On Thursday 29 October 2009 11:32:00 am Paul McNett wrote: > I went ahead and committed my patch, since it doesn't appear to break > anything and I think it is a step in the right direction. > > Actually, I think we need to bite the bullet and convert our INSERT and > UPDATE statements to send the parameters to the dbapi drivers, instead of > converting to string ourselves. Carl had a patch for that a couple years > ago, that couldn't be applied because right after he submitted it, we made > a lot of incompatible changes to dCursorMixin, and we never got an updated > patch from him. > > I think that BLOB support is important, and am surprised it's gone this > long without anyone noticing it's been broken.
Paul, Larry was able to get Dabo to accept the binary data. However, we followed the idea that each DB adapter had a "fixBinaryData()" method that made the correct conversion to allow Dabo to save/update/insert. Other code changes were done too. But my point is I still think in the end having a backend method is going to be a requirment. For example in the case of Postgres def fixBinaryData(self, val): return psycopg2.Binary(val) FireBird has something called BLOB_load() Oracle appears to several different types of binary data types each with different functions to store the data. MsSQL has several ways too. And so does MySQL. I don't how we can avoid the not using the special backend methods. Your thoughts? Johnf _______________________________________________ Post Messages to: [email protected] Subscription Maintenance: http://leafe.com/mailman/listinfo/dabo-users Searchable Archives: http://leafe.com/archives/search/dabo-users This message: http://leafe.com/archives/byMID/[email protected]
