- Zope-Dev - Zope lists

Re: [Zope-dev] TrackerBase Product fails (2.4.0b3, Tracker from CVS)
by Steve Alexander 27 Jul '01

27 Jul '01

Here's the diffs I forwarded onto the tracker maintainer. However, Tracker isn't a "supported product", and I guess updating the CVS version to work with 2.4 is some way down the priority list. When the CVS is open to checkins, the community will be able to take up some of this slack :-) The diff attached is not ideal, but it works. As Chris Withers pointed out on the recent Xron thread on zope-dev, I probably shouldn't be adding things from the SearchIndex module to the catalog. Rather, I should be adding instances of classes from the Products/PlugInIndexes product. That's easy to change, though. -- Steve Alexander Software Engineer Cat-Box limited

1 0

TrackerBase Product fails (2.4.0b3, Tracker from CVS)
by Romain Slootmaekers 27 Jul '01

27 Jul '01

Yo dudes, I tried to install the TrackerBase product on my 2.4.0b3 Zope version.... with the tracker from CVS (just got it this morning) the README is quite specific: 1) make sure the TrackerBase is available as a product the install dir is ../lib/python/Products/TrackerBase ok. 2) make a simlink in Extensions to TrackerMethods.py ok. 3) Copy Tracker.zexp to your import dir ok. ( 4) import Tracker.zexp from Control_Panel/Products (I first restarted the Server) ok ("Tracker successfully installed") Now I tried to add a Tracker object to a Folder, id=x title=y and I get this exception from the server: Error Type: TypeError Error Value: Catalog addIndex now requires the index type to be resolved prior to adding; create the proper index in the caller. Traceback (innermost last): File .../lib/python/ZPublisher/Publish.py, line 223, in publish_module File .../lib/python/ZPublisher/Publish.py, line 187, in publish File .../lib/python/Zope/__init__.py, line 226, in zpublisher_exception_hook (Object: RoleManager) File .../lib/python/ZPublisher/Publish.py, line 171, in publish File .../lib/python/ZPublisher/mapply.py, line 160, in mapply (Object: addTracker) File .../lib/python/ZPublisher/Publish.py, line 112, in call_object (Object: addTracker) File .../lib/python/Products/ExternalMethod/ExternalMethod.py, line 274, in __call__ (Object: addTracker) (Info: ((<Factory instance at 81267c8>, 'x', <h3>form</h3><table><tr>... Ok, since the error message is very consise, I suspect this to be a known problem .... So do I fix this ? TIA, Sloot.

1 0

TAL : here/title_or_id fail when place in root
by kent sin 27 Jul '01

27 Jul '01

When the page template is placed in root and called, it complain that title_or_id is not defined. But it is find to be called in other folder. To reproduce : create the template in root name index_html. Then call it in root http://localhost:8080/index_html It fail, but assume fldr is a folder, then http://localhost:8080/fldr succeed. __________________________________________________ Do You Yahoo!? Make international calls for as low as $.04/minute with Yahoo! Messenger http://phonecard.yahoo.com/

1 0

cgi.py vulnerability = opera multipart handling
by Christian Theune 27 Jul '01

27 Jul '01

Hi. It seems that my longrunning-unanswered problem with Opera and Zope (see earlier Posts) is the same as the cgi.py problem. My Question: These Requests, that cause the DoS, are they malformed or valid? I need to know because if they are malformed, i have to blame the opera team, if they are valid, i have to blame someone at ZC to fix it otherwise. Regards -- Christian Theune - ct(a)gocept.com gocept gmbh & co.kg - schalaunische strasse 6 - 06366 koethen/anhalt tel.+49 3496 3099112 - fax.+49 3496 3099118 mob. - 0178 48 33 981 reduce(lambda x,y:x+y,map(lambda x:chr(ord(x)^42),tuple('zS^BED\nX_FOY\x0b')))

3 4

View permission with DTML Methods & Documents
by Randall F. Kern 26 Jul '01

26 Jul '01

I must be losing my mind, I don't think this could possibly be broken in 2.4... 1) Create a DTML Method (or Document) 2) Goto it's security tab, and clear the View permission except for Managers (i.e. uncheck Acquire, and check Manager) 3) Close your browser (logging out) 4) Goto the URL for your new method With 2.4 on linux or Win2K, I can see the document (as anonymous). With 2.3, it raised 'Unauthorized', so a login box appeared. What's going on? -Randy

1 0

GADFLY Database: Can it lose data if the Zope process is aborted?
by Joachim Werner 26 Jul '01

26 Jul '01

Hi! We have a very strange case of data loss: A client of us is using a Gadfly DB to store some news items. He reported that he lost a couple of his most recent data entries when we restarted our servers after applying the latest Hotfix. I know that Gadfly is not to be used for serious applications etc., but I thought that this was because of performance issues, not because it can lose data. Has anybody had similar experiences? Our first tests say that Gadfly stores all INSERTs into the DB file right after they have been commited. So the only explanation for data loss would be some strange caching that is not always used ... Joachim

3 4

RE: [Zope-dev] ZClass Instances
by Tim McLaughlin 26 Jul '01

26 Jul '01

How about just using meta_type? on the zclass this'll be 'Z Class' and on an instance of the ZClass, it'll be that meta_type. _____________________________________________________ Tim McLaughlin iterationZERO - www.iterationzero.com 703-481-2233 > > Message: 6 > From: "Edgardo Zulian" <ezulian(a)elsitio.com.ar> > To: <zope-dev(a)zope.org> > Subject: [Zope-dev] ZClass Instances > Date: Wed, 25 Jul 2001 09:41:15 -0300 > > Hi, > > I have a zclass, which have a propertysheet with some properties like > multiple lines, strings, etc.. which have some default values that i > assigned to them. > When i create an instance of that zclass and i change the value of a > property, then the instance does not reflect any more the > changes that i > make to the zclass, which is okey. > > So the question is: > There is a way to know using an external method or with a dtml, if an > instance is different from the zclass? > > Thanks.. > > Edgardo

3 4

RE: [Zope-dev] GADFLY Database: Can it lose data if the Zope proc ess is aborted?
by sean.upton＠uniontrib.com 26 Jul '01

26 Jul '01

Hmm, I think it stores everything to flat files on disk, but it can only perform ops and has to keep the entire db in memory at run-time, which means you mave some problems if you have a 1GB database, for example, in that (I think) all the python processes' memory spaces running Zope would have to contain all the Gadfly tables, unless some sort of shared memory setup was used - either way, it's likely not ideal for anything larger than perhaps some commonly used lookup tables, at least that is what I remember reading at some point quite a long time ago... Sean -----Original Message----- From: Andy McKay [mailto:andym@ActiveState.com] Sent: Thursday, July 26, 2001 11:23 AM To: Joachim Werner; zope-dev(a)zope.org Subject: Re: [Zope-dev] GADFLY Database: Can it lose data if the Zope process is aborted? I thought gadfly didn't store data between restarts. Or am I losing it? Cheers. -- Andy McKay. ----- Original Message ----- From: "Joachim Werner" <joe(a)iuveno-net.de> To: <zope-dev(a)zope.org> Sent: Thursday, July 26, 2001 11:17 AM Subject: [Zope-dev] GADFLY Database: Can it lose data if the Zope process is aborted? > Hi! > > We have a very strange case of data loss: A client of us is using a Gadfly > DB to store some news items. He reported that he lost a couple of his most > recent data entries when we restarted our servers after applying the latest > Hotfix. > > I know that Gadfly is not to be used for serious applications etc., but I > thought that this was because of performance issues, not because it can lose > data. Has anybody had similar experiences? Our first tests say that Gadfly > stores all INSERTs into the DB file right after they have been commited. So > the only explanation for data loss would be some strange caching that is not > always used ... > > Joachim > > > _______________________________________________ > Zope-Dev maillist - Zope-Dev(a)zope.org > http://lists.zope.org/mailman/listinfo/zope-dev > ** No cross posts or HTML encoding! ** > (Related lists - > http://lists.zope.org/mailman/listinfo/zope-announce > http://lists.zope.org/mailman/listinfo/zope ) > _______________________________________________ Zope-Dev maillist - Zope-Dev(a)zope.org http://lists.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope )

1 0

Announce: TransactionAgents 0.0.1
by Steve Alexander 26 Jul '01

26 Jul '01

Hi folks, Please forgive the cross-posting. The discussion of Transactions and ZODB has spilled over a number of mailing lists. Just be thankful I'm not copying in the cmf list! Here's what I've been working on recently: http://www.cat-box.net/steve/TransactionAgents-0.0.1.tar.gz Inside the archive file are three products. I'm using all three, and they work for me. You'll need Python 2.1 and Zope 2.4. As usual, here's a warning: Do not use this stuff on important systems. Do not rely on the API staying the same. Do not rely on any of this finding its way into ZPatterns or Zope or ZODB. The products are as follows: PlugIns: Exactly what I found just now in CVS at eby-sarna.com ZPatterns: I took exactly what I found just now in CVS at eby-sarna.com, fixed one showstopper bug (see my message to the ZPatterns list) made a few changes for it to play nicely with Zope 2.4, and altered the parts of ZPatterns that use transactions to use the TransactionAgents product. For kicks, the version.txt of this release is ZedPatterns-0-4-3p2-transactionagents TransactionAgents: A runtime patch to ZODB.Transaction.Transaction to implement four types of Agent. 1: The Rule Agent. Runs on a commit 2: The Indexing Agent. Runs on a commit, after the Rule Agents are done. (*) 3: The Boundary Agent. Runs when the outermost transaction is done (either aborted or commited). 4: The Post-Commit Agent. Runs in its own transaction after the outer transaction is successfully committed. (*) Things get interesting when you commit a subtransaction from a rule agent or an indexing agent. It works; see the test suite for an example of what happens it works. An important thing is that exceptions in Rule Agents and Indexing Agents are not caught, but propagate just like they would in normal methods. Exceptions raised in Boundary Agents and Post Commit Agents get swallowed and ignored (except for a warning to the log). This is important because running Rules and updating Indexes are usually a vital part of the transactional logic of an application. (Someone please translate that into plain language!). See also the test suite in the tests/ subdirectory. The comments and docstrings there should help to explain what is going on. For ZPatterns enthusiasts, here's the release notes for ZedPatterns-0-4-3p2-transactionagents: ---- 26 July 2001 Welcome to ZedPatterns-0-4-3p2-transactionagents. This is an experimental and unofficial release of ZPatterns, to test out new ideas, and clear a path to compatibility with Zope 2.4 and Python 2.1. I've included an unaltered copy of the PlugIns product, as it is required by ZPatterns. UnTar the contents of the archive directly in your Products directory. Changes from ZPatterns 0-4-3p2: * SkinScript works with Zope 2.4. * Fixed __set_attr__ bug involving _v_parent * Requires Python 2.1, Zope 2.4. * Requires TransactionAgents product. * Agents now run just before transactions get commited rather than during a tpc_commit. * New SkinScript for running a trigger as an IndexingAgent. It goes like this:: WHEN OBJECT CHANGED INDEX CALLING expression Indexing agents run after normal agents/triggers (called Rule Agents), although there are special rules for when a subtransaction is committed while an agent is running. See the TransactionAgents product's test suite for details. * ZPatterns doesn't go into an infinite loop if you commit a subtransaction from a trigger. * Support for list comprehensions in SkinScript. * Per-transaction cached _v_ attributes now last until the outermost transaction has finished. This will improve execution speed by a miniscule amount in some circumstances. ---- Credit where it's due... The ideas behind TransactionAgents were discussed by Phillip Eby and Shane Hathaway a couple of months ago on the Zope-Dev mailing list. I used Casey Duncan's Hotfixes HowTo as a guide to writing the runtime patch. I read up on test suites from Michel P and Chris McD's ZDG. Phillip Eby and Ty Sarna came up with the original ZPatterns product, maintain the difinitive version of it, and are the ultimate arbiters of taste and decency as to what features and patches make it into a real ZPatterns release. All bugs and mistakes in ZedPatterns and TransactionAgents are mine alone. Patches and feedback welcome. -- Steve Alexander Software Engineer Cat-Box limited

1 0

Take the cgi-vulnerability patch serious!
by Joachim Werner 26 Jul '01

26 Jul '01

Hi! Normally I do not comment on security patches for Zope because they fix very minor issues. The recent patch announced on http://www.zope.org/Products/Zope/Hotfix_2001-07-25/security_alert is different. We tested the exploit script provided at sourceforge, and it immediately pushed any of our servers we tested it on to > 90% system load. With two or three calls of the script, any Zope server (including all other services running on the server) can be brought to a halt. So please take care of your servers! The exploit is posted with the bug report, and anybody who nows how to copy&paste and start a python script can use it to stop any Zope server in the world that is unprotected. Moreover, there seems to be an Opera bug that has the same effect ... Cheers iuveno AG Joachim Werner CEO

2 1