- Zope - Zope lists

Re: CatalogPathAwareness, CatalogAwareness and ZClass CatalogAware (2.3.3)
by Didier Georgieff 15 Nov '01

15 Nov '01

Hello, To respond/updtae this post, the explanation for this problem seems the right one. BUT even when i register the new CatalogPathAwareness.py in __init__, the OLD CatalogAwareness seems used for every new instance i create. ? Is there something hard coded in my ZClass (created before this change) so it doen't use the new __init__ ? ? If yes, could it be changed besause this ZClass contains A LOT of things.? ? if not is it safe to change the CatalogAwareness.py with adapting the url() ? Will i break something else ?? Any advice more than welcome. Le 14 Nov 2001, à 17:57, Didier a écrit : > All CatalogAware auto-indexing is fine when i dont use VHM, and > when i have (thru Apache and VHM) my root as a folder the > autocataloging get the wrong physical path : > in ZCatalog/__init__.py > > there is registring for the old CatalogAwareness instead of the new > CatalogPathAwareness (with the getURL, getPath wich i use > without anu problem in my application) > > ? What happens if i just change the ZCatalog/__init__ > import ZCatalog, Catalog, CatalogPathAwareness, Vocabulary, > ZClasses > createZClassForBase( CatalogPathAwareness.CatalogAware, > globals() > , 'CatalogAwareBase', 'CatalogAware' ) -- Didier Georgieff Chef de projet technique SIT du Bas-Rhin tél : 03.88.76.78.35 - fax : 03.88.76.78.34 email : didier.georgieff(a)agriculture.gouv.fr

1 0

CatalogPathAwareness, CatalogAwareness and ZClass Catalogaware
by Didier Georgieff 15 Nov '01

15 Nov '01

Hello, On 2.3.3 i just discovered a problem with CatalogAware ZClass and VHM. All CatalogAware auto-indexing is fine when i dont use VHM, and when i have (thru Apache and VHM) my root as a folder the autocataloging get the wrong physical path : Example for the same object : www.mysite.com/folder1/item1 is well autoZCataloged as /folder1/item1 (no VHM) but myvirtualsite.com/item1 is missAutoZcataloged as /item1 (with VHM) If i try to "find"/Catalog the objects they are **well** cataloged (with WHM and without VHM). After looking at the list, i found few informations about VirtualHostFix but normally it's fixed and i do think the problem is elsewhere. So : in ZCatalog/__init__.py there is registring for the old CatalogAwareness instead of the new CatalogPathAwareness (with the getURL, getPath wich i use without anu problem in my application) import ZCatalog, Catalog, CatalogAwareness, Vocabulary, ZClasses createZClassForBase( CatalogAwareness.CatalogAware, globals() , 'CatalogAwareBase', 'CatalogAware' ) My questions are : ? Why is there no reference to the new CatalogPathAwareness for CatalogAware ZClass ? I suspect this is the reason why CatalogAware fails and "find" works fine on CatalogAware instance. ? Is this a bug or done by purpose ? ? What happens if i just change the ZCatalog/__init__ with import ZCatalog, Catalog, CatalogPathAwareness, Vocabulary, ZClasses createZClassForBase( CatalogPathAwareness.CatalogAware, globals() , 'CatalogAwareBase', 'CatalogAware' ) Will the thunder will strike me ?? Any help or tip would be more than welcome. Thanks in advance. -- Didier Georgieff Chef de projet technique SIT du Bas-Rhin tél : 03.88.76.78.35 - fax : 03.88.76.78.34 email : didier.georgieff(a)agriculture.gouv.fr

1 0

External Images
by Norman Khine 15 Nov '01

15 Nov '01

Hello, I have a public folder in usr/local/apache/htdocs/posters/ of my apache setup containing something like 3,000 images. Now my question is how do I go about in using this information within zope, without actually having to import all the images into zope? I already have a database which has indexed all the filenames for the images, I can create a ZSQL Method which displays this, but I can figure how to specify the external file system, so that when I have a method to display an image, the image id is picked up from the ZSQL Method, this adds it to <img src="usr/local/apache/htdocs/posters/image1.jpg" border="0" alt="" /> I can create content using html tags within a method, so that includes links to an external images resource. e.g. <img src="http://localhost/catalog/posters/61_imageThree.jpg" border="0" alt="" /> which works. BUT is this the done thing???? Hope it makes sense.. Thanks Norman

3 2

first logging
by Gaetan GUYODO 15 Nov '01

15 Nov '01

Hi, I'm new user of zope. I'm under SuSe 7.2, with a zope default installation, but I don't know what are the default loggin and password for zope access... anyone can help me ? thanks, Gaetan

3 3

Fw: [Zope] name based virtual hosts with squid
by sebas ICK 15 Nov '01

15 Nov '01

For the mailing list archives ... sebas. ----- Original Message ----- From: "Richard Barrett" <R.Barrett(a)ftel.co.uk> To: "sebas ICK" <sebas(a)ick-kupplungen.de> Sent: Wednesday, November 14, 2001 11:59 AM Subject: Re: [Zope] name based virtual hosts with squid At 16:00 14/11/2001 -0800, sebas ICK wrote: >I want to use squid as a front end server indeed and hide zope and apache >behind it. Squid can do some security checks and advanced caching for >frequently requested site. > >I wonder why it is this badly (maybe not at all) documented how to handle >multiple backend servers with squid. Well, I guess that's the trouble with open software. >This a cleaned up version of the *relevant* part of my squid.conf: > >------------------------------- > ># set port where squid listens to, enable logging >http_port 131.174.168.120:80 >emulate_httpd_log on > ># set security checks off for debugging purposes >http_access allow all > ># define virtual hosts for zope and apache >httpd_accel_uses_host_header on >httpd_accel_host zopedev.dhs.org #redirect this to apache on port 90 >httpd_accel_port 90 >httpd_accel_host zopedev-alpha.dhs.org #redirect this to zope on port 8080 >httpd_accel_port 8080 > I'll say from the start that my experience is with using Apache to serve its own content and transparently proxy for Zope servers. My interest in Squid has been using it as a regular (non-transparent) outbound proxy server for browsers running on machines inside our firewall. Still her goes. Looking at the Squid documentation (http://squid.visolve.com/squid24s1/contents.htm) I think I see a couple of problems with your squid.conf: 1. under httpd_accel_host it says: "This tag is used to set the hostname of the accelerated server. It's possible to have only one destination server, so you can have only one occurrence of this line. If you are going accelerate more than one server, or transparent traffic, you will have to use the word virtual instead of a hostname here" Note: "only one occurence", you are trying to specify two 2. under httpd_accel_port it says: "Accelerated requests can only be forwarded to one port: there is no table that associates accelerated hosts and a destination port. Squid will connect to the port that you set the httpd_accel_port value to ... If you want virtual port support then specify the port as "0" Note: "one port", you are trying to specify two Can we be clear about what you are trying to achieve. I'll make the following assumptions. If they are wrong then the argument collapses and will have to be remade on whatever set of objectives you do wish to pursue: 1. You want to run Squid on a machine with the IP number 131.174.168.120, listening on port 80. 2. You want Squid to act as a transparent proxy for Zope and Apache servers. From what you've said, it isn't clear whether these other servers are to be running on the same or different machines to Squid. 3. Each HTTP Request directed at 131.174.168.120:80 is to be passed transparently to one of the "hidden" Zope/Apache servers with the server chosen by analysis of the request headers. 4. Whether the hidden servers are running on the same machine as Squid or not will affect how incoming URLs must be structured. A separate Redirector program that you must identify to Squid has to take the incoming URL and, based on its contents, rewrite it to a URL suitable for one of the hidden servers. If the other servers are on the same machine but differentiable by the Host header of the request, the Redirector can potentially use that to rewrite the URL to reflect the different port number used by each server. If however, the servers are on different machines and cannot be differentiated by Host header then some part of the path in the URL will need to allow the identification of which server is to get the rewritten request. By my reckoning your squid.conf has to contain something like this: # set port where squid listens to, enable logging http_port 131.174.168.120:80 emulate_httpd_log on # set security checks off for debugging purposes http_access allow all # we are using virtual hosts httpd_accel_uses_host_header on httpd_accel_host virtual httpd_accel_port 0 # and defining the redirector which will rewrite the incoming URLs redirector_access allow redirect_program /path/to/you/redirector/program You need to consider the type of HTTP clients that will be in use. The Host header is an HTTP/1.1 feature which some older clients may not provide and a dependency on it in your redirector might bring you grief. A safer bet might be to use the path in the incoming URL as the means of selecting the server as this is HTTP version independent. Another point to consider is the issue of dealing with re-writing redirects from the hidden servers, achieving the equivalent of Apache's ProxyPassReverse directive with Squid. Possibly a bigger problem is checking for and rewriting returned content from the real servers which contains non-relative URLs that do not take account of the proxying that is taking place. For instance, how, with Apache, do you intend to resolve the sort of issues that are overcome by the use of Zope's Virtual Host Monster. >------------------------------- > >Maybe someone got another approach? > >kind regards, > >sebas. > If you are looking for simple solutions I'd urge you to consider using Apache as the primary server for its own content and which can also transparently proxy for Zope using ProxyPass, ProxyPassReverse and RewriteCond/RewriteRule in the Apache httpd.conf. Zope's VHM solves much of the content rewriting issue at source. Best of luck Richard >----- Original Message ----- >Sent: Wednesday, November 14, 2001 3:10 AM >Subject: Re: [Zope] name based virtual hosts with squid > > >If I understand you correctly, you want to hide Apache and Zope behind >Squid. If you want Squid to rewrite incoming URLs as appropriate to either >the Apache or Zope server you will need to provide what is referred to in >Squid documentation as a Redirector program. This program/script, which >you must supply, can rewrite the incoming URLs using whatever >transformation rules you choose to program. The effect is equivalent to >using ProxyPass and RewriteCond/RewriteRule with Apache. > >See the Squid documentation http://www.squid-cache.org/Doc/FAQ/FAQ-15.html >for more information. > >Then again why use Squid? You could achieve similar objectives by having >Apache ProxyPass/Rewrite directives to rewrite URLs to the Zope server. See >this list's archives for postings on this subject. Then again Squid would >probably work better with a high traffic levels. > >At 22:13 13/11/2001 -0800, Sebastian Kügler wrote: > >Hi group, > > > >Here's my problem, maybe someone's got better idea's ... > > > >I want to use squid to server apache and zope on one port. domains under > >zope.domain.org have tp be served by zope and domains with www.domain.org >by > >apache. I have the following in my squid.conf; > > > >----------------------------------------------------- > > > >httpd_accel_uses_host_header on > ># added port 80 here to listen to port 80 (doH!) edited by sebas > > http_port 80 > > emulate_httpd_log on # use NCSA https style logfile > > > ># http_access allow all > > httpd_accel_host zopedev.dhs.org > > httpd_accel_host virtual > > httpd_accel_port 90 # port where Apache is running > > > ># added port 80 here to listen to port 80 (doH!) edited by sebas > > http_port 80 > > emulate_httpd_log on # use NCSA https style logfile > ># this line replaces 'http_access deny all' > ># http_access allow all > > httpd_accel_host zopedev-alpha.dhs.org > > httpd_accel_host virtual > > httpd_accel_port 8080 # port where Zope is running > > > >------------------------------------------------------- > > > >As some of you might already have guessed: it does not work in the right > >way, squid does only seem to use the last http_port specified, just forgets > >about our first one. > > > >I tried different things with that config file, but none of them did the > >trick. > > > >Maybe some of you got it working, if so please tell me how. > > > >regards, > > > >sebas. > > > > > > > >_______________________________________________ > >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

Path keeps every directory I go through
by Gordon Holtslander 15 Nov '01

15 Nov '01

I have a series of directories in my site called maturity, membership, ministry, missions, home, etc. As I move from directory to directory on the site using the common standard_html_header I get this very bizarre path developing. http://localhost:8080/membership/maturity/ministry/missions/home/today.html I'd kind of like to have it start over every time I click on the header icons. Am I doing something wrong or is this just acquisition in a way I wasn't expecting it. Talk to you soon, Gord Gordon W. Holtslander Director of Development & Information Technology The Lung Association - Saskatchewan gordon.holtslander(a)sk.lung.ca <mailto:gordon.holtslander@sk.lung.ca> V: 306.343.9511 C: 306.241.2201 F: 306.343.7007

2 1

Permissions problems
by <"Nico De Boer" 15 Nov '01

15 Nov '01

I have some problems with adding my new product into a site. When I want to add my product in my test-folder, it's no problem. But when I want to add my product in a website, it isn't shown in the Add a Product list. I think it is a problem with the Zope-permissions, but I have looked in several source code and have made it just the same. This is my __init__.py file: import FormMail def initialize(context): try: context.registerClass( instance_class=FormMail.FormMail, permission='Add MailFormulier', constructors = ( FormMail.FormMail_Add, FormMail.manage_FormMail_Add, ), ) except: import sys, traceback, string type, val, tb = sys.exc_info() sys.stderr.write(string.join(traceback.format_exception(type, val, tb), '')) del type, val, tb And this are my __ac_permissions: __ac_permissions__=( ('View management screens', ('manage_main',)), ('Change permissions', ('manage_access',), ('Owner', 'Manager')), ('Change MailFormulier', ('manage_edit',''), ('Owner', 'Manager')), ('View', ('index_html', 'view_form','manage_form'), ('Anonymous', 'Owner', 'Manager')), ) Can anyone help me solving my problem?

1 0

Re: [Zope] ZClass instance with an other ZClass definition
by Didier Georgieff 15 Nov '01

15 Nov '01

Le 14 Nov 2001, à 22:33, Dieter Maurer a écrit : > Didier Georgieff writes: > > ... calling ZClass method fails ... > > It fails because it asks for login (even if i'm manager) and fails with an > > Unauthorized in ZopeSecurityPolicy.py > Did you enabled the necessary permissions in "Permission mapping"? This what i was suspected, but coudn't find what was the right permissions mapping settings. So, at the moment I'm solving this problem by moving the ZClass1 instance from the Zclass2 definition to each ZClass2 instances. This works. Thanks a lot Dieter. Didier.

1 0

JASC Namo a Visual Web Editor for Zope?
by Harlow Pinson 15 Nov '01

15 Nov '01

Hi. Can someone confirm this... I just downloaded and installed the JASC Namo WebEditor and it seems to let me visually edit Zope objects that do not have file extensions, and also not clobber my code. http://list.jasc.com/UM/T.asp?A1.31.42.1.2373433 If so, after fighting with Dreamweaver as a visual Zope editor, this would make my day, and make Zope interface development much more accessible to designers. Any other input about quirkyness of this product with Zope would be appreciated. I'm just starting to look at it. Harlow Pinson Indepth Learning Email: hpinson(a)indepthl.com Web: http://www.indepthl.com Voice: 505-994-2135 FAX: 208-475-7678

3 2

redirecting exUserFolder login
by Mark Logan 15 Nov '01

15 Nov '01

When a user tries to login to exUserFolder, the 'docLogin.dtml' method is displayed. I need to redirect the user to 'https:' + URL[5:] if they try to access the login form via unencrypted http. However, I can't seem to do this. I've tried putting the redirect in docLogin, that doesn't work (unless the user navigates directly to docLogin, which they almost certainly never will). I've tried putting the redirect right before every "raise 'LoginRequired', self.docLogin(...)" call in the exUserFolder.py. Neither of these work. I confess I don't have the best understand of how the redirection to login screens works. I can say that it seems that the HTTP 'location' header is getting set, but the response code is always 500, which means that the browser ignores the 'location'. I am passing a true value for 'lock' in my redirect calls, i.e. REQUEST.RESPONSE.redirect('http:' + URL[5:], 1) But none of this works. Any clues? Thanks, Mark

1 1