Andrew Dunstan wrote: > PF wrote: >> >> >> Well, it's tested, and doesn't crash. However, depending on the length >> param the first or second record will be replaced with binary junk >> and/or be blank. >> >> >
This time I have tested the patch, and it works well for me (tested against pg 8.1). The patch is against cvs HEAD from GBorg - is that still the right location? Can we use this as the basis to get rid of the length param altogether? IIRC we now insist on a libpq version that is modern enough to provide PQgetCopyData() which requires no length param. cheers andrew
dbdpatch
Description: Binary data
