February 2001 - Zope

Re: [Zope] Images and uploading
by Harald Scheirich 28 Feb '01

28 Feb '01

What might help in your case is the ExtFile/ExtImage product. It deposits binaries in a directory on the server but also creates Zope objects to access that file. It uses the Python Imaging Library for the handling of Images. I assume PIL does not use file endings but header info from the files to detect the filetype you probably still need to actually create the object to do some checking or subclass ExtImage in Python to provide some more error checking . Harald At 12:13 PM +0100 2/28/01, Paul Zwarts wrote: >So the question is, how do I get an attribute of an object while its >still masquerading in the REQUEST as an upload instance....? Other >reasons being is that on my old system, (its a link databse) hackers >tried to upload .exe's thinking they could get a server-side trojan >running, so I provide some kind of error checking....

2 3

Re: [Zope] Proposal: Include tino's patched <dtml-in> in future Zope
by Itai Tavor 28 Feb '01

28 Feb '01

Toby Dickenson wrote: >On Tue, 27 Feb 2001 09:56:17 +1100, Itai Tavor <itai(a)optusnet.com.au> >wrote: > >>citing some irrelevant concern about polluting the namespace. > >namespace pollution is less serious if you have a greater number of >namespaces, and more serious if you have fewer. > >dtml is the only language I know which only provides *one* namespace. Ok, so namespace pollution might be a valid argument against the prefix patch. But as Chris and Tino pointed out, it's irrelevant for sequence_xxx - and that's the patch that has been repeatedly requested on the list and rejected by DC. Chris Withers wrote: >Sorry to jump in here, but can we please not let this argument rob us of yet >another chance to get a nice alternative to such abominations as: ><dtml-var "_['sequence-item']*3"> > ><dtml-var "sequence_item 3"> is much nicer :-) > >Tino, is there any chance you could split your dtml-in patch into two so that >the contraversial bit can be argued over while not impeding the most >useful bit? > >please?! > >Chris Hey, that's just what I wanted to say! I vote that that sequence_xxx aliases be added to Zope 2.3.1, while discussions on the prefix patch can continue. I personally don't care about this one - I never met the need for it. But I use sequence_xxx names in preference to _['sequence_xxx'] all the time, and I've seen no good argument against adding it to the core. But then, votes only count in a democracy. Toby Dickenson wrote: >IMO the *real* answer is to not implement logic in dtml. The problems (and >the need for this patch) go away. People keep telling us DTML coders this. I would really like to see some of your code, because I just don't get it. Formatting HTML is a whole lot easier in DTML than in python. A DTML document looks mostly just like the HTML it generates, plus some DTML tags. It's easy to read and it's easy for non-programmers to modify. HTML in python is a long ugly list of print statements or a bunch of messy variable substitutions on string constants. I just can't see what can possibly be simpler than a dtml-in loop. dtml-in might be used for logic, and that would be a bad idea. But it's also used a whole lot for display formatting - and for that it's the best tool around. People keep talking about phasing DTML out. But if the alternative is formatting HTML in python, I'm outta here. Itai -- -- Itai Tavor -- "Je sautille, donc je suis." -- itai(a)optusnet.com.au -- - Kermit the Frog -- -- -- -- "If you haven't got your health, you haven't got anything" --

2 1

RE[2]: Zope Good/Bad News Article
by Thierry Nabeth 28 Feb '01

28 Feb '01

Hello Brian and Paul, I would like to remind you a relately nasty bug that considerably reduce the potential of Code Reuse for the product. The bug is the following. Because of Export/Import (that you have to use for instance if you upgrade the Zope version), "you can not define in a product a class that inherits another class that is defined in another product." !!!! (if you try, you just kind of create some Zombi class). Which means that you can not use inheritance between components that are defined in a products !!! Thierry Nabeth INSEAD CALT http://www.insead.edu/CALT/ PS: You can see the reason if you examin the exported package in xml -------- exported package extract---- <record ...> (Class TB definition that inherits from class TA of another product) .... <tuple id="24671.6"> <global id="24671.3" name="PersistentClass" module="ZClasses.ZClass"/> <persistent> <string id="24671.4" encoding="base64">AAAAAAAAYFE=</string> ###### ==> class in other package. ##### </persistent> <global id="24671.5" name="SimpleItem" module="OFS.SimpleItem"/> </tuple> -------- As you see, the module of class TA is not define. Actually, a TA zombi is defined in the Product in which TB is defined. You should have instead something for TA such as <global id="24671.3" name="TA" module="TA_Product"/> The problem comes from the (bad) use of the Python Marchalling/unMarchaling functions.

1 0

Re: [Zope-dev] RE: [Zope] Zope Good/Bad News Article
by Oliver Bleutgen 28 Feb '01

28 Feb '01

> One idea we've kicked around is to make the eventual > "components area" on zope.org such that component users > can provide some +1 -1 type feedback that is incorporated > into the way that components are listed on the site. This > would let the most popular / best components appear at > the top of listings etc. Have you had a look at http://www.php.net ? They have a "PHP Annotated Manual" (right under Documentation) where users can add comments to the pages. I always found this very helpful, but it may be too much work for DC to read over every comment (php's annotations are moderated). Perhaps just allowing members of zope.org for the comments and/or letting the maintainer of the product moderate it might solve that. This might be a nice suggestion to the rating system and could all be made configurable by the product maintainer. cheers, oliver

1 0

Images and uploading
by Paul Zwarts 28 Feb '01

28 Feb '01

Hi all, I've progressed on with LocalFS and feel comfortable with it, but Im having difficulties with uploading images. I really need to get the content_type attribute of the image, so I can store in the dbms with its record what file extension it is... I thought of querying the getProperty of the REQUEST stream, but it doesnt seem to work the way I expected. What happens with: Case A) <dtml-call expr="LinkImgMnt[_.str(SESSION['key2'])].manage_upload(id=_.string.replace(ltitle,' ','+'), file=REQUEST['image'])"> <dtml-var "LinkImgMnt[_.str(SESSION['key2'])][_.string.replace(ltitle,' ','+')].getProperty('content_type')"> Case B) <dtml-call expr="LinkImgMnt[_.str(SESSION['key2'])].manage_upload(id=_.string.replace(ltitle,' ','+') + '.jpg', file=REQUEST['image'])"> <dtml-var "LinkImgMnt[_.str(SESSION['key2'])][_.string.replace(ltitle,' ','+') + '.jpg'].getProperty('content_type')"> So, obviously LinkImgMnt is the LocalFS mount. and I need to give the names of the files + for whitespace. The vars after that is my test. Case A) I try to query the content type of the file AFTER I've deposited it. But I note that without providing the extension in the manage_upload, I get application/octet-stream. Which is totally acceptable and proof of working, but not desired outcome. Its a GIF that went through _image_. Case B) Doing the same thing, only appending .jpg to the id, I get the right answer back... at least, I get AN answer. But since _image_ is a gif, then obviously Im jury rigging it to give me what I asked for... So the question is, how do I get an attribute of an object while its still masquerading in the REQUEST as an upload instance....? Other reasons being is that on my old system, (its a link databse) hackers tried to upload .exe's thinking they could get a server-side trojan running, so I provide some kind of error checking.... Help much appreciated... Paz

1 0

Re: zope & backup
by Martijn Pieters 28 Feb '01

28 Feb '01

On Tue, Feb 27, 2001 at 06:15:28PM -0500, Turgut Durduran wrote: > I saw your posting in a mailing list about Zope. > > I am extremely unfamiliar with Zope. However, my site admins upgraded > Zope on their site and I lost all my data. They made a copy of the old > files and old Data*.fs etc. However, they can't figure out how to restore > those to something 'readable'. > > Is there a way to do that? Hi Turgut, You can best ask this question on the Zope Mailinglists, not to an individual. It could well be that I don't have time, or knowledge to help you, but the mailinglists ae read by people from all around the world, 24 hours a day. I sent a copy of this mail to the zope(a)zope.org list, and I urge you to post any follow-ups to that list. To quickly answer your question: all you need to get your site running again are the Data.fs file (just that one file), and any extra Products and External Methods you have. You may run into problems with an upgraded Zope though, depending on the Products you use and the Zope versions you are upgreading from and to. Hope this helps! -- Martijn Pieters | Software Engineer mailto:mj@digicool.com | Digital Creations http://www.digicool.com/ | Creators of Zope http://www.zope.org/ ---------------------------------------------

1 0

Subject: Re: [Zope] dynamic & automatic 3 level navigation bar ?
by Tino Wildenhain 28 Feb '01

28 Feb '01

Hi Georg, I recommend the following precondition: either use a hierarchy of zclasses derrived from folder for better structure of the site or at least do the following: write a index_html method high up in the path, which would call everything needed like standard_html_header and footer, navigation bars and last but not least: a content object. lets assume you are using "content" (a DTML-Document) for the sake of example. This way you are very flexible with the way your site looks and feels. Authors only have to provide folders and one content object inside them. The navigation would only address the folders. Their index_html inherited from the root will display the content according to your rules. Now lets come to the navigation bar: You can call it from index_html like that: <dtml-with "PARENTS[-1]> <dtml-var navigation_bar> </dtml-with> And here the framework of navigation_bar (the method has to be in your zope root) <dtml-in "objectValues('Folder')" skip-unauthorized> <dtml-if title> <dtml-comment> just to make sure we are showing anything </dtml-comment> <dtml-if "sequence_item in PARENTS"> <dtml-var title> <dtml-comment> we are in parents so this part shows an open branch </dtml-comment> <dtml-var navigation_bar> <dtml-comment> we have opened this folder so show its content</dtml-comment> <dtml-else> <a href="<dtml-var absolute_url>"><dtml-var title></a> </dtml-if in PARENTS> </dtml-if title> </dtml-in Folder> HTH Tino Wildenhain --On Mittwoch, 28. Februar 2001 08:41 +0100 Georg Hoermann <schorsch(a)ecology.uni-kiel.de> wrote: > Hallo Zope-world, > > I have a problem I think someone *must* have already > solved. > I am (relatively) new to zope and I want to offer zope as > an easy to use simple > CMS for our scientists and students. The put their stuff/files > into zope and thats it. No fuzzing around with navigation bars, > frames and so if they do not want. > They will create rather deep file-hierarchies and the navigation bar > taken from standard_header. This is where my problem comes in: > > I need a simple (text is enough) navigation bar which shows a menu of > (all) the entries/folders at the actual, the parent and the > child level. So even at level 20 there would only be a > simple 3-level navigation bar with the actual position > in the middle which adapts itself dynamically to the position > within the file system. > > All the navigation bars I encountered up to now (including the > tree tag) clutter the screen if the hierarchy becomes to deep/large, > must by manually created and maintained or do not adapt > to the hierarchy level. > > Is there a solution or could anyone help me with this ? > If someone could program it I promise to write a decent > documentation and do the polishing 8-). > I learned a lot about zope > looking at the code of the existing products but > I am fed up in the mixture of many differents manuals, > 164 HowTos, the source code and the zope book. > > Thanks & Greetings > > Georg > > > -- > Georg Hoermann, Ecosystem Research Center, Kiel University, Germany > Tel. #49-431-880-4029, fx: -4083 Home: 0451/47 70 32, -33 fax, 0172/431 > 57 15 schorsch(a)ecology.uni-kiel.de > > _______________________________________________ > 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

authenticate against postgesql, UserDB/mysqluserfolder
by Olaf Marc Zanger 28 Feb '01

28 Feb '01

hi there, as integrator i'd like to use a postgresql table to authenticate. (since i'm quite clueless in python i can not do a lot there). i installed userdb on zope 2.3 and be able to update, edit, list users. even the groups are set right. unfortunately it does not authenticate me when i try accessing the content (it works with a normal acl_users there is no error message at all. since mysqluserfolder works i wonder if there is any chance to adapt that one to postgresql thanks for help olaf -- soli-con Engineering Zanger, Dipl.-Ing. (FH) Olaf Marc Zanger Lorrainestrasse 23, 3013 Bern / Switzerland fon:+41-31-332 9782, mob:+41-76-572 9782 mailto:olaf.zanger@soli-con.com, http://www.soli-con.com

1 0

Zope security management
by John R. Daily 28 Feb '01

28 Feb '01

I'm increasingly frustrated with the Zope security management framework, and I'd like to know if there is a way to work around some of my problems, and/or whether this will be addressed in the future. Or, perhaps I'm looking at all this from the wrong perspective. Essentially, I'd like a way to eliminate a role in certain directories. For example, if anonymous users should be granted no access to a "/private" folder, I want to lock down /private and all sub-directories against anonymous access. The only solutions I've found are inadequate. What I've found: * At the root folder, find those permissions which are enabled for the anonymous role, and remove them in /private by de-selecting the "inherit permissions" checkbox and re-enable appropriate roles. * In /private, de-select _all_ "inherit permissions" checkboxes and re-enable appropriate roles. The first is inadequate because of the lack of control over what permissions are enabled for anonymous users at the root folder. If a particular permission is added to the root folder the next day, anonymous users now have a permission in /private which they should not have. The second, besides being extremely tedious and error-prone, removes the flexibility of defining globally what permissions roles should play across all of the server. What am I missing? -- -- John R. Daily jdaily(a)progeny.com Systems Programmer Progeny Linux Systems Master of the ephemeral epiphany

4 5

Strange error/traceback
by Phil Harris 28 Feb '01

28 Feb '01

Hi all, Anyone got any idea what this means: argument 1: expected string without null bytes, string found I would send the traceback but this is happening in a few places, various places which seem to have no connection except that they are all External Methods. I'm stumped. tia Phil phil.harris(a)zope.co.uk

2 4