17 Translating Programs To Ada

 

  <--Last Chapter Table of Contents Next Chapter-->  

 


17.1 c2ada: Translating Your Programs


c2ada is available from http://www.skinner.demon.co.uk/aidan/programming/


17.2 Interfaces.C package


Ada Package Description C Equivalent
int C integer int
unsigned C unsigned integer unsigned
char_array(n) C character array char [n]
long_long C long long long long
etc
The Interfaces.C and Interfaces.C.Extensions packages provide basic type definitions and conversions functions for C programs.

Gnat 3.12 introduces a new boolean type, C_bool, which behaves as a proper C boolean value: 0 is false and any other value is true.

One thing to remember about this package is that C strings are defined as an array of characters, and Ada will raise a CONSTRANT_ERROR exception if two arrays of characters are not exactly equal length, even if a smaller array is being assigned to a larger one. For example,

s : char_array( 1..80 ) := To_C( "Fred Smith" ); -- bad
This example will raise the exception because the string is 11 characters long (10 characters plus a null character), but the array being assigned to is 80 characters. You can get around these kind of errors with dynamic allocation.

The following program demonstrates some of the types and functions in the Interfaces.C packages.

with text_io, unchecked_deallocation,
Interfaces.C.Extensions;


use text_io, Interfaces.C, interfaces.C.Extensions; procedure ctest is -- my types -- -- pointer to C string and deallocation procedure for same type stringptr is access all char_array; procedure free is new unchecked_deallocation( char_array, stringptr ); -- types from standard Ada package Interfaces.C i : int; -- integer u : unsigned; -- unsigned integer l : long; -- long ul : unsigned_long; -- unsigned long c : char; -- a character sp : stringptr; -- ptr to a string (array of characters) f : C_float; -- a float d : double; -- a double wc : wchar_t; -- 16-bit wide character -- additional types from Gnat package Interfaces.C.Extensions ll : long_long; -- long long ull: unsigned_long_long; -- unsigned long long vp : void_ptr; -- void pointer begin Put_Line( "This is an example of Interfaces.C" ); New_Line; sp := new char_array'( To_C( "This is a string" ) ); Put_Line( "The C string s is '" & To_Ada( sp.all ) & "'." ); Free( sp ); end ctest;

This is an example of Interfaces.C

The C string s is 'This is a string'.

17.3 Interfaces.C.Pointers Package

One C feature that Ada programs lack is pointer arithmetic. In C, you can move pointers forward and backwards through an array by using simple arithmetic operations. For example, adding two to an character pointer move the pointer two characters forward in a string. Decrementing an integer pointer moves the pointer back one index position in an integer array.

Since pointer arithmetic is important in many C programs, especially sorts, Ada 95 provides a standard generic package called Interfaces.C.Pointers which implements access types that can use pointer arithmetic.

To instantiate the package, you need to specify the elements that will be in your arrays, an unbounded array that will contain the elements, the range of index values, and a default terminator value used by some of the package's subprograms.

For example, to create C-style pointers for the unbounded Char_Array (C string) type in Interfaces.C, you could instantiate the package with

package StringPtrs is new Interfaces.C.Pointers( 
   Index => size_t, -- the index range is size_t
   Element => char, -- the array contains chars
   Element_Array => char_array, -- the unbounded type
   Default_Terminator => char'val( 0 ) -- the terminator value, ASCII.NUL
);
use StringPtrs; -- need this to make + and - visible

strptr : StringPtrs.Pointer;

The use clause is very important.  Without it, the arithmetic operators would not be directly available because they would be hidden inside the StringPtrs package.

Pointers created using Interfaces.C.Pointers are access types, and can be used like any other access type.

  Put_Line( "strptr is pointing to the character " & strptr.all );
However, unlike other access types, they have new pointer arithmetic features. Addition and subtraction is performed the same way as in C by specifying how many positions in the array to move. To move strptr ahead 2 index positions in a string, add 2 to it:
  strptr := strptr + 2;
Since Ada has no increment or decrement operators, two procedures are provided to move a pointer forward or backward by one array position:
  Increment( strptr );  -- forward one position
  Decrement( strptr );  -- back one position
The package also makes four additional subprograms available: The Virtual_Length function returns the length of an array, up to end of the array or until a terminator is found. [If no terminator, do you get a storage error?--KB] The Value function returns a slice from the array, from the position of the pointer to the end of the array or until a terminator is found. It can also slice a specific number of elements from an array.

Copy_Array copies a slice of a specific number of elements from one pointer to another. Copy_Terminated Array copies a slice from the pointer position until a terminator is found.

The following program demonstrates C pointers to integer arrays.

with Ada.Text_IO, Interfaces.C.Pointers;

use Ada.Text_IO;

procedure point is

-- To use Interfaces.C.Pointers, you need to define an unbounded
-- array type. In this case, we'll create an unbounded array
-- called IntegerArrays with a maximum index range of 1 to 9.
-- BiggestArray is the largest IntegerArrays array possible,
-- with an index range of 1 to 9.  IntegerArrays must have
-- aliased elements because we will be accessing them with an
-- access type.

subtype PointerRange is integer range 1..9;

type IntegerArrays is array PointerRange range <> )
  of aliased integer;

type BiggestArray is new IntegerArrays( PointerRange );

package IntPtrs is new Interfaces.C.Pointers(
  Index => PointerRange, -- the index range
  Element => Integer, -- what the array contains
  Element_Array => IntegerArrays, -- the unbounded type
  Default_Terminator => 0 ); -- the terminator value

use IntPtrs; -- need this to make + and - visible

procedure ShowArray( ia : IntegerArrays ) is

-- show the contents of any IntegerArrays array

begin

  for i in ia'first..ia'last-1 loop
    Put( i'img );
    Put( " =>" );
    Put( ia( i )'img );
    Put( "," );

  end loop;

  Put( ia'last'img );
  Put( " => " );
  Put_Line( ia( ia'last )'img );
end ShowArray;

ia, ia2 : BiggestArray;  -- two integer arrays
ip, ip2 : IntPtrs.Pointer; -- two pointers to integer arrays

begin

  Put_Line( "This program demonstrates C-style pointers provided" );
  Put_Line( "by Interfaces.C.Pointers" );
  New_Line;

  -- initialize and display the contents of the array

  for i in PointerRange'first..PointerRange'last-1 loop
    ia( i ) := i*2;
  end loop;

  ia( PointerRange'last ) := 0;
  Put_Line( "The array is: " );
  ShowArray( IntegerArrays( ia ) );
  -- must typecast ia because ShowArray is expecting an IntegerArrays

  Put_Line( "Zero is our terminator in this example" );
  New_Line;

  -- set the pointers to the first elements in the arrays

  ip := ia( ia'first )'access;
  ip2 := ia2( ia'first )'access;

  -- ip works like a normal access type

  Put_Line( "Our pointer is set to first position in the array");
  Put_Line( "The element is " & ip.all'img );
  New_Line;

  -- increment example

  Increment( ip );
  Put_Line( "Incrementing the pointer, it now points at " &
    ip.all'img );
  New_Line;

  -- decrement example

  Decrement( ip );
  Put_Line( "Decrementing the pointer, it now points at" &
    ip.all'img );
  New_Line;

  -- addition example

  ip := ip + 3;
  Put_Line( "Addition moves the pointer forward." );
  Put_Line( "Moving forward three elements, it now points at"
    & ip.all'img );
  New_Line;

  -- subtraction example

  ip := ip - 2;
  Put_Line( "Subtraction moves the pointer backwards." );
  Put_Line( "Moving backwards two elements, it now points at"
    & ip.all'img );
  New_Line;

  -- Virtual_Length examples

  Put_Line( "Virtual_Length gives the length from the pointer to the" );

  Put_Line( "default terminator.  The length from this position is" &
  Virtual_Length( ip )'img & " positions" );
  Put_Line( "Virtual_Length can also use an arbitrary terminator." );
  Put_Line( "The length from the pointer to the first 14 is" &
  Virtual_Length( ip, 14 )'img & " positions" );
  New_Line;

  -- Value examples

  Put_Line( "Value returns the array slice from the pointer position to" );
  Put_Line( "the terminator.  The array value from this position is" );
  ShowArray( Value( ip ) );

  Put_Line( "Value can also return a slice of a specific length." );
  Put_Line( "The next four elements are" );
  ShowArray( Value( ip, Length => 4 ) );
  New_Line;

  -- Copy_Terminated_Array example

  Put_Line( "Our second array contains" );
  ShowArray( IntegerArrays( ia2 ) ); -- must typecast here
  New_Line;

  Put_Line( "Copy_Terminated_Array copies elements from one pointer to" );
  Put_Line( "another, up to and including the terminator.  Copying to" );
  Put_Line( "the second array " );
  Copy_Terminated_Array( ip, ip2 );
  ShowArray( IntegerArrays ( ia2 ) ); -- must typecast here
  New_Line;

  -- Copy_Array example

  Put_Line( "Copy_Array copies a specific number of elements.");
  Put_Line( "Copying 4 elements from 3 positions ahead, the new");
  Put_Line( "array contains" );
  Copy_Array( ip+3, ip2, 4 );
  ShowArray( IntegerArrays( ia2 ) ); -- must typecast here
  New_Line;

end point;


This program demonstrates C-style pointers provided
by Interfaces.C.Pointers

The array is:

1 => 2, 2 => 4, 3 => 6, 4 => 8, 5 => 10, 6 => 12, 7 => 14, 8 => 16, 9 => 0

Zero is our terminator in this example

Our pointer is set to first position in the array

The element is 2

Incrementing the pointer, it now points at 4

Decrementing the pointer, it now points at 2

Addition moves the pointer forward.
Moving forward three elements, it now points at 8

Subtraction moves the pointer backwards.
Moving backwards two elements, it now points at 4

Virtual_Length gives the length from the pointer to the
default terminator.  The length from this position is 7 positions

Virtual_Length can also use an arbitrary terminator.
The length from the pointer to the first 14 is 5 positions

Value returns the array slice from the pointer position to
the terminator.  The array value from this position is

1 => 4, 2 => 6, 3 => 8, 4 => 10, 5 => 12, 6 => 14, 7 => 16, 8 => 0

Value can also return a slice of a specific length.

The next four elements are
1 => 4, 2 => 6, 3 => 8, 4 => 10

Our second array contains

1 => 12, 2 => 134531961, 3 => 12, 4 => 1, 5 => 134560412, 6 => 134891560, 7 => 134575980, 8 => 0, 9 => 134560432

Copy_Terminated_Array copies elements from one pointer to another, up to
and including the terminator.
Copying to the second array
1 => 4, 2 => 6, 3 => 8, 4 => 10, 5 => 12, 6 => 14, 7 => 16, 8 => 0, 9 => 134560432

Copy_Array copies a specific number of elements.

Copying 4 elements from 3 positions ahead, the new array contains
1 => 10, 2 => 12, 3 => 14, 4 => 16, 5 => 12, 6 => 14, 7 => 16, 8 => 0, 9 => 134560432

17.4 Interfaces.C_Streams package

Ada Package Description C Equivalent
fopen Open a text file (C stream) fopen
fclose Close a text file (C stream) fclose
fread Read bytes from a text file (C stream) fread
etc.
Although the basic Ada types are identical to their C counterparts, the IO libraries are not guaranteed to write data in a format that is readable from other languages. Text files are fine, but to write binary files that can be accessed by C, you'll need to read and write the files using C file handing libraries.

The Interfaces.C_Streams package provides a thin binding to the C stdio library. This is comparable to the gnat.os_lib library, but the binding is "thinner" and covers all C stream operations. Some stdio library functions aren't covered because they can't be represented by Ada. Gnat guarantees these functions will be available, no matter what platform gnat is running under, even if it isn't UNIX-based.

It is also possible to call stdio directly. See the discussion above.

c_streams uses "stream" to refer to a Linux text file.

procedure clearerr(stream : FILEs);
Clear any error associated with the stream.

function fclose(stream : FILEs) return int;
Close a stream.

function fdopen(handle : int; mode : chars) return FILEs;
Open a stream by a handle (UNIX file descriptor).

function feof(stream : FILEs) return int;
Check for the end of stream.

function ferror(stream : FILEs) return int;
Return any error associated with the last stream operation.

functionfflush(stream : FILEs) return int;
Finish writing any outstanding data to the stream.

function fgetc(stream : FILEs) return int;
Read one character from the stream. Characters will be ASCII values between 0 and 255, and can be converted to a character with character'val.

function fgets(strng : chars; n : int; stream : FILEs) return chars;
Read a string from the stream. Note this is not an Ada string.

function fileno(stream : FILEs) return int;
Return the fine number associated with a stream for use with standard Linux file operations.

function fopen( filename : chars; Mode : chars) return FILEs;
Open a stream.

function fputc(C : int; stream : FILEs) return int;
Write one character to a stream. Convert the character to an integer using character'val.

function fputs(Strng : chars; Stream : FILEs) return int;
Write a string of characters to the stream.

function fread( buffer : voids; size : size_t; count : size_t; stream : FILEs) return size_t;
Read count bytes into a buffer of length size and return number of bytes actually read.

function freopen( filename : chars; mode : chars; stream : FILEs) return FILEs;
Reopen the stream with a new mode.

function fseek( stream : FILEs; offset : long; origin : int) return int;
Move offset bytes from the specified origin point.

function ftell(stream : FILEs) return long;
Get stream offset for fseek.

function fwrite( buffer : voids; size : size_t; count : size_t; stream : FILEs) return size_t;
Write count bytes from a buffer of count length and return the number of bytes actually written.

function isatty(handle : int) return int;
[NQS--determine if stream is a TTY device?--KB]

procedure mktemp(template : chars);
Create a random name for a temporary file.

procedure rewind(stream : FILEs);
Move to the start of the stream.

function setvbuf(stream : FILEs; buffer : chars; mode : int; size : size_t) return int;
[NQS--used to know what this did--KB]

procedure tmpnam(string : chars);
[Difference with tmpnam?--KB]
The parameter must be a pointer to a string buffer of at least L_tmpnam bytes (the call with a null parameter is not supported).

function tmpfile return FILEs;
[NQS--KB]

function ungetc(c : int; stream : FILEs) return int;
Back up one character in the stream.

function unlink(filename : chars) return int;
Delete a stream file.

The following are related utility functions added by ACT. They are not standard UNIX functions like the above.

function file_exists(name : chars) return int;
Returns 0 if a file doesn't exist, 1 if it does.

function is_regular_file(handle : int) return int;
Return 1 if given handle is for a regular file, or 0 for some other kind of file.

procedure set_binary_mode( handle : int);
Read text without translation. Only works if compiled with text_translation_required.

procedure set_text_mode( handle : int);
Translate text. Only works if compiled with text_translation_required.

procedure full_name(nam : chars; buffer : chars);
Return the full path of a file as a C string.

The following program demonstrates some of the c_stream functions.

with text_io, unchecked_deallocation, Interfaces.C_Streams;
use text_io, Interfaces.C_Streams;

procedure cstreamtest is

  fd : FILEs;
  line2write : constant string := "This is a test";
  cline2write: constant string := "This is a test" &
    ASCII.NUL;
  path : constant string := "testfile.xxx";
  cpath : constant string := path & ASCII.NUL;
  fileMode : constant string := "w";

  amountWritten : size_t;
  result : int;

begin

  Put_Line( "This is an example of Interfaces.C_Streams" );
  New_Line;

  fd := fopen( cpath'address, fileMode'address );
  if ferror( fd ) = 0 then
    Put_Line( "Opened " & path & " with fopen");
    Put_Line( "Writing '" & line2write & "' with fwrite" );
    amountWritten := fwrite( line2write'address, -- what to write
                             1, -- size of elements
                             line2write'length, -- how many to write
                             fd ); -- to which file

    Put_Line( "Wrote" & amountWritten'img & " characters" );
    New_Line;
    Put_Line( "Writing with fputs" );
    result := fputs( cline2write'address, fd );
    Put_Line( "Result was" & result'img );
    New_Line;

    result := fclose( fd );
    Put_Line( "Closed " & path & " with fclose");
    Put_Line( "Result was" & result'img );
    New_Line;

    result := unlink( cpath'address );
    Put_Line( "Deleted " & path & " with unlink");
    Put_Line( "Result was" & result'img );
  end if;

end cstreamtest;

This is an example of Interfaces.C_Streams

Opened testfile.xxx with fopen
Writing 'This is a test' with fwrite
Wrote 14 characters

Writing with fputs
Result was 1

Closed testfile.xxx with fclose
Result was 0

Deleted testfile.xxx with unlink
Result was 0

17.5 Ada and C Files

Gnat provides several interfacing packages to allow Ada to read and write C files. These are a "thicker binding" than Interfaces.C_Streams.

17.6 A Word on Interfaces.Fortran

Gnat provides interfacing packages for languages besides C. Interfaces.Fortran contains types and subprograms to link Fortran language programs to your Ada programs. The GCC Fortran 77 compiler is g77.

As with gcc, most of the Fortran data types correspond identically with an Ada type. A Fortran real variable, for example, is the same as an Ada float, and a double precision variable is an Ada long_float. Other Ada compilers may not do this: if portability is an issue, always use the types of Interfaces.Fortran.

Gnat 3.12 introduces a proper Fortran logical type that behaves according to Fortran semantics.

Fortran subprograms may be imported into Ada using pragma import:

procedure MyFortranSubroutine;
pragma import( Fortran, MyFortranSubroutine );
Variables may be likewise imported.
RealVar : float;
pragma import( Fortran, RealVar );

g77 adds an undescore to subroutine names, so ifyou are importing from g77 you'll need to include the name of the subroutine with a trailing underscore in pragma import.

pragma import (Fortran, MyFortranSubroutine, "MyFortranSubroutine_");

 

17.7 P2Ada - Pascal to Ada

There is a Pascal to Ada translator called P2Ada. It supports some of the TurboPascal conventions. It's available from http://www.mysunrise.ch/users/gdm/gsoft.htm#p2ada
 

 


  <--Last Chapter Table of Contents Next Chapter-->