Nikki Locke wrote:
Joe Wilson wrote:
Looks reasonable. Consider filing a ticket with your patch.
A few "make test" tests that rely on the old short name behavior
for SELECT * with joins will report failure. Whether the old way is
necessarily correct in some tests is a matter of debate.
As will all production code which uses column names on selects with joins!
This patch had better not be included in the release code - perhaps it
could be enabled with a PRAGMA or something?
I can see, how this would break, when short_column_names flag is turned
on ( because in that case, no matter what the column names shouldn't be
prefixed with table names ) . This fix is not the right one, I guess !
- Sandeep.
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------