From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.4
X-Google-Thread: a07f3367d7,c19e8df8a75221d0
X-Google-Attributes: gida07f3367d7,public,usenet
X-Google-NewGroupId: yes
X-Google-Language: ENGLISH,ASCII
Path: 
 g2news2.google.com!postnews.google.com!p36g2000vbn.googlegroups.com!not-for-mail
From: Martin <martin.dowie@btopenworld.com>
Newsgroups: comp.lang.ada
Subject: Re: Q: Line_IO
Date: Mon, 31 Aug 2009 01:28:36 -0700 (PDT)
Organization: http://groups.google.com
Message-ID: 
 <7225bda9-8757-4c5c-bb44-b3be21a1e1f9@p36g2000vbn.googlegroups.com>
References: <4a9b045a$0$31875$9b4e6d93@newsspool3.arcor-online.net>
NNTP-Posting-Host: 20.133.0.8
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-Trace: posting.google.com 1251707317 20579 127.0.0.1 (31 Aug 2009 08:28:37
 GMT)
X-Complaints-To: groups-abuse@google.com
NNTP-Posting-Date: Mon, 31 Aug 2009 08:28:37 +0000 (UTC)
Complaints-To: groups-abuse@google.com
Injection-Info: p36g2000vbn.googlegroups.com; posting-host=20.133.0.8;
	posting-account=g4n69woAAACHKbpceNrvOhHWViIbdQ9G
User-Agent: G2/1.0
X-HTTP-UserAgent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.1.2)
	Gecko/20090729 Firefox/3.5.2,gzip(gfe),gzip(gfe)
Xref: g2news2.google.com comp.lang.ada:8070
Date: 2009-08-31T01:28:36-07:00
List-Id: <comp.lang.ada>

On Aug 30, 11:59=A0pm, Georg Bauhaus <see.reply...@maps.futureapps.de>
wrote:
> Text_IO seems fairly slow when just reading lines of text.
> Here are two alternative I/O subprograms for Line I/O, in plain Ada,
> based on Stream_IO. =A0 They seem to run significantly faster.
>
> However, there is one glitch and I can't find the cause:
> output always has one more line at the end, an empty one.
> Why? =A0If you have got a minute to look at this, you will
> also help us with getting faster programs at the Shootout.
> These read lines by the megabyte.
>
> generic
> =A0 =A0Separator_Sequence : in String; =A0-- =A0ends a line
> package Line_IO is
>
> =A0 =A0pragma Elaborate_Body;
>
> =A0 =A0--
> =A0 =A0-- =A0High(er) speed reading and writing of lines via Stream I/O.
> =A0 =A0-- =A0Made with Unix pipes in mind.
> =A0 =A0--
> =A0 =A0-- =A0Assumptions:
> =A0 =A0-- =A0- Lines are separated by a sequence of characters.
> =A0 =A0-- =A0- Characters and stream elements can be used interchangeably=
.
> =A0 =A0-- =A0- Lines are not longer than internal buffer size.
> =A0 =A0--
> =A0 =A0-- =A0I/O exceptions are propagated
>
> =A0 =A0procedure Print(Item : String);
>
> =A0 =A0function Getline return String;
>
> end Line_IO;
>
> with Ada.Streams.Stream_IO;
> with Ada.Unchecked_Conversion;
>
> package body Line_IO is
>
> =A0 =A0use Ada.Streams;
>
> =A0 =A0Stdout : Stream_IO.File_Type;
> =A0 =A0Stdin : Stream_IO.File_Type;
>
> =A0 =A0-- writing
>
> =A0 =A0procedure Print (Item : String) is
>
> =A0 =A0 =A0 subtype Index is Stream_Element_Offset range
> =A0 =A0 =A0 =A0 Stream_Element_Offset(Item'First)
> =A0 =A0 =A0 =A0 .. Stream_Element_Offset(Item'Last + Separator_Sequence'L=
ength);
> =A0 =A0 =A0 subtype XString is String (Item'First
> =A0 =A0 =A0 =A0 .. Item'Last + Separator_Sequence'Length);
> =A0 =A0 =A0 subtype XBytes is Stream_Element_Array (Index);
> =A0 =A0 =A0 function To_Bytes is new Ada.Unchecked_Conversion
> =A0 =A0 =A0 =A0 (Source =3D> XString,
> =A0 =A0 =A0 =A0 =A0Target =3D> XBytes);
> =A0 =A0begin
> =A0 =A0 =A0 Stream_IO.Write (Stdout, To_Bytes (Item & Separator_Sequence)=
);
> =A0 =A0end Print;
>
> =A0 =A0-- ----------------
> =A0 =A0-- reading
> =A0 =A0-- ----------------
> =A0 =A0-- Types etc., status variables, and the buffer. =A0`Buffer` is at=
 the
> =A0 =A0-- same time an array of Character and and array of Stream_Element
> =A0 =A0-- called `Bytes`. =A0They share the same address. =A0This setup m=
akes the
> =A0 =A0-- storage at the address either a String (when selecting result
> =A0 =A0-- characters) or a Stream_Element_Array (when reading input bytes=
).
>
> =A0 =A0BUFSIZ: constant :=3D 8_192;
> =A0 =A0pragma Assert(Character'Size =3D Stream_Element'Size);
>
> =A0 =A0SL : constant Natural :=3D Separator_Sequence'Length;
>
> =A0 =A0subtype Extended_Buffer_Index is Positive range 1 .. BUFSIZ + SL;
> =A0 =A0subtype Buffer_Index is Extended_Buffer_Index
> =A0 =A0 =A0range Extended_Buffer_Index'First .. Extended_Buffer_Index'Las=
t - SL;
> =A0 =A0subtype Extended_Bytes_Index is Stream_Element_Offset
> =A0 =A0 =A0range 1 .. Stream_Element_Offset(Extended_Buffer_Index'Last);
> =A0 =A0subtype Bytes_Index is Extended_Bytes_Index
> =A0 =A0 =A0range Extended_Bytes_Index'First
> =A0 =A0 =A0.. (Extended_Bytes_Index'Last - Stream_Element_Offset(SL));
>
> =A0 =A0subtype Buffer_Data is String(Extended_Buffer_Index);
> =A0 =A0subtype Buffer_Bytes is Stream_Element_Array(Extended_Bytes_Index)=
;
>
> =A0 =A0Buffer : Buffer_Data;
> =A0 =A0Bytes =A0: Buffer_Bytes;
> =A0 =A0for Bytes'Address use Buffer'Address;
>
> =A0 =A0Position : Natural; -- start of next substring
> =A0 =A0Last =A0 =A0 : Natural; -- last valid character in buffer
>
> =A0 =A0function Getline return String is
>
> =A0 =A0 =A0 procedure Reload;
> =A0 =A0 =A0 -- =A0move remaining characters to the start of `Buffer` and
> =A0 =A0 =A0 -- =A0fill the following bytes if possible
> =A0 =A0 =A0 -- =A0post: Position in 0 .. 1, and 0 should mean end of file
> =A0 =A0 =A0 -- =A0 =A0 =A0 =A0Last is 0 or else the index of the last val=
id element in
> Buffer
>
> =A0 =A0 =A0 procedure Reload is
> =A0 =A0 =A0 =A0 =A0Remaining : constant Natural :=3D Buffer_Index'Last - =
Position + 1;
> =A0 =A0 =A0 =A0 =A0Last_Index : Stream_Element_Offset;
> =A0 =A0 =A0 begin
> =A0 =A0 =A0 =A0 =A0Buffer(1 .. Remaining) :=3D Buffer(Position .. Buffer_=
Index'Last);
>
> =A0 =A0 =A0 =A0 =A0Stream_IO.Read(Stdin,
> =A0 =A0 =A0 =A0 =A0 =A0Item =3D> Bytes(Stream_Element_Offset(Remaining) +=
 1 ..
> Bytes_Index'Last),
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 Last =3D> Last_Index);
> =A0 =A0 =A0 =A0 =A0Last :=3D Natural(Last_Index);
> =A0 =A0 =A0 =A0 =A0Buffer(Last + 1 .. Last + SL) :=3D Separator_Sequence;
>
> =A0 =A0 =A0 =A0 =A0Position :=3D Boolean'Pos(Last_Index > 0
> =A0 =A0 =A0 =A0 =A0 =A0and then Buffer(1) /=3D ASCII.EOT =A0 -- ^D
> =A0 =A0 =A0 =A0 =A0 =A0and then Buffer(1) /=3D ASCII.SUB); -- ^Z
>
> =A0 =A0 =A0 end Reload;
>
> =A0 =A0 =A0 function Sep_Index return Natural;
> =A0 =A0 =A0 -- =A0position of next Separator_Sequence
> =A0 =A0 =A0 pragma Inline(Sep_Index);
>
> =A0 =A0 =A0 function Sep_Index return Natural is
> =A0 =A0 =A0 =A0 =A0K : Natural :=3D Position;
> =A0 =A0 =A0 begin
> =A0 =A0 =A0 =A0 =A0pragma Assert(K >=3D Buffer'First);
> =A0 =A0 =A0 =A0 =A0pragma Assert(Buffer(Buffer_Index'Last + 1 .. Buffer'L=
ast)
> =A0 =A0 =A0 =A0 =A0 =A0=3D Separator_Sequence);
>
> =A0 =A0 =A0 =A0 =A0while Buffer(K) /=3D Separator_Sequence(1) loop
> =A0 =A0 =A0 =A0 =A0 =A0 K :=3D K + 1;
> =A0 =A0 =A0 =A0 =A0end loop;
>
> =A0 =A0 =A0 =A0 =A0return K;
> =A0 =A0 =A0 end Sep_Index;
>
> =A0 =A0 =A0 Next_Separator : Natural;
> =A0 =A0begin =A0-- Getline
> =A0 =A0 =A0 pragma Assert(Position =3D 0 or else Position in Extended_Buf=
fer_Index);
> =A0 =A0 =A0 pragma Assert(Last =3D 0 or else Last in Buffer_Index);
>
> =A0 =A0 =A0 if Position =3D 0 then
> =A0 =A0 =A0 =A0 =A0raise Stream_IO.End_Error;
> =A0 =A0 =A0 end if;
>
> =A0 =A0 =A0 Next_Separator :=3D Sep_Index;
>
> =A0 =A0 =A0 if Next_Separator > Buffer_Index'Last then
> =A0 =A0 =A0 =A0 =A0-- must be sentinel
> =A0 =A0 =A0 =A0 =A0Reload;
> =A0 =A0 =A0 =A0 =A0return Getline;
> =A0 =A0 =A0 end if;
>
> =A0 =A0 =A0 if Next_Separator <=3D Last then
> =A0 =A0 =A0 =A0 =A0declare
> =A0 =A0 =A0 =A0 =A0 =A0 Limit : constant Natural :=3D Natural'Max(0, Next=
_Separator - SL);
> =A0 =A0 =A0 =A0 =A0 =A0 -- there was trouble (Print) when Integer Limit c=
ould be
> negative
> =A0 =A0 =A0 =A0 =A0 =A0 -- (for 2-char SL and Next_Separator =3D 1)
> =A0 =A0 =A0 =A0 =A0 =A0 Result : constant String :=3D Buffer(Position .. =
Limit);
> =A0 =A0 =A0 =A0 =A0begin
> =A0 =A0 =A0 =A0 =A0 =A0 Position :=3D Limit + SL + 1;
> =A0 =A0 =A0 =A0 =A0 =A0 return Result;
> =A0 =A0 =A0 =A0 =A0end;
> =A0 =A0 =A0 else
> =A0 =A0 =A0 =A0 =A0-- the separator is among the characters beyond `Last`
> =A0 =A0 =A0 =A0 =A0declare
> =A0 =A0 =A0 =A0 =A0 =A0 Limit : constant Positive :=3D Last;
> =A0 =A0 =A0 =A0 =A0 =A0 Result : constant String :=3D Buffer(Position .. =
Limit);
> =A0 =A0 =A0 =A0 =A0begin
> =A0 =A0 =A0 =A0 =A0 =A0 Position :=3D 0; =A0-- next call will raise End_E=
rror
> =A0 =A0 =A0 =A0 =A0 =A0 return Result;
> =A0 =A0 =A0 =A0 =A0end;
> =A0 =A0 =A0 end if;
>
> =A0 =A0 =A0 raise Program_Error;
> =A0 =A0end Getline;
>
> begin
> =A0 =A0-- (see <ILmdnWHx29q5VMrZnZ2dnUVZ_sedn...@megapath.net> for names
> =A0 =A0-- of standard I/O streams when using Janus Ada on Windows.)
>
> =A0 =A0Stream_IO.Open (Stdout,
> =A0 =A0 =A0Mode =3D> Stream_IO.Out_File,
> =A0 =A0 =A0Name =3D> "/dev/stdout");
> =A0 =A0Stream_IO.Open (Stdin,
> =A0 =A0 =A0Mode =3D> Stream_IO.In_File,
> =A0 =A0 =A0Name =3D> "/dev/stdin");
>
> =A0 =A0-- make sure there is no line separator in `Buffer` other than the
> sentinel
> =A0 =A0Buffer :=3D Buffer_Data'(others =3D> ASCII.NUL);
> =A0 =A0Buffer(Buffer_Index'Last + 1 .. Buffer'Last) :=3D Separator_Sequen=
ce;
> =A0 =A0Position :=3D Buffer_Index'Last + 1; =A0-- See also
> `Getline.Reload.Remaining`
> =A0 =A0Last :=3D 0;
> end Line_IO;
>
> --
> -- A small test program.
> --
> with Line_IO;
> with Ada.Text_IO;
>
> procedure Test_Line_IO is
> =A0 =A0Want_Text_IO : constant Boolean :=3D False;
>
> =A0 =A0-- pick the correct one for your input files
> =A0 =A0UnixLF =A0: constant String :=3D String'(1 =3D> ASCII.LF);
> =A0 =A0MacCR =A0 : constant String :=3D String'(1 =3D> ASCII.CR);
> =A0 =A0OS2CRLF : constant String :=3D String'(1 =3D> ASCII.CR, 2 =3D> ASC=
II.LF);
>
> =A0 =A0package LIO is new Line_IO(Separator_Sequence =3D> UnixLF);
>
> begin
> =A0 =A0if Want_Text_IO then
> =A0 =A0 =A0 loop
> =A0 =A0 =A0 =A0 =A0declare
> =A0 =A0 =A0 =A0 =A0 =A0 A_Line : constant String :=3D Ada.Text_IO.Get_Lin=
e;
> =A0 =A0 =A0 =A0 =A0begin
> =A0 =A0 =A0 =A0 =A0 =A0 LIO.Print(A_Line);
> =A0 =A0 =A0 =A0 =A0 =A0 null;
> =A0 =A0 =A0 =A0 =A0 =A0 pragma Inspection_Point(A_Line);
> =A0 =A0 =A0 =A0 =A0end;
> =A0 =A0 =A0 end loop;
> =A0 =A0else
> =A0 =A0 =A0 loop
> =A0 =A0 =A0 =A0 =A0declare
> =A0 =A0 =A0 =A0 =A0 =A0 A_Line : constant String :=3D LIO.Getline;
> =A0 =A0 =A0 =A0 =A0begin
> =A0 =A0 =A0 =A0 =A0 =A0 LIO.Print(A_Line);
> =A0 =A0 =A0 =A0 =A0 =A0 null;
> =A0 =A0 =A0 =A0 =A0 =A0 pragma Inspection_Point(A_Line);
> =A0 =A0 =A0 =A0 =A0end;
> =A0 =A0 =A0 end loop;
> =A0 =A0end if;
>
> end Test_Line_IO;

Nice one...I'll try these out on Win23 and see what happens :-)

But surely "Put_Line" and "Get_Line" are preferable subprogram
names?...

Cheers
-- Martin