[Info-gnuprologjava] Problems with loading prolog-file (ensureLoaded); falling into recursion?

[Tim Hemig]

Hi There!

I want to load previously generated knowledge bases into the Environment. There 
is one general file, and after that I include different rulesets depending on 
what I find in the first import. Background is the following:
I want to load annotated text into prolog, and use it for textmining and 
further automated annotating. An example ruleset with a typical parent and 
grandparent relation does load, and I can consult it, so I think the mechanics 
are valid, I guess.
My problem-file is just a file full of facts about the different tiers of my 
annotated text, read from an XML file before (Exmaralda-file if interested).

But as soon as I import the generated code my java process does not come back 
from the call to environment.ensureLoaded(AtomicTerm.get("filename"));

The only difference between the two test files I use (one gets loaded, the 
other not) is that the negative one does enclose every Atom with ' around it. 
in the positive file I only use lower case words.

Are there known reasons for not returning from this call? I do not believe 
there is some endless recursion to blame, since I only want to parse facts. I 
also cannot work with error messages since there are none to find. the 
java-processes continue to use a good amount of system- and user-cpu-time until 
the process is killed.

Open for any suggestion, and pleas do not hesitate to request more specific 
information, if the situation is unclear.
Thank you in advance and best regards,
  Tim Hemig

PS: A short example of the doc/3 predicate. doc(X,Y,Z) reads "document has the 
token Z at index Y in tier/level X":

doc('TXT','0','Kirchenaustritt').
doc('TXT','1','Ein').
doc('TXT','2','selbstbestimmter').
doc('TXT','3','Kirchenaustritt').
doc('TXT','4','ist').
doc('TXT','5','nach').
doc('TXT','6','Vorstellung').
doc('TXT','7','der').
doc('TXT','8','Kirchen').
doc('TXT','9','nicht').
doc('TXT','10','wirksam;').