Re: Ada.Strings.Fixed.Count raises Storage_Error

[Victor Porton <porton@narod.ru>]

I've reported the bug:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71674

Victor Porton wrote:
> GNAT 6.1.1 on Debian Linux "testing":
> 
> $ ./test_count
>           0          1
> 
> raised STORAGE_ERROR : stack overflow or erroneous memory access
> 
> Let's decide who of us will report the bug, so that the report won't
> happen to be duplicate?
> 
> Georg Bauhaus wrote:
>> On 12.05.16 13:36, Xavier Petit wrote:
>> 
>>> with Ada.Strings.Fixed, Ada.Integer_Text_IO;
>>> use  Ada.Strings.Fixed, Ada.Integer_Text_IO;
>>>
>>> procedure Test_Count is
>>>     S : constant String (Positive'Last - 2 .. Positive'Last) := "Ada";
>>> begin
>>>     Put (Count (Source => S,                         Pattern => "AA"));
>>>     Put (Count (Source => S (S'First .. S'Last - 1), Pattern => "A"));
>>>     Put (Count (Source => S,                         Pattern => "A"));
>>> end;
>>>
>>> Output on gnat-gcc 5.3.1 :
>>>
>>>            0          1
>>>
>>> raised STORAGE_ERROR : stack overflow or erroneous memory access
>> 
>> Compiling such that the RTS is recompiled will result in
>> 
>> raised CONSTRAINT_ERROR : a-strsea.adb:102 overflow check failed
>> 
>> The message points at a line in the implementation of
>> Ada.Strings.Search.Count, for the third of the calls. It reads
>> 
>>       Ind := Ind + 1;
>> 
>> where Ind looks seems to be go along with the Index of your S,
>> IINM. Effectively (PL1 = Pattern'Length - 1 being 0), then loop
>> there boils down to this:
>> 
>>     while Ind <= Source'Last loop
>>        if Pattern = Source (Ind .. Ind) then
>>           ...
>>        else
>>           Ind := Ind + 1;
>>        end if;
>>     end loop;
>> 
>> And this would explains the overflow.
>> 
>> Just the slip-up that Ada subtypes were designed to help prevent?
>> 
>> 
-- 
Victor Porton - http://portonvictor.org