Received: from sss.sss.pgh.pa.us (sss.pgh.pa.us [206.210.65.6]) by hub.org (8.9.2/8.9.1) with ESMTP id MAA12086 for ; Sat, 6 Mar 1999 12:21:42 -0500 (EST) (envelope-from tgl@sss.pgh.pa.us) Received: from sss.sss.pgh.pa.us (localhost [127.0.0.1]) by sss.sss.pgh.pa.us (8.9.1/8.9.1) with ESMTP id MAA04181; Sat, 6 Mar 1999 12:20:50 -0500 (EST) To: pgsql-sql@hub.org cc: pgsql-hackers@postgreSQL.org Subject: Re: Number of parameters in a sql function In-reply-to: Your message of Fri, 5 Mar 1999 23:00:26 -0500 (EST) <199903060400.XAA05262@hub.org> Date: Sat, 06 Mar 1999 12:20:49 -0500 Message-ID: <4179.920740849@sss.pgh.pa.us> From: Tom Lane "Stan Ng" writes: > I'm running into a limit problem with CREATE FUNCTION. Whenever I exceed > the limit on parameters, somewhere around 12 for me, psql bombs on me > with the "pqReadData() -- backend closed the channel unexpectedly" error > message. Is there any way to increase the parameter limit for > functions? It shouldn't be letting you get that high; the system limit is 8 parameters for a function. (It looks to me like parse_func.c should be, but is failing to, verify that the length of the parameter lists it is passed are within bounds --- or else gram.y itself should complain if the list is overlength.) Increasing the limit would be rather a major project I suspect. A quick look around the sources found a lot of dependencies, some being symbolic constants (not all the same one :-() and some being hard coded "8"s. Ick. Also, you would certainly be facing an initdb and database reload, because the number of argument-type columns in pg_proc would have to increase. It'd be good to clean this up and replace pg_proc's fixed-size proargtypes column with a variable-size "array of oid" column. Any volunteers? regards, tom lane