2017-09-07 0:57 GMT+02:00 Simon Slavin <[email protected]>: > > > On 6 Sep 2017, at 11:31pm, Cecil Westerhof <[email protected]> wrote: > > > 2017-09-07 0:20 GMT+02:00 Richard Hipp <[email protected]>: > > > >> On 9/6/17, Cecil Westerhof <[email protected]> wrote: > >> > >>> Maybe this is correct, but it is certainly confusing. > >> > >> The constraint check occurs before the implicit conversion. > > > > Should that not be the other way around? > > This is an important point. But I’d say not. Constraint checks are there > to make sure that the programmers are doing the Right Thing, not that the > DBMS is doing the Right Thing. So it’s the source value which is checked, > not the value stored in the database. > > To do it the other way around suggests that SQLite needs to check that > SQLite is doing the Right Thing, which would be a waste of cycles, and a > sign that the developers need to consult a psychiatrist. >
I would not see it as database checking, but that is just my opinion. ;-) But it would be a good idea to mention this. I just tested my assumption, but maybe someone else ‘knows’ he only has to check for real. -- Cecil Westerhof _______________________________________________ sqlite-users mailing list [email protected] http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
