Re: Is this a (Meridian) compiler bug?

[abbasi@smaug.enet.dec.com (Nasser ABBASI)]

In article <5385@ns-mx.uiowa.edu>, bonak@herky.cs.uiowa.edu (Esmail Bonakdarian) writes...
>It the following subtype declaration (line 3) legal?  The LRM seems to
>specify a type_mark after the 'is' reserved word (and an optional
>constraint).
> 
>The Meridian compiler compiles this code without any problems though the
>Telesoft compiler I have access to does not like it. I can fix the code by
>changing line 3 to
> 
>subtype ROW_TYPE is INTEGER range 1 .. ROWS;
> 
>Am I reading the LRM incorrectly or have I stumbled across a bug in this
>compiler? I would like somebody to confirm this for me before I contact
>Meridian (by the way, do they have an e-mail address?).
> 
>Thanks.
> 
>Esmail
>------------------
>    Meridian AdaVantage v4.1  Source file: bug.ada  Page 1
>    Sun Apr  7 1991 19:25:17
> 
> 
>        1  procedure BUG is
>        2
>        3  ROWS: constant := 5;
>        4
>        5  subtype ROW_TYPE is 1 .. ROWS;
>        6
>        7  begin
>        8    null;
>        9  end;
> 
>    Compilation Summary
>         Ada library used: ada.lib
>         Compile options:
>           (none)
>         9 lines compiled.
>         9 lines listed.
>         No errors detected.
>         No warnings reported.



The Above  should not compile successfully, VAX/VMS ADA compiler showes
error:



BUG    9-Apr-1991 01:31:21    VAX Ada V2.1-28                     Page   1

    1 	procedure BUG is
    2 	ROWS: constant := 5;
    3 	subtype ROW_TYPE is 1 .. ROWS;
............................1
%ADAC-E-FOUNDEXP, (1) Found number 1 when expecting one of { identifier string-literal character-literal }
%ADAC-I-IGNOREDECLL, (1) Declaration ignored due to syntactic errors

    4 	begin
    5 	 null;
    6 	 end;


/naser