"Ian Bell" <ib@xxxxxxxxxxxxxxxxxxx> writes:
> I just tried your suggestion and it generates the same error. For
> verification, here is the function run in the debugger.
I just realized that "syntax error at position N" is *not* the exact
spelling you would get from a syntax error in PG's main parser, which
I think is what we were all assuming this'd trace down to somehow.
What it is is the (very poorly chosen) message you get if either
ltree_in() or lquery_in() doesn't like what it's been handed.
So basically, the problem here is that TestLtree() is being called
with a string that isn't a valid ltree representation. You haven't
shown us what you are doing on that end, but for example I can
duplicate the error message with
# select text2ltree(' ');
ERROR: syntax error at position 0
regards, tom lane
