> I don't see an easy way to get around this, though, and it's not too
> concerning. Amusingly it does completely break the SQLSTATE and SQLERRM
> tests we added a few days ago :)
Yes, it's true. But with simple plpgsql parser isn't possible expect
miracles :). I think so there is some space for modifications -
read_sql_construct - add control for availability sql params. You can
solve all errors where is expr_until_semi.
This is exactly same situation like
create function ...
begin
execute 'select '||a;
end; $$
it generate runtime error
>
> BTW, another easy improvement in this area is changing the RAISE format
> string to allow it to be an expression, rather than only a string literal.
>
No problem. But it's maybe big change. And I don't see using format string
too much limiting.
Pavel
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
