-<xlog server="chat.freenode.net" channel="#existdb" date="2008-04-01">

<message nick="perig" time="11:44:27.209+01:00">still too few applications for GSoC IMHO...</message>

<message nick="wolf77" time="11:45:20.811+01:00">Yeah, I just posted another announcement on the news</message>

<message nick="perig" time="11:46:36.785+01:00">There were 8 (including one categorized as "ineligible") last year, that resulted in 2 "slots"</message>

<message nick="mb211" time="12:52:21.522+01:00">hello</message>

<message nick="wolf77" time="12:52:49.831+01:00">Hi there</message>

<message nick="mb211" time="12:53:16.226+01:00">i have a question i hope someone might be able to help with</message>

<message nick="mb211" time="12:54:33.051+01:00">i can't get xdb:extract to work, i keep getting told it's "not defined in module namespace"</message>

<message nick="mb211" time="12:55:02.25+01:00">xdb:store, xdb:remove, etc, all work fine</message>

<message nick="wolf77" time="12:55:21.366+01:00">mmmh, I don't think there's a function xdb:extract... where did you find it?</message>

-<message nick="mb211" time="12:55:40.502+01:00">

<a href="http://exist-db.org/ant-tasks.html#N10452">http://exist-db.org/ant-tasks.html#N10452</a>

</message>

<message nick="mb211" time="12:56:06.728+01:00">that would explain it!</message>

<message nick="wolf77" time="12:56:30.47+01:00">ahhhh</message>

<message nick="wolf77" time="12:56:38.902+01:00">the ant tasks are not XQuery functions</message>

<message nick="mb211" time="12:57:51.791+01:00">is there something i can use from within the xquery code to extract a resource from the database?</message>

<message nick="wolf77" time="12:58:49.359+01:00">you can get a reference to a document with doc("/db/test/mydoc.xml")</message>

<message nick="wolf77" time="12:58:59.312+01:00">and then just return it from the query</message>

<message nick="wolf77" time="12:59:10.768+01:00">let $doc := doc("/db/test/mydoc.xml") return $doc</message>

<message nick="mb211" time="12:59:44.354+01:00">what i'm trying to do is create a CSV file from a query result</message>

<message nick="wolf77" time="13:00:03.99+01:00">a query can return plain text as well</message>

<message nick="mb211" time="13:00:49.602+01:00">yeah, the main problem is i was hoping to do it without using XSLT, so everything was in XQuery and output as HTML</message>

<message nick="wolf77" time="13:01:46.66+01:00">yes, that should be ok. so you don't want to return the CSV, but store it somewhere?</message>

<message nick="mb211" time="13:02:46.351+01:00">well it doesn't have to be stored</message>

<message nick="mb211" time="13:04:11.597+01:00">what i'm doing is building a web interface in XQuery with the eXist db. The db contains metadata records and the interface allows you to query these</message>

<message nick="mb211" time="13:05:11.097+01:00">the queries are generated from what the user does with the HTML form options. then it is output back to the page</message>

<message nick="wolf77" time="13:05:33.888+01:00">yes, that sounds ok</message>

<message nick="mb211" time="13:05:34.599+01:00">for each dynamically generated query i wanted the option of having the results as a CSV file</message>

<message nick="wolf77" time="13:05:48.919+01:00">ok, you can just return CSV instead of the html</message>

<message nick="mb211" time="13:06:05.922+01:00">so my initial idea was to create a link (in HTML) on the output page to a file</message>

<message nick="wolf77" time="13:06:06.232+01:00">just change the mime-type of the query result:</message>

<message nick="wolf77" time="13:06:24.807+01:00">declare option exist:serialize "media-type=text/text";</message>

<message nick="mb211" time="13:07:59.666+01:00">ok, but how do i output the csv to a new page? i will be displaying the results in an HTML table on the output page so i need the media-type as HTML</message>

<message nick="wolf77" time="13:08:55.863+01:00">ok, but that HTML page will contain a link to a different XQuery script which returns the cvs</message>

<message nick="mb211" time="13:09:43.597+01:00">that sounds like a good way to do it</message>

<message nick="mb211" time="13:12:52.684+01:00">one other thing i'd like to know ... is it possible to assign a variable to the return of a function? ie $myVariable = local:myfunction() ?</message>

<message nick="mb211" time="13:13:51.895+01:00">i want to be able to save the result of a user generated query, but i'm not sure how to do it (i'm very new to XQuery and all this, in case you hadn't guessed!!)</message>

<message nick="mb211" time="13:14:08.858+01:00">ok, great</message>

<message nick="wolf77" time="13:14:55.251+01:00">one approach to evaluate a user-generated query is to use the util:eval function</message>

<message nick="mb211" time="13:16:35.937+01:00">the application i'm making won't let users type their own XQueries though</message>

<message nick="wolf77" time="13:17:06.449+01:00">sure, but I guess you need to assemble a query from what they passed in via parameters</message>

<message nick="mb211" time="13:17:08.081+01:00">it presents them with HTML forms depending on the type of information they want to get</message>

<message nick="mb211" time="13:17:14.259+01:00">yeah, thats it</message>

<message nick="mb211" time="13:18:25.887+01:00">it took quite a while to create a simple XPath to a node from user input!</message>

<message nick="mb211" time="13:19:06.602+01:00">there's definitely a big jump from creating nice static queries to mapping them out from HTML forms!</message>

<message nick="mb211" time="13:20:03.409+01:00">thanks a lot for your help</message>

<message nick="mb211" time="13:20:32.419+01:00">it's great to be able to talk to someone who can help!</message>

<message nick="wolf77" time="13:20:40.229+01:00">welcome and good luck ;-)</message>

<message nick="mb211" time="13:20:57.783+01:00">thank you :-)</message>

<message nick="mb211" time="13:21:27.924+01:00">... i may have more questions down the line ;-)</message>

<message nick="wolf77" time="13:21:55.491+01:00">if we have time to answer, we will do it</message>

<message nick="mb211" time="13:23:42.537+01:00">great ... hoping i've not distracted you too much with random questions!</message>

<message nick="ljo" time="16:40:44.152+01:00">Off for home in a minute too. Hopefully my net connections is stable from now on again ...</message>

<message nick="perig" time="19:20:01.463+01:00">marpom : I'm Pierrick</message>

<message nick="marpom" time="19:20:41.787+01:00">hi Pierick :)</message>

<message nick="marpom" time="19:21:15.473+01:00">Yes, I think I have little experience comparing to other people that are already working on the plugin</message>

<message nick="marpom" time="19:21:39.996+01:00">I have not even started.</message>

<message nick="perig" time="19:22:04.67+01:00">what lets you think that ?</message>

<message nick="marpom" time="19:23:21.995+01:00">I do not know XQuery, I have written just a few lines. I think the project my be to demanding for me.</message>

<message nick="marpom" time="19:23:42.072+01:00">*might be</message>

<message nick="perig" time="19:24:44.126+01:00">well, I rarely earn 4500 USD without a little bit of demand from projects :-)</message>

<message nick="marpom" time="19:26:18.204+01:00">Oh yes, you are right about that. Sorry, are the two student already writhing the Eclipse plugin or they are applying to GSoC to write it?</message>

<message nick="marpom" time="19:26:56.346+01:00">*students</message>

<message nick="perig" time="19:27:22.812+01:00">wolf77 would tell you more... So far, we indeed have another Eclipse application, much less detailed than yours though...</message>

<message nick="wolf77" time="19:27:40.486+01:00">I know two people working on an eclipse plugin, but this is not in the context of GSoC</message>

<message nick="wolf77" time="19:28:04.128+01:00">and I don't see why their efforts shouldn't be combined with yours.</message>

<message nick="wolf77" time="19:28:57.31+01:00">you could concentrate on things they do not cover</message>

<message nick="perig" time="19:29:44.134+01:00">At least, you have another good reason to chat with the developers ;-)</message>

<message nick="wolf77" time="19:29:49.762+01:00">but well, it is true that we already had a few people here expressing their interest in an eclipse plugin</message>

<message nick="marpom" time="19:30:46.839+01:00">I see, excuse me for a while</message>

<message nick="wolf77" time="19:30:58.114+01:00">from my point of view, there are other projects which I would like to see covered</message>

<message nick="wolf77" time="19:31:43.546+01:00">e.g. the XQuery debugger. if you have Java experience ... I would very much like to see someone apply for the debugger ;-)</message>

<message nick="marpom" time="20:16:51.854+01:00">perig and wolf77: coming back to my GSoC application, I really would like to withdraw it. However, I would like to keep in touch with you. I hope I will still be welcome in your IRC channel.</message>

<message nick="wolf77" time="20:17:46.288+01:00">And we cannot convince you to work on something else? Like the XQuery debugger?</message>

<message nick="wolf77" time="20:19:36.248+01:00">You see, I really think your application wasn't bad at all ;-)</message>

<message nick="perig" time="20:21:21.04+01:00">So do I : it was becoming pretty interesting :-)</message>

<message nick="marpom" time="20:21:35.369+01:00">Thank you. That's nice to hear :) The project you mension is surely very interesting, however I have also other projects to do. As regards XQuery, I would not like to promise anything about it, because I don't know how much time it would take me to learn it.</message>

<message nick="marpom" time="20:22:26.459+01:00">I wrongly counted the time I have.</message>

<message nick="perig" time="20:22:49.991+01:00">I'm not even sure you have to be an XQuery guru to writ an XQuery debugger...</message>

<message nick="wolf77" time="20:23:48.06+01:00">Well, I don't think you need to know that much XQuery. But anyway, if you think it doesn't fit into your time plan... I surely understand that. I often have difficulties to calculate time requirements myself.</message>

<message nick="marpom" time="20:25:31.09+01:00">You are cool guys, thank you :) Thanks wolf77 for understanding my decision</message>

<message nick="perig" time="20:26:18.103+01:00">wouldn't be anybody interested at your university ?</message>

<message nick="marpom" time="20:26:41.605+01:00">I will post some info on our group forum.</message>

<message nick="marpom" time="20:28:15.573+01:00">For now I don.t know anybody else participating. Maybe there are some on other specializations. </message>

<message nick="perig" time="20:29:54.337+01:00">curiously enough, we have very few students from european universities...</message>

<message nick="perig_" time="21:50:09.927+01:00">How to enforce that with the new in-memory nodes ?</message>

<message nick="perig_" time="21:50:13.552+01:00">If $node, or the context item if the second argument is omitted, is a node in a tree whose root is not a document node [err:FODC0001] is raised.</message>

<message nick="wolf77" time="21:51:44.536+01:00">it would not have been possible with the old in-memory nodes as well</message>

<message nick="wolf77" time="21:53:08.059+01:00">btw., I just found that the checks for hasOne in ExtArrayNodeSet.add() render the whole idea of allowing duplicates for performance obsolete</message>

<message nick="wolf77" time="21:54:00.35+01:00">or wait, maybe not ... gotta recheck this</message>

<message nick="wolf77" time="21:54:26.496+01:00">no, should be ok. the performance leak has to be elsewhere.</message>

<message nick="wolf77" time="21:55:08.102+01:00">I'm currently hunting for the "bug" posted by Andrzej.</message>

<message nick="perig_" time="21:56:11.368+01:00">the performance one ?</message>

<message nick="wolf77" time="21:56:39.567+01:00">yes. a really bad one. a simple query takes 9 seconds while the more complex workaround completes in 200ms</message>

<message nick="wolf77" time="22:24:25.379+01:00">very strange... giving up for today</message>

<message nick="wolf77" time="22:25:25+01:00">btw, I guess the only way to detect "root is not a document node" would be to add a flag to DocumentImpl (notADocument)</message>

<message nick="wolf77" time="22:25:43.165+01:00">but I doubt it's worth the effort</message>

<message nick="wolf77" time="22:26:59.769+01:00">how else would you detect it?</message>

<message nick="perig_" time="22:27:08.521+01:00">that might be useful for the descendant-or-self axis</message>

<message nick="perig_" time="22:28:34.497+01:00">in fact I really wonder if that simply doesn't mean "constructed/computed nodes" and thus "in-memory nodes"</message>

<message nick="wolf77" time="22:29:33.578+01:00">yeah, if you create a fragment with document { }, the flag DocumentImpl.isADocument will be set, otherwise not </message>

<message nick="perig_" time="22:30:51.063+01:00">hehe : here is the discriminant info</message>

<message nick="perig_" time="22:31:33.951+01:00">as a general rule, is looks like we should have a kind of "wrapped" flag for in-memory nodes</message>

<message nick="wolf77" time="22:32:00.257+01:00">what does "wrapped" mean here?</message>

<message nick="perig_" time="22:32:09.399+01:00">for instance, an empty text node is perfectly valid unwrapped but must be ignored if wrapped</message>

<message nick="perig_" time="22:32:21.546+01:00">"contained" by comething else ;-)</message>

<message nick="wolf77" time="22:32:47.371+01:00">err, but the text node could "know" that by looking at the document object</message>

<message nick="wolf77" time="22:33:27.656+01:00">or rather: the document should know that it contains nothing but a text node</message>

<message nick="perig_" time="22:34:05.627+01:00">once attached yes. I was speaking about a top-down approach ; such a text node shouldn't be constructed if it knows it is wrapped. Petty tricky :-)</message>

<message nick="perig_" time="22:34:37+01:00">At least, that's what I probably experience with our constructor failures</message>

<message nick="wolf77" time="22:35:29.221+01:00">well, I saw some failures like that as well. but this is indeed not easy to fix - compared to other xqts failures we have.</message>

<message nick="perig_" time="22:37:44.875+01:00">the problem is that I sometimes fear possible DB corruptions, namely adjacent text nodes</message>

<message nick="perig_" time="22:38:05.793+01:00">those bugs are well identified though...</message>

<message nick="wolf77" time="22:39:59.328+01:00">anyway, I would first like to get the new in-memory processing stable enough so it could be propagated to a stable branch</message>

<message nick="wolf77" time="22:41:02.123+01:00">so please handle this part of the code like a raw egg ;-)</message>

<message nick="perig_" time="22:41:15.231+01:00">splash :-)</message>

<message nick="wolf77" time="22:41:53.623+01:00">hehehe</message>

<message nick="wolf77" time="22:42:39.395+01:00">okay, I have to leave. Good night!</message>

</xlog>