- Zope-Dev - Zope lists

Zope Performance - Windows 2000 .. Newbie
by Bryan Baszczewski 05 Mar '01

05 Mar '01

I am getting some serious Zope lag time (15-20 seconds) for simple HTML pages when also trying to hit the same port with data requests. The data requests are <= 50 records of text data and SQL Server is optimized (meaning the querys dont take that long, just the connection to Zope). We are using port 8080 on Windows 2000 and was wondering if there is some kind of queue or prioritizing of Zope requests? If so, how can I work around this? Thanks in advance!

1 0

Apology! (was: [Zope] Problems with DTML-IN patch!!!)
by Dieter Maurer 05 Mar '01

05 Mar '01

John Morton writes: > On Fri, 2 Mar 2001 20:55:40 +0100 (CET) Dieter Maurer <dieter(a)handshake.de> > wrote: > > > Chris Withers writes: > > > John Morton wrote: > > > > > > > > It doesn't appear to be defined in DT_In.py in 2.2.5, but it is ment > ion in > > > > it's doc string, and is available in the dtml-in namespace. I never > got > > > > around to tracking down where it was added in. > > "DT_InSV.py" (for DTML In Sequence Variables) defines it, if > > it is defined (when I recently looked through the sources, it has > > not been). > > I can't find them in either DT_In.py or DT_InSV.py in 2.2.5, yet they do > actually work fine... You are right! "{next,previous}-sequence-start-number" and friends are still there. Nobody made the stupidity to remove them. Please (Brian) accept my apology for my quite rude accusation. I was just not smart enough to understand the ingenious code in "DT_In" and "DT_InSV". Let's review the situation: "DT_In" only defines "next-sequence-start-index". But what is pushed onto the DTML namespace is in fact a "DT_InSV.sequence_variables" object. Its "__getitem__" is used to look up sequence variables. The "__getitme__" code is quite sophisticated: If the key is not defined, it splits it at the last '-' into prefix and suffix (in our case, suffix is "number"). If the class "sequence_variables" has a method whose name is given by suffix (this is the case for "number" and many of its friends), then the corresponding "index" key is looked up and the method applied to it. This way, "next-sequence-start-number" and friends come into existence, without that one easily recognized how this happens. Sorry for the false accusations! Dieter

1 0

RE: [Zope-dev] is there a hook for before the transaction is comm itted
by Tim McLaughlin 05 Mar '01

05 Mar '01

I want to be able to raise an event from the property sheets to the product if a property changes. I'm calling it manage_afterChange. I want it only to fire once per transaction so that multiple changes are aggregated. Here's the declaration I have in mind. manage_afterChange(oldItems, newItems) oldItems: dict of id - values before modifications newItems: dict of new values at end of transaction This would allow an elegant "reindex" or notification system for objects. It would also allow for validation (ie. raise error if you don't like newItems). Anyway, that's it. I'm halfway there, just can't get a method to be called on my object _just_ before commit. Cheers, Tim ps. I'm not so sure that __getstate__ will do this for me, but that's probably cause I have no idea what __getstate__ does. It just doesn't sound right ;-) -----Original Message----- From: Toby Dickenson [mailto:tdickenson@devmail.geminidataloggers.co.uk] Sent: Monday, March 05, 2001 12:08 PM To: Tim McLaughlin Cc: 'zope-dev(a)zope.org' Subject: Re: [Zope-dev] is there a hook for before the transaction is committed On Mon, 5 Mar 2001 10:56:38 -0500, Tim McLaughlin <tim(a)BCSWebservices.net> wrote: >Is there a hook for before the transaction is committed for objects which >subclass Persistent? __getstate__ ? But why would you want that? Toby Dickenson tdickenson(a)geminidataloggers.com

2 1

Re: [Zope-dev] DB connection pooling
by Dieter Maurer 05 Mar '01

05 Mar '01

Erik Stephens writes: > I've been working on an implementation of connection pooling for the > MySQLDA. Basically, Zope connections will be pooled by connection > string to reduce the number of unecessary connections to MySQL. Isn't that trivial with Zope (thanks to acquisition)? Just instantiate a single MySQL Database Connection per connect string (sufficiently high above in the site hierarchy). Use it below through acquisition. Dieter

2 1

How is memory shared between Zope threads?
by Erik Stephens 05 Mar '01

05 Mar '01

Hello all, I have noticed incosistencies between Zope threads when objects change. It appears that some threads see the new image of the object while others see the old image. Can anyone enlighten me on how Zope ensures that all threads are seeing the same images? Or at least point me to the modules in the source that handle these matters? Naively, I can think of 2 approaches: a shared memory approach or a separate memory per thread with notifications. Zope behaves like each thread as its own copy of in-memory pieces and sometimes these per/thread memory blocks do not get updated. System info: Zope 2.1.6, Redhat 6.2, PC arch Thanks, Erik

3 4

Weird Re-Indexing Errors after move to Zope 2.3.1b1
by Chris Withers 05 Mar '01

05 Mar '01

Hi, Since moving one of our servers from Zope 2.2.4 to 2.3.1b1, we've been frequently getting this weird error when trying to do something that requires an object to be re-indexed in the Catalog: Error Type: AttributeError Error Value: 'tuple' object has no attribute 'append' Traceback: Traceback (innermost last): <snip> File /usr/local/zope/cases/Products/aProduct/aProduct.py, line xxx, in index_something (Object: Traversable) File /usr/local/zope/Zope-2.3.1b1-src/lib/python/Products/ZCatalog/ZCatalog.py, line 411, in catalog_object (Object: Traversable) File /usr/local/zope/Zope-2.3.1b1-src/lib/python/Products/ZCatalog/Catalog.py, line 382, in catalogObject File /usr/local/zope/current/lib/python/SearchIndex/UnTextIndex.py, line 394, in index_object File /usr/local/zope/current/lib/python/SearchIndex/UnTextIndex.py, line 305, in insertReverseIndexEntry AttributeError: 'tuple' object has no attribute 'append' Any ideas what's going on here or how to fix it? cheers, Chris

3 6

Re: [Zope-dev] ZCatalog hackery
by Dieter Maurer 04 Mar '01

04 Mar '01

R. David Murray writes: > ... > The issue is that parts of the database get updated periodically from an > external source. That is, the author and book tables get replaced > wholesale. But the Book and Author objects are cataloged using > ZCatalog, doing a full text index on a combination of fields from > the external database and the html chunks. As far as I can see > at the moment, this means that every object has to get uncataloged > and recataloged, meaning every time the database update happens > the zodb inflates considerably, even though very little data has > actually changed. Maybe, you can determine whether something relevant changed and call "catalog_object" only in this case. Dieter

1 0

ZPatterns, DynPersist.dll and Zope 2.3.0
by morten＠esol.no 04 Mar '01

04 Mar '01

Hi guys, I've got a problem making a version of the DynPersist.dll file work on windows. The message when trying to load the DynPersist module says (paraphrasing) "A unit attached to the system doesn't work". Anyone else experienced this? Also, I read that users of > Zope 2.2.x could skip this module, is there any quick hack to fix this? Thanks, Morten

2 2

Site Access Rewrite Rule documentation?
by R. David Murray 04 Mar '01

04 Mar '01

Is there any real documentation for Site Access rewrite rules anywhere? There isn't anything in the help system of 2.3.1b1, and the docs under SiteAccess on zope.org have a couple of examples but no technical explanation of how the parts of the example interact with Zope. My goal is pretty simple. I have a Specialist under /store/marketItems. I want my content managers to be able to use URLs like: /Books/someisbnnumber to refer to the object that is really at: /store/marketItems/someisbnnumber I could use a redirect, but why suffer the transaction overhead? I figured I could have a folder Books with an access rule like this: stack = context.REQUEST['TraversalRequestNameStack'] if stack[0]=='Books' and (len(stack)==1 or stack[1][:6]=='manage'): stack[0:1] = ['store','marketItems'] I figured I needed the manage thing so that I could edit the access rule. When I try to use this, no rewriting appears to happen. So I changed it to: if not stack[0][:6]=='manage': stack[0:1] = ['store','marketItems'] This didn't do anything for the URL '/Books', and gave me a notfound error on /Books/marketItems if I put something after the /Books. Changing it to stack[1:1] gives me a list index out of range. I can't figure out where *that* would be coming from. So, clearly I don't have a clue what stack is supposed to contain or what changing it does. Can anyone point me to docs or give me some help? You can also tell me, "that's stupid, just solve your problem this way"... --RDM

2 2

PlugIns: updated dtml files for ZPatterns+PlugIns
by Steve Alexander 04 Mar '01

04 Mar '01

Hi folks, I've updated the dtml files for Zope 2.3 for the PlugIns product. You can find them here: http://www.cat-box.net/steve/main.dtml http://www.cat-box.net/steve/PlugInGroup.dtml The update is to fix some badly formed HTML that makes the Zope Management Interface unusable with recent Mozilla builds. The same problem affects Zope 2.3 from public CVS. See Collector report 2020: http://classic.zope.org:8080/Collector/2020/view -- Steve Alexander Software Engineer Cat-Box limited

1 0