-<xlog server="chat.freenode.net" channel="#existdb" date="2008-02-05">

<message nick="wolf77" time="10:16:46.279Z">ljo and I were isolated on a different irc server</message>

<message nick="wolf77" time="10:17:26.794Z">perig_: if you have an expression:</message>

<message nick="DiZzZz" time="10:17:35.966Z">:-) xdrone as well?</message>

<message nick="DiZzZz" time="10:17:45.459Z">netsplit.... great</message>

<message nick="wolf77" time="10:18:01.941Z">let $ordered := for $n in ... order by $n/nn return $n</message>

<message nick="wolf77" time="10:18:09.592Z">return $ordered[1]</message>

<message nick="wolf77" time="10:18:35.577Z">should the positional predicate return the first item in the ordered sequence or the first one in document order?</message>

<message nick="wolf77" time="10:20:34.879Z">(right now eXist returns the first in document order)</message>

<message nick="perig_" time="10:23:37.938Z">Well, nearly sure the document order should disappear here</message>

<message nick="perig_" time="10:24:07.127Z">Have to theck, but that leads us back to ArrayNodeSet revival</message>

<message nick="perig_" time="10:33:13.439Z">mmh... can't reproduce ?</message>

<message nick="wolf77" time="10:33:43.64Z">I'm using the example posted by Sebastian Rahtz</message>

<message nick="wolf77" time="10:34:01.495Z">"puzzling error examining ..."</message>

<message nick="wolf77" time="10:34:35.26Z">it results in the well-known ClassCastException</message>

<message nick="perig_" time="10:37:54.331Z">yeah : trying to simplify that :-)</message>

<message nick="wolf77" time="10:38:36.538Z">I partially fixed it yesterday by re-writing the commented out section in OrderedValueSequence.</message>

<message nick="wolf77" time="10:38:56.255Z">but the original question remains</message>

<message nick="perig_" time="10:40:31.644Z">yes. I have understood , we cleanly have some problems here (see bug tracker) ; that's why I've deferred most of this stuff until we have a new architecture for in-memory nodes</message>

<message nick="perig_" time="10:41:25.047Z">BTW, I have a long standing patch in OrderedValueSequence</message>

<message nick="wolf77" time="10:42:51.294Z">would it help - for a start - if we generate node ids for the in-memory DOM? it would then be easier to map the stored nodes to the in-memory ones.</message>

<message nick="perig_" time="10:44:05.555Z">O yes ! I think think it might be a good idea</message>

<message nick="wolf77" time="10:45:06.297Z">well, I can try that. we will need the code anyway if we want to make the two DOMs compatible.</message>

<message nick="perig_" time="10:49:04.821Z">Won't it make the merge complicated ? (I already think it would be difficult)</message>

<message nick="wolf77" time="10:49:31.828Z">Not sure. Adam will need this code to finish his stuff.</message>

<message nick="wolf77" time="10:49:51.305Z">I guess it will just be one additional method in DocumentImpl</message>

<message nick="perig_" time="10:50:07.927Z">Sure, but since I never see traunk ports to his branch...</message>

<message nick="wolf77" time="10:50:35.314Z">He will have difficulties indeed.</message>

<message nick="wolf77" time="10:58:24.632Z">and there's a 2nd issue here, this time related to context sequences:</message>

<message nick="wolf77" time="10:58:27.285Z">{$q[1]/tei:forename}: {count($q)}: {data($q[1]/@notBefore)}: {data($q[last()]/@notAfter)}</message>

<message nick="wolf77" time="10:58:50.717Z">the result of count($q) becomes the context sequence for the following data() call</message>

<message nick="wolf77" time="10:59:10.654Z">I would say this is certainly wrong, isn't it?</message>

<message nick="perig_" time="11:02:34.141Z">yes, we still have some issues in this area</message>

<message nick="perig_" time="11:05:11.524Z">... and I've been wondering if we shouldn't push/pop the context sequences</message>

<message nick="wolf77" time="11:07:16.124Z">well, in this case, the outer <p> element constructor uses a PathExpr for its content. And PathExpr will chain the context sequences. This is not desirable here though.</message>

<message nick="wolf77" time="11:08:52.805Z">so I will just add a flag to PathExpr to make it keep the outer context sequence</message>

<message nick="wolf77" time="11:32:06.658Z">let $ordered := for $a in $n/a order by $a/text() return $a</message>

<message nick="wolf77" time="11:32:13.678Z">return $ordered[last()]</message>

<message nick="wolf77" time="11:32:42.647Z">saxon returns <a>H</a></message>

<message nick="wolf77" time="11:33:08.512Z">so the positional predicate selects from the ordered set.</message>

<message nick="perig_" time="12:15:08.995Z">hehe, I wonder if the lastest patch hasn't some positive effect on the XQTS</message>

<message nick="wolf77" time="12:15:22.844Z">unfortunately not :-(</message>

<message nick="wolf77" time="12:15:53.175Z">but it also has no negative effect. there are just no matching tests. ;-)</message>

<message nick="perig_" time="12:19:13.377Z">really ? I would have though no... never mind</message>

<message nick="perig" time="15:17:20.575Z">wolf77's been killed ?</message>

<message nick="DiZzZz" time="15:17:48.213Z">hihihihihhi</message>

<message nick="wolf77" time="15:17:49.254Z">No, I can still see you</message>

<message nick="perig" time="15:18:00.359Z">miracle ! :-)</message>

<message nick="DiZzZz" time="15:18:04.645Z">"wolf77 [n=XIRCProx@62.232.6.162] has quit: "Killed by sagan.freenode.net (Nick Collision)"</message>

<message nick="wolf77" time="15:19:13.168Z">hehe, ljo and I were sitting on this node the whole day</message>

<message nick="wolf77" time="15:20:47.276Z">-l starts an embedded connection</message>

<message nick="wolf77" time="15:20:51.862Z">so -s is correct</message>

<message nick="perig" time="15:24:35.065Z">is Jayesh really subscribed ?</message>

<message nick="wolf77" time="15:25:49.617Z">I manually approved his last message ;-)</message>

<message nick="wolf77" time="15:26:07.911Z">maybe I should subscribe him by force...</message>

<message nick="perig" time="15:26:31.223Z">is "approved" the correct word ?</message>

<message nick="wolf77" time="15:27:16.234Z">well, the mailing list manager asks to "approve or deny" the request</message>

<message nick="wolf77" time="15:28:01.075Z">which doesn't mean I personally approve what he wrote ;-)</message>

<message nick="perig" time="15:29:17.249Z">Some fear the machines ; we now know that humans are worse :-)</message>

<message nick="perig" time="15:30:53.419Z">Well, I plan to introduce a configuration option that wouldn't bind deprecated functions. Would help the XQTS...</message>

<message nick="wolf77" time="15:32:36.87Z">I also saw in the XQTS we lack some checks for duplicate function names.</message>

<message nick="perig" time="15:33:24.825Z">yes : this part of the code needs to be reconsidered </message>

<message nick="perig" time="15:33:41.738Z">... but I think it might be better to wait for the new parser</message>

<message nick="wolf77" time="15:33:54.075Z">could be, yes</message>

<message nick="DiZzZz" time="15:43:31.168Z">I pinged him as well (Jayesh)</message>

<message nick="DiZzZz" time="15:44:06.717Z">double function names... there is a report available + test</message>

<message nick="DiZzZz" time="15:45:52.921Z">I have a confirmation about xmlprague ; no event this year. We'll be pinged in may for ideas for 2009 session (march 2009)</message>

<message nick="wolf77" time="15:47:20.39Z">ok, we have to organize something by ourselves then</message>

<message nick="newbie2" time="16:51:00.238Z">hello</message>

<message nick="newbie2" time="16:51:56.955Z">is these possible to run XQuery script with & by some replace sequence?</message>

<message nick="newbie2" time="16:52:16.301Z">for example &</message>

<message nick="newbie2" time="16:52:56.525Z">or it's forbidden by database at all</message>

<message nick="Mephisto" time="16:55:19.259Z">you actually mean you have a query with a string like .. "Tom & Jerry" that doesn't work?</message>

<message nick="Mephisto" time="16:58:17.982Z">"Tom &amp; Jerry"</message>

<message nick="newbie2" time="16:58:22.949Z">but i should mention that i have an old version of database</message>

<message nick="newbie2" time="16:58:52.479Z">i have an exception, something like: "Ampersand should be escaped ..."</message>

<message nick="newbie2" time="16:58:59.929Z">besides</message>

<message nick="newbie2" time="16:59:16.101Z">that script change current user password</message>

<message nick="newbie2" time="17:00:46.304Z">Mephisto: what do you so surprised? is that should work?</message>

<message nick="newbie2" time="17:00:47.275Z">java.rmi.RemoteException: query execution failed: expecting '"', found '&'</message>

<message nick="newbie2" time="17:00:57.709Z">java.rmi.RemoteException: query execution failed: XPST0003 : Ampersands (&) must be escaped (following character was not a name start character).</message>

<message nick="newbie2" time="17:01:07.743Z">return $l</message>

<message nick="newbie2" time="17:01:25.397Z">Mephisto: do you know something about it?</message>

<message nick="Mephisto" time="17:01:29.692Z">no, you don't need C-style escaping</message>

<message nick="Mephisto" time="17:01:44.603Z">try what i posted</message>

<message nick="newbie2" time="17:02:03.799Z">return $l</message>

<message nick="newbie2" time="17:02:20.572Z">the second script raise first exception</message>

<message nick="newbie2" time="17:03:26.521Z">another major thing</message>

<message nick="newbie2" time="17:03:32.87Z">i run script via SOAP</message>

</xlog>