September 1999 - Zope

Re: [Zope] ANNOUNCE: beehive releases new calendar product
by Anthony Baxter 30 Sep '99

30 Sep '99

>>> Mark Pratt wrote > "The beehive Calendar is a calendar for the > intranet. It can be used to make announcements > and show the goals of the team members and the > respective deadlines. Additionally, each team > member can manage their daily planer." Hm - it looks like it was meant to be run on Windows as it refers to a filename (TeamGoals.dtml) in a case-insensitive manner. Oops. :( It does this in beehiveCal.py in 2 places. For me, the icon image is also broken, I haven't worked out why yet. Anthony

3 2

javascript problems with zope/apache
by Fredrick Rybarczyk 30 Sep '99

30 Sep '99

Hi all. my box: zserver 8080 , apache 80 I need to pass information from a window, opened with javascript, back to a hidden input tag in within a named form in the original window. The new window is opened with a url to a cgiscript on my apacheserver. The script generates a page where the user can create a CERIF classification on a webresource. So, when the user finishes, avalue is returned back to the original form. This works like charm on our other pages. When I try to do the same from Zope, the window opens but the input tag don't get updated. Both servers run from the same box. I tried to fib with document.domain and got some interesting results. If I set " document.domain = 'our.domain'; ", Zope won't open the window. I've also tried " document.domain = 'our.domain:8080'; " which resulted in that zope opened the CERIF window but the input tag still didn't get updated. Should different portnumbers break the "Same origin principle" in the javascript security model? Is there a way to get around this problem with javascript without having the client to change his browser configuration? Do I have to set up apache to handle my zopepages ? :-( Could anyone please help me out before I dig myself deeper into this pit. -Fredrick -- _____________________________________________________ Fredrick Rybarczyk voice +46 46 222 9369 Research Engineer fax +46 46 222 3682 LUB NetLab http://www.lub.lu.se/~fredrick/

1 0

BMP on Zope list; Please: NO!
by Robert O'Connor 30 Sep '99

30 Sep '99

Hi Sunit Joshi, Please do not submit such large attachments to the ZOPE list. Your message was 1.7 Megabytes! It almost exceeded my mail provider's 2 Meg Limit and besides, when sent on the list to say about 600 people this makes the mail server send over 1 GIGAbyte and slow down considerably. Better to place your gif on a web page somewhere with a reference in the email for those interested to go take a look. PS: Mail administrator: Isn't there a limit to the size of an attachment allowed on the list? -Bob OConnor bob(a)rocnet.com > --__--__-- > > Message: 2 > From: "Joshi, Sunit" <sjoshi(a)ingr.com> > To: zope(a)zope.org > Subject: [Zope] Help! Publishing Methods > Date: Wed, 29 Sep 1999 11:11:37 -0500 >

2 1

how to access files outside the database
by zope 30 Sep '99

30 Sep '99

How can I access files that do not lie inside the Zope database. I need that feature because I have to write an image database where the images are stored in a common directory and only the filenames are stored in the RDBMS. If I address an image wit the <IMG...> tag zope searches in its internal object hierarchy. Is there any possibility? Thanks in advance Philipp Auersperg

2 1

Any alternative to &arg in URL?
by John Hermes 30 Sep '99

30 Sep '99

I am working on a site using both POST and GET forms and I use URL args frequently in the target: action="subfolder/index_html?arg1=spam&arg2=eggs&. .." Using the '&' character in URLs is not legal HTML (4.0-strict). Does Zope recognize other arg separators besides '&'? I don't want to give up the way Zope parses and places the args into the namespace for me. Any thoughts would be much appreciated! John Hermes jhermes(a)infoglobe.com

3 2

Re: [Zope] Formatted Documents
by Amos Latteier 30 Sep '99

30 Sep '99

At 10:58 AM 9/30/99 +1000, Grant Street wrote: >I noticed that when you log in to the Zope site you can add >a "Formatted Document". >Is this product available ? I looked in bothe Zope sites and I'm feeling a >bit confused? >I've been looking for somthing like this for a while. Formatted Document is not currently available. You could whip one up pretty easily, using Z Classes, just like I did. Basically define a couple properties, say 'content' of type 'text', and 'format' of type 'string'. One property will hold the content of the document, and one will specify the formatting. Then in the index_html method of the ZClass do something like this, <dtml-var standard_html_header> <dtml-if "format=='structured-text'"> <dtml-var content fmt="structured-text"> </dtml-if> <dtml-if "format=='html'"> <dtml-var content> </dtml-if> <dtml-var standard_html_footer> You can add any other formatting options you want. The way I allowed for acquired formatting methods is like so, <dtml-if "format=='acquired'> <dtml-var "_.getitem(format_method,0)(this(),REQUEST)"> </dtml-if> This says to get the object whose name is given by the variable 'format_method' then call it with two arguments: this(), which is the formatted document, and the REQUEST. BTW, this is a rather low performance solution to a multiply formattable document. You should look at developing a full Zope Product that is a lot smarter if you care about high performance formatting. Good luck! -Amos

1 0

ZPublishing
by Chui.Tey＠env.qld.gov.au 30 Sep '99

30 Sep '99

I am trying to send some dynamically generated binary data to the output buffer. How do I do it in python with Zope? As I understand it HTMLFile(xxxx) returns a dtml document object. I could dump the data to file and then try to send it using the binary equivalent of the HTMLFile class. Is there such an animal? I noticed there is the ImageFile class, but the guess_encoding module does not list the content type I need to publish. Is it a good idea to add mine there or can I override this? Isn't there anyway I could just write to the output buffer? Thanks.

2 1

RE: [Zope] Allaire Spectra
by chas 30 Sep '99

30 Sep '99

>That said, are there other particular lessons that you think we >should learn from CF/Spectra? Let me go off on a tangent for a moment : I got up to speed on CF instantaneously because I knew what it was supposed to do (and what I intended to do) before installing the software. So, perhaps that's part of the reason people have so much trouble. Zope is a new 'product category' for many. Here's part of an email somebody sent me privately after I mentioned zope on another mailing list : [snip] >Can you explain to me what Zope does for you and what the advantage to using >it is? I just don't get it. I've read all of the docs, the web site, installed >it and poked at the admin interface, but I just don't understand why it's >better to use zope than to just use apache and a few hand-hacked HTML files. >Perhaps my difficulty arises from the fact that I don't have any particular >problem to solve with zope. I'm just wondering that it is good for. [/snip] I have 6 other emails like this - people wondering why they just don't get it, whether it's worth sticking it out or whom have just given up, and feel pathetic about it. And I did exactly the same 6 months ago. And therein lies a problem. When I installed CF I knew I wanted to query and update relational databases. And I now know exactly what I want to do with Zope.... BUT most people don't. How to solve this ? Well, this reminded me a lot of the first time I got on Domino/Notes... I was like "What a waste of space. What's the point ? I just don't get it". It was only upon seeing professional apps built by people who truly knew what they were doing that I finally went 'aaaah' and it sunk in. But that took months. (I still maintain it's a waste of space though <g>) OK, so we've got the products that people are uploading and Kevin's KmNews has been a lifesaver. But that still doesn't make a complete application - a complete application for all of us must be a finished website. And so I guess we're just paying the price for being the trailblazers. Maybe a year from now, with decent docs and lots of different apps to check out, the learning curve will be greatly reduced. Also judge your target audience (or at least their expectations). Going back to the Notes/Domino analogy : in a Notes environment, you have the application designers and the system administrators and the end users.. and each has it's own software (the Notes Designer, Notes Administrator and the standard Notes client respectively). At the moment we seem to be throwing everybody into the same funnel which is causing some grief... splitting the lists may well help in that respect. This is one place where you can see Allaire were really good - they understood their target audience and wrote documentation accordingly : assuming no knowledge, step-by-step, foolproof. By comparison, have you noticed the number of people who post to this list saying "I'm sorry, I feel incredibly stupid, I just don't get it". Very rarely are these unintelligent people... zope really has an amazing way of making intelligent people feel dim. (whereas CF could be said to cause the reverse <g>). This is a real shame. Most common reason is that the docs don't contain EVERYTHING. We'll be trying to solve a problem and can't find a solution. We post a question to the list. Somebody tells us the solution. But the solution can't be found in the docs anywhere. It's trying to do a jigsaw puzzle but failing, throwing your hands in the air and saying "i give in, i can't do it" then your friend pulls a missing piece out of his pocket and saying "is this what you're looking for ?" Unfortunately, I can't put my finger on the exact solutions to these symptoms - work to do, chas

4 3

zope.cgi file under NT installation?
by Chris Fassnacht 30 Sep '99

30 Sep '99

I'm trying to get the ZServer (and Zope) to run behind IIS on NT 4.0. The How-To's, list messages, and readme's refer to modifying the zope.cgi file. Unfortunately, I've got an NT installation of Zope that I installed direct from precompiled binaries, and I can't find a zope.cgi file anywhere. I've seen some examples of zope.cgi posted on the list, but these were all for linux installations. Does anybody have an example zope.cgi from an NT installation, and that is intended to run behind IIS, not Apache? Or, barring that, would anyone be willing to help make suggestions on modifying an existing zope.cgi file from a linux installation. Most of the paths are obvious to change, but I can't, for instance, find the pcgi.soc, pcgi.pid, or pcgi.log files. One example I have is this one: #!/usr/local/zope/Zope-2.0.0b6-linux2-x86/pcgi/pcgi-wrapper PCGI_NAME=Zope PCGI_MODULE_PATH=/usr/lib/python1.5/Zope PCGI_PUBLISHER=/usr/local/zope/Zope-2.0.0b6-linux2-x86/pcgi/pcgi_publisher.p y PCGI_EXE=/usr/bin/python PCGI_SOCKET_FILE=/usr/local/zope/Zope-2.0.0b6-linux2-x86/var/pcgi.soc PCGI_PID_FILE=/usr/local/zope/Zope-2.0.0b6-linux2-x86/var/pcgi.pid PCGI_ERROR_LOG=/usr/local/zope/Zope-2.0.0b6-linux2-x86/var/pcgi.log PCGI_DISPLAY_ERRORS=1 BOBO_REALM=/usr/local/httpd/cgi-bin/Zope.cgi BOBO_DEBUG_MODE=1 INSTANCE_HOME=/usr/local/zope/Zope-2.0.0b6-linux2-x86 Also, there seem to be several ways to get Zope running behind another server: - single-threaded without using ZServer - multithreaded using ZServer - using pcgi_publisher.py - using ZServer, and not using pcgi_publisher.py - using Alex Staubo's ISAPI filter? - perhaps other, obsolete, methods. I believe the most up to date method, and the one I should be using, is the multithreaded method using ZServer to handle the pcgi calls. Am I correct in assuming that the following how-to from BrianH (http://www.zope.org/Members/brianh/iis_howto) is the most recent documentation for this process? Also, assuming I already have Zope up and running on the same server running IIS, is this a sufficient set of instructions, or are there details that might be likely to hang someone (i.e., me) up? Thanks very much. Chris Fassnacht How-To: Zope 2, IIS, and PCGI Created by brianh. Last modified on 1999/09/15. Recipe for setting up Zope with IIS: 1. (If you're not using a source copy of Zope) Copy over the Zope.cgi file from a non-Windows Zope dist. Put it in your Zope directory top level, and rename it something like zope.pcgi. 2. Edit zope.pcgi; fix the paths, and add a PCGI_PORT directive specifying some hitherto unused port (for example, 8090). 3. Copy zope.pcgi and pcgi\bin\pcgi-wrapper.exe into IIS's scripts directory. 4. Add a new extension mapping to IIS for PCGI. In the Microsoft Management Console, select your site's Properties, and then under the 'Home Directory' tab click on 'Configuration'. Here you should be able to add a new extension linking .pcgi to 'pcgi-wrapper %s'. 5. Add the start up option '-p zope.pcgi' when starting Zope. If you installed Zope as a Windows NT Service, you will have to edit the registry key: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\ <Service Name>\Parameters\start Add the start up option '-p zope.pcgi' (plus any other z2.py command line options you feel like adding at the time). 6. Make sure that IIS and Zope are both started up, and try accessing your server via http://localhost/scripts/zope.pcgi/ This should give you the same thing as http://localhost:8080/

2 1

manage_addFile() => TypeError
by Terrel Shumway 30 Sep '99

30 Sep '99

I want to upload a File into a Folder under DTML control This code <dtml-call "manage_addFile(rep.this(),'newid',file)"> where rep is a folder, and file is from <input type='file' name='file'> produces this error Error Type: TypeError Error Value: argument 1: expected read-only character buffer, ImplicitAcquirerWrapper found Traceback (innermost last): <snip> File E:\opt\z\lib\python\DocumentTemplate\DT_Let.py, line 145, in render (Object: newid=get_new_id) File E:\opt\z\lib\python\DocumentTemplate\DT_Util.py, line 321, in eval (Object: manage_addFile(rep.this(),newid,file)) File <string>, line 0, in ? File E:\opt\z\lib\python\OFS\Image.py, line 115, in manage_addFile (Object: ElementWithAttributes) File E:\opt\z\lib\python\OFS\Image.py, line 169, in __init__ (Object: <Folder instance at 00FFD620>) File E:\opt\z\lib\python\OFS\Image.py, line 252, in _get_content_type (Object: <Folder instance at 00FFD620>) File E:\opt\z\lib\python\OFS\content_types.py, line 162, in guess_content_type (Object: ElementWithAttributes) File E:\opt\z\bin\lib\mimetypes.py, line 57, in guess_type (Object: ElementWithAttributes) File E:\opt\z\bin\lib\urllib.py, line 759, in splittype (Object: ElementWithAttributes) File E:\opt\z\bin\lib\re.py, line 119, in match TypeError: (see above) The part I don't understand is in OFS/Image.py, line 252 if type(body) is not type(''): body=body.data content_type, enc=guess_content_type( getattr(file, 'filename',id), body, content_type) return content_type What is this getattr() that takes three arguments? My guess is that id is the default value in case 'filename' is not found... no Id should be a string, not an ImplicitAquirerWrapper. I give up. I even tried to pass the content type explicitly too, but File.__init__() still tries to "guess" the content type. (why? I suppose mimetypes.guess_type() is smarter than the programmer/user <g>) Anyway. I don't care anymore *why* it doesn't work. I want to know what *will* work. I would prefer to use OFS.Image.File (a core component), but I am about ready to use the FileObject from Confera. Am I doing something terribly wrong?

1 0