Re: unsigned type

[anon@anon.org (anon)]

        A String literal is define in RM 4.2 ( 4, 11 ) with the String 
type defined in RM 3.6.3 ( 4 ) and its index type (Positive) is define 
in RM 3.6.3 ( 3 ) and RM 3.5.4 ( 13 ).  With the Positive base range 
being Natural defined by RM 3.5.4 ( 13..14 ).

So to create a "null range" Strings aka a null String literal, the type and 
range must be check! In the following hierarchical outline, based on RM 3.2:

  1. RM 4.2 ( 11 )  : Check if Lower bound > Base_Range'First.
                      (Lower bound must be a member of the "subtype" 
                       or Base_Range type in order to preform the 
                       comparison, so Lower Bound type is checked)

  2. RM 3.6.1 ( 4 ) : Check if both indexes values are within subtype 
                      or range type. 
                      (Upper bound is now verified to be a valid 
                       type of the range)

  3. RM 3.5         : Check if Lower bounds > Upper bounds
                      (verify the "null range" definition)

  4. Else it is not a valid null range String and a Constraint_Error 
          will be raised. 


So, is "A : String ( 1..0 )" legal.  

Yes maybe! Yes, because RM 4.2 ( 11 ) states that "for the evaluation of 
a null string literal, a check is made that its lower bound is greater 
than the lower bound of the base range of the index type.  And the Lower 
bound ( 1 ) > ( 0 ) aka Natural'First which fills the definition in 
RM 4.2 ( 11 ). And the Lower bound ( 1 ) > Upper bound ( 0 ) RM 3.5 ( 4 ). 

But there is a problem from RM 3.6.1 ( 4 ) it states that both indexes must 
be a member of the subtype for the arrays range. Since, Strings are just a 
predefined one-dimensional character array using Positives as indexes. That 
means for a String that both indexes must be a member of the Positive type 
making the ( 1..0 ) invalid, since 0 is not define as a Positive type. So, 
from RM 3.6.1 ( 4 ) this null range String is not valid.


Now, is "B : String := ( -5 .. -7 => ' ' ) ;" legal? 

First, it violates RM 4.2 ( 11 ), that is, the lower bound ( -5 ) is not 
greater than the lower bound of the base range of the index type 
Natural aka Natural'First or 0. So, from RM 4.2 ( 11 ) this should raise 
a Constraint_Error. But it does not!

And then both indexes are not valid members of either the index type 
Positive or even its base range type aka Natural. So, from RM 3.6.1 ( 4 ) 
this is an error.

But, it does comply with the "null range" in  RM 3.5 ( 4 ), where the 
lower bounds must be greater than upper bounds, but that does not 
replace the other two requirements, for a null String literal.

So, this should raise a Constraint_Error. 



In <4a4e7705$0$31863$9b4e6d93@newsspool3.arcor-online.net>, Georg Bauhaus <rm.tsoh.plus-bug.bauhaus@maps.futureapps.de> writes:
>anon wrote:
>
>> Since the RM does not directly 
>> give one permission to use an illegal index value, then one should assume 
>> that all index must be valid before checking for a null array. 
>
>To "assume" formal properties like index validity seems inadequate:
>this is about index ranges, not index values, and about what the RM
>has to say about the formal notion of a "null range" (as has now been
>said many times).
>
>> Actually, 
>> Ada RM 0 ( 29 ) states that it is illegal and the program shall create 
>> an exception if the program "access an array component by using an 
>> incorrect index value".
>
>There isn't any access to an array component through a null range.
>For indexing an array component, one must have an index value.
>There is no value in a null range.  Therefore, there is no
>access to an array component.
>
>
>> Also, a number of ACATS test suggest this as well 
>> and the program below will illustrate this. With that stated from the RM 
>> the program shall 
>> 
>>  if S'First in S'Range and then S'Last in S'Range then 
>>    if S'First <= S'Last then
>>      -- access/create an array bound by ( S'First .. S'Last )
>>    else
>>     -- return a null created array access
>>    end if ;
>>  else
>>   raise Constraint_Error ;
>>  end if ;
>> 
>> You have to look at the whole RM to find out if there are other sections 
>> that may clarify a statements or allow an exception. But there is no 
>> RM rules that allows an out of bounds index to be used.
>
>Maybe the verse and illustrations given by J.-P. Rosen and
>Adam Beneschan have slipped:  There are RM rules regarding
>null ranges.  The construct "Greater .. Smaller" as used in, say,
>initializing a null string refers to null ranges, so not to any
>index value at all.
>