From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-0.8 required=5.0 tests=BAYES_00,INVALID_DATE
	autolearn=no autolearn_force=no version=3.4.4
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!lll-crg!ames!ucbcad!ucbvax!MITRE-BEDFORD.ARPA!rh
From: rh@MITRE-BEDFORD.ARPA (Rich Hilliard - Mitre Software Center)
Newsgroups: comp.lang.ada
Subject: Re: Restrictions on Ada Main Programs
Message-ID: <8612121635.AA28148@mitre-bedford.ARPA>
Date: Fri, 12-Dec-86 11:35:13 EST
Article-I.D.: mitre-be.8612121635.AA28148
Posted: Fri Dec 12 11:35:13 1986
Date-Received: Mon, 15-Dec-86 19:35:23 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The MITRE Corp., Bedford, MA
List-Id: <comp.lang.ada>


Last week, I posted a question on Ada main programs (see end of this
message).  Geoff Mendal has pointed out to me that I goofed, he
writes:

	The implementation is correct to reject your generic
	instantiation as a main program.  A careful reading of LRM
	10.1(8) shows that only a SUBPROGRAM that is a library unit
	can be a main program.  While 10.1(2) allows generic
	instantations to be library units, a subprogram is not the
	same as a generic instantiation.  Using the syntagmas in
	10.1(2) and the prose in 10.1(8) we find that only a "proper"
	subprogram (not an instantiation) can be a main program.


Now, more questions:

There ARE compilers which accept generic instantiations as main
programs - that's what got me wondering about this in the first place.
Are such implementations non-conforming?

What is the rationale for distinguishing generic instantiations from
"first-class" library units?



> Subject: Restrictions on Ada Main Programs 
> Date: Mon, 01 Dec 86 11:01:53 EST
> From: Rich Hilliard - Mitre Software Center <rh>
> 
> I always thought I understood Ada RM 10.1 (8):
> 
> 	... An implementation may impose certain requirements on the
> 	parameters and on the result, if any, of a main program (these
> 	requirements must be stated in Appendix F).  In any case,
> 	every implementation is required to allow, at least, main
> 	programs that are parameterless procedures, and every main
> 	program must be a subprogram that is a library unit.
> 
> 
> In particular, I thought "parameters" referred to run-time parameters
> - those supplied in a call.  Now I find an implementation that rejects
> generic instantiations of otherwise "parameterless" procedures owing
> to the presence of a generic formal parameter (see below) and citing
> this paragraph.  Is this a legitimate interpretation or merely an
> implementor error?  
> 
> 
> 	generic 
> 	  Default : Integer := 10;
> 	procedure Generic_Install;
> 
> 
> 	procedure Generic_Install is 
> 	  Site_Selectable_Parameter : Integer;  
> 	begin
> 	  Site_Selectable_Parameter := Default;
> 	end Generic_Install;
> 
> 
> 	with Generic_Install;
> 	procedure Install is new Generic_Install;
> 	--  	  ^^^^^^^
> 	-- RM 10.1(8): instantiation install can not be a main program
> 	-- A parameterless integer function or procedure is required