On 01/11/12 14:32, Mark Brand wrote:
My point is that the proposed lint mode should worry more about 2) than
1).
I didn't think they are mutually exclusive. The class of probable errors
detected overlaps.
But the "missing" AS is not a probable error. It's standard and
well-established. So I should probably have said "..lint mode should
worry about 2) and ignore 1).."
... as Oracle at least doesn't allow it.
All lint mode stuff would be warnings not errors, so as a developer you
can choose which ones to ignore, black/whitelist or pay attention to and
cause a fault in your test suite.
Why should the "missing" AS even be a warning in lint mode? SQL-92 does
not mandate it. One of the most important implementations does not even
accept it. Do you know of any SQL implementation that requires an AS here?
Mark
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
