May 2002 - Zope - Zope lists

URGENT -Version Blow-Up
by Dan Shafer 13 May '02

13 May '02

I have managed to get myself into a gnarly position from which I cannot find an escape route. The site I'm building is currently absolutely unusable as a result. I was making extensive changes to a page, so I created a version (called "adding scripts" -- notice the space, which I think may be a culprit here). I started using the version, made the changes to the page in question, saved changes, tested the page and, after a few more tweaks, got it where I wanted it. So I went to the version page figuring I'd commit the changes. It said there were no unsaved changes. I stopped using the version and when I went back to the ZMI, the changed page had a version lock on it and there was no unlocked/original version of the page showing up. That's when I noticed that even though I had quit using the version in question, the page still showed that I was using it.That's also when I noticed that the top of the page that showed what version I was using called the version simply "adding" rather than "adding scripts." After several more start using - stop using sequences, I could not get out of this morass. I restarted the Zope server. No effect. I quit using the version and deleted the version object. I know, bad move, but I did not receive a warning from Zope about this. It just happily deleted the version, an action which I later found out was undoable. So now Zope thinks I am working in a non-existent version and it won't let me use the site past a certain point (even though the process I'm testing with doesn't even use the changed page). I get a version lock error. This is clearly a Catch-22. Is there any way out of this? This is the first thing I've run into in the ZMI that seems to me just not to work very well or to be documented clearly or something. I've used versions in the past successfully, and this just has me bewildered. And my client very angry. HELP!!!!!!!!!!!

2 2

ZEO Client with a custom INSTANCE_HOME
by Igor Stroh 13 May '02

13 May '02

Hi there, I'm trying to setup a "discless" ZEO client. The INSTANCE_HOME is pointing to '/var/run/zope'[1] (/var is mounted on a ramdisk). If I try to start zope with `/usr/sbin/zopectl -I /var/run/zope/ start -P 8000 -u www-data -D` I get the errors in [2]. The custom_zodb.py looks like [3]. Afaik all file permissions/modes are set correctly, what else might be the reason? ZEO is running smooth, everything works if I try to run Zope with a default INSTANCE_HOME. If I remove the old INSTANCE_HOME and supply a custom one I get those erros in [1]... some more info: Zope version: 2.5.1-1 (from debian woody) python version: 2.1.3 Thnx in advance, Igor [1]: NFSROOT@test1:~# ls -l /var/run/zope/ /var/run/zope/var total 12 -rw-r--r-- 1 www-data www-data 39 May 13 17:54 access -rw-r--r-- 1 www-data www-data 90 May 13 17:54 custom_zodb.py -rw-r--r-- 1 root root 245 May 13 18:38 custom_zodb.pyc drwxr-xr-x 2 www-data www-data 0 May 13 19:02 var /var/run/zope/var: total 0 -rw-r--r-- 1 www-data www-data 0 May 13 18:33 Z2.log [2]: NFSROOT@test1:~# /etc/init.d/zope start Starting Zope...------ 2002-05-13T17:19:19 INFO(0) client Trying to connect to server: ('<some IP>', 7700) ------ 2002-05-13T17:19:19 INFO(0) ClientStorage Connected to storage ------ 2002-05-13T17:19:19 INFO(0) ZODB Opening database for mounting: '142508672_1021297398.478297' ------ 2002-05-13T17:19:19 INFO(0) ZODB Mounted database '142508672_1021297398.478297' at /temp_folder ------ 2002-05-13T17:19:19 INFO(0) Zope Default Object Creation /usr/lib/zope/lib/python/../../import/Examples.zexp examples import file could not be found. ------ 2002-05-13T17:19:19 INFO(0) ZServer HTTP server started at Mon May 13 19:19:19 2002 Hostname: test1.domain.tld Port: 8080 ------ 2002-05-13T17:19:19 INFO(0) ZServer FTP server started at Mon May 13 19:19:19 2002 Hostname: test1 Port: 8021 ------ 2002-05-13T17:19:19 ERROR(200) ZServer Cannot write PID file. ------ 2002-05-13T17:19:19 PANIC(300) z2 Startup exception Traceback (innermost last): File /usr/sbin/zope-z2, line 627, in ? File /usr/lib/zope/ZServer/PCGIServer.py, line 275, in __init__ File /usr/lib/zope/ZServer/medusa/asyncore.py, line 306, in bind error: (2, 'No such file or directory') [3]: NFSROOT@fs0:/var/run/zope# cat custom_zodb.py import ZEO.ClientStorage Storage=ZEO.ClientStorage.ClientStorage(('<some IP>', 7700)) -- ScanPlus GmbH NOC Ulm - Germany - Griesbadgasse 7-13 - D 89073 Ulm TEL +49 731 920 13 100 - FAX +49 731 920 13 290 eMail: support(a)scan-plus.de Amtsgericht Ulm - HRB3220 - Geschaeftsfuehrer: Juergen Hoermann

2 2

two submit buttons/ one method
by John Hunter 13 May '02

13 May '02

I have a form with two submit buttons. Rather than call two different methods as in <form action="" method="post"> ... <input type="submit" name="some_func:method" value="Button 1"> <input type="submit" name="some_other_func:method" value="Button 2"> </form> I would like each button to call the same function, but with a different value for a variable. Something like: <form action="" method="post"> ... <input type="hidden" name="somevar" value="someval"> <input type="submit" name="some_func:method" value="Button 1"> <input type="hidden" name="somevar" value="someotherval"> <input type="submit" name="some_func:method" value="Button 2"> </form> But this doesn't work. Any suggestions? Thanks, John Hunter

7 11

RE: [Zope] PIL Question
by Ross Lazarus 13 May '02

13 May '02

the img tag expects the server to deliver an existing image file from somewhere. What you need to do is a little more complex (!). You COULD write the image into a zope object (bad idea for lots of reasons - eg zodb bloat) or to a localFS (better but still sucky) and deliver it from there using an img tag. AFAIK, this is the ONLY way to embed your image into a page with text and stuff. I use PIL to make images and either email them or return them for the browser to display or to save. Email is good if the generation takes a long time obviously. To get the browser to display a returned image binary, you must set the mime type in the http headers to image/jpeg using or whatever and returning the binary image - the browser will display it correctly. If you want the image to be saveable, it gets worse. IE and netscape are both broken if you try the file-disposition header, it doesn't work as advertised. There's a work around which I use - namely setting the mime type to application/octet-stream > > I'm trying to create an external method that will take variables that exist in a ZSQL call and will output an image > > based on that: > > import Image, ImageDraw, StringIO, os.path > > def PILtest(): > image = Image.new('RGB',[25,25]) > ## IMage Stuff Happens here! > return image > > Where this would occur in a DTML-Method of: > > <dtml-in SQL_COMMAND> > <img src="PILtest">  > </dtml-in> > > But I can't get PILtest to output the image itself. Save() doesn't seem to be of help directly. Anyone have any suggestions? > > Jason -- Ross Lazarus Visiting Associate Professor in Medicine, Channing Labs, 181 Longwood Ave., Boston MA 02115, USA. Fax:+617 525 0958, Office: +617 525 2730

3 4

ZSQL questions
by McDonnell, Larry 13 May '02

13 May '02

Hi, First I want to thank the people who helped me get this populating form thing working (Colin Fox, Tom Passin, and Jon Erickson). I have a couple questions on how ZSQL methods and their behavior. From a form I created, I input the data and hit the submit button. Under superuser the data is added to the db but if I am not superuser, the ZSQL method the you would see in test mode appears where you would need to file in the data again. Why does it work under superuser and no where else? Is it a permission thing? Also on the same subject, how do I suppress the message telling me "no data matched" while it is adding a new record to the db? Thanks. Larry McDonnell Proton Energy Systems 50 Inwood Rd. Rocky Hill, CT 06067 (860) 571-6533 ext. 531 Email:lmcdonnell@protonenergy.com www.protonenergy.com

2 1

Virtualhost monster problems
by Wayne Pascoe 13 May '02

13 May '02

Hi all, I've seen similar things discussed on the list, but not quite this question. I have setup the virtualhostmonster in my root folder. In my apache config, I have the following: <VirtualHost *> RewriteEngine On RewriteCond %{HTTP_HOST} ^.*$ RewriteRule ^/(.*) http://127.0.0.1:8080/VirtualHostBase/http/%{HTTP_HOST}:80/$1 [L,P] <IfModule mod_proxy.c> ProxyRequests On ProxyVia On </IfModule> </Virtualhost> The rewriterule has wrapped to two lines in this e-mail, but in my config it is on one. I have got an A record, zaphod-content.mydomain.com pointed at the IP Address that the Apache httpd server is listening on. In Zope, I have added a folder named zaphod-content.mydomain.com off of root. This has an index_html document in it. When I connect to http://zaphod-content.mydomain.com I get the normal localhost:8080/ page. If I go to http://zaphod-content.mydomain.com/zaphod-content.mydomain.com I get what I expect to. zaphod-content is not the primary name of this machine. The primary name is zaphod. Going to this also gives me the content I would expect to see at localhost:8080/ Any ideas why the contents of zaphod-content.mydomain.com are not being served as / when I request it as zaphod-content.mydomain.com ? TIA, -- Wayne Pascoe <wayne.pascoe(a)ehsbrann.com> Give me enough medals, and I'll win any war. - Napolean

2 3

Creating soft links within ZODB ?
by Oliver Pabst 13 May '02

13 May '02

Hi, I have following Tree: A |-- B |-- | -- C |-- D |-- | -- E is it passible to create E as being a "link" of C, i.e. a copy of all entries in C. So that if a property "p1" is set in B and in D with different values, a call of an object in C (and in its linked copy in E) will see two different values of p1 (respectively as in B and in D) -- --- --- Oliver Pabst .-------------------------. mailto:olpa@sybcom.de : : phone :+49 681 56600600 : project department : SYBCOM GmbH fax :+49 681 56600660 : : http://www.sybcom.de

2 1

Virtual links...
by Josef Meile 13 May '02

13 May '02

Hi! Recently I uploaded a hole website on a folder inside my zopesite. Lets say the old link is www.oldsite.com and the new location is www.myzope.com/folder1/folder2/website/. I just would like to know how can I redirect all the requests from www.oldsite.com to www.myzope.com/folder1/folder2/website/ without using Apache. Thanks in advanced, Josef.

1 0

TinyTablesPlus - key columns
by Roger Fisher 13 May '02

13 May '02

Hello - I am having difficulty setting up an integer "key" (unique) column: Task:int is accepted (integer, not a key) Task* is accepted (string, key) Task:int* and task*:int do not work as expected - the "*" is apparently ignored after the change/save.

1 0

Implementing Backlinks in ZOPE
by Scott Wilson 13 May '02

13 May '02

How to do backlinks in ZOPE =========================== It may not be the nicest code ever, but here is something I wrote to show backwards links from articles published using Zope. If you get it working (or not :-D) and can think of improvements, let me know... BTW, the whole thing is inspired by backlinks code at http://www.downes.ca/referrers.htm and at http://www.disenchanted.com/dis/linkback.html. I couldn't find anything similar for Zope, hence this lot! Enjoy! - Scott =========================== Step 1: Create the Referrer zClass ----------------------------------- This is a little zClass for holding the referral data. Just create a blank zClass, call it Referrer, and add a propertysheet called 'common' with three attributes (referrer, title, count). Count should be an int with default value of 1, the others are strings. Step 2: Create the referrer script to be called from your index_html method ------------------------------------------------------------------------ --- This takes a look at the Request object, and decides whether or not to create a Referrer. Note that it does this inside the context it is being called from, so your calling object must be an Object Manager! Note that you need to change the Product name here too. Call this Python script from the standard "view" method of the object (usually index-_html): ################## ## getReferrals ## ################## from Products.PythonScripts.standard import html_quote import string request = container.REQUEST RESPONSE = request.RESPONSE # # Get referrer # sReferrer = str(request.get_header('HTTP_REFERER')) # # check against the kill list. Definite room for improvement here # if sReferrer <> '' and string.count(sReferrer,'YOUR-DOMAIN-HERE')==0 and string.count(sReferrer,'frontier.userland')==0 and string.count(sReferrer,'google')==0 and sReferrer <> 'None': # # make the referral, or update count for existing referrals # bExists=0 for oReferrer in context.objectValues('Referrer'): if oReferrer.referrer == sReferrer: oReferrer.propertysheets.common.manage_changeProperties(count = oReferrer.count+1) bExists=1 if bExists==0: id = str(context.ZopeTime().strftime('%Y%m%d%H%M%S')) oNewReferrer = context.manage_addProduct['YOUR-PRODUCT-NAME-HERE'].Referrer.createInObj ectManager(id,request) oNewReferrer.propertysheets.common.manage_changeProperties(referrer = sReferrer) ######### ## end ## ######### Step 3: Create a Robot that substitutes real page titles for the URLs --------------------------------------------------------------------- Add this external method to your Zope installation, and call it periodically on all your new Referrer instances (using cron, Scheduler or whatever). NB, you may want to set Python's Proxy values before doing this. This script fetches the URL from each referrer, and extracts any Title tags. Put it into the Extensions folder. ##################### ## getPageTitle.py ## ##################### ## Fetches the URL provided, and returns the title of the page ## If the page is unavailable, or untitled, ## then returns the URL as the title ## import urllib import re import string def getPageTitle(sUrl): try: oWeb = urllib.urlopen(sUrl).read() try: oRe = re.compile("<title>(.*?)</title>",re.I|re.S) oMatch = oRe.search(oWeb) if oMatch <> None: sTitle = string.strip(oMatch.group(1)) else: sTitle = sUrl except: sTitle = sUrl if sTitle == '' or sTitle=='None': sTitle = sUrl except: sTitle = sUrl return sTitle ######### ## end ## ######### On my site the updater script I use to call the external method looks like: for oItem in context.content.objectValues('ContentItem'): for oReferrer in oItem.objectValues('Referrer'): sTitle = context.getPageTitle(oReferrer.referrer) oReferrer.propertysheets.common.manage_changeProperties(title = sTitle) (Where "getPageTitle" is the name of the external method.) ============= Scott Wilson Assistant Director Centre For Educational Technology Interoperability Standards (CETIS) +44 (0)1248 383645 web: http://www.cetis.ac.uk newsfeed: http://www.cetis.ac.uk/news.xml

1 0