> On Tue, Feb 4, 2020 at 5:38 PM Doug <dougf....@comcast.net> wrote:
> > > You're twisting my point here. I obviously want the reverse,
> > > I want the database types to "drive" the binding done. 1-to-1.
> > > Because even if binding a different type would work, via
> SQLite's
> > > own implicit conversion, I don't want that, because it's
> hiding a
> > > bug in the code most likely instead. --DD
> > Is the code inadvertently putting quotes (') around in integer
> value [...]?
> I'm talking about "real" binding here:
> https://www.sqlite.org/c3ref/bind_blob.html
> In C/C++, you could mess up your col indexes when binding, or bind
> incorrectly for some other reason, and "strong static typing" is more
> likely to find those, via SQL failures, than SQLite's default
> flexible-typing, that accepts any value in any typed column,
> unless you have these explicit CHECK+typeof constraints. --DD

So you are talking about a bug in your code where you inadvertently called:
  sqlite3_bind_text(sqlite3_stmt*,int,const char*,int,void(*)(void*));
instead of
  sqlite3_bind_int64(sqlite3_stmt*, int, sqlite3_int64);
and you want SQLite to tell you about it.

I have a hard time seeing how you could make that kind of coding error, given 
the different parameters and types in the calls.


sqlite-users mailing list

Reply via email to