Position of "use"

Position of "use"

[Victor Porton]

What is the difference between

with X; use X;
package Y is
end Y;

and

with X;
package Y is
   use X;
end Y;

?

-- 
Victor Porton - http://portonvictor.org

Re: Position of "use"

[J-P. Rosen]

Le 05/07/2014 08:43, Victor Porton a écrit :
> What is the difference between
> 
> with X; use X;
> package Y is
> end Y;
> 
> and
> 
> with X;
> package Y is
>    use X;
> end Y;
> 
> ?
> 
A use clause (as well as any declaration) is usable from where it
appears up to the associated "end".

In the first case, it will allow short names for other with/use clauses,
while in the second case it will be effective only inside the package.

Personal note: it is quite rare to need use clauses in context clauses
(the first case), and my personal policy is to use use clauses, but
restricted to the innermost scope where they are useful. Therefore, I
tend to prefer the second solution.

-- 
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52, Fax: +33 1 45 29 25 00
http://www.adalog.fr

Re: Position of "use"

[Adam Beneschan]

On Saturday, July 5, 2014 12:12:59 AM UTC-7, J-P. Rosen wrote:
> Le 05/07/2014 08:43, Victor Porton a �crit :
> 
> > What is the difference between
> 
> > with X; use X;
> > package Y is
> > end Y;
> 
> > and
> 
> > with X;
> > package Y is
> >    use X;
> > end Y;

> > ?
> 
> A use clause (as well as any declaration) is usable from where it
> appears up to the associated "end".
> 
> 
> In the first case, it will allow short names for other with/use clauses,
> while in the second case it will be effective only inside the package.

Not true.  This is illegal:

    with Ada;
    use Ada;
    with Strings.Unbounded;
    package X is ...

You can't use this as a substitute for "with Ada.Strings.Unbounded".

The visibility rules for context clauses are separate from visibility rules in other cases.  At the moment, I am not sure there is any difference at all between Victor's two examples; I'll have to study the rules carefully to make sure.  It's possible that there is a difference if there is a child package (package X.Child is ...), or if there is a pragma in the context clause.  I'll try to look at it later.

                             -- Adam

Re: Position of "use"

[Pascal Obry]

Le lundi 07 juillet 2014 à 09:45 -0700, Adam Beneschan a écrit :
> The visibility rules for context clauses are separate from visibility
> rules in other cases. At the moment, I am not sure there is any
> difference at all between Victor's two examples; I'll have to study
> the rules carefully to make sure. It's possible that there is a
> difference if there is a child package (package X.Child is ...), or if
> there is a pragma in the context clause. I'll try to look at it later.

I suppose it is different if there is multiple units in a single file?

-- 
  Pascal Obry /  Magny Les Hameaux (78)

  The best way to travel is by means of imagination

  http://v2p.fr.eu.org
  http://www.obry.net

  gpg --keyserver keys.gnupg.net --recv-key F949BD3B

Re: Position of "use"

[Peter Chapin]

On 2014-07-07 13:08, Pascal Obry wrote:

> Le lundi 07 juillet 2014 à 09:45 -0700, Adam Beneschan a écrit :
>> The visibility rules for context clauses are separate from visibility
>> rules in other cases. At the moment, I am not sure there is any
>> difference at all between Victor's two examples; I'll have to study
>> the rules carefully to make sure. It's possible that there is a
>> difference if there is a child package (package X.Child is ...), or if
>> there is a pragma in the context clause. I'll try to look at it later.
> 
> I suppose it is different if there is multiple units in a single file?

In that case, doesn't each unit have it's own context clause even so?

with A;
package P1 is
  ...
end P1;

with B;
-- The 'with A' above is not relevant here (yes?)
package body P2 is
  ...
end P2;

Re: Position of "use"

[Adam Beneschan]

On Monday, July 7, 2014 10:08:37 AM UTC-7, Pascal Obry wrote:

> I suppose it is different if there is multiple units in a single file?

Absolutely not.  Ada's rules do not treat compilation units in the same file any differently than compilation units in different files.

                             -- Adam

Re: Position of "use"

[J-P. Rosen]

Le 07/07/2014 18:45, Adam Beneschan a écrit :
>> In the first case, it will allow short names for other with/use clauses,
>> > while in the second case it will be effective only inside the package.
> Not true.  This is illegal:
> 
>     with Ada;
>     use Ada;
>     with Strings.Unbounded;
>     package X is ...
> 
> You can't use this as a substitute for "with Ada.Strings.Unbounded".
Yes of course, I typed a bit too fast for "with", but it still applies
to "use" clause as context clauses.

-- 
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52, Fax: +33 1 45 29 25 00
http://www.adalog.fr

Re: Position of "use"

[Adam Beneschan]

On Monday, July 7, 2014 10:26:36 PM UTC-7, J-P. Rosen wrote:

> > Not true.  This is illegal:

> >     with Ada;
> >     use Ada;
> >     with Strings.Unbounded;
> >     package X is ...

> > You can't use this as a substitute for "with Ada.Strings.Unbounded".
> 
> Yes of course, I typed a bit too fast for "with", but it still applies
> to "use" clause as context clauses.

with Ada;  use Ada;
with Ada.Strings.Unbounded;
use Strings.Unbounded;
package X is ...

Still illegal, according to both GNAT and ICC Ada.

                             -- Adam

Re: Position of "use"

[Adam Beneschan]

On Friday, July 4, 2014 11:43:46 PM UTC-7, Victor Porton wrote:
> What is the difference between
> 
> with X; use X;
> package Y is
> end Y;
> 
> and
> 
> with X;
> package Y is
>    use X;
> end Y;
> 
> ?

OK, I've looked into it, and there's no difference.  I'm assuming that there may be more code in the specification of Y, but that in the second example, "use X" is the first thing in this specification.  

RM 8.4(6-7) describe the "scope" of a use_clause, which is the portion of code that is affected by the "use".  If it's in the context clause, the scope is the entire declarative region of the package.  If it's inside the declarative region, the scope is the part of the declarative region starting from the "use" and ending at the end of the declarative region; if the "use" clause is the first thing in the region (as in the second example above), that means that the scope is the entire declarative region of the package, same as the first example.  (Note that in both cases, the declarative region will include child packages, if any.  This follows from the rules in RM 8.1.)

The scope doesn't include the context clause itself.  (The context clause is the "with" and "use" statements that occur before the "package/procedure/function" of a top-level library unit, and it may include pragmas.)  RM 10.1.6 says that the visibility rules don't apply to the context clause, and it contains special rules for context clauses.  Nothing in those rules says that "use" makes anything visible in a context clause; therefore, a "use" in a context clause has no effect on any other "with", "use", or "pragma" in the context clause, or in any other context clause (i.e. the context clause on a child package or subunit). 

Therefore, the effect of a "use" that is the first thing in a package specification has the exact same effect as a "use" appearing in the context clause.

                           -- Adam

Re: Position of "use"

[Victor Porton]

Adam Beneschan wrote:
> On Friday, July 4, 2014 11:43:46 PM UTC-7, Victor Porton wrote:
>> What is the difference between
>> 
>> with X; use X;
>> package Y is
>> end Y;
>> 
>> and
>> 
>> with X;
>> package Y is
>>    use X;
>> end Y;
>> 
>> ?
> 
> OK, I've looked into it, and there's no difference.  I'm assuming that
> there may be more code in the specification of Y, but that in the second
> example, "use X" is the first thing in this specification.
> 
> RM 8.4(6-7) describe the "scope" of a use_clause, which is the portion of
> code that is affected by the "use".  If it's in the context clause, the
> scope is the entire declarative region of the package.  If it's inside the
> declarative region, the scope is the part of the declarative region
> starting from the "use" and ending at the end of the declarative region;
> if the "use" clause is the first thing in the region (as in the second
> example above), that means that the scope is the entire declarative region
> of the package, same as the first example.  (Note that in both cases, the
> declarative region will include child packages, if any.  This follows from
> the rules in RM 8.1.)
> 
> The scope doesn't include the context clause itself.  (The context clause
> is the "with" and "use" statements that occur before the
> "package/procedure/function" of a top-level library unit, and it may
> include pragmas.)  RM 10.1.6 says that the visibility rules don't apply to
> the context clause, and it contains special rules for context clauses. 
> Nothing in those rules says that "use" makes anything visible in a context
> clause; therefore, a "use" in a context clause has no effect on any other
> "with", "use", or "pragma" in the context clause, or in any other context
> clause (i.e. the context clause on a child package or subunit).
> 
> Therefore, the effect of a "use" that is the first thing in a package
> specification has the exact same effect as a "use" appearing in the
> context clause.
> 
>                            -- Adam

Possible reason why these two equal variants of syntax were introduced by 
Ada designers:

It would be different if we'd replace package with a procedure or a 
function, because use can apply or not apply to procedure parameters.

The same applies to generic packages.

-- 
Victor Porton - http://portonvictor.org

Re: Position of "use"

[anon]

Using a global "use_clause" can sometime cause the compiler to hide all 
routines of one or more libraries. It is better to place most use_clause 
in a local routine or within the declare section of a block_statement.

--
--  Example File: Test_2.adb
--
--  Showing where compiler could hide visibility access to packages.
--
with Ada.Text_IO ;  --  There are a number of routines that are 
with GNAT.IO ;      --  common in both packages so, placement of
                    --  the use_clause is critical, if used.

procedure body Test_2 is 

--  Both Ada.Text_IO and GNAT.IO packages must be access by using 
--  "selected_component" at this point

  procedure Test_2b is 
    use GNAT.IO ; --  Grants global visibility to package GNAT.IO. but
                  --  limits Ada.Text_IO to a "selected_component" 
                  --  access
    begin
      null ;
    end ;

  --  Allow all following routines will have visibility to 
  --  Ada.Text_IO could also cause visibility problems with other 
  --  packages.
  use  Ada.Text_IO ; 

  procedure Test_2a is 
      use GNAT.IO ; --  Will cause the compiler to hide packages, 
                    --  so in this example any reference must be by 
                    --  "selected_component" for both Ada.Text_IO
                    --  and GNAT.IO packages
    begin
      null ;
    end ;

begin
  null ;
end ;


For Ada compilers, where you can have more than one compilation_unit 
within one file (RM 10.1 -- 10.1.1), it allows all preceding 
compilation_unit to have access the same library with only one 
"with_clause" without sharing global visibility to that package.

Note: GNAT limits the compilation_unit to one per file.  But GNAT still 
allows this type of program design. Putting the "with_clause" in the 
specification file (.ads) allows access for the corresponding body file 
(.adb) in some cases without having the "with_clause" in the body file.


--
--  Example of a multiple compilation_unit File: Test_Lib.ada
--
--  Showing different access/visibility to package Text_IO

--
--  compilation_unit 1 : Access to Text_IO is forbidden in this 
--                       package
package Test_Init is
--  ...
end Test_Init ;
--
with Text_IO ;  -- Access is Allowed
--  use Text_IO ;  --  Visibility would be global to all
--                 --  specification and body units if this 
--                 --  statement is used.

--
--  compilation_unit 2 : Limits visibility to Text_IO to a 
--                       "selected_component", unless the 
--                       global "use_clause" is used.
--
--              example :  Text_IO.Col 
--
package Test_55 is
  procedure Init ;
--  ...
end Test_55 ;
--
--
--  compilation_unit 3 : Visibility of package depend upon 
--                       placement of the "use_clause"
--
package body Test_55 is

--  use Text_IO ;  -- Visibility is global within this package body

  procedure Init is
      use Text_IO ;  --  Visibility is local to procedure
    begin
      Set_Col ( 5 ) ; 
      Set_Line ( 2 ) ; 
    end Init ;

  procedure Error is
    begin
      declare
        use Text_IO ;  --  Visibility is local to this block statement
      begin
       Put ( "Error" ) ;
      end ;
    end Error ;
--    ...
end Test_55 ;
--

--
--  compilation_unit 4 : Limits visibility to Text_IO to a 
--                       "selected_component", unless the 
--                       global "use_clause" is used.
--
--              example :  Text_IO.New_Line ;
--
package Test_Cal is
--    ...
end Test_Cal ;



In <lp86r0$ssl$1@speranza.aioe.org>, Victor Porton <porton@narod.ru> writes:
>What is the difference between
>
>with X; use X;
>package Y is
>end Y;
>
>and
>
>with X;
>package Y is
>   use X;
>end Y;
>
>?
>
>-- 
>Victor Porton - http://portonvictor.org

Re: Position of "use"

[Adam Beneschan]

On Wednesday, July 9, 2014 3:36:09 AM UTC-7, an...@att.net wrote:

> For Ada compilers, where you can have more than one compilation_unit 
> within one file (RM 10.1 -- 10.1.1), it allows all preceding 
> compilation_unit to have access the same library with only one 
> "with_clause" without sharing global visibility to that package.
> 
> Note: GNAT limits the compilation_unit to one per file.  But GNAT still 
> allows this type of program design. Putting the "with_clause" in the 
> specification file (.ads) allows access for the corresponding body file 
> (.adb) in some cases without having the "with_clause" in the body file.

I'm not sure what you're trying to say here, but I want to reiterate: according to the Ada rules, putting multiple compilation units in the same file has *no* effect on the visibility rules (or any other rules).  If you have package specifications for two different packages, P1 and P2, in the same file, a "with" or "use" clause on P1 does not apply to P2, and vice versa.

A "with" clause on a specification of any package automatically applies to the body, and it also automatically applies to any child unit specifications or bodies.  This is due to the rules in RM 8.1 and has nothing to do with GNAT, and nothing to do with whether the specification and body are in the same file or not.

Thus, in this example, the comments are wrong:

> --
> --  compilation_unit 4 : Limits visibility to Text_IO to a 
> --                       "selected_component", unless the 
> --                       global "use_clause" is used.
> --
> --              example :  Text_IO.New_Line ;
> --
> package Test_Cal is
> --    ...
> end Test_Cal ;

Text_IO is not visible at all, since there is no "with Text_IO" on Test_Cal.  The "with Text_IO" earlier in the file, on "package Test_55", has no effect.

                          -- Adam

Re: Position of "use"

[anon]

Actually, the comments for compilation_unit 4 are correct under multiple 
compilation_unit type files, which GNAT does not support. It was tested 
under AdaEd-1.11.0a compiled for SuSE Linux, before I posted a condensed 
version.

Note: GNAT (Adacore or GNU) does not support multiple 
compilation_unit (s) within one file. 

For multiple compilation_unit, likes those files found in ACVC, the 
external visibility starts with the "with_clause" and that clause 
grants all succeeding compilation_unit(s) access to that package or 
routine no matter how many compilation_unit there are. In other words 
you only need one "with_clause" for Ada compilers that support multiple 
compilation_unit in the same file. 

But the direct visibility is allowed by the "use_clause". Exception is 
where two or more routines from two sources are the same. 

Example is GNAT.IO.Put ( S : String ) and 
           Ada.Text_IO.Put ( Item : in String )

In this case the compiler requires help to link to the correct package 
with requires a "selected_component" type statement to correctly 
identify the correct routine.


Under Ada the following should work but some versions of GNAT over the 
years have not match the formal_parameter of two or more routines 
correctly when using two unique packages.

     Put ( S => <string> ) ;    -- GNAT.IO
     Put ( Item => <string> ) ; -- Ada.Text_IO

so it better just to use 

           GNAT.IO.Put ( <String> ) -- or
           Ada.Text_IO.Put ( <String> )
or

           GNAT.IO.Put ( S => <String> ) -- or
           Ada.Text_IO.Put ( Item => <String> )


In <33fbd821-cb30-4be9-93c1-17002470413e@googlegroups.com>, Adam Beneschan <adambeneschan@gmail.com> writes:
>On Wednesday, July 9, 2014 3:36:09 AM UTC-7, an...@att.net wrote:
>
>> For Ada compilers, where you can have more than one compilation_unit=20
>> within one file (RM 10.1 -- 10.1.1), it allows all preceding=20
>> compilation_unit to have access the same library with only one=20
>> "with_clause" without sharing global visibility to that package.
>>=20
>> Note: GNAT limits the compilation_unit to one per file.  But GNAT still=
>=20
>> allows this type of program design. Putting the "with_clause" in the=20
>> specification file (.ads) allows access for the corresponding body file=
>=20
>> (.adb) in some cases without having the "with_clause" in the body file.
>
>I'm not sure what you're trying to say here, but I want to reiterate: accor=
>ding to the Ada rules, putting multiple compilation units in the same file =
>has *no* effect on the visibility rules (or any other rules).  If you have =
>package specifications for two different packages, P1 and P2, in the same f=
>ile, a "with" or "use" clause on P1 does not apply to P2, and vice versa.
>
>A "with" clause on a specification of any package automatically applies to =
>the body, and it also automatically applies to any child unit specification=
>s or bodies.  This is due to the rules in RM 8.1 and has nothing to do with=
> GNAT, and nothing to do with whether the specification and body are in the=
> same file or not.
>
>Thus, in this example, the comments are wrong:
>
>> --
>> --  compilation_unit 4 : Limits visibility to Text_IO to a=20
>> --                       "selected_component", unless the=20
>> --                       global "use_clause" is used.
>> --
>> --              example :  Text_IO.New_Line ;
>> --
>> package Test_Cal is
>> --    ...
>> end Test_Cal ;
>
>Text_IO is not visible at all, since there is no "with Text_IO" on Test_Cal=
>..  The "with Text_IO" earlier in the file, on "package Test_55", has no eff=
>ect.
>
>                          -- Adam

Re: Position of "use"

[AdaMagica]

On Thursday, July 10, 2014 3:27:08 AM UTC+2, an...@att.net wrote:
> Actually, the comments for compilation_unit 4 are correct under multiple 
> compilation_unit type files, which GNAT does not support. It was tested 
> under AdaEd-1.11.0a compiled for SuSE Linux, before I posted a condensed 
> version.

That sounds quite unbelievable to me; looks like AdaEd has a severe bug - if your claim is correct.

> Note: GNAT (Adacore or GNU) does not support multiple 
> compilation_unit (s) within one file. 
>
> For multiple compilation_unit, likes those files found in ACVC, the 
> external visibility starts with the "with_clause" and that clause 
> grants all succeeding compilation_unit(s) access to that package or 
> routine no matter how many compilation_unit there are. In other words 
> you only need one "with_clause" for Ada compilers that support multiple 
> compilation_unit in the same file. 

Please show us such a file or at least give the reference in ACATS 3.1.

Re: Position of "use"

[J-P. Rosen]

Le 10/07/2014 11:50, AdaMagica a écrit :
> That sounds quite unbelievable to me; looks like AdaEd has a severe bug - if your claim is correct.
AdaEd is roughly 30 years old, and not updated since... It's quite
unbelievable that anybody is still using it, and it is certainly not a
trustable reference nowadays.

-- 
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52, Fax: +33 1 45 29 25 00
http://www.adalog.fr

Re: Position of "use"

[Adam Beneschan]

On Thursday, July 10, 2014 6:10:10 AM UTC-7, J-P. Rosen wrote:

> > That sounds quite unbelievable to me; looks like AdaEd has a severe bug - if your claim is correct.
> 
> AdaEd is roughly 30 years old, and not updated since... It's quite
> unbelievable that anybody is still using it, and it is certainly not a
> trustable reference nowadays.

True, but even younger compilers that are actively being maintained aren't completely trustable references.  GNAT lets things slip through that should be errors.  

                               -- Adam

Re: Position of "use"

[Tero Koskinen]

10.7.2014 18:57, Adam Beneschan wrote:
> GNAT lets things slip through that should be errors.  

To GNAT's defence, this kind of bugs are pretty hard to find,
especially afterwards if enough care has not been taken
during the initial implementation.

Many times you only notice "accepts invalid code" bug
if another compiler rejects the code.[1]

Some of my recent examples:
1)
GNAT compiles Strings Edit 2.9 cleanly, while ICCAda rejects
the code:
>
> "strings_edit-integers-subscript.ads", line 31: Error: Undefined >
identifier:
>
>           Number. [RM 4.1(11), 8.3, 8.6(28)]
>

(See http://build.ada-language.com/job/Strings_Edit_ICCAda/3/console for
full log.)


2)
ICCAda and GNAT both compile Debug package from Adalog
( http://www.adalog.fr/compo2.htm#Debug ) while Janus/Ada rejects
the code:
> In File C:\work\adalog-debug\DEBUG.ADS at line 206
> --------------
>   205:     type Debug_String is new String;
>   206:     procedure Free is new Ada.Unchecked_Deallocation
> (Debug_String, Debug_String_Access);
>-------------------------------------------------------------------------^
> *ERROR* Formal and actual must both be constrained or unconstrained
> (6.4.10) [RM 12.5.4(3)]


Of course, in these cases, one could probably argue which compilers
are correct and which are not; luckily I a user, who does not need
to worry about the compiler implementation. :)

Yours,
 Tero

[1] Yes, there is ACATS, but as you can see, it does not cover everything.

Re: Position of "use"

[Jeffrey Carter]

On 07/10/2014 06:10 AM, J-P. Rosen wrote:
> Le 10/07/2014 11:50, AdaMagica a écrit :
>> That sounds quite unbelievable to me; looks like AdaEd has a severe bug - if your claim is correct.
> AdaEd is roughly 30 years old, and not updated since... It's quite
> unbelievable that anybody is still using it, and it is certainly not a
> trustable reference nowadays.

Anon has never been a trustable reference.

-- 
Jeff Carter
"Spam! Spam! Spam! Spam! Spam! Spam! Spam! Spam!"
Monty Python's Flying Circus
53

Re: Position of "use"

[anon]

--
--  Just because one may think its wrong, does not make it so.
--  The RM does not limit scope of the "with_clause". And does not 
--  fully check the simple_name for duplicates.
--
--  This makes it a RM problem not a compiler bug. The "With_Clause"
--  scope is defined in RM 10.1.2 (5), and does suggest that the 
--  "With_Clause" for a multiple compilation files are global.
--
--  This example works under GNAT.
--
with Ada.Text_IO ;

--  Some will say that the compiler should give a warning for 
--  duplicate entry but the RM permit the following statements

with Ada.Text_IO ;
with Ada.Text_IO ;

procedure t is

  use  Ada ;
  use  Ada.Text_IO ;  
  use  Text_IO ;  

  -- Should give a warning for duplicate but the RM permit this
  use  Text_IO ;

begin
  Ada.Text_IO.New_Page ;
  Text_IO.Put ( "T" ) ;
  New_Line ;
end ;



In <f8345aa5-d4e9-4919-8b32-541336a2e436@googlegroups.com>, AdaMagica <christ-usch.grein@t-online.de> writes:
>On Thursday, July 10, 2014 3:27:08 AM UTC+2, an...@att.net wrote:
>> Actually, the comments for compilation_unit 4 are correct under multiple 
>> compilation_unit type files, which GNAT does not support. It was tested 
>> under AdaEd-1.11.0a compiled for SuSE Linux, before I posted a condensed 
>> version.
>
>That sounds quite unbelievable to me; looks like AdaEd has a severe bug - if your claim is correct.
>
>> Note: GNAT (Adacore or GNU) does not support multiple 
>> compilation_unit (s) within one file. 
>>
>> For multiple compilation_unit, likes those files found in ACVC, the 
>> external visibility starts with the "with_clause" and that clause 
>> grants all succeeding compilation_unit(s) access to that package or 
>> routine no matter how many compilation_unit there are. In other words 
>> you only need one "with_clause" for Ada compilers that support multiple 
>> compilation_unit in the same file. 
>
>Please show us such a file or at least give the reference in ACATS 3.1.

Re: Position of "use"

[Georg Bauhaus]

On 15.07.14 07:56, anon@att.net wrote:
> The "With_Clause"
> --  scope is defined in RM 10.1.2 (5), and does suggest that the
> --  "With_Clause" for a multiple compilation files are global.

10.1.2 (5) refers to a declaration (such as a package declaration),
and to this declaration's declarative region.

10.1.1 lists compilation_unit as either a library_item or a subunit.
(For example, a package spec, or a separate procedure body.)

8.1, at which Adam Beneschan has hinted, expounds declarative_region,
mentioned in 10.1.2.
For any such declaration, as I understand it, a with_clause is in scope of
the corresponding declarative_region, but nowhere else. So, for example,
a with_clause above a package spec will be in scope of a package spec,
its body, a separate procedure body (subunit) declared somewhere in
the package, and children of the package.

But, for example, another package (other than children of the first,
or subunits) that happens to be declared in the same file establishes
another declarative region. Any with_clause, then, that is atop any
prior declarative region is not in scope of this package's declarative
region.


GNAT's gnatchop will confirm this when run on the two units below placed
in the same one file. I do get a warning about redundant with clauses,
when the compiler analyses the resulting unit associated with the name Foo.
I do get two files, foo.adb and bar.adb. But gnatchop does not place
a with_clause ("with Ada.Text_Io") in bar.adb.  Correspondingly, any
use_clause placed anywhere in the declarative region of Bar will not
work. Nor will Ada.Text_Io.Put_Line name a procedure, as Ada.Text_IO will
not have been with-ed there.

with Ada.Text_Io;
with Ada.Text_Io ;

procedure Foo is
begin
    Ada.Text_Io.Put_Line ("Foo");
end Foo;

procedure Bar is
begin
    Ada.Text_Io.Put_Line ("Bar");
end Bar;

Re: Position of "use"

[Simon Wright]

Georg Bauhaus <rm.dash-bauhaus@futureapps.de> writes:

> But, for example, another package (other than children of the first,
> or subunits) that happens to be declared in the same file establishes
> another declarative region. Any with_clause, then, that is atop any
> prior declarative region is not in scope of this package's declarative
> region.

But, in Ada83, a *pragma*'s scope extended over the whole compilation
(all the compilation units).

GNAT (gnatchop) doesn't seem to support this; the modern equivalent
would be configuration pragmas, ARM 10.1.5(1)[1].

I think perhaps Mr Anon is confusing context clauses with pragmas.

[1] http://www.adaic.org/resources/add_content/standards/12rm/html/RM-10-1-5.html#p8

Re: Position of "use"

[Jeffrey Carter]

On 07/15/2014 10:01 AM, Simon Wright wrote:
>
> I think perhaps Mr Anon is confusing context clauses with pragmas.

"Mr"? What evidence do you have that "anon" is an adult, human male? :)

-- 
Jeff Carter
"Ah, go away or I'll kill ya."
Never Give a Sucker an Even Break
100

Re: Position of "use"

[Simon Wright]

Jeffrey Carter <spam.jrcarter.not@spam.not.acm.org> writes:

> On 07/15/2014 10:01 AM, Simon Wright wrote:
>>
>> I think perhaps Mr Anon is confusing context clauses with pragmas.
>
> "Mr"? What evidence do you have that "anon" is an adult, human male? :)

I suppose I just think that women usually have more sense. Gender
profiling, then.

Re: Position of "use"

[G.B.]

On 15.07.14 19:01, Simon Wright wrote:
> But, in Ada83, a*pragma*'s scope extended over the whole compilation
> (all the compilation units).
>
> GNAT (gnatchop) doesn't seem to support this; the modern equivalent
> would be configuration pragmas, ARM 10.1.5(1)[1].

There's

   8.2 Operating gnatchop in Compilation Mode

in the User's Guide saying gnatchop prepends configuration pragmas
to all chopped files if run in this mode.

$ gnatchop -h

   -c       compilation mode, configuration pragmas follow RM rules

Re: Position of "use"

[Adam Beneschan]

On Tuesday, July 15, 2014 10:01:12 AM UTC-7, Simon Wright wrote:

> But, in Ada83, a *pragma*'s scope extended over the whole compilation
> (all the compilation units).

I don't know that this was universally true; I don't see a general rule to that effect.  The rules vary from pragma to pragma.  STORAGE_UNIT and SYSTEM_NAME (since removed) had to appear at the beginning of a compilation and applied to every compilation unit in the compilation (this seems to anticipate the "configuration pragma" rules of Ada 95 and beyond).  ELABORATE, however, appears in a context clause (as does the "use" clause) but only applies to one compilation unit.  Most of the pragmas were only allowed inside declarations.  LIST(ON) and LIST(OFF) apply to the entire compilation, of course, but since it controls only the listing output and has no effect on semantics or generated code, it's really a special case.  Implementation-defined pragmas follow whatever rules the vendor decides they should.

I think the situation is pretty much the same in Ada 95+, except a little more formalized.  There's no one rule defining the scope of pragmas, but there are rules for categories of pragmas (in particular "configuration pragmas").

                            -- Adam

Re: Position of "use"

[Simon Wright]

Adam Beneschan <adambeneschan@gmail.com> writes:

> On Tuesday, July 15, 2014 10:01:12 AM UTC-7, Simon Wright wrote:
>
>> But, in Ada83, a *pragma*'s scope extended over the whole compilation
>> (all the compilation units).
>
> I don't know that this was universally true; I don't see a general
> rule to that effect.  The rules vary from pragma to pragma.
> STORAGE_UNIT and SYSTEM_NAME (since removed) had to appear at the
> beginning of a compilation and applied to every compilation unit in
> the compilation (this seems to anticipate the "configuration pragma"
> rules of Ada 95 and beyond).  ELABORATE, however, appears in a context
> clause (as does the "use" clause) but only applies to one compilation
> unit.  Most of the pragmas were only allowed inside declarations.

-gnat83 isn't perfect, and even with it gnatchop -c doesn't treat
 Storage_Unit as a configuration pragma.

Re: Position of "use"

[anon]

Adam.

  Note: I was not talking or using option -gnat83, or 
pragma Ada_83. Only the "with_clause" and "use_clause" using
AdaEd and GNAT.

I used AdaEd and GNAT and their RM so anyone see what I am 
saying, because only AdaEd and GNAT compilers are downloadable 
and free and both have been validated, while the others 
compilers requires purchasing, and no guarantee they will work 
with current OS. Also, unsure if the sold IBM Apex compiler is 
even available to the general public.

Now, "-gnat83" basically just set the restrictions mode for 83,
like the pragma Ada_83.

  But, in the GNAT_RM  

   pragma Ada_83. states:

     "However, there is no guarantee that code that is processed 
     correctly by GNAT in Ada 83 mode will in fact compile and 
     execute with an Ada 83 compiler, since GNAT does not enforce 
     all the additional checks required by Ada 83."


Then there the change to "copy by type to the second stack" aka 
"Return by type" so, a number of Ada 83, 95, and early Ada 2005 
programs will no longer compile as initially written.

Which suggest that unless you using a pre-"Return by type" compiler 
that the pragma Ada_83, Ada_95 or Ada_05, or command options -gnat83, 
-gnat95, -gnat05 will not truly guarantee the code to be processed 
correctly. 

This one one reason I wish Adacore would make available for 
download the compiler source package for each Ada standard.
such as 
  GNAT 0.xx: initial version of gnat for Ada 83, Ada 95 code removed
  GNAT 3.1x: gnat for Ada 95
  GNAT 2011: gnat for Ada 2005

Or course it up to the downloader to find a way to compile the 
code. And add the correct gcc back_end.



Also most people use GNAT by either 
            gnat make <program_name>

or when binding to other languages.
            gnat compile <program_name>.adb
            gnat build <program_name>.ali
            gnat link <program_name>.ali  <non-Ada object code list>

while the other 10+ routines like chop are not normally used.

Plus, GNAT "Chop" splits the file by looking for the end of 
the "library_unit" or "secondary_unit". Any statement after the 
end of the unit begins on the next file, this can include comments 
after unit.  Also, "Chop" does not check for configuration pragmas 
because GNAT does not support theses pragmas ( System_Name, 
Memory_Size, Storage_Unit ).



In <f02e3f5c-dcfc-4d69-9caa-ee0f14723f70@googlegroups.com>, Adam Beneschan <adambeneschan@gmail.com> writes:
>On Tuesday, July 15, 2014 10:01:12 AM UTC-7, Simon Wright wrote:
>
>> But, in Ada83, a *pragma*'s scope extended over the whole compilation
>> (all the compilation units).
>
>I don't know that this was universally true; I don't see a general rule to =
>that effect.  The rules vary from pragma to pragma.  STORAGE_UNIT and SYSTE=
>M_NAME (since removed) had to appear at the beginning of a compilation and =
>applied to every compilation unit in the compilation (this seems to anticip=
>ate the "configuration pragma" rules of Ada 95 and beyond).  ELABORATE, how=
>ever, appears in a context clause (as does the "use" clause) but only appli=
>es to one compilation unit.  Most of the pragmas were only allowed inside d=
>eclarations.  LIST(ON) and LIST(OFF) apply to the entire compilation, of co=
>urse, but since it controls only the listing output and has no effect on se=
>mantics or generated code, it's really a special case.  Implementation-defi=
>ned pragmas follow whatever rules the vendor decides they should.
>
>I think the situation is pretty much the same in Ada 95+, except a little m=
>ore formalized.  There's no one rule defining the scope of pragmas, but the=
>re are rules for categories of pragmas (in particular "configuration pragma=
>s").
>
>                            -- Adam

Re: Position of "use"

[Adam Beneschan]

On Wednesday, July 9, 2014 6:27:08 PM UTC-7, an...@att.net wrote:
> Actually, the comments for compilation_unit 4 are correct under multiple 
> compilation_unit type files, which GNAT does not support. It was tested 
> under AdaEd-1.11.0a compiled for SuSE Linux, before I posted a condensed 
> version.

The language rules are clear.  10.1.2(5) says that the scope of a with_clause that appears on a library unit declaration consists of the entire declarative region of the *declaration*, which includes all children and subunits (and, by 8.1(8), includes the package body if the with_clause appears on a package specification).  If a file contains two compilation units, they are two separate declarations.  Therefore, unless the second declaration is a body or child unit or subunit of the first one, the scope of a with_clause on the first unit in a file does not include the second unit.

The fact that one or another compiler accepts it means nothing.  Compilers have bugs.  Compiler bugs that allow illegal code to be accepted are given less attention, in my experience, than compiler bugs that reject legal Ada code or generate incorrect object code.  That's certainly the case with GNAT, which has a number of bugs in which illegal code is accepted (which unfortunately means that open-source Ada projects that get put on the Internet sometimes have illegal Ada in them).  But to determine whether a construct is legal or an identifier is visible or whatever, you have to check the language standard; you cannot count on a compiler to get it 100% right.


> For multiple compilation_unit, likes those files found in ACVC, the 
> external visibility starts with the "with_clause" and that clause 
> grants all succeeding compilation_unit(s) access to that package or 
> routine no matter how many compilation_unit there are. In other words 
> you only need one "with_clause" for Ada compilers that support multiple 
> compilation_unit in the same file. 

No, that isn't true.  It may *appear* to be true in the ACATS because a later compilation unit is often the completion (body) of an earlier one, or it's a child unit.  And, as I pointed out, with_clauses on package specs *do* apply to their bodies and child units.  But if you can find a case where a with_clause on one package in a ACATS file applies to a later *unrelated* package, please post it.  (But really, don't bother, because there aren't any.  If there were, our compiler would have caught it a long time ago.)

                                  -- Adam