The following bug has been logged online:
Bug reference: 1619
Email address: matthias(dot)schinacher(at)gmx(dot)de
PostgreSQL version: 8.0.1
Operating system: Suse Linux 9.1 (x86)
Description: ECPG allows definition of binary cursors but can't
handle the result
I'm not certain this is a "bug", but seems a un-documented restriction of
one can use a binary cursor definition in a ECPG program, but there is no
(or no documented) way of fetching the data into host variables, as this
allways assumes a text representation.
thus for example, when fetching an integer type field from a binary cursor
into a 'int', the applied conversion implicitly assumes an empty string if
the first byte retrieved is '\0' and will assign a zero value to the 'int',
even if the binary representation received from the server has a non zero
interpretation (.e.g. like receiving the four bytes 00,FF,FF and FF).
maybe it would be a good idea to explicitly mention this in the manual as a
restriction and/or to generate
an error when ecpg encounters a binary cursor?
pgsql-bugs by date
|Next:||From: Nahuel Greco||Date: 2005-04-22 19:39:33|
|Subject: BUG #1620: triggers breaks with alter table (at least with plpythonu)|
|Previous:||From: Kris Jurka||Date: 2005-04-22 13:54:32|
|Subject: Re: BUG #1611: reading a date-field by "ResultSet.getTimestamp()"|