- Zope-Dev - Zope lists

My ZSql method return the result of Query!
by Victor.Zhai＠ogilvy.com 20 Jan '00

20 Jan '00

Hi,all I query from a DTML Method that call the ZSql method like this: ZSQl method named "sqlQueryCustomer": select * from customers where customer_id = <dtml-sqlvar customer_id type=string> It can work by using TEST. But I want to use it through one DTML Document named "QueryCustomer": <The one which I query interface> <dtml-var standard_html_header> <h2><dtml-var title_or_id></h2> <form action="dtPerformQueryCustomer"> <table> <tbody> <tr> <th aligh="left">Customer ID</th> <td><input type="text" name="customer_id"></td> </tr> </tbody> </table> <input type="submit" value="Query Customer"> </form> </table> </form> My DTML Method named "dtPerformQueryCustomer": <dtml-var standard_html_header> <dtml-call sqlQueryCustomer> <h3>This is <dtml-var customer_id> was queried.</h3> <dtml-var standard_html_footer> It can only return "This is XXX was queried".And cannot return the result like "*** **** *** ***". What is the problem, Give me your hand. Thanks a lot. "WorldSecure Server <ogilvy.com>" made the following annotations on 01/20/00 04:55:42 ------------------------------------------------------------------------------ Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not copy or deliver this message to anyone. In such case, you should destroy this message and kindly notify the sender by reply email. Please advise immediately if you or your employer does not consent to email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of the Ogilvy Group shall be understood as neither given nor endorsed by it. ==============================================================================

2 1

fcgi-mangled images: resolved
by John Edstrom 20 Jan '00

20 Jan '00

I found why fasctcgi was mangling images and large data files from Apache on DEC Unix. It was a problem with mod_fastcgi where it used int position counters for the input buffer where it needed longs. All that is necessary is to change some 'int's to 'size_t's in a few files and everything works. I have patches for the Oct 6 mod_fastcgi SNAP if anyone wants them. -- John Edstrom

2 1

Sql interface!
by Victor.Zhai＠ogilvy.com 20 Jan '00

20 Jan '00

Hi.all Are there some information about good interface of Sql method! Thanks a lot! Victor zhai. "WorldSecure Server <ogilvy.com>" made the following annotations on 01/19/00 22:28:13 ------------------------------------------------------------------------------ Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not copy or deliver this message to anyone. In such case, you should destroy this message and kindly notify the sender by reply email. Please advise immediately if you or your employer does not consent to email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of the Ogilvy Group shall be understood as neither given nor endorsed by it. ==============================================================================

1 0

PTK thoughts
by Andrew M. Kuchling 19 Jan '00

19 Jan '00

I've looked over the notes about the Portal Toolkit, have just joined the zope-ptk mailing list, and have a few thoughts. (Still have to read over the back archive, though...) * I find it strange that members are able to publish content at all. For most of the portals I can think of -- Yahoo, My Netscape, etc. -- membership means you have access to various features that run off scripts or databases, and you have some sort of customized display. The set of people who publish content is much, much smaller than the set of all members. Zope.org and community discussion forums are the only cases I can think of where members publish content, and only on Zope.org do people see anything like a management screen. In essence, I'm wondering if Members should have no content management capability at all, and their management capability given to some class intermediate between Member and Contributor. * Would it be possible to generalize the model of creating objects which get approved by a manager, so that you have almost workflow-like features. Looking at the PTK from the point of view of the MEMS Exchange, we'd be interested in having more complex transitions, and transitions that can be triggered by different sets of people. For example, our workflow would look like this: * Member creates a Run object * An MX staff member looks at the Run, makes some comments, and bounces it back to the member, who revises it. * MX staffer looks at the run again, and approves the revised version. * 3 other Members with a certain role look at the run and approve it. Various other people get notified about this. The general abstraction is that there can be an arbitrary set of states, and some arbitrary criterion must be met before transitioning from one state to another, instead of only 2 approved/not approved states. This is obviously more complicated, and probably too ambitious for the first version, but it could make the PTK a useful foundation for intranet applications, where you'd embody the way your business works in Python code. BTW, I just re-read "Philip and Alex's Guide to Web Publishing" recently; highly recommended. Chapters 2 and 3 are especially applicable to the PTK, and chapter 13 is also worth reading for Zopers. It's online at http://www.photo.net/wtr/thebook/, but it's still worth buying a copy. -- A.M. Kuchling http://starship.python.net/crew/amk/ We tend to idealize tolerance, then wonder why we find ourselves infested with losers and nut cases. -- Patrick Hayden

2 1

Re: IMAP Server
by Butch Landingin 19 Jan '00

19 Jan '00

Hi all, I'm not doing anything IMAP-related but I found something which might be of use to you guys: http://horde.org/papers/Scalable_webmail_HOWTO.php3 Maybe you could incorporate some of his ideas into whatever you guys are building? HTH. Regards, Butch __________________________________________________ Do You Yahoo!? Talk to your friends online with Yahoo! Messenger. http://im.yahoo.com

1 0

Fix for ZPublisher.Client
by Hunter Kelly 19 Jan '00

19 Jan '00

Hi, I was having trouble using the load_site script with directories that had spaces in the names. Turns out that ZPublisher.Client.Function does not quote the url when it does the post. If there are problems with doing this, or there is a better way, please let me know. Please CC me on replies as I am not on the Zope-Dev mailing list. Hunter Kelly --- Client.py.orig Wed Jan 19 17:51:48 2000 +++ Client.py Wed Jan 19 17:50:52 2000 @@ -263,7 +263,7 @@ else: d[k]=v else: d[k]=v - rq=[('POST %s HTTP/1.0' % self.rurl),] + rq=[('POST %s HTTP/1.0' % quote(self.rurl)),] for n,v in self.headers.items(): rq.append('%s: %s' % (n,v)) if self.username and self.password:

1 0

RE: [Zope] AW: [Zope-dev] Re: [Zope] ZCatalog Find Items broken i n ZOPE 2.1? FIXED! n ZOPE 2.1? FIXED!
by Michel Pelletier 19 Jan '00

19 Jan '00

> -----Original Message----- > From: Markus Strickler [mailto:mstrickler@gmx.de] > Sent: Wednesday, January 19, 2000 4:10 AM > To: Stephen Pitts > Cc: Jason Spisak; Brian Lloyd; zope-dev(a)zope.org; zope(a)zope.org > Subject: [Zope] AW: [Zope-dev] Re: [Zope] ZCatalog Find Items > broken in > ZOPE 2.1? FIXED! > > > Did you follow the discussion about / and id in ZCatalog? > I haven't checked whether the apply_path bug has been fixed > in the cvs as > well. This was fixed along with an earlier bug in the code. Thanks for the patch. -Michel > > I just got hit by this as well. The culprit is in line 536 of > > lib/python/Products/ZCatalog/ZCatalog.py: > > > > apply_func(ob, (apply_path+'/'+p)) > > > > If apply_path is null, then an extra '/' is prepended to p > (the object's > > path) for no apparent reason. A quick fix is to replace this with: > > > > if apply_path: > > apply_func(ob, (apply_path+'/'+p)) > > else: > > apply_func(ob, p) > > > > A patch to that effect is attached > > -- > > Stephen Pitts > > smpitts(a)midsouth.rr.com > > webmaster - http://www.mschess.org > > > > > _______________________________________________ > 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

RE: [Zope] Very important for CatalogWare zopists
by Brian Lloyd 19 Jan '00

19 Jan '00

> I have submitted this to the Collector and it is major for people who > have items in their ZODB that are Catalog aware. > > To create the bug follow these steps: > > 1.Create a Catalog Aware ZClass > 2.Create an instance of that class > 3.Make sure it is indexed by looking in the Catalog "Catolaged items" > tab > 4.Notice that there is *no* slash in front of the indexed url: > "MyFolder/myobject" > 5.Now clear the Catalog > 6.Go to the "Find Items to Catalog" > 7.Select just your New ZClass meta type so it is quick > 8.Click "Find" > 9.Make sure it is indexed by looking in the Catalog "Catolaged items" > tab > 10. Oh no! It has a different url. A leading slash! > "/MyFolder/myobject" > 11. Click on the link. It's broken! Hi Jason - these steps are equivalent to what is currently done on Zope.org (which is running on a checkout equivalent to 2.1.2), and this does not seem to be a problem there. While I won't totally rule out the possibility of something in Zope, the fact that we don't see this behavior seems to suggest that there might be something going awry in your particular ZClasses. In particular, I'd look for some difference in the way that it registers itself with the catalog at the time that it is created versus when it is "found" by the catalog. The difference could be a difference in either behavior, state or environment (i.e. is the acquisition context, etc. the same in both cases when the "absolute_url" method gets called to generate the url for the catalog? Is there any custom code around this logic?). The object constructor is a good place to start looking for this sort of thing - try to trace through the activity in the constructor versus what happens when the existing object is "found" by the catalog. If possible, try to see what the value of the url given to the catalog is what you expected it to be... > What this means is that if you have a lot of items in you > Catalog don't > ever ever ever use the "Find" machinery until Digital fixes this(It > worked in 2.0.0 Final). You will end up not being able to > edit/reindex > your objects, and the links in your applications may not > function. I have not heard any other reports of a problem like this, so until we can verify that it is a Zope problem I think that this is a little strong. If other Zopistas can reproduce this I would certainly like to hear about it though. Brian Lloyd brian(a)digicool.com Product Manager 540.371.6909 Digital Creations http://www.digicool.com

6 8

IMAP Server
by Michel Pelletier 19 Jan '00

19 Jan '00

Brad, Hi, I think I recall you being interested in writing an IMAP server in Python. I have begun writing an IMAP server, and I was curious if you'd like to join me in the development. I have an IMAP server based on asychat, my own protocol handler, and ZODB. Zope itegration is imminently planned. Below is the readme. I'll be setting up a site on starship tonight/tommorow to host the project. Are interested? -Michel Test: What the hell should I call this thing? This thing is an IMAP server written in Python. It consists of a simple asychat dispatcher, a generic protocol handler class, some mail objects, and an abstract backend storage mechanism. This thing requires the following software: asyncore/asynchat - Standard with Python. Also comes with ZServer (Zope) ZODB - The default namespace storage for this thing requrires ZODB. The backend is totaly plugable however, so I don't see any reason why some other mapping storage cannot be used (like gdbm, sleepycat berkeley, etc...) Currently, ZODB is not shipped seperatly from Zope, so this thing requires Zope to run. In fact, this thing hijacks quite a bit of Zope. To get it fired up, ungzip and untar the file into your Zope top level directory. edit lib/python/start_imap.py, change the path.append to reflect your medusa directory. I know, this is an ugly hack, bear with it for now. A Sample interaction (this is for real! not a simulation!): [michel@localhost python]$ telnet localhost 143 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. * OK IMAP4rev1 v11.241 Welcome 1 capability * CAPABILITY IMAP4 IMAP4rev1 NAMESPACE AUTH=LOGIN XUNDO 1 OK CAPABILITY completed 2 login mike pass 2 OK LOGIN complete 2 create aNewBox 2 OK create completed 3 select aNewBox * 0 EXISTS* 0 RECENT* [UNSEEN 1] Messages 1 is first unseen * [UIDVALIDITY 1] UIDs valid * FLAGS (\Flag1 Flag2)* [PERMANENTFLAGS %s] Limited 3 OK [READ-WRITE] SELECT completed 5 close aNewBox 5 OK close completed 6 namespace * (("~" "/")) (("Public/" "/")) (("" "/")) 6 OK namespace completed 7 logout * BYE BYE logging out 7 BYE LOGOUT completed Connection closed by foreign host. Because it is based on really cool peices of software, this thing has a number of benefits. High performance -- Ok that's not proven yet, but asyncore in general has been proven pretty speedy and the select() model is regarded as a high performance solution. Further, then entire enchelada, if Zope itegration is not used, runs in one process space. It can even run in the same ZServer thread space as a Zope server. ZODB database cache keeps commonly used objects in memory. Also, the entire IMAP mailbox namespace is used as keys to the ZODB dict, so the namespace is kept in memory at all times. This makes listing and searching through mailbox names very fast. Easy possbile Zope itegration -- The storage backend for this server can be realized by a simple layer over a mapping storage like object. So gbdm, berkley, or a bunch of Zope objects can be used to store mail objects. If Zope were used, it would be possible for mail and mailbox objects to have web interfaces, and either interface, IMAP or HTTP (or XML-RPC, WebDAV and FTP) can manipulate the SAME objects. It's the same old classic dream of multi-protocol publishing realized once again in IMAP. This has not been thought out much, specificly how the "security system" (tm) in this thing can interface with Zope's. In the end, it might require medling the entire thingy into Zope's framework, but I want to avoid that if possible. ZODB provides plugable storage For now, this thing depends on ZODB for storage. This has lots of benefits, for one, ZODB supports pluggable storage formats. At the moment there are several, like FileStorage, BerkeleyStorage, DemoStorage and ReadOnly. Digital Creations will likely in the future add other storage types, and this thing might as well let them do all the work. Fine grained multi-tasking Every client request and every response gives asyncore

3 2

tree tag, expand_all and branches.
by Anthony Baxter 19 Jan '00

19 Jan '00

Has anyone started/finished work on making the tree tag's 'expand_all' functionality work when you have 'branches_expr' but not 'branches'? There's 4 different reports in the collector for this bug, but all are 'under review', so I can't see if there's any patches for these in any of them. I can't find anything in my mail about this except a message from Martijn Pieters (in his pre-DC days) last October that he thought this was fixed in CVS - it sure doesn't seem to be fixed right now. Anthony

1 0