August 2002 - Zope - Zope lists

Memory Leaks
by eric.n.dunn＠bankofamerica.com 14 Aug '02

14 Aug '02

Using Zope 2.3.3 on Solaris 6 Have memory leaks. Where should I start looking? What is usually the primary cuprit? Eric Dunn

3 2

Re: [Zope] Zope on Windows, Apache on Linux ???
by Jaroslav Lukesh 14 Aug '02

14 Aug '02

| Odesílatel: Robert Segall <roseg(a)apsis.ch> | Apache does NOT need access to Zope, except via the HTTP port. This happens | to be true of any proxy (which Apache essentially is in this role). Set up | Zope on your NT machine the usual way and have the proxy access it as an HTTP | server. | | Regretably, Zope won't take advantage of multiple CPU's - that's a python | limitation, so you're wasting your NT machine. You may want to consider using | a smaller hardware setup there. If you arer really worried about performance | (which shouldn't be an issue at 10 hits/second unless you have extremely | complex stuff) you may want to look at multiple single-CPU machines running | against a ZEO backend. "NT" is our corporate CitrixMetaframe server. | I should think that the old 486 is more than enough for anything less than | 25-40 pages/second (in the proxy role). However, I would consider adding a | bit more RAM - 16MB is small. In any case make sure Linux runs without a | graphical desktop and with as few services enabled as possible (which you | should do anyway for a machine in the DMZ). 486 is our existing web server (apache) and squid proxy based on RedHat Linux 5.2. Connecting with SSH to that box makes lot of swapping and apache is very slow. | For a lightweight alternative to Apache in the proxy role try our Pound proxy | server at http://www.apsis.ch/pound - it does exactly what you need and was | designed for that, avoiding the overhead of a full-fledged web server. I need few virtual webs to run here. It is able to setup pound to use redirecting, for example www.domain.com --> 192.168.2.2:8080/www/www.domain.com/ another.domain.com --> 192.168.2.2:8080/www/another.domain.com/ www.domain2.com --> 192.168.2.2:8080/www/www.domain2.com/ All virtual domains are name-based virtuals at single external IP od 1.2.3.4 translated into DMZ to address 192.168.1.2 where is that small linux box. It is possible? Many thanks, J. Lukesh

1 0

truncation
by STACCINI PASCAL CHU Nice 14 Aug '02

14 Aug '02

hello I would like to know if there is a way to truncate a string while displaying it in a table using TAL, as the following code allows it in classic DTML code : <dtml-var mystring size=10 etc=", etc.">. actually the TAL code is : <td tal:content=mystring>dummy string</td> thank you for your answer

3 3

RE: [Zope] the quantity of users on site
by Schmidt, Allen J. 14 Aug '02

14 Aug '02

Just a side note on this...sort of off topic but hopefully helpful... Intro: We started using LivePerson on our site about 3 months ago. You know, the "Click Here For LiveHelp" folks. It's working great inside Zope. The code for the button goes in one method and the monitoring code goes in another, both in the root - then both are called from our footer on every page. The thing needs an account ID so it knows what operator console to direct the clicks. I just added a property to the root to handle this ID. Then other LiveHelp operators in different areas of the site will have their own buttons by having their own ID property defined on that folder. Works well. One big benefit of this system is that it is able to show how many users are looking at your site at that very moment. It can track exactly where users go in our site in real-time. Then when they request Help, we know what they have been looking at or searching for. Its a tad expensive but is working very well for us. Lots of possibilities. -Allen -----Original Message----- From: zope-admin(a)zope.org [mailto:zope-admin@zope.org]On Behalf Of Jens Vagelpohl Sent: Wednesday, August 14, 2002 8:22 AM To: Dmitry Matveev Cc: zope(a)zope.org Subject: Re: [Zope] the quantaty of users on site HTTP is a stateless protocol. that means you cannot directly tell how many people are on the site at any given time. most people overlook this very basic fact. that said, you can use workarounds like sessioning or hit counters to produce approximately correct data. read up on the sessioning facilities built into zope. jens On Wednesday, August 14, 2002, at 07:52 , Dmitry Matveev wrote: > Hello All!!! > > How the quantaty of users on site can be determined, at this very moment > in dynamic > (the moment of their departure???) > > -- > Best regards, www.uralfirm.ru > Dmitry Matveev mailto:matveev@uralfirm.ru > _______________________________________________ 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 )

1 0

Transient Object container/persistant add script?
by Michael Fischer v. Mollard 14 Aug '02

14 Aug '02

Hi! I ran into the following problem: I'm using the Zope session machinery and want to use add- and del-scripts with the *temp_folder <http://localhost:9673/temp_folder/manage_workspace>/session_data <http://localhost:9673/temp_folder/session_data/manage_workspace> ** Transient Object Container*. I added them and it works very fine, but after restarting Zope the add/del scripts are gone. How can I make them survive a reboot?

3 4

Trapping for empty fields
by McDonnell, Larry 14 Aug '02

14 Aug '02

Hi, I am using MS Access dbs for storage (company policy). I am starting to create online forms so the company can collect data. There were some dbs already created that they want access via a web browser. Some of the fields are empty and when I try to update the record using my ZSQL Method I get the following error message: Error Type: sql.error Error Value: ('S1000', -3702, "[Microsoft][ODBC Microsoft Access Driver] Field 'units_in_service.Hydration' cannot be a zero-length string.") How do I trap for this error? Examples and pointers will be greatly appreciated. Thanks in Advance. Larry McDonnell Proton Energy Systems 50 Inwood Rd. Rocky Hill, CT 06067 (860) 571-6533 ext. 531 Email:lmcdonnell@protonenergy.com www.protonenergy.com

3 2

Re: [Zope] Problems with list of object references
by Sven Rudolph 14 Aug '02

14 Aug '02

Hello Max, > Then I would probably create a list of the 100 most recent articles. You > can create a product that holds a single list "latestNews". This is what I'm currently doing, as you can see in my initial posting. > Then have a > method something like this to add the news items to: > > def addDateItem(self, itm): > news = self.latestNews > news.append(news.getPhysicalPath()) > if len(news) > 100: > news.remove(100) > self.latestNews = news > > def getLatestNews(self): > # converts paths to objects > return [self.getPhysicalRoot().restrictedTraverse(path) for path in > self.latestNews] I do exactly the same in my buffer product, but with different code. The problem with this method is, that I have to call self.getPhysicalRoot().restrictedTraverse(path) 100 times every time I call getLatestNews, and this is to slow :-( I was thinking about your suggestion to use ZCatalog. Isn't the attribute data_record_id_ some kind of object reference? It seems to be unique and static. Do you now how it is generated? Greetings Sven -- Sven Rudolph, Programmierer GermanMedicalServices.de GmbH Unter den Eichen 5, 65195 Wiesbaden Tel.: 06 11 / 97 46 25 2

2 2

Preffered Startup procedure
by Kevin Sullivan 14 Aug '02

14 Aug '02

Hi, I have been having problems keeping zope running for an extended period of time. I start zope with "./start &" from a terminal. My current instance of zope has been running fine for a few days, but it seems to be because my terminal is still active and available to write logging to. When I start zope and my terminal becomes disconnected zope seems to die fairly quickly. Is there a way to avoid this, or a way to automatically restart zope? Thanks... Kevin Sullivan, MCP, A+ Systems Administrator MIT Economics Dept ksull(a)economics.mit.edu 617-253-8895

5 4

python if/else problem
by Tom Germaine 14 Aug '02

14 Aug '02

Hello All: I am uploading a DTMLDocument using a form and python script, but I'm stumped as to why the if/else clause in the script does not work. The script: import string request = container.REQUEST ident=request.ident myfile=request.myfile mytitle=request.mytitle if ident[3:6]=='do_': client=context.do_ mytitle2 = mytitle + '- AVC, Dean\'s Office' if ident[6:9]=='adm': mytitle3=mytitle2 + ', Administrative' elif ident[6:9]=='hr_': mytitle3=mytitle2 + '-, Human Resources' if ident: client.manage_delObjects([ident]) client.manage_addDTMLDocument(id=ident,file=myfile,title=mytitle3) else: client.manage_addDTMLDocument(id=ident,file=myfile,title=mytitle3) ------ if file 'ident' exists, no problem, it is deleted and DTMLDocument is added if file 'ident' does not exist I get the error: Error Type: BadRequest Error Value: (filename) does not exist if I eliminate the 'if' part and comment out 'else' a new DTMLDocument is added ok What am I missing here? Tom Germaine

2 1

multiple issues: ie vs mozilla/manage_workspace/traceback/phantom directories
by Ricardo Anguiano 14 Aug '02

14 Aug '02

Greetings Zope Community, I have a few issues I hope someone can help me resolve. Here are the vital statistics. If I left anything out please let me know. Pointers to helpful documents are much appreciated. Zope 2.4.1 (binary release, python 2.1, linux2-x86), python 2.1.0, linux2 Zope 2.5.0 (binary release, python 2.1, linux2-x86), python 2.1.2, linux2 ie vs mozilla 1. This is happening on the 2.4.1 site. Occasionally I run across one of my pages which I can edit using IE and not mozilla. When I go to the /manage interface in IE, I get the folders and the tabs across the top like a regular manage_workspace. When I go to the /manage interface using mozilla, the page displays as it would when I normally load it, as if I had clicked on the view tab. What do I need to be able to edit all my pages using mozilla/netscape. I have only seen this on my Zope 2.4.1 site. manage_workspace 2. This is happening on the 2.5.0 site. Currently, I have a broken page. When I enter the URL into a browser I get a Zope traceback. If I try to edit the page using the /manage administrative interface I get a display of the traceback and not the regular manage_workspace. This is a similar problem to #1, but this is happening only on pages with traceback. How do I get the regular interface back when I have a problem like this? traceback 3. This is happening on the 2.5.0 site. What's going on in this traceback? Traceback (innermost last): File /usr/local/Zope/lib/python/ZPublisher/Publish.py, line 150, in publish_module File /usr/local/Zope/lib/python/ZPublisher/Publish.py, line 114, in publish File /usr/local/Zope/lib/python/Zope/__init__.py, line 158, in zpublisher_exception_hook (Object: credits) File /usr/local/Zope/lib/python/ZPublisher/Publish.py, line 98, in publish File /usr/local/Zope/lib/python/ZPublisher/mapply.py, line 88, in mapply (Object: binutils) File /usr/local/Zope/lib/python/ZPublisher/Publish.py, line 39, in call_object (Object: binutils) File /usr/local/Zope/lib/python/OFS/DTMLDocument.py, line 127, in __call__ (Object: binutils) File /usr/local/Zope/lib/python/DocumentTemplate/DT_String.py, line 473, in __call__ (Object: binutils) File /usr/local/Zope/lib/python/DocumentTemplate/DT_In.py, line 654, in renderwob (Object: binutils_contributors) File /usr/local/Zope/lib/python/ZODB/Connection.py, line 493, in setstate File /usr/local/Zope/lib/python/zLOG/__init__.py, line 129, in LOG File /usr/local/Zope/lib/python/ZLogger/ZLogger.py, line 17, in log_write File /usr/local/Zope/lib/python/ZLogger/stupidFileLogger.py, line 27, in __call__ File /usr/local/Zope/lib/python/ZLogger/stupidFileLogger.py, line 79, in stupid_log_write IOError: (see above) phantom directories 4. This is happening on the 2.5.0 site. I have a directory structure which looked like this: http://example.com/foo http://example.com/foo/f1 http://example.com/foo/f2 http://example.com/foo/f3 For some reason links pointing off into f1 and f3 from foo's index_html document come out as tracebacks, when followed. When I manage the foo directory, only f2 is shown in the left frame. If I enter the url manually (/manage/example/foo/f1/manage_main, manage_workspace displays the missing page just fine. The dtml appears ok.) Any ideas why the links are broken and why the directories are missing, but there when explicitly requested? Thanks, -- Ricardo Anguiano anguiano(a)codesourcery.com CodeSourcery, LLC http://www.codesourcery.com

2 2