Details
-
Bug
-
Resolution: Fixed
-
None
-
Critical
-
If an imported TBX was containing empty terms (which is basically non sense) and that term collection was then used for termtagging in asian languages, the termtagger was hanging in an endless loop and was not usable anymore.
-
-
Description
problem
Importing a TBX with the following fragment:
<termEntry> <langSet xml:lang="en"> <tig> <term>XXX</term> </tig> </langSet> <langSet xml:lang="ko"> <tig> <term></term> </tig> </langSet> </termEntry>
was producing the following TBX for termtagging:
<termEntry id="168240a7-bb28-44a8-ac45-6cca9e10b772"> <langSet lang="en"> <tig id="tig_bc84c8d2-1d95-4159-9282-481d691b8b17"> <term id="9aeecc50-0557-4d17-b68d-bc8214b4149a">XXX</term> <termNote type="normativeAuthorization">admittedTerm</termNote> </tig> </langSet> <langSet lang="ko"> <tig id="tig_bc84c8d2-1d95-4159-9282-481d691b8b17"> <term id="b778bc04-28ea-4886-b68e-9c59ceb0ab7f"/> <termNote type="normativeAuthorization">admittedTerm</termNote> </tig> </langSet> </termEntry>
which was then crashing the termtagger instance by setting into a dead loop (but not on importing the TBX, but on usage of it for tagging).
solution
Empty terms are filtered out now on TBX import and on TBX creation (for existing terms and if such is possible somehow via the termportal or at least by accident)