Dannes Wessels | 1 Apr 08:24 2010

Re: nullresource gets created when I create files through WebDAV

Hi,

On Wed, Mar 31, 2010 at 10:42 PM, Casey Jordan <casey.jordan <at> jorsek.com> wrote:
> I am getting some wierd behavior when I try and create files through the
> WebDAV interface. Sometimes it works fine, sometimes the file just contains:
>
> <!-- place holder for null byte sized nullresource XML document -->
> <nullresource/>
>
> and I cannot write to it.
>
> Any ideas why this might be happening?

This 'null resource' is generated because your webdav client attempts
to create a new document on the server. For this first a lock is
obtained on this non existing document, in a second phase the document
is actually written.

The problem is...... how to represent this  non existing XML document?
For binary resources this is not difficult, we use a 0 byte file. For
XML this is more difficult, hence we use the <nullresource/>
placeholder......

"I cannot write it"....... is this caused by the webDAV lock? The
exist-db log should give some hints. Webdav does not give too much
feedback....

cheers

Dannes
(Continue reading)

Dannes Wessels | 1 Apr 08:25 2010

Re: nullresource gets created when I create files through WebDAV

Hi,

On Thu, Apr 1, 2010 at 8:24 AM, Dannes Wessels <dannes <at> exist-db.org> wrote:
> On Wed, Mar 31, 2010 at 10:42 PM, Casey Jordan <casey.jordan <at> jorsek.com> wrote:
>> I am getting some wierd behavior when I try and create files through the
>> WebDAV interface. Sometimes it works fine, sometimes the file just contains:

You could help us by telling us what webdav client you use.

D.

--

-- 
eXist-db Native XML Database - http://exist-db.org
Join us on linked-in: http://www.linkedin.com/groups?gid=35624

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
raja38 | 1 Apr 11:49 2010
Picon
Picon

LDAP configuration change


Hi exist-db members !!!!
 I need to enable LDAP authentication in eXist-db . now i have enabled that
securityManager in conf.xml . then what i have to do . how to test that..

By
Rajamani marimuthu
XForms team
--

-- 
View this message in context: http://old.nabble.com/LDAP-configuration-change-tp28106491p28106491.html
Sent from the exist-open mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
Dannes Wessels | 2 Apr 14:45 2010

Re: LDAP configuration change

Hi,

On Thu, Apr 1, 2010 at 11:49 AM, raja38 <m.rajamani <at> nic.in> wrote:
>  I need to enable LDAP authentication in eXist-db . now i have enabled that
> securityManager in conf.xml . then what i have to do . how to test that..

probably you need to read some documentation first....

regards

Dannes

--

-- 
eXist-db Native XML Database - http://exist-db.org
Join us on linked-in: http://www.linkedin.com/groups?gid=35624

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
John Craft | 2 Apr 16:43 2010

Cumulative Scoring with Lucene Index

I suspect there is no way to do this without getting creative and complicating things, but is there a way to
get the Lucene full-text index to cumulatively score a "document" based on several different criteria? 
An example might better illustrate what I would like to do:

Given a <section> element that contains <title>, <chapter-title>, and one or more <p> elements, I would
like to add to the score of the <section> if there are search hits in <title> and/or <chapter-title>.  The
<section> would have a base score based on the number of matches within the <p> elements.  Added to that
score would be the score from matches in <title>, which should be boosted by some factor to give them more
weight.  Finally, the <chapter-title> score would be added to the running total for the <section> and
matches in <chapter-title> would be boosted by a larger factor than <title>.  The final score for the
<section> "document" would be a total of the three scores: Lucene score for <section> + Lucene score for
<title> multiplied by 2 + Lucene score for <chapter-title> multiplie
 d by 5.

I have tried indexing <section> and the two titles separately, similar to the documentation:

<text qname="section"/>
<text qname="title" boost="2.0"/>
<text qname="chapter-title" boost="5.0"/>

However, this clearly won't work because this creates three "documents" for each <section> element in the
source XML.  I would like to somehow combine the scores easily, if possible, and treat each section + title +
chapter-title as a single entity.  If this isn't possible, I assume I could index the three elements,
search across all three of them, and then iterate over the results and combine them using an  <at> id attribute
to sort of join the search hits back together.  Manually combining the scores sounds like it could be a slow
and tedious process.

Has anyone else tried to achieve the same thing?   Does anyone have a better idea or know of a way to leverage
Lucene in a way I'm not seeing (I am new to Lucene and know little about it).

(Continue reading)

Lucas Jarmin | 2 Apr 17:11 2010
Picon

Indexing A Collection Twice?

Hello!  I was hoping you all could help us out with a conundrum we're having.  We're using the Lucene searching in eXist to search a corpus of just over a million words.  What it seems we want to do is index that same corpus twice, using two different analyzers.  The SimpleAnalyzer works great for most of our needs and covers probably 85% of the use cases--i.e. just searching for words.  But, we also want to have the ability to search for numbers, punctuation and the combination of those with words (for example, a user may want to search for a word followed by an exclamation point).
 
How can we do this?  It seems that the WhitespaceAnalyzer allows us to do those things, but it doesn’t work as well for the normal 85% of the searches, and it doesn't use LowercaseFilter, anyway.  Can we index the same collection twice and use some logic to decide which one it goes to?  Could we do this by implmenting just one custom analyzer?  If so, how do we do that?  I see little documentation about how to make a custom analyzer and then plug it in to everything else.
 
Thanks!
 
 
God Bless,
Lucas Jarmin
Email: ljarmin <at> umich.edu
Twitter: <at> ljarmin
University of Michigan School of Information
MSI Candidate, 2010
------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Exist-open mailing list
Exist-open <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/exist-open
Jacob Myers | 2 Apr 23:46 2010
Picon

Lucene Analyzer

I have recently found some spare time to work on my little phonetic
analyzer/filter set for the lucene stuff and I have come across an
uncertainty. I couldn't find a reference to this in the documentation
or anything, but I am curious if there is any way in the index
configuration to pass information into the analyzer that gets created.

What I am wondering is basically is there any way to do something like:

<analyzer  class="org.apache.lucene.analysis.standard.StandardAnalyzer"
type="Something" />

and be able to get at the type from within my analyzer. I am actually
guessing there isn't really a way to do this unless there were
standard options available or something. So assuming that it isn't
available what is the recommendation for doing configuration on an
analyzer from the index? Just creating another analyzer that does that
exact combination of options or something?

Just wanting to make sure I don't do something absurd when there is an
easier solution to the problem.

Thanks,
Jacob Myers

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
Michael Sokolov | 3 Apr 05:05 2010

Re: Cumulative Scoring with Lucene Index

I think you should be able to achieve this by combining multiple query
"filters" with "or":

//section[ft:query(title, 'term') or  ft:query(., 'term')]

-Mike

> -----Original Message-----
> From: John Craft [mailto:jcraft <at> jonesmcclure.com] 
> Sent: Friday, April 02, 2010 10:44 AM
> To: exist-open <at> lists.sourceforge.net
> Subject: [Exist-open] Cumulative Scoring with Lucene Index
> 
> I suspect there is no way to do this without getting creative 
> and complicating things, but is there a way to get the Lucene 
> full-text index to cumulatively score a "document" based on 
> several different criteria?  An example might better 
> illustrate what I would like to do:
> 
> Given a <section> element that contains <title>, 
> <chapter-title>, and one or more <p> elements, I would like 
> to add to the score of the <section> if there are search hits 
> in <title> and/or <chapter-title>.  The <section> would have 
> a base score based on the number of matches within the <p> 
> elements.  Added to that score would be the score from 
> matches in <title>, which should be boosted by some factor to 
> give them more weight.  Finally, the <chapter-title> score 
> would be added to the running total for the <section> and 
> matches in <chapter-title> would be boosted by a larger 
> factor than <title>.  The final score for the <section> 
> "document" would be a total of the three scores: Lucene score 
> for <section> + Lucene score for <title> multiplied by 2 + 
> Lucene score for <chapter-title> multiplied by 5.
> 
> I have tried indexing <section> and the two titles 
> separately, similar to the documentation:
> 
> <text qname="section"/>
> <text qname="title" boost="2.0"/>
> <text qname="chapter-title" boost="5.0"/>
> 
> However, this clearly won't work because this creates three 
> "documents" for each <section> element in the source XML.  I 
> would like to somehow combine the scores easily, if possible, 
> and treat each section + title + chapter-title as a single 
> entity.  If this isn't possible, I assume I could index the 
> three elements, search across all three of them, and then 
> iterate over the results and combine them using an  <at> id 
> attribute to sort of join the search hits back together.  
> Manually combining the scores sounds like it could be a slow 
> and tedious process.
> 
> Has anyone else tried to achieve the same thing?   Does 
> anyone have a better idea or know of a way to leverage Lucene 
> in a way I'm not seeing (I am new to Lucene and know little about it).
> 
> Thanks for your time and have a good Easter.
> 
> John Craft
> 
> --------------------------------------------------------------
> ----------------
> Download Intel&#174; Parallel Studio Eval Try the new 
> software tools for yourself. Speed compiling, find bugs 
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Exist-open mailing list
> Exist-open <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/exist-open
> 

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
Tucanette | 4 Apr 19:54 2010

Re: Multiple Trigger Problem

Hi,

Thanks a lot for your support!

I have a solution running and tested for about a month.
This is a summary to be used by other users having similar problems.

Unfortunately the solution proposed by Adam Retter did not work for me. 
The MasterTrigger does not return, the thread gets stuck.

What I did instead was to keep MyTrigger class unchanged.
I also kept the VersioningTrigger class unchanged.
In my solution the MasterTrigger extends the VersioningTrigger class and 
has an instance variable of MyTrigger class.

MyTrigger as well as VersioningTrigger can be tested and used separately
or together using the MasterTrigger.

Each of the methods prepare, configure, startElement and finish calls the method of the instance variable
first 
and then the same method of the super class - the VersioningTrigger class.

Below is the simplified code:
--
public class MasterTrigger extends VersioningTrigger {

	private MyTrigger myTrigger;
	
	public MasterTrigger() {
		super();
		myTrigger = new MyTrigger();
		}

	public void configure(DBBroker broker, Collection parent, Map parameters) throws CollectionConfigurationException{
		myTrigger.configure(broker, parent, parameters);
		super.configure(broker, parent, parameters);
	}

	public void prepare(int arg0, DBBroker arg1, Txn arg2, XmldbURI arg3,
			DocumentImpl arg4) throws TriggerException {
		myTrigger.prepare(arg0, arg1, arg2, arg3, arg4);
		super.prepare(arg0, arg1, arg2, arg3, arg4);
	}

	public void startElement(String namespaceURI, String localName, 
			String qname, Attributes attributes) throws SAXException {

		myTrigger.startElement(namespaceURI, localName, qname, attributes);
		super.startElement(namespaceURI, localName, qname, attributes);
	}

	public void finish(int arg0, DBBroker arg1, Txn arg2, XmldbURI arg3,
			DocumentImpl arg4) {
		myTrigger.finish(arg0, arg1, arg2, arg3, arg4);
		super.finish(arg0, arg1, arg2, arg3, arg4);
	}
}
--
In order to participate in a bigger application, 
it would be great to have a transaction manager as well as support for XA transactions.

Regards 

           Simone

-----Ursprüngliche Nachricht-----
Von: Adam Retter [mailto:adam <at> exist-db.org] 
Gesendet: Donnerstag, 4. März 2010 10:28
An: Tucanette
Cc: exist-open <at> lists.sourceforge.net; Evgeny Gazdovsky
Betreff: Re: [Exist-open] Multiple Trigger Problem

On 3 March 2010 09:44, Tucanette <tucanette <at> googlemail.com> wrote:
> Hi,
>
> currently I am using one trigger to ensure referential integrity beyond xml document boundaries.
> That trigger is written in Java.  Additionally I would like to have the versioning feature of exist-db in my application.
>
> I have a few questions:
> 1.      Evgeny is it possible to specify the sequence in which the triggers are fired?
>
> 2.      When will the branch with the new trigger design be available in the trunk or a release?
>
> 3.      Using the current release 1.4 of exist, how do I call from one master trigger the code of two other triggers?
>      Which method is being called, to ensure the life cycle of a trigger with calling prepare, configure,
finish ...?
>      Simply asked from a master trigger, how do I call the two other java trigger?

For your master Trigger create a Java Class which extends
FilteringTrigger, and then from each of the methods that you have to
implement, call the corresponding method on The Versioning Trigger and
your own Trigger.

e.g.

public class MasterTrigger extends FilteringTrigger {

private VersioningTrigger versioningTrigger = new VersioningTrigger();
private MyTrigger myTrigger = new MyTrigger();

public void configure(DBBroker broker, Collection parent, Map
parameters) throws CollectionConfigurationException {
    versioingTrigger.configure(broker, parent, parameters);
    myTrigger.configure(broker, parent, parameters);
}

etc. etc.

>
> Regards
>
>               Simone
>
>
> -----Ursprüngliche Nachricht-----
> Von: Adam Retter [mailto:adam <at> exist-db.org]
> Gesendet: Dienstag, 2. März 2010 14:46
> An: Tucanette
> Cc: exist-open <at> lists.sourceforge.net; Evgeny Gazdovsky
> Betreff: Re: [Exist-open] Multiple Trigger Problem
>
> If you look at the code here for the function createTrigger() -
> https://exist.svn.sourceforge.net/svnroot/exist/trunk/eXist/src/org/exist/collections/CollectionConfiguration.java
>
> It would seem to me that you can only register one trigger per-event.
> You have two triggers that you both want to response to the events
> "store" and "update".
>
> We should change the Triggers code ideally so you can have any number
> of triggers per event, I think Evgeny may have been attempting this in
> his triggers branch of the code base. I am not sure of the status of
> that, how well tested it is or when it will be integrated back into
> trunk - Evgeny?
>
> In the meantime, you could simply create a Trigger in Java that
> encapsulates those two triggers and passes off the events to them one
> after another.
>
> Cheers Adam.
>
> On 28 February 2010 23:02, Tucanette <tucanette <at> googlemail.com> wrote:
>> Hi,
>>
>>
>>
>> I would like to have the versioning trigger and also my own trigger to be
>> fired on the same update event.
>>
>> I am running exist db 1.4.
>>
>>
>>
>> My configuration looks like this:
>>
>> <?xml version="1.0" encoding="UTF-8"?>
>>
>> <collection xmlns="http://exist-db.org/collection-config/1.0">
>>
>>     <triggers>
>>
>>                         <trigger event="store,remove,update"
>> class="org.exist.versioning.VersioningTrigger"><parameter name="overwrite"
>> value="yes"/></trigger>
>>
>>                         <trigger event="store,update"
>> class="de.tucanette.myapp.trigger.ReferentialIntegrityTrigger"/>
>>
>>     </triggers>
>>
>>     <validation mode="yes"/>
>>
>> </collection>
>>
>>
>>
>> However, only the second trigger in the configuration list is being
>> executed.
>>
>> Separately configured each of the triggers works fine.
>>
>>
>>
>> Can someone help me solve the problem?
>>
>>
>>
>> Regards,
>>
>>   Simone
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> Exist-open mailing list
>> Exist-open <at> lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/exist-open
>>
>>
>
>
>
> --
> Adam Retter
>
> eXist Developer
> { United Kingdom }
> adam <at> exist-db.org
> irc://irc.freenode.net/existdb
>
>

--

-- 
Adam Retter

eXist Developer
{ United Kingdom }
adam <at> exist-db.org
irc://irc.freenode.net/existdb

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Exist-open mailing list
Exist-open <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/exist-open
Manfred Staudinger | 4 Apr 23:56 2010
Picon

How not to process an XSL PI in eXist

Hi,

When I use the browser to retrieve a document with
http://localhost:8080/exist/rest/db/index.xml
then the xslt stylesheet from the PI gets processed in exist.
- the serializer element in conf.xml has enable-xsl="no" specified (unchanged).
Anything in addition I should look at? I'm using eXist 1.4.0.

But when I add ?_xsl=no to the URI, the document gets transformed in
the browser as expected.

Regards,
Manfred

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev

Gmane