Michael Glaesemann <grzm@seespotcode.net> writes:
> On May 16, 2006, at 14:17 , Tom Lane wrote:
>> [1] Mostly. There's the infamous continued-string-literal
>> construct...
> Are you referring to this?
> ^
> test=# SELECT 'foo'
> test-# 'bar'; -- valid
Yeah. It seems weird that a newline should be *required* there.
I think I understand why the SQL committee did that: they thoughtSELECT 'a' 'b'
was much too likely to be a syntax error and shouldn't be made into
a valid construct. But it's weird nonetheless, and awkward to
implement.
regards, tom lane