Order of SUBSTR and UPPER in statement

Hi,

I encountered something I can't really explain. I use the following
statement in my application:

COALESCE(UPPER(SUBSTR("Y"."Firma",1,7)),'')

This returns "ERROR:  syntax error at end of input"


However, using the following statement is fine:

COALESCE(SUBSTR(UPPER("X"."Firma"), 1, 7), '')


The fieldtype of "Firma" is character varying.

The only difference is the order of UPPER and SUBSTR. Is it possible
that this changed during some PostgreSQL version update? By the way,
right now I'm using 8.2.

Regards,
Hermann