How to get a compilation error

How to get a compilation error

[Wilhelm Spickermann]

Hi,

I would like to get a compilation *error*, if some compile time 
constant is not suitable for the algorithm in use. Example:

Chunk_Bits : constant :=
  Integer'Min (32, System.Max_Binary_Modulus / 2);
type Check is (Checkitem1, Checkitem2);
for Check'SIZE use 1 - Integer'MIN (96 mod Chunk_Bits, 1);

The last two lines result in a compilation error, if and (I 
hope) only if 96 is not a multiple of Chunk_Bits. That´s what I 
want -- but it´s really ugly. Does anyone have something better?

Wilhelm

Re: How to get a compilation error

[Martin Dowie]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 950 bytes --]

"Wilhelm Spickermann" <wilhelm.spickermann@t-online.de> wrote in message
news:mailman.1005833842.603.comp.lang.ada@ada.eu.org...
Hi,

I would like to get a compilation *error*, if some compile time
constant is not suitable for the algorithm in use. Example:

Chunk_Bits : constant :=
  Integer'Min (32, System.Max_Binary_Modulus / 2);
type Check is (Checkitem1, Checkitem2);
for Check'SIZE use 1 - Integer'MIN (96 mod Chunk_Bits, 1);

The last two lines result in a compilation error, if and (I
hope) only if 96 is not a multiple of Chunk_Bits. That�s what I
want -- but it�s really ugly. Does anyone have something better?


Could try:

   type A_Valid_Range is range 0 .. 0;

   My_Check : constant A_Valid_Range :=
      A_Valid_Range (96 mod (Integer'Min (32, System.Max_Binary_Modulus /
2)));

This will probably raise a warning rather than an error though.
Some compilers will let you specify that warning are to be
treated as errors.

Re: How to get a compilation error

[Jean-Pierre Rosen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1491 bytes --]


"Martin Dowie" <martin.dowie@nospam.baesystems.com> a �crit dans le message news: 3bf3d68c$1@pull.gecm.com...
> "Wilhelm Spickermann" <wilhelm.spickermann@t-online.de> wrote in message
> news:mailman.1005833842.603.comp.lang.ada@ada.eu.org...
> Hi,
>
> I would like to get a compilation *error*, if some compile time
> constant is not suitable for the algorithm in use. Example:
>
> Chunk_Bits : constant :=
>   Integer'Min (32, System.Max_Binary_Modulus / 2);
> type Check is (Checkitem1, Checkitem2);
> for Check'SIZE use 1 - Integer'MIN (96 mod Chunk_Bits, 1);
>
> The last two lines result in a compilation error, if and (I
> hope) only if 96 is not a multiple of Chunk_Bits. That�s what I
> want -- but it�s really ugly. Does anyone have something better?
>
>
> Could try:
>
>    type A_Valid_Range is range 0 .. 0;
>
>    My_Check : constant A_Valid_Range :=
>       A_Valid_Range (96 mod (Integer'Min (32, System.Max_Binary_Modulus /
> 2)));
>
> This will probably raise a warning rather than an error though.
> Some compilers will let you specify that warning are to be
> treated as errors.
To get an error, do the following:
subtype Assert is boolean range True..True;
Check : constant := Assert'Pos(96 rem Chunk_Bits = 0)

Since Check is a named number, an invalid value will result in a compile time error (new in 95).

--
---------------------------------------------------------
           J-P. Rosen (rosen@adalog.fr)
Visit Adalog's web site at http://www.adalog.fr

Re: How to get a compilation error

[Wilhelm Spickermann]

--On Donnerstag, November 15, 2001 16:09:50 +0100 Jean-Pierre 
Rosen <rosen@adalog.fr> wrote:

> To get an error, do the following:
> subtype Assert is boolean range True..True;
> Check : constant := Assert'Pos(96 rem Chunk_Bits = 0)
>
> Since Check is a named number, an invalid value will result in
> a compile time error (new in 95).

Hmm, I didn´t get a compilation error with gnat and I think 
thats correct. 0 is a valid value for Check and False is a valid 
parameter for Assert´POS as it takes Assert´BASE parameters 
according to ARM 3.5.5(3).

Wilhelm

Re: How to get a compilation error

[Wilhelm Spickermann]

--On Donnerstag, November 15, 2001 15:06:41 +0000 Martin Dowie 
<martin.dowie@nospam.baesystems.com> wrote:

> Could try:
>
>    type A_Valid_Range is range 0 .. 0;
>
>    My_Check : constant A_Valid_Range :=
>       A_Valid_Range (96 mod (Integer'Min (32,
> System.Max_Binary_Modulus / 2)));
>
> This will probably raise a warning rather than an error though.
> Some compilers will let you specify that warning are to be
> treated as errors.

Yes, thank you. But it would be nice to be sure, that the guys 
using a package cannot get it to work, if it would fail in their 
environment.

Wilhelm

Re: How to get a compilation error

[martin.m.dowie]

> > Could try:
> >
> >    type A_Valid_Range is range 0 .. 0;
> >
> >    My_Check : constant A_Valid_Range :=
> >       A_Valid_Range (96 mod (Integer'Min (32, System.Max_Binary_Modulus
/
> > 2)));
> >
> > This will probably raise a warning rather than an error though.
> > Some compilers will let you specify that warning are to be
> > treated as errors.
> To get an error, do the following:
> subtype Assert is boolean range True..True;
> Check : constant := Assert'Pos(96 rem Chunk_Bits = 0)
>
> Since Check is a named number, an invalid value will result in a compile
time error (new in 95).

Ooohhh! That's good!!!

Don't you just love this language!

Re: How to get a compilation error

[Jean-Pierre Rosen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 852 bytes --]


"Wilhelm Spickermann" <wilhelm.spickermann@t-online.de> a �crit dans le message news:
mailman.1005881899.15799.comp.lang.ada@ada.eu.org...
>> To get an error, do the following:
>> subtype Assert is boolean range True..True;
>> Check : constant := Assert'Pos(96 rem Chunk_Bits = 0)
>>
>> Since Check is a named number, an invalid value will result in
>> a compile time error (new in 95).
>
>Hmm, I didn�t get a compilation error with gnat and I think
>thats correct. 0 is a valid value for Check and False is a valid
>parameter for Assert�POS as it takes Assert�BASE parameters
>according to ARM 3.5.5(3).
That's right... try:
Check : constant := Assert'Pos(Assert(96 rem Chunk_Bits = 0))


--
---------------------------------------------------------
           J-P. Rosen (rosen@adalog.fr)
Visit Adalog's web site at http://www.adalog.fr

Re: How to get a compilation error

[Gautier Write-only-address]

>From: "Jean-Pierre Rosen" <rosen@adalog.fr>

>To get an error, do the following:
>subtype Assert is boolean range True..True;
>Check : constant := Assert'Pos(96 rem Chunk_Bits = 0)
>
>Since Check is a named number, an invalid value will result in a compile 
>time error (new in 95).

Interesting to observe that neither GNAT 3.13p nor OA 7.2.1
says anything against
------------------------------------------------------
procedure xxx is

  subtype Assert is Boolean range True..True;
  Check1 : constant := Assert'Pos(True) ;
  Check2 : constant := Assert'Pos(False) ;

begin
  null;
end;
------------------------------------------------------
;-)
______________________________________________________
Gautier  --  http://www.mysunrise.ch/users/gdm/e3d.htm

NB: Do not answer to sender address, visit the Web site!
    Ne r�pondez pas � l'exp�diteur, visitez le site ouaibe!


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

Re: How to get a compilation error

[Wilhelm Spickermann]

--On Donnerstag, November 15, 2001 16:09:50 +0100 Jean-Pierre 
Rosen <rosen@adalog.fr> wrote:

> Since Check is a named number, an invalid value will result in
> a compile time error (new in 95).

I´ve used this idea to make my solution less ugly. Instead of 
writing:

type Check is (Checkitem1, Checkitem2);
for Check'SIZE use 1 - Integer'MIN (96 mod Chunk_Bits, 1);

I´ve used now:

Check : constant := 1 / (1 - Integer'MIN (96 mod Chunk_Bits, 
1));

Thanks
  Wilhelm

Re: How to get a compilation error

[Wilhelm Spickermann]

--On Freitag, November 16, 2001 09:19:11 +0100 Jean-Pierre Rosen 
<rosen@adalog.fr> wrote:

> That's right... try:
> Check : constant := Assert'Pos(Assert(96 rem Chunk_Bits = 0))

Thats it! Thank You.
  Wilhelm

PS: I had tried this after Your first mail and thought it 
wouldn´t work, because gnat told me "non-static expression used 
in number declaration". Well, I should have read all the 3 error 
messages from gnat: it was only a follow-up-error of "static 
expression raises ""Constraint_Error""".

Re: How to get a compilation error

[Pi]

Jean-Pierre Rosen wrote :

> 
> "Wilhelm Spickermann" <wilhelm.spickermann@t-online.de> a �crit dans le
> message news: mailman.1005881899.15799.comp.lang.ada@ada.eu.org...
> >> To get an error, do the following:
> >> subtype Assert is boolean range True..True;
> >> Check : constant := Assert'Pos(96 rem Chunk_Bits = 0)
> >>
> >> Since Check is a named number, an invalid value will result in
> >> a compile time error (new in 95).
> >
> >Hmm, I didn�t get a compilation error with gnat and I think
> >thats correct. 0 is a valid value for Check and False is a valid
> >parameter for Assert�POS as it takes Assert�BASE parameters
> >according to ARM 3.5.5(3).
> That's right... try:
> Check : constant := Assert'Pos(Assert(96 rem Chunk_Bits = 0))
> 

or

Check : constant Boolean := Boolean'Pred(96 rem Chunk_Bits = 0);

or 

Check : constant := Boolean'Pos(Boolean'Pred(96 rem Chunk_Bits = 0));

no need for an Assert-subtype ;-)
Only that Check will always be false when it compiles.

TMTOWTDI

-- 
3,14159265359

Re: How to get a compilation error

[Wilhelm Spickermann]

--On Freitag, November 16, 2001 05:22:34 -0500 Pi 
<pi3_1415926536@yahoo.ca> wrote:

> Check : constant Boolean := Boolean'Pred(96 rem Chunk_Bits =
> 0);

Looks really good!

Thanks
 Wilhelm

> --
> 3,14159265359

--
13h25m11.6s,-11°09´41.0"

Re: How to get a compilation error

[Preben Randhol]

On Fri, 16 Nov 2001 21:39:53 -0800, R. Tim Coslet wrote:
>> This message is in MIME format. Since your mail reader does not understand
> this format, some or all of this message may not be legible.
> 
> --Boundary_(ID_jt2nYvWuvjIUcJVM7XMYIw)
> Content-type: text/plain; charset=ISO-8859-1
> Content-transfer-encoding: quoted-printable
> 
> The problem with this is it allocates a constant.
> 
> The named number solution allocates no memory for a constant in the
> executable.
> 
> --=20
>        R. Tim Coslet
>        r_tim_coslet@pacbell.net


Turn off HTML when you post to usenet.

Preben
--
                 �For me, Ada95 puts back the joy in programming.�