- Zope - Zope lists

Re: [Zope] Zope woes continue - server going down regularly. (possible solution)
by Tres Seaver 08 Dec '99

08 Dec '99

Michel Pelletier <michel(a)digicool.com> wrote: > > -----Original Message----- > > From: Cayce Ullman [mailto:c_ullman@yahoo.com] > > Sent: Wednesday, December 08, 1999 11:26 AM > > To: zope(a)zope.org; panda(a)skinnyhippo.com > > Subject: [Zope] Zope woes continue - server going down regularly. > > (possible solution) > > > > > > >>> Alternatively, the server just hangs and doesn't > > respond > > >>> (but ps -aux shows the process is running). > > >> > > >>Is it spinning? (consuming 100% CPU resources?) or > > hung? (consuming > > >>none..) > > > > I noticed a couple of weeks ago that my Zope, > > started spinning when I hit a certain page 5 or > > 6 times. I noticed that this page was different > > from the others as it tried to do a Zope.app(). > > This was interesting cause the actual Zope.app() > > call worked, except it would eventually hose the > > server. I figured I was just doing something wrong > > and figured out another way to do it. Anyway, > > here's how to reproduce the problem : > > > > Create a external method that looks like this : > > > > import Zope > > > > def helloworld(self): > > x = Zope.app() > > return "Hello World" > > This is because the connection gets opened, but never closed. Zope 2.x > is hardwired to have only 7 available database connections. We'll > probably add a knob in 2.2 that lets you configure the number of db > connections. In either case, the problem is not the hardwiring, the > problem is that your code has a bug in it, because you do not free the > resource you are consuming. Do an: > > x.close() > > at the end of your method and you won't run out of connections. There is stuff in ZODB.ZApplication.py which looks as though it should not be necessary to close a ZODB connection manually: conn=db.open(version) # arrange for the connection to be closed when the request goes away cleanup=Cleanup() cleanup.__del__=conn.close REQUEST._hold(cleanup) Am I misreading this? Or do we have a refcount cycle in here somewhere? -- ================================================================================ Tres Seaver tseaver(a)palladion.com http://www.palladion.com Palladion Software Houston, Texas, USA 713-523-6582

1 0

RE: [Zope] Zope stability
by Michel Pelletier 08 Dec '99

08 Dec '99

> -----Original Message----- > From: LD Landis [mailto:ldl@LDL.HealthPartners.COM] > Sent: Wednesday, December 08, 1999 3:14 PM > To: michel(a)digicool.com > Subject: Re: [Zope] Zope stability > > > Michel, > > Not that you don't have anything else to do... but it would be real > handy to have a "tutorial" on how to use all of this stuff... If one > has been created... sorry, and TIA for the pointer! (Also, > what's all > that stuff about the other ports that are being listened to as well? > isn't that a remote "control terminal" or some such) What should happen is someone needs to expand on ken's how-to. It's like this, debugging is an art above and beyond the debugger, if you don't know *how* to debug, no debuger, no matter how visual, will aid you. So really, what is needed is a doc that explains how to fire up into the debugger, and that's easy. If you're lost from there, you need to learn how to debug (and how to use a command line debugger like pdb (which is a lot like all other forms of command like debugger I've ever used). This is how you get into the debugger. Shut down your Zope. Go to the lib/python directory Run python: [michel@korak python]$ python1.5.2 Python 1.5.2 (#1, Jul 5 1999, 14:47:37) [GCC egcs-2.91.66 19990314/Linux (egcs- on linux2 Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam import Zope and ZPublisher Py$ import ZPublisher, Zope Now you can issue requests to Zope just like a web browser: Py$ ZPublisher.Zope('index_html') will call and return the 'index_html' object in the root folder. You can enter the debugger with the d=1 argument: Py$ ZPublisher.Zope('index_html', d=1) * Type "s<cr>c<cr>" to jump to beginning of real publishing process. * Then type c<cr> to jump to the beginning of the URL traversal algorithm. * Then type c<cr> to jump to published object call. > <string>(0)?() pdb> This fires you into the python debugger, pdb. First off, hit 's' to step into the debugger. For your convinience, two breakpoints are set for you. Hit 'c' once to continue to the first breakpoint, which is at the beginning of the URL traversal to find your object. I ususaly skip beyond this as it is really only useful to look at the request before it traversed and to debug ZPublisher. Hit 'c' again to continue past the first breakpoint to the second, which is rigt before the object you are looking for is called. This is most useful to debug application specific logic. -Michel

1 0

Zope direction and 2.2 roadmap...
by Brian Lloyd 08 Dec '99

08 Dec '99

Hi all. Now that Zope 2.1.0 is out the door, we at DC want to let the community know our direction and the "roadmap" leading to 2.2. In a nutshell, our direction and focus will be on usability. Making it easier to get things done with Zope has been a recurring theme on the lists lately, and we will be working to make progress on this front. For 2.2, here is the short list of we will be working on: o Online Help System - Amos has been doing great work on a real honest-to-goodness online help system for Zope that will make its debut in 2.2. o Python Methods - work will be done to include Evan Simpson's Python Methods in the Zope core. This should help a lot in avoiding the headaches of convoluted DTML logic. o Site Objects - will make it easier to deploy "virtual sites" in the same Zope with far less effort than is required now. o The ability to associate ZClasses as "brains" for relational database data instead of having to write Python classes. o And of course, fixes for bugs and issues in the Collector. We are expecting to have these things available in at least a pre-release by the time of the Zope track at the Python Conference. There are other usability initiatives underway as well. Rob and Amos have been doing a _lot_ of work coming up with a much better and much more scalable approach to documentation that should have a big impact on the quality and breadth of documentation. Work will also continue on making Zope.org a better resource with things like ZTopics (Yahoo-style directory views). And there are some other projects percolating (which I can't talk about yet :) that will have a big impact on usability. Brian Lloyd brian(a)digicool.com Software Engineer 540.371.6909 Digital Creations http://www.digicool.com

1 0

RE: [Zope] Zope stability
by Michel Pelletier 08 Dec '99

08 Dec '99

> -----Original Message----- > From: Teodor Cimpoesu [mailto:teo@digiro.net] > Sent: Wednesday, December 08, 1999 11:26 AM > Cc: zope(a)zope.org > Subject: Re: [Zope] Zope stability > > > Hi all, <snip> > > Michel Pelletier wrote: > > to integrate them with Zope, I've never gotten either to work well > > debugging Zope. > is idle IDLE? yeah. > [sic] There are only CVS snapshots [iirc] for this beast > and I'll have to build it from source to give it a try :\ > > But anyways, you guys @ DC what debugging tool are using when hacking > into Zope? The Zope debugger, which is based on pdb. My most common form of debugging is to find the point in the code I want to stop at, and insert the code: import pdb pdb.set_trace() restart Zope, hit a page, and the controlling terminal is now in the debugger (you must be in -D debug mode in order to prevent Zope from detaching from the terminal) -Michel

1 0

RE: [Zope] New site using Zope technology
by Michel Pelletier 08 Dec '99

08 Dec '99

> -----Original Message----- > From: Bruno Mattarollo [mailto:brunomadv@ciudad.com.ar] > Sent: Wednesday, December 08, 1999 1:24 PM > To: zope(a)zope.org > Subject: [Zope] New site using Zope technology > > > Hello... > > I am glad to announce our first public beta of the site > we have been > developing for the last month. It's all done in Zope served > through Apache. Were you just teasing us or are you going to tell us where it is? ;) -Michel

1 0

RE: [Zope] String to object ID
by Chris McDonough 08 Dec '99

08 Dec '99

Now this is cool. Chris McDonough mailto:chrism@digicool.com Digital Creations http://www.digicool.com Publishers of Zope http://www.zope.org > -----Original Message----- > From: Evan Simpson [mailto:evan@tokenexchange.com] > Sent: Monday, December 06, 1999 10:45 AM > To: zope(a)zope.org; Goodrichs > Subject: Re: [Zope] String to object ID > > > I missed some of the prior discussion, so I hope this isn't > irrelevent. > > Are you trying to access an object given a string such as > 'hie/dee/hoe'? If > so, try this: > > DTML Method FetchObj: > <dtml-let obj="[_]"> > <dtml-in expr="_.string.split(fetchStr, '/')"> > <dtml-call expr="obj.append(obj.pop()[_['sequence-item']])"> > </dtml-in> > <dtml-return expr="obj[0]"> > </dtml-let> > > example call: <dtml-var expr="_.render(FetchObj(_.None, _, > fetchStr='hie/dee/hoe').id)"> > > Some notes about the above: > o We start with a list containing the global namespace object. > o For each path element, we pop the current object, find the > sub-object, and > put it back in the list. > o You could use REQUEST.set instead of the list foolery if > you wanted to. > o _.render is necessary since the .id of folders is a string, > while that of > methods is a method :-P > > If you use PythonMethods, the above can be more simply written as: > > PythonMethod FetchObj: > <params>_, fetchStr, attr=None</params> > obj = _ > for p in string.split(fetchStr, '/'): > obj = obj[p] > if attr is None: > return obj > else: > return render(_.getattr(obj, attr)) > > > <dtml-var expr="FetchObj(_, 'hie/dee/hoe', 'id')"> > > ----- Original Message ----- > From: Goodrichs <nbd95(a)macconnect.com> > > Well I still don't know how to reference a Zope object from > a string, but > I > > have a interim solution for now. > > > > _______________________________________________ > Zope maillist - Zope(a)zope.org > http://lists.zope.org/mailman/listinfo/zope > No cross posts or HTML encoding! > (Related lists - > http://lists.zope.org/mailman/listinfo/zope-announce > http://lists.zope.org/mailman/listinfo/zope-dev ) >

2 1

RE: [Zope] Errors
by Michel Pelletier 08 Dec '99

08 Dec '99

> -----Original Message----- > From: Daniel G. Rusch [mailto:drusch@globalcrossing.com] > Sent: Wednesday, December 08, 1999 2:17 PM > To: zope(a)zope.org > Subject: [Zope] Errors > > > What is the definition of (or where can I find one): > 1. KeyError > 2. AttributeError > 3. NameError You asked this identical question on 12/3/99 at 5:53 PM, and three people answered you quite accurately. -Michel

1 0

RE: [Zope] Acquire variable for ZCatalog indexing?
by Michel Pelletier 08 Dec '99

08 Dec '99

> -----Original Message----- > From: Art Hampton [mailto:arth@pacsg.css.mot.com] > Sent: Wednesday, December 08, 1999 1:23 PM > To: zope(a)zope.org > Subject: [Zope] Acquire variable for ZCatalog indexing? > > > I asked this previously, and got no replies. I'll give it one more > shot. > > When making a Z Catalog of a certain class of object, is it > possible to > acquire a variable from parent objects? Do the objects catalog themselves, or are you using Find? Acquisition is based on the context in which you access objects. If you you find, for example, you will be accessing the objects in the context of the catalog, not in the context of their parents. If objects catalog themselves when created/moved (ie, CatalogAwareness) they index themselves in the context of the request that created them, in the constructor of your ZClass, you can essentially manipulate in what context you want your object to be indexed. It's really kinda complicated actually. -Michel

2 1

Errors
by Daniel G. Rusch 08 Dec '99

08 Dec '99

What is the defintion of (or where can I find one): 1. KeyError 2. AttributeError 3. NameError Thanks, DR

1 0

Zope 2.1 does NOT work with NT Apache!
by Rik Hoekstra 08 Dec '99

08 Dec '99

> > I apologize for writing such a strong title, but I can't get Zope 2.1 to > work with NT Apache at all, and my previous message (more extensive) > describing my problems went unanswered. I realize installation issues > are an impopular topic. > > The binary distributation is insufficient, because no Zope.cgi is > included. Yes, you're right. You have to get a Zope.cgi from a non-binary distribution, I'm afraid. Why this is, I cannot tell. It worked under IIS. I have no experience with apache, but I do have with setting up Zope (2.0.1, but it should be the same) under IIS. Probably IIS is (even) more tricky, but the service and pcgi part should be the same > > The documention is insufficient, because it is unix-specific; it talks > about passing command line options to Zope, but Zope on NT can be ran as > a Service, and how to pass command line options then? > Um, you;ll have to hack the registry for that. Look in HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\ <Service Name>\Parameters\start To get pcgi working for IIS you have to add '-p X:\IIS\CGI-BIN\zope.pcgi' If I understood it rightly, the service is more or less responding to the same commandline options as z2.py (in fact it is just another way to start up z2.py). > There may be a bug, as a manually constructed Zope.cgi makes the Zope > installation 'spin'. You might want to use the utilities in the pcgi directory for checking your pcgi setup against pcgifile.py. Check out <zope-home>pcgi\win32\README.NT for (some) more information. > > See my previous mail to this list for more information. > > I can't get it to work at all, and I need to get it to work within the > week. Any help would be appreciated. I suppose Kamon Ayeva's draft on win32 installation could help you out at least for a part, or did you already consult that (even though it is mostly about IIS)? see http://lists.zope.org/pipermail/zdp/1999-October/000830.html good luck Rik

2 1