On Thu, Mar 7, 2013 at 7:36 AM, Michael Black <mdblac...@yahoo.com> wrote: > Personally I think this behavior is horrid. Is there some scenario where > this wouldn't be a latent bug?
I don't like it either. I also share Simon's (and yours!) opinion regarding your patch: if cast be fixed at all (and it shouldn't be, in SQLite3, for compatibility reasons) then it should be fixed to raise an error or return NULL, not 0. > Here's a patch against 3.7.14.1 which behaves the way it should IMHO > Though I think it should actually throw an error when not parseable > correctly. Right. But it's almost certainly too late to fix this in SQLite3. Maybe with a PRAGMA? The current behavior of CAST seems best suited for a function, not for CAST. > Why is this behavior allowed now? Someone from the SQLite team might be able to tell you, but my guess is that this is academic now. Nico -- _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
