BUG in DB XML? (when dealing with modules, user-defined functions)

Question

Hi,
this post can be related to Different results for the semantically the same queries but here examples are even simpler.
I have the module with one user-defined function:
module namespace tnt = "http://tntbase.mathweb.org/ns";
declare function tnt:filter-path($docs as document-node()*, $path as xs:string) as document-node()*{
  $docs[dbxml:metadata('tnt:path') = $path]
};Then I have a query:
import module namespace tnt = 'http://tntbase.mathweb.org/ns'  at 'file:/path/to/module/user-func.xq';
(: this variable is IMPORTANT in the query:)
declare variable $len := 3;
(: here goes the same function as in the module, but with different name :)
declare function tnt:filter-path2($docs as document-node()*, $path as xs:string) as document-node()*{
  $docs[dbxml:metadata('tnt:path') = $path]
(:the query itself :)
tnt:filter-path(collection("xml_content.dbxml"), substring("a:/doc.xml", $len)) If I execute the query I get *0* results, if I use function tnt:filter-path2 (instead of module's tnt:filter-path) then I get the expected *1* result.
Also I experimented a bit with a query and got the following observations (independent from each other):
1) If I substitute $len by value *3* or *3.0*, then I get right results in both cases, i.e. if the query itself looks:
tnt:filter-path(collection("xml_content.dbxml"), substring("a:/doc.xml", 3)) 2) If I change the module's function so that it uses value comparison instead of general comparison, then I get the right results in both cases as well:
module namespace tnt = "http://tntbase.mathweb.org/ns";
declare function tnt:filter-path($docs as document-node()*, $path as xs:string) as document-node()*{
  $docs[dbxml:metadata('tnt:path') eq $path]
};Please, help me out! I'm pretty sure it's a problem of DB XML and I'm struggling with it for a long time in this or that shape.
Thanks a lot in advance,
Vyacheslav

Answer 1

I upgraded from 2.5.13 to 2.5.16 and the bug disappeared. The reason why I didn't do it before was that I hadn't seen any relevant for me fixes in the release notes of 2.5.16 comparing to 2.4.13 (Actually I already have had two patches for two relevant issues in the 2.5.13 version)
Sorry for the false alarm.
Vyacheslav