Raising IO_EXCEPTIONS.END_ERROR in a loop

Raising IO_EXCEPTIONS.END_ERROR in a loop

[Scott Ingram]

The following fragment compiles, and even runs correctly through
one iteration but blows up when re-reading a string from standard
input.  The exception that is raised makes me think that the
string variable used is still visible the second time through--
but shouldn't it pass out of existence at the end of the block?
or what am I misunderstanding?

procedure somesuch is
 begin
      loop
         exit when Continue = 'n' or Continue = 'N';
         Get_Variable:
         declare
            Variable_Name : String (1..OS_Line_Length);
            Name_Length : Natural;
         begin --Get_Variable
            Put ("Enter the name of the variable to examine: ");
            Get_Line (Variable_Name, Name_Length);
               Put_Line ("The variable is set to " &
Variable_Name(1..Name_Length));
         end Get_Variable;
         Put ("Examine another (y|Y)? ");
         Get_Immediate (Continue);
      end loop;
   end somesuch;

-- 
Scott Ingram
Sonar Processing and Analysis Laboratory
Johns Hopkins University Applied Physics Laboratory

Re: Raising IO_EXCEPTIONS.END_ERROR in a loop

[Stephen Leake]

Scott Ingram wrote:
> 
> The following fragment compiles, and even runs correctly through
> one iteration but blows up when re-reading a string from standard
> input.  The exception that is raised makes me think that the
> string variable used is still visible the second time through--
> but shouldn't it pass out of existence at the end of the block?
> or what am I misunderstanding?
>  <see patched code below>

You need to be more precise about what happens - "blows up" is not
something my computer does very often :). Does it raise an exception?
crash your OS? Are you running on Windows or some other OS? You are
using Get_Immediate, which is (unfortunately), system-dependent. What
compiler are you using?

Jumping way out on a limb, I assume you have the same system I do
(Windows 95, GNAT 3.09) (yes, I'll upgrade to GNAT 3.10 one of these
days; I was holding out for an ada-aware debugger :( ). I compiled and
ran your code, after adding the missing declarations. It works fine,
except on this system Get_Immediate doesn't return until the user hits
<return>, so you need to add Skip_Line. 

Hope this helps.
> --
> Scott Ingram
> Sonar Processing and Analysis Laboratory
> Johns Hopkins University Applied Physics Laboratory
with Ada.Text_IO; use Ada.Text_IO;

procedure Test is
   OS_Line_Length : constant := 80;
   Continue : Character := 'Y';
begin
   loop
      exit when Continue = 'n' or Continue = 'N';
      Get_Variable:
      declare
         Variable_Name : String (1..OS_Line_Length);
         Name_Length : Natural;
      begin --Get_Variable
         Put ("Enter the name of the variable to examine: ");
         Get_Line (Variable_Name, Name_Length);
         Put_Line ("The variable is set to " &
                   Variable_Name(1..Name_Length));
      end Get_Variable;
      Put ("Examine another (y|Y)? ");
      Get_Immediate (Continue);

      
      Skip_Line; -- added this line --
   end loop;

end Test;

-- 
- Stephe

Re: Raising IO_EXCEPTIONS.END_ERROR in a loop

[Scott Ingram]

Stephen Leake wrote:
> 
> Scott Ingram wrote:
> >
> > The following fragment compiles, and even runs correctly through
> > one iteration but blows up when re-reading a string from standard
(extra lines snipped)
> 
> You need to be more precise about what happens - "blows up" is not
> something my computer does very often :). Does it raise an exception?
> crash your OS? Are you running on Windows or some other OS? You are
> using Get_Immediate, which is (unfortunately), system-dependent. What
> compiler are you using?
>

Sorry, thought that "Raising IO_EXCEPTIONS.END_ERROR ..." as a subject
line was specific enough to elucidate how it was blowing up :-)
Also, I realize that Get_Immediate is system dependent:  but I didn't
realize HOW dependent, since I have only been using in a loop as
illustrated when I wanted to hand feed some value in.  Up until now,
it appeared to behave the same way (for my purposes) on several
varieties
of Unix (including Linux.)

> Jumping way out on a limb, I assume you have the same system I do
> (Windows 95, GNAT 3.09) (yes, I'll upgrade to GNAT 3.10 one of these
> days; I was holding out for an ada-aware debugger :( ). I compiled and
> ran your code, after adding the missing declarations. It works fine,
> except on this system Get_Immediate doesn't return until the user hits
> <return>, so you need to add Skip_Line.
> 
> Hope this helps.
(more stuff snipped)

Yes, it does.  I didn't think to try it on a sufficiently different
system
than the two where the problem showed up!  Thanks.


> --
> - Stephe

-- 
Scott Ingram
Sonar Processing and Analysis Laboratory
Johns Hopkins University Applied Physics Laboratory