- Zope-Dev - Zope lists

[commercial]zope python expertise required
by gnurpreet＠softhome.net 23 Feb '04

23 Feb '04

Software architect for Python/Zope/Plone development. Innovative (free/open source) software company in Rome, Italy, develops software for international political organizations. Proven and multi-year experience, management skills a plus. Relocation to Rome preferred, but also remote consulting and mixed are considered. Competitive compensation. jobs-mgmt(a)partecs.com Programmer for Python/Zope/Plone development. Innovative (free/open source) software company in Rome, Italy, develops software for international political organizations. Proven and multi-year experience, management skills a plus. Relocation to Rome preferred, but also remote consulting and mixed will be considered. Competitive compensation. jobs-tech(a)partecs.com ------------------------------------ GNUrpreet Singh Pre Sales Consultant ParTecs - Participatory Technologies www.partecs.com

1 0

Re: Zope 2.8
by Evan Simpson 20 Feb '04

20 Feb '04

Christian Theune wrote: > - Does a proposal for the post-traverse-hook have a chance for 2.8? (The > code is here completely working on 2.7, I only need to write tests.) Please, write tests and merge to the HEAD. I see no reason why this shouldn't go into 2.8. Cheers, Evan @ 4-am

1 0

RE: [Zope-dev] Proposal (was How to make Zope fail nicely under highload?)
by Bjorn Stabell 20 Feb '04

20 Feb '04

Toby wrote: > On Thursday 19 February 2004 11:18, Bjorn Stabell wrote: [...] > > The solution is a bit different from what we discussed on the list > > before; basically do away with the Medusa request queue > > altogether and use listen()'s backlog instead. > > Last week you were keen that medusa keep processing requests > so that it can > return a 503 error quickly. I guess your view on that requirement has > changed? This is still something that have to be tested to see what works best. Managing the work with the listen() backlog instead of Medusa's request queue has the benefit of being less resource intensive (Medusa doesn't have to do anything). If Zope is behind a HTTP accelerator/surrogate or load balancer, the surrogate's connect() should fail if the listen queue is full, and it may return 503 anyways (not tested). In the end, the most important point is to put an upper bound on the request backlog (and therefore the response time). Bye, -- Bjorn

1 0

Zope 2.8
by Christian Theune 19 Feb '04

19 Feb '04

Hi, I'm posting this again is i didn't got too much feedback on the first time (a single +1) with some meta questions: - Is the Zope 2.8 roadmap already fixed? - Does a proposal for the post-traverse-hook have a chance for 2.8? (The code is here completely working on 2.7, I only need to write tests.) Cheers, Christian -- Christian Theune, gocept gmbh & co. kg http://www.gocept.com - ct(a)gocept.com fon: 03496 3099112 fax: 03496 3099118 mobile: 0179 7808366

1 0

Importing Custom Python Classes (fwd)
by D. Bickle 19 Feb '04

19 Feb '04

How do you import your own python class modules into Zope? Here is what I've done: In lib/python/Products I created a directory called MyScriptModules which contains the following: First I import the required Modules: from Product.PythonScripts.Utility import allow_module, allow_class from AccessControl import ModuleSecurityInfo, ClassSecurityInfo from Globals import InitializeClass Then I add the security declarations: from ClassModule.ClassModule import Class1 allow_module('ClassModule.ClassModule') allow_class(Class1) I created a ClassModule directory in the Products directory and placed ClassModule.py class module in it. I then rebooted the Zope server. Now everytime I try this: from ClassModule.ClassModule import Class1 I get this error: import of ClassModule.ClassModule is unauthorized. I have read somewhere that in order to create new Products in Zope you have to meet some Zope API requirements like assigning certain class attributes. Is this true? If so where can I find some documentation? Is there a better way to import custom Python class modules than the way described above? Thanks in advance.

1 0

The number of threads in z2.py file.
by asd asdda 19 Feb '04

19 Feb '04

i'm using Zope-2.6.3, Red Hat-8.0, CMF-1.4.2, and CMFPlone-2.0-RC3. If the number of threads=4, the zope can start but cannot load into ZMI. After i change the number of threads=1, it can load into ZMI. Why? __________________________________ Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online. http://taxes.yahoo.com/filing.html

2 1

some breakage on the head
by Jeremy Hylton 19 Feb '04

19 Feb '04

I was doing some branch merging and apparently got mislead by stray .pyc or .so files being left behind after .py and .c files were removed. So there's some broken code on the head; I'll fix ASAP. Jeremy

2 1

[ZConfig] "import" inhomogenous
by Dieter Maurer 19 Feb '04

19 Feb '04

The "import" element of ZConfig schemas is inhomogenous: * with a "package" attribute it imports a component from a Python package * with an "src" attribute it imports a schema from an URL It would be better if the type of the imported object (schema or component) were orthogonal to the location from where the object is found (in a package or via an URL). A use case is the import of the section types defined in "Zope.Startup.zopeschema.xml" for other ZEO applications that require access to the same ZODB database(s). It is far more convenient to access this schema via "package"+"file" than an URL. The attached patch adds an additional "type" attribute to the "import" element. It can take values "schema" or "component" and determines the type of object to be imported. Its default value ("") is handled to provide backward compatibility. -- Dieter

2 2

Proposal (was How to make Zope fail nicely under high load?)
by Bjorn Stabell 19 Feb '04

19 Feb '04

I wrote up a simple proposal for how to change Zope to work better under heavy load: http://zope.org/Members/bjorn/proposals/MakeZopeFailNicelyUnderHeavyLoad The solution is a bit different from what we discussed on the list before; basically do away with the Medusa request queue altogether and use listen()'s backlog instead. Comments are welcome. Regards, -- Bjorn

3 2

Image ranges
by Florent Guillaume 19 Feb '04

19 Feb '04

Is there a reason why OFS.Image.Image does not __implement__ HTTPRangeInterface like OFS.Image.File does ? As far as I can see there's no reason it shouldn't, as their index_html method (and HEAD inherited from webdav.Resource) are the same. Florent -- Florent Guillaume, Nuxeo (Paris, France) +33 1 40 33 79 87 http://nuxeo.com mailto:fg@nuxeo.com

2 1