Re: Service program signatures, and how to avoid unnecessarily recompiling programs that call service programs

Every service program I create has the following pieces: 1 (or more) modules, named the same as the service program (if more than one with an index number tacked on to the end), binder source named the same as the service program and stored in QSRVSRC, prototypes of exported procedures named the same as the service program and stored in QPROTOSRC, and a binding directory named the same as the service program. You will notice that all these parts are named the same as the service program. This helps keep them together. In addition to all the special purpose binding directories used to build my service programs, I keep a general purpose binding directory that contains all of my service programs. This is used to build all of my program objects.

Mark Murphy
Atlas Data Systems
mmurphy@xxxxxxxxxxxxxxx


-----Dan <dan27649@xxxxxxxxx> wrote: -----
To: "RPG programming on the IBM i (AS/400 and iSeries)" <rpg400-l@xxxxxxxxxxxx>
From: Dan <dan27649@xxxxxxxxx>
Date: 09/13/2017 07:58PM
Subject: Re: Service program signatures, and how to avoid unnecessarily recompiling programs that call service programs


I've made a notation in the binder language source indicating that all new
procedures should always be added to the end of the export list, and I've
included an explicit SIGNATURE parameter.

I specify this binder language source in the CRTSRVPGM command. I also
created a binding directory by the same name as the binder language source
member; the directory references the service program only. Is it typical
to name the binding directory the same as binder language source member? I
don't remember why I did that, whether I read it somewhere or made an
assumption on my part, and I'm not really sure I understand the
relationship between the two. Unless otherwise required, I think I am
inclined to name them differently, so that I don't confuse them. Thoughts?

- Dan