit magic to me ;-) So thanks in advance if anybody can help me :) cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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] SiteAccess improvements
Hi! > I haven't been very deep down into site access things, mainly because I get > confused each time I try to set up a SiteRoot. :-) > So it may be that some of the features I'd like are already availiable, but > that I just didn't know. :-) > > Anyway, this is what I would like: > > The virtual hosts of Zope is managed from the Control Panel and not from > several objects in several directories. This would mean that if you enter > incorrect information, you don't have to reboot the server with > SUPPRESS_SITEROOT=1 set in the environment to fix the stupid errors you > made, you just go into the Control Panel and fix it.It also means that a > customer of a web hotel can't fiddle around with his own SiteRoot. God eh? > :-) This Control Panel thingy should also have a possibility to turn on and > off individual virtual host mappings without removing the settings, as well > as a global on/off. Basically this is possible using our VirtualSiteRoot Product, which can be found at http://www.zope.org/Members/comlounge/ It does not actually install the config in the control panel but in the parent directory. It also lets you setup different roots for management and public access. So it's e.g. possible to use the management screens via https:// and the public access is still working via http://. Also included is some plugin mechanism e.g. for automatically inserting session ids in the URL. cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] Zope crash in select_trigger
Hi! I once had similar problems and as far as I remember also with SuSE 6.4. I think it stopped when using SuSE 7.1 (and 7.0 had the problems, too I think). I was using python 1.5.2 and various Zope versions.. so maybe the pymalloc hint would have worked for me then, too. Actually the crashes were some sort of non deterministic and did not depend on load etc. Unfortunately I don't know any details anymore.. I might test this again sometime and give you a note. (maybe there also some old mails from me in the archive explaining some more details but as I am on a text terminal right now I haven't looked..) greetings! Christian On Tue, Nov 06, 2001 at 12:00:40PM -0800, Mark Mitchell wrote: > > > --On Tuesday, November 06, 2001 02:58:11 PM -0500 Chris McDonough > <[EMAIL PROTECTED]> wrote: > > > Hi mark, > > > > Which OS, and which Zope version? I dont have an answer at all right > > now, but I'd like to start collecting info. If anybody else has this > > problem, can they also provide platform and Zope version? > > >From the control panel: > > Zope version: (Zope 2.4.2 (binary release, python 2.1, linux2-x86), python > 2.1.0, linux2) > Python version: 2.1 (#2, Jun 22 2001, 10:20:42) [GCC 2.7.2.3] > System Platform: linux2 > > The system is SuSE Linux 6.4 with their most recent kernel patch, i.e, > the 2.2.19 kernel. > > -- > Mark Mitchell[EMAIL PROTECTED] > CodeSourcery, LLChttp://www.codesourcery.com > > ___ > Zope-Dev maillist - [EMAIL PROTECTED] > 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 ) -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] cycles and mem leaks and Zpatterns..
Hi! As some of you might know, I still have my mem leak problem.. I've now looked again at some code and discovered some cycle in ZPatterns.. Means: I have a Price class and some PriceWinner class, which reference each other via some SkinScript. Could this lead to a leak then? I am not sure at the moment if the backlink to the price is used somewhere.. At least I will try to eliminate it now and maybe replace the price attribute in the winner class by some method which returns the price.. (hope, this is no problem then..) regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] memory eating Zope..
Hi! > It looks like you're leaking requests. This is often the consequence of > setting an attribute on the REQUEST object which is aquisition wrapped. > For instance: > > REQUEST.adtmldoc = REQUEST['PARENTS'][0] > > When requests leak, references to all the objects referenced within the > request (acquisition wrappers, etc.) leak as well. This can be > disastrous and gets worse when the site is busy (when there are more > requests). ok, that's what I experienced.. the day the site got some more traffic it also suddenly grow memorywise. I also didn't know that doing above is a bad thing to do.. (though I don't actually understand why this is a problem and why it's not freed then.. but on the other hand I am not that much into the inner workings of acquisition ;-) > It used to be the case that tracking this problem down was tediously > difficult, and involved basically a binary search for leaks in all > methods on the site. But now, finding the cause is likely a job for > Shane's LeakFinder product: > http://www.zope.org/Members/hathawsh/LeakFinder . Ah, will try that out :) And actually I directly found some method which fetches the actual user object from a ZPatterns rack and sets this in the REQUEST at every request. So maybe that's already the problem then.. will try to remove that.. > Note that this is likely caused by some sort of brainbending circular > reference problem that might be helped by Python 2.X's cyclic garbage > collector. Well, will also try out python 2.x then... (btw. are all 2.x python versions working without problems with Zope? Don't want to get a new problem when resolving the old one ;-) Thanks! Christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] memory eating Zope..
Hi there! I got another problem with my Zope server and that seems to be a memory hole.. I am using Zope 2.3.2 with Python 1.5.2 now on SuSE 7.0 (this seems to solve the hanging problem, which appeared on 7.1) and I use Python Script, ZSQL methods, ZMySQLDA and ZPatterns. Now using top I see over time how the memory consumption of the zope processes grows, which means I have to restart once a day as it then becomes too slow.. When looking at the refcount in the debug section I see the following: Acquisition.ImplicitAcquirerWrapper: 7509 DocumentTemplate.DT_Util.Eval: 719 Products.ZPatterns.Expressions.Expression: 716 ZPublisher.HTTPRequest.HTTPRequest: 457 ZServer.HTTPResponse.ZServerHTTPResponse: 450 ZPublisher.BaseRequest.RequestContainer: 447 So Acquisition.ImplicitAcquirerWrapper seems to be referenced quite often (this is in an early state..) So does anybody know what this might cause? Or maybe how to track that problem down? (seems to be a long way with that server to become it stable.. ;-) Thanks in advance! Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] dtml-in bug?
Hi! ok, some more people told me now about it.. ;-) It's not that intuitive nevertheless.. (IMHO) ;-) cheers and thx, Christian On Mon, May 28, 2001 at 05:20:39PM +0200, Tino Wildenhain wrote: > Hi Christian, > > --On Montag, 28. Mai 2001 15:38 +0200 Christian Scholz <[EMAIL PROTECTED]> > wrote: > > > Hi! > > > > Dunno if it's again my "strange" setup, but I have some problem with > > dtml-in or maybe I simply didn't get something right.. > > > > Well, I used the following dtml code: > > > > > > > > a > > > > > > > *rustle* try orphan=0 :-) > > Greetings > Tino > > > > and I thought I will get three a's as output.. instead I get 6.. > > If I do size=3 it's ok (3 output), if I do size=4 or 5 or 6 I get > > them all.. > > (actually it happened with some more complex list of Result objects but > > this seems to do the same here) > > > > Can someone explain that to me..? > > > > (it's Zope 2.3.2, Python 1.5.2, SuSE Linux 7.0) > > > > cheers, > > Christian > > > > -- > > COM.lounge http://comlounge.net/ > > communication & design [EMAIL PROTECTED] > > > > ___ > > Zope-Dev maillist - [EMAIL PROTECTED] > > 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 ) > > > > -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] dtml-in bug?
Hi! Dunno if it's again my "strange" setup, but I have some problem with dtml-in or maybe I simply didn't get something right.. Well, I used the following dtml code: a and I thought I will get three a's as output.. instead I get 6.. If I do size=3 it's ok (3 output), if I do size=4 or 5 or 6 I get them all.. (actually it happened with some more complex list of Result objects but this seems to do the same here) Can someone explain that to me..? (it's Zope 2.3.2, Python 1.5.2, SuSE Linux 7.0) cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] getPersistentItemIDs not cooperating?
Hi! > It works.. partially. For some reason it looks like getPersistentItemIDs > does not always return a *complete* list. I need to run this method > several times to completely exhaust the Rack's storage. Thoughts? The > only way I ever create Track objects is via a different method that is > only accessed once (a long time ago!) Well, I experiences something similar when changing a Racks storage from persistent to non-persistent. The included objects are marked as orphaned then and should be deleted with the Clear-Button. This also works, but only the half of them gets deleted every time I click on that button.. So I also have to press it several times.. Sounds like the same problem.. cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
just for general information: Apache was not the problem.. PCGI also shows the problem.. a Zope binary release did not help it too bad.. ;-) -- christian On Fri, May 18, 2001 at 05:36:47PM +0200, Christian Scholz wrote: > Hi! > > Another test without apache inbetween is still running now.. > Could apache and it's proxypass be the problem? > > I am simply using ProxyPass and ProxyPassReverse in front > of Zope in a simple virtual site environment.. > > Apache version is 1.3.19 > > Where is the bug then? In Zope or in Apache? > I also have other servers running like that for ages > without problems though it's another host and thus another > apache version. > > cheers, > Christian > > > -- > COM.lounge http://comlounge.net/ > communication & design [EMAIL PROTECTED] > > ___ > Zope-Dev maillist - [EMAIL PROTECTED] > 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 ) -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! Another test without apache inbetween is still running now.. Could apache and it's proxypass be the problem? I am simply using ProxyPass and ProxyPassReverse in front of Zope in a simple virtual site environment.. Apache version is 1.3.19 Where is the bug then? In Zope or in Apache? I also have other servers running like that for ages without problems though it's another host and thus another apache version. cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! Hm, another strange thing.. I've now tried to start Zope via the normal ./start script. (I've done it via some own script with INSTANCE_HOME set). What I've got was: bash-2.04# ./start -- 2001-05-18T14:59:49 PANIC(300) z2 Startup exception Traceback (innermost last): File /opt/zopeneu/z2.py, line 566, in ? File , line 1, in ? File /opt/zopeneu/lib/python/Zope/__init__.py, line 94, in ? File /opt/zopeneu/lib/python/ZODB/__init__.py, line 85, in ? ImportError: undefined symbol: PyObject_Init So what does this want to tell me? ;-) I've also done the following before that: - delete Data.fs in my instance home - create a new user - create a folder - put an image (test.gif) into that folder - started the test on that image => hang I've also removed any products I don't need which still have been installed.. -- christian On Fri, May 18, 2001 at 10:49:23AM -0400, Chris McDonough wrote: > Is there any evidence of any sort of relationship between the broken pipe > error and the hang? Do they happen at the same time? Or do they have no > discernable relationship with one another? How about memory utilization? > Is the process leaking memory? > > ----- Original Message - > From: "Christian Scholz" <[EMAIL PROTECTED]> > To: "Chris McDonough" <[EMAIL PROTECTED]> > Cc: "Christian Scholz" <[EMAIL PROTECTED]>; "Tino Wildenhain" > <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> > Sent: Friday, May 18, 2001 10:13 AM > Subject: Re: [Zope-dev] another Zope hanging.. > > > > Hi! > > > > Just FYI: The server still hangs once a day and another one is doing > > the same.. > > It even does this when simply accessing one page or image over and over > > again. > > More precisely I've started the server and did > > > > ab -n 10 -c 10 http://foo.bar/we/pics/someimg.gif > > > > (ab being the apache benchmarking utility) > > > > and after a while I get the broken pipe error. > > > > I've also changed my python version to 2.0 and tested it with this > > but still the same effect. Also changing the server (and thus > > hardware and system) did not help it. > > I've also removed the mysqlda from the zope installation but also this > > did not help.. > > I will try now to create the smallest version of the site that still > hangs.. > > > > Dunno if anybody has some idea, I actually have none.. > > > > I'll keep you informed.. > > > > -- christian > > > > PS: strangely it works on the development server but also here I am doing > > some tests with ab right now.. and this one is still 1.5.2 > > > > > > On Thu, May 03, 2001 at 02:16:30PM -0400, Chris McDonough wrote: > > > > Well, my problem might be that in my case most sql statement are done > > > inside > > > > some attribute provider of ZPatterns.. This is using ZSQL methods > > > internally > > > > but actually those then won't show up I guess.. > > > > > > Even if you weren't using ZPatterns, the error probably wouldn't jump > out > > > and say "here I am!" So I don't think there's much difference between > using > > > ZPatterns and not using ZPatterns. The process of detecting when > something > > > hangs is just like any other troubleshooting process, it's a matter of > > > exclusion. If you notice that the request named "foobargorf/fleafang" > > > *always* hangs, you investigate what it does, and try to reproduce it. > If > > > it's incidental, so be it, and move on to the next theory. > > > > > > > > > > > I've also seen that some more recent version of the mysql stuff is > around > > > > and I am using this now.. > > > > > > > > The problem's also that I just have a few methods to invoke from the > > > outside > > > > which do lots of things by calling other object. Thus I might not > really > > > see > > > > what's really causing the problem.. I hope though that my upgrade will > > > show some > > > > benefit.. > > > > > > The -M log perhaps won't show you the actual operation that's causing > the > > > hang, but it will show you the entry point into a routine which causes > the > > > hang. It's your job from there to track down the cause. This is just > like > > > debugging a program. You get an error somewhere, and you need to track > it > > > back to its root, which may be six levels up the call stack buried in > some > > > godforsaken regex. ;-) > > > > > > > > > > > > > -- > > COM.lounge http://comlounge.net/ > > communication & design [EMAIL PROTECTED] > > > -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! > Is there any evidence of any sort of relationship between the broken pipe > error and the hang? Do they happen at the same time? Or do they have no > discernable relationship with one another? How about memory utilization? > Is the process leaking memory? yes, they happen at the same time.. I've also stripped it down and removed evertyhing and put some test.gif up there.. Then running ab against it created again a broken pipe (quite soon). looking via top the memory utilization seems ok.. -- christian PS: with this test.gif it seems to hang very quickly.. PPS: ab tests on my development host are still running without problems.. strange.. > > - Original Message ----- > From: "Christian Scholz" <[EMAIL PROTECTED]> > To: "Chris McDonough" <[EMAIL PROTECTED]> > Cc: "Christian Scholz" <[EMAIL PROTECTED]>; "Tino Wildenhain" > <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> > Sent: Friday, May 18, 2001 10:13 AM > Subject: Re: [Zope-dev] another Zope hanging.. > > > > Hi! > > > > Just FYI: The server still hangs once a day and another one is doing > > the same.. > > It even does this when simply accessing one page or image over and over > > again. > > More precisely I've started the server and did > > > > ab -n 10 -c 10 http://foo.bar/we/pics/someimg.gif > > > > (ab being the apache benchmarking utility) > > > > and after a while I get the broken pipe error. > > > > I've also changed my python version to 2.0 and tested it with this > > but still the same effect. Also changing the server (and thus > > hardware and system) did not help it. > > I've also removed the mysqlda from the zope installation but also this > > did not help.. > > I will try now to create the smallest version of the site that still > hangs.. > > > > Dunno if anybody has some idea, I actually have none.. > > > > I'll keep you informed.. > > > > -- christian > > > > PS: strangely it works on the development server but also here I am doing > > some tests with ab right now.. and this one is still 1.5.2 > > > > > > On Thu, May 03, 2001 at 02:16:30PM -0400, Chris McDonough wrote: > > > > Well, my problem might be that in my case most sql statement are done > > > inside > > > > some attribute provider of ZPatterns.. This is using ZSQL methods > > > internally > > > > but actually those then won't show up I guess.. > > > > > > Even if you weren't using ZPatterns, the error probably wouldn't jump > out > > > and say "here I am!" So I don't think there's much difference between > using > > > ZPatterns and not using ZPatterns. The process of detecting when > something > > > hangs is just like any other troubleshooting process, it's a matter of > > > exclusion. If you notice that the request named "foobargorf/fleafang" > > > *always* hangs, you investigate what it does, and try to reproduce it. > If > > > it's incidental, so be it, and move on to the next theory. > > > > > > > > > > > I've also seen that some more recent version of the mysql stuff is > around > > > > and I am using this now.. > > > > > > > > The problem's also that I just have a few methods to invoke from the > > > outside > > > > which do lots of things by calling other object. Thus I might not > really > > > see > > > > what's really causing the problem.. I hope though that my upgrade will > > > show some > > > > benefit.. > > > > > > The -M log perhaps won't show you the actual operation that's causing > the > > > hang, but it will show you the entry point into a routine which causes > the > > > hang. It's your job from there to track down the cause. This is just > like > > > debugging a program. You get an error somewhere, and you need to track > it > > > back to its root, which may be six levels up the call stack buried in > some > > > godforsaken regex. ;-) > > > > > > > > > > > > > -- > > COM.lounge http://comlounge.net/ > > communication & design [EMAIL PROTECTED] > > > -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! Just FYI: The server still hangs once a day and another one is doing the same.. It even does this when simply accessing one page or image over and over again. More precisely I've started the server and did ab -n 10 -c 10 http://foo.bar/we/pics/someimg.gif (ab being the apache benchmarking utility) and after a while I get the broken pipe error. I've also changed my python version to 2.0 and tested it with this but still the same effect. Also changing the server (and thus hardware and system) did not help it. I've also removed the mysqlda from the zope installation but also this did not help.. I will try now to create the smallest version of the site that still hangs.. Dunno if anybody has some idea, I actually have none.. I'll keep you informed.. -- christian PS: strangely it works on the development server but also here I am doing some tests with ab right now.. and this one is still 1.5.2 On Thu, May 03, 2001 at 02:16:30PM -0400, Chris McDonough wrote: > > Well, my problem might be that in my case most sql statement are done > inside > > some attribute provider of ZPatterns.. This is using ZSQL methods > internally > > but actually those then won't show up I guess.. > > Even if you weren't using ZPatterns, the error probably wouldn't jump out > and say "here I am!" So I don't think there's much difference between using > ZPatterns and not using ZPatterns. The process of detecting when something > hangs is just like any other troubleshooting process, it's a matter of > exclusion. If you notice that the request named "foobargorf/fleafang" > *always* hangs, you investigate what it does, and try to reproduce it. If > it's incidental, so be it, and move on to the next theory. > > > > > I've also seen that some more recent version of the mysql stuff is around > > and I am using this now.. > > > > The problem's also that I just have a few methods to invoke from the > outside > > which do lots of things by calling other object. Thus I might not really > see > > what's really causing the problem.. I hope though that my upgrade will > show some > > benefit.. > > The -M log perhaps won't show you the actual operation that's causing the > hang, but it will show you the entry point into a routine which causes the > hang. It's your job from there to track down the cause. This is just like > debugging a program. You get an error somewhere, and you need to track it > back to its root, which may be six levels up the call stack buried in some > godforsaken regex. ;-) > > > -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! > > most problems with hanging zope come from external RDBMS-access. > > Lots do, but lots don't... > > > You wont see anything useful in the logs because there are > > some objects published before the *sql-method is hanging. > > If you see a request come into ZServer, and you see the request go into > ZPublisher from ZServer, but ZPublisher never returns any data to ZServer, > that's an indication that something is wrong. A SQL method hanging would > look like this, it's the same thing as any other Zope component hanging. > > The "big-M" log is your friend, even in this case. As a matter of fact, > it's *especially* your friend in this case because it will be patterned > against one or a few methods, which is an easy problem to track down. Well, my problem might be that in my case most sql statement are done inside some attribute provider of ZPatterns.. This is using ZSQL methods internally but actually those then won't show up I guess.. I've also seen that some more recent version of the mysql stuff is around and I am using this now.. The problem's also that I just have a few methods to invoke from the outside which do lots of things by calling other object. Thus I might not really see what's really causing the problem.. I hope though that my upgrade will show some benefit.. (strangely on my development server these problems did not arise.. might have to check for any differences between the two..) I will keep cheking the -M anyway (as it seems all I can do.. ;-) cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! > Alright... so you're getting an OSError 11, and you're using Zope 2.3.1. > > errno 11 is "temporarily unavailable" which means about as much as "foobar > feeblefar" as far as I'm concerned ;-). great :) > Though you've said the -M log isn't helpful here, I think it is. If this is > happening when a certain method is hit, or at a certain time of day, or when > a certain number of connections are reached, etc., the -M log in combination > with requestprofiler can help determine this. Otherwise it's a guessing > game. ok, just thought that maybe some exception can be caught.. I will try to use requestprofiler now, though I checked the -M log before and around that time there was no obvious problem.. (also it's not loaded at the moment and it's happening here and then, don't see a pattern there..) I will check (and maybe change the mysql adapter, as I recently had problems with it on a sun machine, too). thanks anyway :) Christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Maybe another general question: How are exceptions handled which appear in your own products and you don't catch them? Is there some except somewhere "on top" ? How does this look if you're using ZPatterns with your own attribute providers (as I do with my sql attribute provider). Are these all handled somehow? (usually I get some sort of traceback on the console/stupid log is something goes wrong there and I don't handle it myself.. hope it's in general..) cheers, Christian On Thu, May 03, 2001 at 11:23:54AM -0400, Chris McDonough wrote: > What Python version are you using? > > - Original Message - > From: "Christian Scholz" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Thursday, May 03, 2001 11:05 AM > Subject: [Zope-dev] another Zope hanging.. > > > > Hi! > > > > I also now have some problems with a hanging zope. > > > > I got something like this in my stupid logfile: > > > > 2001-05-03T13:44:39 ERROR(200) ZServer uncaptured python exception, > closing channel > (exceptions.OSError:[Errno 11] Die Ressource ist zur Zeit nicht verf?gbar > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|poll|83] > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|handle_read_event|335] > [/opt/Zope-2.3.1-src/ZServer/medusa/select_trigger.py|handle_read|77] > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|recv|287] > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|recv|475]) > > > > Now my question is simply if I can catch this exception somewhere and get > some traceback etc. > > Actually I am quite confused from the inner workings of medusa so I dunno > where to start > > searching. > > > > Can anyone help? > > > > (I know about the -M logging but unfortunately this does not help that > much..) > > > > cheers, > > Christian > > > > -- > > COM.lounge http://comlounge.net/ > > communication & design [EMAIL PROTECTED] > > > > ___ > > Zope-Dev maillist - [EMAIL PROTECTED] > > 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 ) > > > > > ___ > Zope-Dev maillist - [EMAIL PROTECTED] > 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 ) -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] another Zope hanging..
Hi! > What Python version are you using? That's 1.5.2 MySQLDA is MySQL-python-0.3.3.tar.gz and ZMySQLDA-2.0.4.tar.gz (if this makes any problems..) Zope is 2.3.1 as suerly have noticed.. :) cheers, christian > - Original Message - > From: "Christian Scholz" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Thursday, May 03, 2001 11:05 AM > Subject: [Zope-dev] another Zope hanging.. > > > > Hi! > > > > I also now have some problems with a hanging zope. > > > > I got something like this in my stupid logfile: > > > > 2001-05-03T13:44:39 ERROR(200) ZServer uncaptured python exception, > closing channel > (exceptions.OSError:[Errno 11] Die Ressource ist zur Zeit nicht verf?gbar > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|poll|83] > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|handle_read_event|335] > [/opt/Zope-2.3.1-src/ZServer/medusa/select_trigger.py|handle_read|77] > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|recv|287] > [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|recv|475]) > > > > Now my question is simply if I can catch this exception somewhere and get > some traceback etc. > > Actually I am quite confused from the inner workings of medusa so I dunno > where to start > > searching. > > > > Can anyone help? > > > > (I know about the -M logging but unfortunately this does not help that > much..) > > > > cheers, > > Christian > > > > -- > > COM.lounge http://comlounge.net/ > > communication & design [EMAIL PROTECTED] > > > > ___ > > Zope-Dev maillist - [EMAIL PROTECTED] > > 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 ) > > > -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] another Zope hanging..
Hi! I also now have some problems with a hanging zope. I got something like this in my stupid logfile: 2001-05-03T13:44:39 ERROR(200) ZServer uncaptured python exception, closing channel (exceptions.OSError:[Errno 11] Die Ressource ist zur Zeit nicht verf?gbar [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|poll|83] [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|handle_read_event|335] [/opt/Zope-2.3.1-src/ZServer/medusa/select_trigger.py|handle_read|77] [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|recv|287] [/opt/Zope-2.3.1-src/ZServer/medusa/asyncore.py|recv|475]) Now my question is simply if I can catch this exception somewhere and get some traceback etc. Actually I am quite confused from the inner workings of medusa so I dunno where to start searching. Can anyone help? (I know about the -M logging but unfortunately this does not help that much..) cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] ZPatterns Rack question..
Hi! First of all one note to the sql attribute provider: My problem seems to be solved now. Actually it was quite easy.. dunno why it took the long way.. thanks anyway to Phillip and Steve! But I have another problem now: I have some rack with persistent objects in it. Now I want to migrate my site to production. Problem: If I do this more often and the objects are within some Rack, then also these get copied to the production site and the old objects will get deleted there. Is there any way to first backup the contents of some rack? Or do I need to construct it all as external storage? (basically most things are in an sql db already, I just put image data etc. into the rack directly as I don't need relational access to it..) Actually I know how to read the objects from the BTree but I dunno how to store them again into another rack (which has the same storage attribs of course). Can I simply use _setRack()? I guess not, eh? :) Thanks for any hints in advance! regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns AttributeProvider question
Hi!
So here's more :)
> >def _objectAdding(self,client,
> >_tmap={ None:AddedStatus,
> >DeletedStatus:ChangedStatus,
> >ChangedStatus:AddedStatus
> >}
> >):
> >t = client._getTokenFor(self)
> >s = t.status
> >t.status = _tmap.get(s,s)
> >
> ># we need to do a commit here in order to store the record into
> the database!
> > # and we need to have it stored as an AttributeFor() of the primary key will
> > # return None otherwise (as this also just makes a query).
> >client.commitSubtransaction()
>
> This is broken. Don't do this! commitSubtransaction() is an
> application-level operation. Only. If you have to use it inside of
> provider-level code or code called from SkinScript, chances are that you
> are doing something horribly wrong. You run the risk of causing a
> recursive commit operation in the Zope transaction machinery. Don't do this!
well, it always looked no-good to me, I just didn't know how to solve my problem.
So my problem I was trying to fix was actually the following:
- When creating a new object via newItem(), the Rack checks if self.loadAttrib is
defined
before creating a new one.
if loadAttrib is returned by the attribute provider then the Rack assumes it already
exists and raises an exception.
- My Attribute Provider now does not know if an object exists or not before looking
into
the database (I tried different ways to check it but somehow they haven't worked).
So it does that. If it finds a record with the given id it returns it, if not it
returns
the default value (which means "does not exist" to the rack).
- When adding a new object in my provider it does an insert via an ZSQL method
to store it inside the database. This is done in _objectAdded().
Unfortunately this is too late as before this is called, we already need to have
it in the database as the request to my primary key will return the default still.
So I moved this insert to _objectAdding().
Unfortunately this new record seems not to appear in the database before commit()
(dunno why, maybe some new transaction code in ZMySQLDA? haven't checked this).
That's why I did a commit() there which then called _objectAdded().
So my problem is basically that I need to know in _AttributeFor() whether this
object is already created or not.. I will check now again with asking the status
of the client but I think I also tried this before.
The best would be to collect all data before _objectAdded() and store it at once
and not as it's done now by first creating an empty record and doing the rest
via _objectChanged().
But I am open for any other idea :)
> >def _SetAttributeFor(self,client,name,value):
> >"""Set the attribute and return true if successful"""
> > # attribs stores the attributes we care about
> >if name in self._attribs.keys():
> >client._getCache()[name]=value
> >client.__dict__[name]=value;
> > client._p_changed = 1
> >return 1
> >return None
>
> This also looks broken to me. Why are you changing both the cache and the
> persistent state? There's no reason to do both.
ok, this is old code where I wanted to test some other solution to fix my problem.
I will remove the additional line.. (__dict__ might be all I need, it's already a cache
if I remember right?!?)
> I think you're seriously over-engineering this. I don't see anything that
> you're doing here that couldn't be done more easily with SkinScript
> statements. If all you want to do is store the data in an SQL database, a
> few SkinScript triggers would suffice.
Well, as mailed yesterday I know this. I've also done this a lot but was then
tired of always typing those paramater lists (for each ZSQL method, for each SkinScript
call twice, ...).
This is of course always the tradeoff between easy to use plugins which are somehow
restricted and the more flexible ones which do need lots of configuration.
I would like to have the first one for the day-to-day-stuff and the latter
for more advanced operations, like linking specialists together etc.
> Now, if you're trying to get a row to be inserted in an SQL database at the
> actual time of the newItem() call, there is a trick you can use to do that
> from SkinScript also. Do this:
>
> INITIALIZE OBJECT WITH someAttr1=default1,
> someAttr2=default2,...,dummyAttributeName=myInsertQuery(primaryKey=self.id)
>
> Where myInsertQuery() is an SQL insert query that takes primaryKey as a
> parameter, and dummyAttributeName is some attribute name you don't care
> about. You should also set any default values for fields that will be
> inserted. Your insert query will be called, initializing the row in the
> database, and then use something like this:
>
> WHEN OBJECT ADDED,CHANGED STORE foo,bar,baz USING
> UpdateMethod(widget_id=self.id, foo_field=self.foo, wbar=self.bar)
>
>
> To update t
Re: [Zope-dev] ZPatterns AttributeProvider question
Hi! Just a quick note as it's quite late already (more tomorrow): > In the early days of ZPatterns, I assumed that I would create SQL > providers, LDAP providers, and suchlike gizmos. Later, it became clear > that it was more useful to have a simple "glue" language to allow > harnessing the full power of Zope in the context of events happening to > objects. I still toy with the idea of making an "SQL attribute provider", > however, that would be based on the ZSQLMethod object and add some > ZPatterns hooks to it, but it's not a big priority. Its main value would > be to cut down on some repetitive typing between one's SQL statements and > SkinScript statements. Well, exactly this was the intention of programming this as I was a bit bored by all this typing, exactly when the attribute lists are getting longer. So right now I define it once in the ZClass, press the "Read ZCLass propsheet" button and the my attribute provider sets everything up itself.. (it also saves me from typing errors..) Actually I did it with SkinScript before but thought some simple plugin would be nicer.. But thanks for explaining the inner workings in some more detail (maybe I should put this into some howto or so? or it might be put into the ZPatterns wiki..), as this really should help (not only me I guess :) I will comment more on this tomorrow.. (at least I got some idea how I might solve my problem... though I thought this some times before already.. ;-) And the other possible option would be (as Steve suggested once on IRC) to create some wizard which sets up all the SkinScript methods and ZSQL methods in one go.. Anyway, more about it tomorrow. good nite.. Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns AttributeProvider question
Hi!
A little update..
Actually I was mistaken and SetAttributeFor() is called. Just _objectChanged()
is not called in that case.. and thus my attributes are not stored to the database.
Also commit() is not called at all in this case. It get's called once when the
subtranscation is committed but not after some data has changed (which happens
after that).
Printing _v_status_ in SetAttributeFor() also gives me "Changed", thus this seems
right.
Maybe some more code will help (just the important pieces):
def _objectAdded(self,client):
""" store cached data in database """
... here we do an sql query to insert things in database ...
def _objectChanged(self,client):
""" store cached data in database """
... here we do an sql query to update things in database ...
def _objectAdding(self,client,
_tmap={ None:AddedStatus,
DeletedStatus:ChangedStatus,
ChangedStatus:AddedStatus
}
):
t = client._getTokenFor(self)
s = t.status
t.status = _tmap.get(s,s)
# we need to do a commit here in order to store the record into the database!
# and we need to have it stored as an AttributeFor() of the primary key will
# return None otherwise (as this also just makes a query).
client.commitSubtransaction()
def _SetAttributeFor(self,client,name,value):
"""Set the attribute and return true if successful"""
# attribs stores the attributes we care about
if name in self._attribs.keys():
client._getCache()[name]=value
client.__dict__[name]=value;
client._p_changed = 1
return 1
return None
So _objectChanging() is not overridden, the rest I have ommitted..
So as you also can see I already do a subtransaction commit but it does
not seem to help for this problem..
regards,
Christian
PS: It might also be that I am simply blind ;-) But somehow ZPatterns always confuses
me,
especially the inner workings ;-)
--
COM.lounge http://comlounge.net/
communication & design [EMAIL PROTECTED]
___
Zope-Dev maillist - [EMAIL PROTECTED]
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 )
Re: [Zope-dev] ZPatterns AttributeProvider question
Hi! > >Yes, as said above it registers for handlers and attributes. Also > >_objectChanged and _SetAttributeFor() etc. are called. They're just not > called > >directly after creating the object (when being in the same request that is). > > _objectChanged() is a transaction-commit message. You'll only get it when > the (sub)transaction commits. You'll receive _objectChanging() just > *before* the object changes. Yep, that's how I understood it. That's also working. > As for _SetAttributeFor(), it should be called at the time > manage_changeProperties executes, not at object creation. Are you sure > that you have it properly named and that it isn't still being defined by > one of your ancestor classes instead? The "AttributeProvider" and > "GTMixin" base classes in ZPatterns implement _SetAttributeFor() by simply > noting the change for later. If you don't override this behavior, your > class won't see _SetAttributFor() itself. Well, it actually is working, just not directly after creating the new object. So what I do now as workaround is dtml method 1: - creates new object via newItem() - redirects to dtml method 2 dtml method 2: - retrieves the newly created object with getItem() - calls manage_changeProperties on that object This is working. Putting the manage_changeProperties directly after the newItem() is not working. I've also tried retrieving the new object directly after newItem() and calling manage_changeProperties on that but this also did not work. (so it is named correctly as otherwise it wouldn't be called either in dtml method 2). So the main problem is: SetAttributeFor() is not called (and also _objectChanging() is not called) in the same request cycle after creating a new object. They're called however in the "next" request after retrieving it with getItem(). Maybe it helps if I publish the source somewhere? It's gonna be some public product anyway. regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns AttributeProvider question
Hi! > > Good evening everybody! > > > > I have some question regarding attribute and agent programming for ZPatterns. > > > > I have some provider which is registered for the "handlers" and "attributes" > > methods and some attributes. > > > > My problem is when trying to create a new object and directly editing it, e.g. > > > > obj=newItem() > > obj.propertysheets.data.manage_changeProperties(foobar=13) > > > > the propertysheet is defined and my provider is also registered for > > handling the property "foobar". Unfortunately my SetAttributeFor() method > > is never called. > > > Have you defined a namesForRegistration method in your provider? Yes, as said above it registers for handlers and attributes. Also _objectChanged and _SetAttributeFor() etc. are called. They're just not called directly after creating the object (when being in the same request that is). cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] ZPatterns AttributeProvider question
Good evening everybody! I have some question regarding attribute and agent programming for ZPatterns. I have some provider which is registered for the "handlers" and "attributes" methods and some attributes. My problem is when trying to create a new object and directly editing it, e.g. obj=newItem() obj.propertysheets.data.manage_changeProperties(foobar=13) the propertysheet is defined and my provider is also registered for handling the property "foobar". Unfortunately my SetAttributeFor() method is never called. Also defined in my provider are _objectAdded(), _objectAdding(), _objectChanged() (but not _objectChanging()). Here's what they do: _objectAdded() is storing the data collected via SetAttributeFor() into an sql database as new record _objectAdding() is doing the same as the original method but also committing the transaction as otherwise the new record will not be stored into the database (basically it calls _objectAdded() via the transaction mechanism). I needed to do that as otherwise a getattr on the primary key (==loadAttrib) failed as I try to retrieve the record inside the _AttributeFor() method. _objectChanged() doing the same as objectAdded() but updating the sql database. So what's happening is that _objectAdding() and _objectAdded() get called from the newItem() call but _SetAttributeFor() is not called. When doing the manage_changeProperties() inside a new method which is called manually (and does a getItem()) everything is fine. I've also tried to do obj=newItem() newid=obj.id obj2=getItem(newid) obj2.propertysheets.data.manage_changeProperties(foobar=13) but this also did not call _SetAttributeFor(). So does any ZPatterns wizard has any idea on what I might do wrong? (I can also send someone the source code if this message seems too confusing ;-) regards, Christian Scholz PS: What I am trying to create is an sql attribute provider for ZPatterns which is some replacement for manually creating the zsql method and SkinScript methods. Unfortunately it seems more difficult than I thought at first ;-) -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] Re: zope nautilus cabal
Hi! > Basically, 'access contents information' isn't a great permission. If you > turn if off, life gets horrible, if you leave it on, bits hang out. I'd > prefer to see something like: > - Access Contents Information via HTTP > - Access Contents Information via FTP > ...etc... This sounds more as if some "external interface" concept is needed, where you can add/disable modules for handling the XMLRPC stuff and the webdav stuff and the normal management interface stuff and the ftp stuff.. And this again also touches the area of model-view again, which seems a bit far away unfortunately .. -- mr topf ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns Attribute Provider Problem
Well, sorry for mailing to quickly, but removing
_objectChanging() from my class solved the problem. My version simply
wasn't setting the right status to the transaction.
(though I think that I had a reason for adding it some time ago..)
But maybe it's working now.. (but it seems so)
-- christian
On Mon, Apr 09, 2001 at 07:31:41PM +0200, Christian Scholz wrote:
> Hi everybody!
>
> I am trying to write some attribute provider for storing data inside
> an SQL table as I am bored of typing parameter lists over and over
> again (e.g. in every ZSQL method, in every SkinScript method etc.)
>
> The problem now is that adding a new row to the table (and thus adding
> a new object to the rack) is working well but updating an object does not.
> Basically the main problem is that _objectChanged is not called in my provider.
>
> What I did is the following:
>
> - If an attribute of the propsheet is changed, _SetAttributeFor is called
> and stores it into the cache (as I don't want to update the database for
> every attribute change).
>
> - _objectChanged is then supposed to create an sql query and execute it.
> This - as said - is actually not happening, as the whole method is not called.
>
> - My namesForRegistration looks like this:
>
> def namesForRegistration(self,container):
> """ return names for registration """
> return {
> 'provides':('handlers','attributes'),
> 'getattr': self._attribs.keys(),
> 'setattr': self._attribs.keys(),
> 'delattr': self._attribs.keys()
> }
>
> with self._attribs containing the attributes my provider is responsible for.
>
> So is there any reason why _objectChanged is not called? Actually I am a bit
> confused when reading the ZPatterns source which should call it.. ;-)
> So which are the conditions under which it's supposed to be called?
> (_objectAdded() as said before is called actually..)
>
> I hope someone can help me :)
>
> cheers,
> Christian
>
> --
> COM.lounge http://comlounge.net/
> communication & design [EMAIL PROTECTED]
>
> ___
> Zope-Dev maillist - [EMAIL PROTECTED]
> 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 )
--
COM.lounge http://comlounge.net/
communication & design [EMAIL PROTECTED]
___
Zope-Dev maillist - [EMAIL PROTECTED]
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 )
[Zope-dev] ZPatterns Attribute Provider Problem
Hi everybody!
I am trying to write some attribute provider for storing data inside
an SQL table as I am bored of typing parameter lists over and over
again (e.g. in every ZSQL method, in every SkinScript method etc.)
The problem now is that adding a new row to the table (and thus adding
a new object to the rack) is working well but updating an object does not.
Basically the main problem is that _objectChanged is not called in my provider.
What I did is the following:
- If an attribute of the propsheet is changed, _SetAttributeFor is called
and stores it into the cache (as I don't want to update the database for
every attribute change).
- _objectChanged is then supposed to create an sql query and execute it.
This - as said - is actually not happening, as the whole method is not called.
- My namesForRegistration looks like this:
def namesForRegistration(self,container):
""" return names for registration """
return {
'provides':('handlers','attributes'),
'getattr': self._attribs.keys(),
'setattr': self._attribs.keys(),
'delattr': self._attribs.keys()
}
with self._attribs containing the attributes my provider is responsible for.
So is there any reason why _objectChanged is not called? Actually I am a bit
confused when reading the ZPatterns source which should call it.. ;-)
So which are the conditions under which it's supposed to be called?
(_objectAdded() as said before is called actually..)
I hope someone can help me :)
cheers,
Christian
--
COM.lounge http://comlounge.net/
communication & design [EMAIL PROTECTED]
___
Zope-Dev maillist - [EMAIL PROTECTED]
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 )
Re: [Zope-dev] FTP interface being worked on?
Hi! > By the way -- is it me, or is the current Import/Export interface > broken? I tried to select multiple objects to export, but I can only > get the first one to actually be exported. It's not just you.. but never thought about whether this might be a bug ;-) -- christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] More Core Session Tracking Woes :-(
Hi! > > FYI, you are not the only one.. I also experienced that. That's also the reason > > why I am back to using SQLSession again.. > > How hard was that to swap in? depends. Of course the API is not the same.. So it depends if you've factored out the session stuff in some dtml or python method which you simply have to change. (and if you can factor it out, of course.. I basically just looked one variable up in the session and copied it to REQUEST, so I simply had to change one header file). Also I haven't written a plugin for my Virtual Site Root yet to make use of the Core Session Tracking package so I cannot use it in siterooted environments yet anyway.. cheers, MrT -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] More Core Session Tracking Woes :-(
Hi! > The setup: > -Zope 2.2.4 > -Core Session Tracking 0.7 > -Cookie session ID manager (path=/,lifetime=0) > -Internal Session Data Manager (timeout set to 60 minutes) > > This seems to randomly (but sometimes very frequently) loose session data items > :-( FYI, you are not the only one.. I also experienced that. That's also the reason why I am back to using SQLSession again.. (actually I was too lazy to track it down..) -- mr topf -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] Adding a property to a batch of objects ?
Hi! > I wanted to do sonething like this: > > > (or objectItems). then just say or you might change the ... to (hope this is right, have not tested it.) First form directly builds a list of the objects instead of the Ids, second form first looks up the id with _['sequence-item'] and then looks up the object with that id in the local object (your folder). Hope this helps.. -- mr topf -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] Playing with DateTime
Hi! > IMHO, the best approach would be to make mxDateTime available separately > from DateTime in the _ variable. That would avoid breaking any code, but > allow anyone who wanted to use mxDateTime that option from within Zope. > > I think a product that adds mxDateTime to the _ variable would be your > best bet. Well, my opinion, too. On the long run this might get the preferred way of dealing with time stuff so that the use of DateTime gets depreceated. But I guess DC needs to decide this.. I also dunno if there are any problems with mxDateTime being used (e.g. security concerns, whatever) as I havent't looked to closely at this. -- christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] getPhysicalPath?
Hi! > Ahh... thanks Brian. Thats a new one to me :) To me, aswell. Is this documented somewhere? ;-) And can someone explain to us where the differences between aq_chain and getPhysicalPath() are? Actually getPhysicalPath() seems also to walk up aq_parent. Or am I missing something? regards, Christian > > > The aq_chain attribute (computed at time of access) provides > > a list (in reverse order) of the objects in the acquisition > > path. > > > > Hope this helps! > > > > > > Brian Lloyd[EMAIL PROTECTED] > > Software Engineer 540.371.6909 > > Digital Creations http://www.digicool.com > > > > > > > > > > ___ > > Zope-Dev maillist - [EMAIL PROTECTED] > > 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 ) > > > > > > ___ > Zope-Dev maillist - [EMAIL PROTECTED] > 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 ) -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] ZClass propertysheets
Hi! I am just wondering if it's possible to get hold of the properties of a propertysheet of a ZClass without instanciating it. So when instanciating I would just do map=object.propertysheets[1].propertyMap() and get all information of my ZClass instance object. So is this also possible on the class definition? (actually I would assume it is stored somewhere as I can edit it via the management interface. I just didn't manage to get it from the source code..) (next question might be how to get the class when only having a meta type or the path to it.. but I think I will just search the Zope source..) kind regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] initializing objects in ZPatterns
Hi! > To get company_address stored, you can probably do something like: > > your_object.manage_changeProperties(company_address=your_object.company_address) > > or possibly your_object.propertysheets.NameOfSheet.manage_changeProperties... > depending on what kind of object you're using. I've now tried WHEN OBJECT ADDED CALL self.propertysheets.data.manage_changeProperties(company_address=self.company_address) which gives me some Unauthorized error on the method call. (when doing it manually in some dtml method after newItem() it works ok, though. And I am logged in as Manager..) Sometimes I wish Zope would print more about the details of such an response, e.g. which permissions might be missing.. Anyway, I will keep experimenting with it (also tried to subclass directly from DataSkin, to override some method which create the company address) but then I noticed that it always needs to be a ZClass (except I would also implement my own Rack, I guess..) Wonder when the basic app will be running.. ;-) ok, nice evening everybody! :) -- christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] initializing objects in ZPatterns
Hi! > > I just encountered another problem with ZPatterns (well not exactly with ZPatterns > > but the way I use it.. ;-) > > > > I have some Specialist with a normal rack which stores data persistently in the >ZODB. > > Everytime I am creating a new object I want to create an object of another >specialist > > on the fly and store it's id inside my first object. > > > > Thus I have setup a SkinScript with the following content: > > > > > > INITIALIZE OBJECT WITH company_address=addresses.createAddress() > > > > WITH addresses.getItem(self.company_address) COMPUTE company_addr=RESULT > > > > STORE company_address IN SELF > > > > The createAddress is called actually and it returns the ID of the new object > > (thus the object is created). Unfortunately the id is not stored inside > > company_address. After reading the new object again this value is still empty > > (and thus also no address object is return by the address specialist). Same happens > > when just using some dummy property and filling it with a fixed string. This > > also disappears.. > > To get company_address stored, you can probably do something like: > > your_object.manage_changeProperties(company_address=your_object.company_address) > > or possibly your_object.propertysheets.NameOfSheet.manage_changeProperties... > depending on what kind of object you're using. > > You could have this as a trigger: > > WHEN OBJECT ADDED CALL > self.manage_changeProperties(company_address=self.company_address) Well, the first one seems to work, but the SkinScript trigger does not unfortunately. So I will try to incorporate this into my other code (which is a bit more tricky as I automated all the property handling stuff). But thanks anyway! > As for whether it is a bug, the implementation of the INITALIZE clause just puts > its attributes directly into the DataSkin's attribute-cache. It appears to be > designed for the behaviour that you see. > > You could get the behaviour that you want by making the implementation actually > set the attributes instead. I'm not sure what the other implications of doing > that are, though. > > If you're interested, the code is in > > lib/python/Products/ZPatterns/SkinScript/Components.py > about line 71. ok, thanks, will have a look at that. cheers, Christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] initializing objects in ZPatterns
Hi! I just encountered another problem with ZPatterns (well not exactly with ZPatterns but the way I use it.. ;-) I have some Specialist with a normal rack which stores data persistently in the ZODB. Everytime I am creating a new object I want to create an object of another specialist on the fly and store it's id inside my first object. Thus I have setup a SkinScript with the following content: INITIALIZE OBJECT WITH company_address=addresses.createAddress() WITH addresses.getItem(self.company_address) COMPUTE company_addr=RESULT STORE company_address IN SELF The createAddress is called actually and it returns the ID of the new object (thus the object is created). Unfortunately the id is not stored inside company_address. After reading the new object again this value is still empty (and thus also no address object is return by the address specialist). Same happens when just using some dummy property and filling it with a fixed string. This also disappears.. So what do I have to do to get it stored? (the other data which is handled via the default plugins is stored and all the attributes are defined inside the ZClass..) Hope, anyone knows how this might be fixed.. regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns and SQL
Hi! > >So I shouldn't define it inside the ZClass propertysheet? > >Do I understand this right? > >(I remember having problems when using properties not defined in a > >sheet somewhere.. or am I mistaken completely somehow? ;-) > > You can't do this with a load attribute, because a default value will > exist. The easiest way to deal with this is simply to have a seperate > attribute name just for loading, or use an attribute that isn't on a > property sheet. Ok, I now removed customer_id (which won't be set anyway as this is the id set by the database automatically) from the propertysheet and used this as loading attribute. Seems to work now :) (though I am still confused a little, as I tried one non-existing attribute and it didn't work. But this one also was not defined inside the SkinScript and the database). But it's working now, so it's ok :) Thanks again! Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns and SQL
Hi! > >So as the item is an instance of my ZClass containing also an company_name > attribute > >it will always return the item and because of that the Rack thinks it's > already there. > >So why is this happening? What do I need to change? > > You need to use an attribute which the object has *if and only if* it > exists in the database. If the class has the attribute defined, all > instances exist, and you can't even load it with SkinScript because > ZPatterns uses __getattr__ to redefine attributes, and that doesn't work if > the attribute already exists in the class. So I shouldn't define it inside the ZClass propertysheet? Do I understand this right? (I remember having problems when using properties not defined in a sheet somewhere.. or am I mistaken completely somehow? ;-) > >When changing the loadAttrib to something else (non existing), the newItem > is working > >but instead getItem is not working anymore as it returns None only.. > (somehow the zsql > >method is not even called thus the skinscript seems not to work in that > case (I guess > >because the loadAttrib is not defined in it somewhere)). > > Right. The load attribute has to be one defined in a provider, and which > will end up being "NOT_FOUND" if the object doesn't exist in the table > you're querying. I will experiment a little with this information. Hope it helps :) Thanks! -- christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] ZPatterns and SQL
Hi there! I have a little problem with my Specialist and DataSkins regarding storage inside an sql database. Mainly I don't know what to put under the "loading by accessing attribute" in the Storage tab of the Rack. I have the following SkinScript: WITH QUERY getCustomer(customer_id=self.id) COMPUTE company_name, customer_no WHEN OBJECT ADDED CALL addCustomer( company_name=self.company_name, customer_no=self.customer_no ) WHEN OBJECT CHANGED STORE company_name,customer_no USING updateCustomer(customer_id=self.id, company_name=self.company_name, customer_no=self.customer_no ) I've now set the loadAttrib to "company_name" and retrieving objects works fine. Unfortunately when doing a newItem() with some random new id (it's not used as the database should provide the id) I get an error that this id already exists. Of course I've checked it in the database and it's not there. After some code reading I found in retrieveItem: a = self.loadAttrib item = None if a: item = self._RawItem(key) if hasattr(item.aq_base, a): return item So as the item is an instance of my ZClass containing also an company_name attribute it will always return the item and because of that the Rack thinks it's already there. So why is this happening? What do I need to change? When changing the loadAttrib to something else (non existing), the newItem is working but instead getItem is not working anymore as it returns None only.. (somehow the zsql method is not even called thus the skinscript seems not to work in that case (I guess because the loadAttrib is not defined in it somewhere)). Does anybody know any help? (somehow I once had a working specialist with an sql table but I dunno what happened to make it not work anymore). My ZPatterns version is the one of stevea and my Zope is 2.3 kind regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns question
Hi! > > Well, virtual in the sense as a specialist is no real folder but can > > provide content from different sources. Thus what I mean is some mechanism > > which emulates objectIds() etc. so it looks to the user (and the ones > > using it via dtml) like a normal folder object. > > Somehow like the Customizer but without the need for actually creating > > Zope objects. Something inbetween Specialist and Customizer this would be > > I guess. > > yeah, this is exactly what I'm after too. I'd like the virtual objects > the specialist is responsible for to have normal Zope management > screens. Well, I think this shouldn't be too difficult to create. Just a subclass of Specialist and add some Contents-Tab, define objectIds() etc. in the Methods tab and call it in the management screen of "Contents". Though the more advanced things like Copy/Paste/Rename might then not be available.. But for me actually it would be sufficient to get the contents. > Also, much mroe trickily, I'd like them to be able to contain each > other, although Steve A's __bobo_traverse__ trick might help with this > bit... What do you mean? Nesting Specialists (or these virtual folder specialists)? cheers, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns question
Hi Steve! On Thu, Dec 28, 2000 at 10:45:46PM +, Steve Alexander wrote: > Steve Spicklemire wrote: > > > > > CS> 2. Is it planned to provide something like a virtual folder > > CS> which acts like a normal object manager but is controlled via > > CS> ZPatterns (so actually something like Folder with Customizer > > CS> Support just without the "anchor" in ZODB. (would also > > CS> require some mechanism asked for in 1.) > > > > Hmmm... I'm not sure what you're after here. Why not just use > > a Specialist? In what sense do you want it to be virtual? > > (Are you looking for a dynamic traversal interface that > > would allow you to map URLs to objects that are managed by > > ZPatterns? Can you give an example?) > > Reading this just after reading the source to Specialists.py, I had a > thought; and tried it out; and it works! :-) > > You can use SkinScript to define __bobo_traverse__ for a particular kind > of DataSkin in a Specialist. > > For example: > >WITH SELF COMPUTE __bobo_traverse__=traversal_method that's very cool :) I just created some base class for this from which my ZClasses are derived but nice to know that it's also possible without this base class. (though for this application it's easier with this base class as it's easier to configure then) cheers, Christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
Re: [Zope-dev] ZPatterns question
Hi! Actually should answer to these posts... ;-) On Thu, Dec 28, 2000 at 03:11:37PM -0500, Steve Spicklemire wrote: > > Hi Christian, > > Well, nobody else answered that I saw... so I'll take a crack > at your questions > > >>>>> "CS" == Christian Scholz <[EMAIL PROTECTED]> writes: > > CS> Hi there! > > CS> Finally I managed to get a basic understanding of how to do > CS> things with ZPatterns ;-) So seems quite cool :) (and > CS> hopefully I find some time to write some basic howto about it) > > CS> But I have some little questions: > > CS> 1. Is it possible to retrieve the set of known IDs from a > CS> specialist? Or would I need to add my own method to it which > CS> does this (and change it accordingly if I switch to another > CS> storage method)? > > If you store persistently you can use the Rack's method: > > "defaultRack.getPersistentItemIDs()" Thanks, also found this out in the meanwhile. > but a couple of notes: 1) this returns a BTree object, not a simple > list, so you can't iterate through it in DTML. You'll need to copy > it to a simple list for that.. and 2) If you change to a different > storage you'll need to create your own method (ZSQL Method?). What I've > found is that if you have a large number of objects you'll either want > to query a Catalog, or an SQL database to get Ids that match some criteria > that limit the number of hits to something that makes sense to display > in a browser. I am now using some method I create inside the specialist for it. > CS> 2. Is it planned to provide something like a virtual folder > CS> which acts like a normal object manager but is controlled via > CS> ZPatterns (so actually something like Folder with Customizer > CS> Support just without the "anchor" in ZODB. (would also > CS> require some mechanism asked for in 1.) > > Hmmm... I'm not sure what you're after here. Why not just use > a Specialist? In what sense do you want it to be virtual? Well, virtual in the sense as a specialist is no real folder but can provide content from different sources. Thus what I mean is some mechanism which emulates objectIds() etc. so it looks to the user (and the ones using it via dtml) like a normal folder object. Somehow like the Customizer but without the need for actually creating Zope objects. Something inbetween Specialist and Customizer this would be I guess. > (Are you looking for a dynamic traversal interface that > would allow you to map URLs to objects that are managed by > ZPatterns? Can you give an example?) yes, something like this comes close I guess. e.g. I have some sql database consisting of people's addresses and I want to create some specialist which queries this database (or how ever this specialist is configured right now in order to get the object ids) and shows it as if it's a normal folder. (seems to me more transparent at some point). But I assume it's not suited for all applications e.g. if the number of objects gets a little bigger. And I guess I can create this myself if I subclass from Specialist and add the rest of the ObjectManager's interface to it. > CS> 3. Is it possible to use ZPatterns also without some exta > CS> ZClass defined in a Product? At least if I don't need methods > CS> for my objects but simple want to define the attribute they > CS> know. Would be nice as I could then hold everything together > CS> in one place (the specialist that is) without requiring to > CS> install something also in the Products folder of the Control > CS> Panel. > > I believe you need to either create a Python subclass, or a ZClass > subclass of DataSkin. "Raw" DataSkins don't have the right permissions > to allow for TTW access. If you're creating a 'product' anyway... just > make a 'dummy' class that you can use for storage. ok, but I still need to define the attributes I want to handle inside a propertysheet of this ZClass, right? Actually what I would like is something which directly resides inside the Specialist and not somewhere outside in some Product folder. So everything is close together in one place. > CS> That's it for now, I will keep experimenting then.. :) > > Good Luck! Oh, it's working better than I thought :) but thanks, christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] ZPatterns question
Hi there! Finally I managed to get a basic understanding of how to do things with ZPatterns ;-) So seems quite cool :) (and hopefully I find some time to write some basic howto about it) But I have some little questions: 1. Is it possible to retrieve the set of known IDs from a specialist? Or would I need to add my own method to it which does this (and change it accordingly if I switch to another storage method)? 2. Is it planned to provide something like a virtual folder which acts like a normal object manager but is controlled via ZPatterns (so actually something like Folder with Customizer Support just without the "anchor" in ZODB. (would also require some mechanism asked for in 1.) 3. Is it possible to use ZPatterns also without some exta ZClass defined in a Product? At least if I don't need methods for my objects but simple want to define the attribute they know. Would be nice as I could then hold everything together in one place (the specialist that is) without requiring to install something also in the Products folder of the Control Panel. That's it for now, I will keep experimenting then.. :) cheers, Christian -- Christian Scholz MrTopf@IRC COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] some suggestion..
Hi! While trying to code some python Product I had the need for extending manage_delObjects in some Folderlike object. I simply want to call the the ObjectManager's version of it and then do my own stuff afterwards (like deleting some other dependant objects). Unfortunately when calling manage_delObjects() one cannot decide if some error occured or everything went ok as in both cases I get back some string (be it some html formatted error message or the manage_main dialog). So what about splitting these methods in two, e.g. one _delObjects() and one manage_delObjects(). The first one would the the actual work like testing for error conditions and deleting the objects and the other one would be the one called by the management interface. manage_delObjects() would then call _delObjects() and checks for errors afterwards generating some error message or the next dialog. The errors could then be passed via exceptions or similar methods from _delObjects() to manage_delObjects(). This would make subclassing a lot easier and would be at least a start in separating layout from logic. Just an idea :) cheers, Christian PS: Maybe it would even make XMLRPC more usable as there I also get back only html formatted error messages instead of something making more sense for a script (like an error code or something). -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] Redirection and Authentication
Hi! I had a little problem and just wanted to ask if someone knows an explanation.. The goal: I want to show a list of newsitems to the user. If an administrator wants to change it he should be able to log in and see the same list but with edit- buttons. The setup is: /folder/list- public accessible dtml method which show the list /folder/edit- protected dtml method The edit method is protected so that the login requester pops up and asks for a passwort. Originally I was simply redirecting back to the list page inside the edit document by (the only line in the document) In list I then check for a login with .. show edit button here ... The problem is now that after the Redirect AUTHENTICATED_USER is again set to Anonymous (which I tested by printing AUTHENTICATED_USER). If I do no redirect but put a normal link back to the list document, everything works as expected (thus showing the edit button). So has anyone an explanation? (Version is 2.2.0) (right now I directly include the list again in the edit document by using dtml-var. Works for this little thing but I assume this only being a workaround..) (I also remember having some strange problems with a redirect from python some time ago. Back then it did not commit the database transaction in Oracle. After putting a get_transaction().commit() before the redirect made things work again.) cheers, Christian -- Christian Scholz MrTopf@IRC COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] dcoracle problem..
Hi!
Maybe someone has had this before and knows a solution:
I am having a stored procedure proc1 and I am calling it twice:
dbc=DCOracle.Connect("secret!")
c=dbc.cursor()
dbc.procedures.my_pkg.proc1(c)
print c.fetchall()
c.close()
dbc.procedures.my_pkg.proc1(c)
print c.fetchall()
In proc1 it looks basically like
OPEN mycursor FOR
SELECT f1,f2,f3
FROM mytable
Thus the cursor get's opened and filled (mycursor is actually a cursor
variable).
The expected result for the above python script is that it's printing two times
the same list of results. Actually it doesn't, the second one is empty.
When creating a new cursor before the second call like
c2=dbc.cursor()
dbc.procedures.my_pkg.proc1(c2)
print c2.fetchall()
I get
Traceback (innermost last):
File "./testcall.py", line 16, in ?
print c2.fetchall()
File "/opt/python/lib/python1.5/site-packages/DCOracle/ociCurs.py", line 397, in
fetchall
v=f(1)
File "/opt/python/lib/python1.5/site-packages/DCOracle/ociCurs.py", line 375, in
fetchone
raise error, 'not querying'
oci.error: not querying
Thus it seems as if the cursor is not opened..
Has anyone an idea about this.. (I was also looking at the code but it's a little
complex..)
regards,
Christian
___
Zope-Dev maillist - [EMAIL PROTECTED]
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 )
[Zope-dev] DCOracle patch
Hi! I encountered the following bug in the recent DCOracle this week: When trying to use stored procedures with more than 1 out parameter like create or replace procedure testproc ( pname out string, pid out int) AS BEGIN pname:='test1'; pid:=100; END; / and calling it from python with pname,pid= dbc.procedures.testproc() an exception will be raised: Traceback (innermost last): File "./testcall.py", line 9, in ? x= dbc.procedures.testproc() File "/opt/python/lib/python1.5/site-packages/DCOracle/ociProc.py", line 149, in __call__ raise t, v, tb NameError: t This is due to the fact that an return from __call__ is only executed when exactly one parameter is returned by Oracle. I fixed this by adding a line returning the whole tuple. BTW: The following raise will do the NameError because the variables will only be defined when the for loop is raising an exception. Without the return it was returning from the loop without an exception and thus raise had no valid namespace. And here is the patch: Common subdirectories: DCOracle.orig/CVS and DCOracle/CVS diff -c DCOracle.orig/ociProc.py DCOracle/ociProc.py *** DCOracle.orig/ociProc.pyThu Jan 14 18:44:56 1999 --- DCOracle/ociProc.py Sun Jul 2 15:54:27 2000 *** *** 143,148 --- 143,149 if f is not None: v=f(v) r.append(v) if len(r)==1: return r[0] + return r except: t,v,tb=sys.exc_info() raise t, v, tb finally: tb=None Go to your DCOracle directory in site-packages/DCOracle (or the source) and run patch -p1 http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] Re: The new SiteAccess and __no_before_traverse__?
Hi! > - Original Message - > From: "Christian Scholz" <[EMAIL PROTECTED]> > > I am playing around with SiteAccess and Zope 2.2b1 right now and I wonder > > if it's still possible to disable SiteAccess (actually an AccessRule) by > > using __no_before_traverse__? > > It seems not to and I am a bit trapped right now.. ;-) > > This has changed, and I need to make it clearer in the documentation. There > is no longer a global "disable all __before_traverse__ hook" URL. Instead, > there are specific SiteAccess object-disabling environment variables. Actually I looked up the documentation and there is still mentioned this hook. > You need to start Zope with environment variable SUPPRESS_ACCESSRULE set. > There is a similar SUPPRESS_SITEROOT variable for SiteRoots. Hm, is there a way of doing this without having to restart Zope. In some environments this might not be the best solution (e.g. if you want to move your application to a production server and have to set up a special SiteRoot and then notices that you mistyped something.. ) > Fire up a command shell and type (*nix): > > SUPPRESS_ACCESSRULE=1 > export SUPPRESS_ACCESSRULE Ok, thanks, that's working :) > > Even deleting SiteAccess from the Products-Directory didn't work as then > > Zope encountered a missing doc string error for the directory when trying > > to access it. > > That's because the fundamental mechanism is built into Zope 2.2, rather than > grafted on by SiteAccess 2. It does look like we'd better wrap the hook > call in a try..except pass, though. It's a little confusing the way it is right now.. Thanks for the fast answer :) -- christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] The new SiteAccess and __no_before_traverse__?
Hi! I am playing around with SiteAccess and Zope 2.2b1 right now and I wonder if it's still possible to disable SiteAccess (actually an AccessRule) by using __no_before_traverse__? It seems not to and I am a bit trapped right now.. ;-) Even deleting SiteAccess from the Products-Directory didn't work as then Zope encountered a missing doc string error for the directory when trying to access it. Hope, someone can help.. otherwise I might have to use 2.1 again until this is fixed.. regards, Christian -- COM.lounge http://comlounge.net/ communication & design [EMAIL PROTECTED] ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
[Zope-dev] Re: [Zope] Please test this Zope 2.2 beta 1 ZSQLMethods fix...
Hi! > I think that you have hit it on the head - we're trying *hard* to > make Zope harder to fool :^) Try adding the following to your > class statement, for example: > > class MyClass: > > __allow_access_to_unprotected_subobjects__=1 Just a question: Is this documented somewhere.. I've seen some more of these things here and wonder if there's some explanation what exactly has changed in 2.2 and which constants can be used to control these things.. (well, my classes seem to work under 2.2, so no problem for me right now but one never knows.. ;-) greetings, Christian ___ Zope-Dev maillist - [EMAIL PROTECTED] 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 )
