November 1999 - Zope-Dev

ctsybasemodule
by Brusowankin, David 19 Nov '99

19 Nov '99

Hi, I noticed that you guys are maintaining ctsybasemodule for now. What is the latest version of ctsybasemodule ? Can I get it ? And, although I am not guaranteeing anything for NOW, I may be able to find some time to help in its maintainence. Please send me the bug list as well. Thanks in advance, David ---------------------------------------------------------------------------- ------------ David Brusowankin Goldman Sachs Information Technology 1 NY Plaza 38th Floor New York, NY 10004 Voice: 212.357.0317 Email: david.brusowankin(a)gs.com

1 0

Writing custom folderish products
by Martijn Faassen 19 Nov '99

19 Nov '99

Hi there, I'm writing a product (in Python, not with ZClasses) which consists of: form - a folderish object to which you can add.. fields - a number of SimpleItemish objects that can be added to a form What I have working: * I can make the form behave like a folder by deriving it from OFS.ObjectManager * I have a number of fields that can be added anywhere in the Zope tree What I want to reach is this situation: * I want the form to behave like an folder, but only fields should be addable. and * Fields should not be addable anywhere but in a form. What set of incantations would accomplish this? Any example source? Or Zope source code I should study? Anything? Any hint in the right direction would be appreciated. Regards, Martijn

4 3

Replacing/Augmenting dtml-sqltest from a Product?
by Bradley McLean 19 Nov '99

19 Nov '99

Please forgive the somewhat naive question. I'm a seasoned C/C++, Java, Perl programmer new to Python and Zope. Hopefully someone can save me a little time. I've modified Shared/DC/ZRDB/sqltest.py and sqlgroup.py to provide some additional semantics that I find very useful. Specifically, I've added range and keyword operations, and comparisons other than equality to sqltest, and I've made a parameterized version of sqlgroup that allows selection of 'and' or 'or' linkages from a variable, making it easier to provide sophisticated search options on a Zope page. I'd like to make these modifications available for public review and comment. The best way I can think of to do this is to make a new Product that provides the new versions and replaces them in DC/ZRDB/DA/SQL. Can anyone give me a brief shortcut on how to do that? I can figure out how to make a Product, but figuring out how to either replace these classes, or to create a duplicate of SQL Method which uses them is not entirely obvious. TIA, -Brad

5 5

Max objects before ZCatalog?
by David Kankiewicz 19 Nov '99

19 Nov '99

What is the magic number for performance? Between the individual lookup, doing a 'large catalog' ZCatalog search, and creating many ZCatalogs. I'm accessing a few variable (~ 2-8) in 20-5000 (~max) subobjects and testing them against the current objects variables. The design is for something that most scale and can easily to be changed (I have this half done). The idea is really close to ZTopics, with my own idea of a minimalistic custom implementation. Is there anyone who has experience or has done performance testing and would like to comment? Thanks in advanced, David

1 0

RE-POST: getting Zap to work with cgi-bin directory?
by Michael Olivier 19 Nov '99

19 Nov '99

I'm reposting because I got no responses to this, and we're stuck. I would greatly appreciate any light you can shed on whether Zap should work with its cgi-bin like standard apache does, and how to troubleshoot problems. Ours isn't working and it isn't clear to us what else to try. --Michael Michael Olivier wrote: > > Hi folks, > > We've got Zope and Zap running but are having trouble with Zap's cgi-bin > directory. Specifically, we're wanting to run some pcgi/bobo stuff but > can't get even the perl test scripts (printenv, test-cgi) to work. > > >From the error log I can see that it's mapping /localhost/cgi-bin to > /usr/local/zope/Zap/htdocs/cgi-bin/ -- > it's looking for cgi-bin below htdocs. I'd like to have it as a sibling > directory with htdocs. If I actually put it inside htdocs > (htdocs/cgi-bin), it > displays the scripts but doesn't execute them. I tried adding this to > the zap.conf file: > > <Directory /usr/local/zope/Zap/cgi-bin> > AllowOverride None > Options ExecCGI > </Directory> > > ...but this didn't seem to have any effect... > > Can someone advise please on how to debug this situation? Doesn't zap > come preconfigured for cgi-bin use, or not? > > Many thanks, > Michael

1 0

mozilla and zope, update.
by Anthony Baxter 19 Nov '99

19 Nov '99

For those who are interested, Mozilla is very very close to being usable for working in Zope. At the moment, the killer is the open bug 6144 - Moz loses track of the auth header, and doesn't re-send it in. This means that in the management interface, you're not authorised, and so manage_workspace defaults to index_html, rather than manage_main. It's currently listed to be fixed before the next milestone release, M12, which is scheduled for 15th December. I'll post an update when it's fixed. Anthony

1 0

killing zombies
by Robin Becker 19 Nov '99

19 Nov '99

After putting the site I've developed on a live test basis I notice that I'm getting some zombie connections. I'm using the medusa active channel display code as in the code below. I notice that there are kill_zombies method for zhhtp_channel, but how should I use that method? My approach uses a similar approach, but is probably wrong. The zombie killers aren't happening in Zope as it stands. Can I ask why not? can I get the same thing to happen for other channels eg ftp etc? #status.py from ZServer.medusa.status_handler import channel_list_producer from ZServer.medusa.asyncore import socket_map from ZServer.HTTPServer import zhttp_channel from string import replace from time import time def __build(s): r='' while 1: t = s.more() if t=='': break r = r + t return r def get_active_channels(): r='' s=channel_list_producer('aaaa') r = r + replace(__build(s),'<a href="aaaa">Status Report</a>','') return r def kill_zombies(self): now = int (time()) for channel in socket_map.keys(): if channel.__class__ == zhttp_channel: if (now - channel.creation_time) > channel.zombie_timeout: channel.close() -- Robin Becker

1 0

Got "expand_all" to work in a tree tag with leaves
by Andrew Wilcox 18 Nov '99

18 Nov '99

I've been battling the tree tag all day to get expand_all to work, and it almost won, but then I hacked the source. {insert evil manical laughter here} Simple use of the tree tag, listing the titles of objects within folders:  <a href="<dtml-var URL0>?expand_all=1">Expand all</a>     Where listObjects is a simple DTML Method that lists the titles of the objects within the folder, with a little indenting:  <dtml-in "objectValues()"> <dtml-if "_['sequence-index'] != 0"> </dtml-if>     <dtml-var title_or_id> </dtml-in>  Now, if you hit the "expand all", it expands all the folders except those at the lowest level, the ones that just contain (non-folder) objects. My first thought was that maybe it wasn't really considering the lowest level folders to be a branch, so I spent a while trying different functions for the "branches" attribute. That didn't work. Take a look at tpValuesId inside of TreeTag.py, which builds a list of all the expandable subitems for expand_all: try: items=getattr(self, branches)() except AttributeError: items=() for item in items: try: if getattr(item, branches)(): If I'm reading this right, the first call on branches gets the subitems which are considered a branch (defaulting to isPrincipiaFolderish over in ObjectManager). But then the second call on branches checks if the subitem itself has a non-zero number of branches. I commented out the second call there, and lo and behold! Suddenly my tree was fully expanding. So, what did I break? :) Andrew

1 0

Re: [Zope-dev] more on sessions in general.
by Anthony Baxter 18 Nov '99

18 Nov '99

>>> "Stuart 'Zen' Bishop" wrote > There is a need for someone to write ZSession or DumbSession that simply > stores the information in a in-memory dictionary (or var/session for ZEO > compatibility?), which would then define the API. Yep. Don't look at me - SQL Sessions do what I need. > How would you 'slot in' a session? > > Perhaps it just calls getSession(this(),REQUEST), which would be aquired > from higher-on-up. This method could defined in SQLSessionFolder, or just > a DTML/external method the site manager left in the root folder, or we could > patch OFS/Folder.py to give it a 'Session' property defining the class of the Right now, you simply <dtml-call Session>. The Session is the SQLSession object - it contains the SQL methods used by the session. It creates REQUEST.SESSION, which is a dictionary-like thing. > It would be nice if REQUEST.SESSION was filled in for you if the session > class has been defined, but someone else will need to comment on its > feasibility :-) That's what I do now. It seemed appropriate - it felt like it should be on an equivalent level to RESPONSE. Anthony -- Anthony Baxter <anthony(a)interlink.com.au> It's never too late to have a happy childhood.

1 0

ZServer FastCGI patch for IRIX (and others?)
by Phillip J. Eby 18 Nov '99

18 Nov '99

Below is a patch to let ZServer's FastCGI support work with FastCGI client implementations that don't properly close the STDIN sub-stream. On the IRIX/Apache/cgi-fcgi combo used where I work, the client side doesn't seem to know when it's done sending STDIN data, and thus the ZServer side simply hangs forever waiting for it. The patch below makes ZServer ignore STDIN altogether if there is no Content-Length, and also stop looking for more STDIN data if Content-Length or more bytes have already been read. --- FCGIServer.py Thu Nov 18 17:08:24 1999 +++ Zope2/ZServer/FCGIServer.py Thu Nov 18 17:05:02 1999 @@ -429,6 +429,8 @@ elif rec.recType == FCGI_PARAMS: if rec.contentLength == 0: # end of the stream self.remainingRecs = self.remainingRecs - 1 + self.content_length=string.atoi(self.env.get('CONTENT_LENGTH','0')) + if self.content_length==0: self.remainingRecs = self.remainingRecs - 1 else: self.env.update(rec.values) @@ -438,6 +440,8 @@ self.remainingRecs = self.remainingRecs - 1 else: self.stdin.write(rec.content) + self.content_length = self.content_length - rec.contentLength + if self.content_length<=0: self.remainingRecs = self.remainingRecs - 1 # read some filter data elif rec.recType == FCGI_DATA: @@ -448,7 +452,7 @@ # We've processed the record. Now what do we do? - if self.remainingRecs: + if self.remainingRecs > 0: # prepare to get the next record self.setInitialState()

1 0