Re: [Pypgsql-devel] PostgreSQL 7.4 cursor oddness (solved)

[Mike C. F. <mcf...@ro...>]

Found it.  At some point I must have accidentally hit a delete key and 
out-dented a line in PgSQL.py while trying to track down a different error:

                if len(parms) == 1 and \
                   (type(parms[0]) in [DictType, ListType, TupleType] or \
                    isinstance(parms[0], PgResultSet)):
                    parms = _quoteall(self.__unicodeConvert(parms[0]))
                    self.res = self.conn.conn.query(_qstr % parms)
                else:
                    parms = self.__unicodeConvert(parms)
                    parms = tuple(map(_quote, 
self.__unicodeConvert(parms)));
                self.res = self.conn.conn.query(_qstr % parms)

so the query wound up being executed twice if a request met a particular 
requirement, and once otherwise.  The reason I couldn't reproduce the 
error in a script is that I was passing in the fully-expanded SQL 
queries in order to be sure I was executing exactly what was getting to 
the engine.  Those would always take the "else" branch and thus only 
create the cursor once (and, incidentally, this explains the 
previously-seen problem of mysterious duplicate inserts which triggered 
the whole update to PostgreSQL 7.4 in the first place).

Sigh.  Two days gone for a stupid reformatting/random keypress error.  I 
should beat myself about the head for that one.  Sorry for bothering 
everyone with this.  Night all,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/