On Sun, Jul 18, 2004 at 09:51:58AM -0400, Jeff Urlwin wrote: > > > > > On Wed, Jul 14, 2004 at 09:59:40PM +0100, Tim Bunce wrote: > > > > > > > t/30long................DBD::Oracle::db ora_lob_append failed: > > > > ORA-00600: internal error code, arguments: [122231], [], > > [], [], [], > > > > [], [], [] (DBD ERROR: OCILobWriteAppend) at t/30long.t line 317. > > > > Um, that's actually using the OCI OCILobWriteAppend call > > because you're using an Oracle 9 client but you're talking to > > an Oracle 8 database. > > > > I think I'm just going to flag ora_lob_append() as not > > working for Oracle 8 databases and skip the test in that case. > > I'm wondering (out loud, at this point), if it's ok to have a one-time call > to determine the SERVER version at runtime and just adjust this dynamically. > I.e, if compiled for version 9, have a DBH level variable that detects the > server version and if < 9, just use our OCILobWriteAppend workaround and > move on. If >= 9, use Oracle's OCILobWriteAppend. This would be a one-time > penalty per call to the append... > > What do you think?
I'm nervous of the added complexity but I'd be happy to look at a patch :) Meanwhile I'll leave the test skipped if the client or server is v8. Tim.
