Martijn van Oosterhout wrote:
<snip>
Just "setof record" will do. As for the implicit declaration of
variable names, that's harder. I don't know if you can do that without
making things very ambiguous.
I know setof record will do if I explicitly name OUT parameters.
But I want Postgre to figure out for himself what parameters to return
as out parameters.
I don't see why it would make things very ambiguous.
Postgre now allows only one function with same set of input parameters.
So if you have function1(in int, out varchar) of course you can't have
another function1(in int)
What it could lead to is problems at running that were not seen at
designing function.
You should get out what you put in, but it won't remember spacing
outside the function body because the lexer eats that.
I have a file with the function as I want it and edit that. That way I
can have the function exactly how I like it.
Have a nice day,
Yeah, I understand pg needs only the body to execute the function, and
having another field in
pg_catalog.pg_proc besides prosrc maybe seems like wasting space but it
would make my life
as a developer a lot easier.
Regards,
Rikard
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
