tiny-cobol-users

RE: [Tiny-cobol-users] Some fixes for indexed file...

[Andrew C. <an...@an...>]

Hi Rildo,

Yes I think that the compiler should generate an error if the key is is
missing on the Select statement for an indexed file.

The whole point of an indexed file is that it is indexed on a primary key
that must be unique.

Regards
Andrew
 
On Thu, 16 Dec 1999, Rildo Pragana wrote:

> Date: Thu, 16 Dec 1999 00:26:34 -0200 (EDT)
> From: Rildo Pragana <rpr...@ac...>
> Reply-To: tin...@so...
> To: tin...@ma...
> Subject: RE: [Tiny-cobol-users] Some fixes for indexed file...
> 
> Hi Andrew,
> 
> I have got your mail from our archive, for I'm still without pop3 :(
> 
> Well, if the "key" clause is needed, then Glen's code is buggy. It tried
> to open the indexed file without it. Please see his test suite.  He will
> only come back saturday.
> If you agree on that, I can enforce this condition in the compilation,
> generating a warning/error.
> 
> regards,
>  ----------------------------------------------------------------
>  Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
>  P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
>  Brazil 54792-990              +55-81-459-1776
> 
> 
> _______________________________________________
> Tiny-cobol-users mailing list
> Tin...@li...
> http://lists.sourceforge.net/mailman/listinfo/tiny-cobol-users
> 

-----------------------------------------------------------------------------

Andrew Cameron
Internet : an...@an...
X.400    : C=ZA G=Andrew S=Cameron Admd=TELKOM400

----------------------------------------------------------------------------

[Tiny-cobol-users] Updated cvs server

[Rildo P. <rpr...@ac...>]

Hi,

I have updated now our cvs server with the latest changes for ncurses
screen i/o, including a small fix after the xoom snapshot was uploaded.
Please disregard it and go to our sourceforge server.

All test.code programs run well (expect for t13, tht already had a
problem). Some of them sows garbage in the screen, but because of
displaying non-initialized fields.

best wishes,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

[Tiny-cobol-users] Are you curious?

[Rildo P. <rpr...@ac...>]

Hi,

Our cobol compiler ran in it's past for allowing HandTerms to communicate
with a msdos(TM) computer, for doing data-entry, point-of-sale operations,
access control in parking, and other applications.

I have sent to my home page two scanned images of HandTerm's folder. If
you want to know a little the prehistoric origins of our current software,
go see it:

http://members.xoom.com/rpragana/HandTerm-1.jpg
and
http://members.xoom.com/rpragana/HandTerm-2.jpg

regards,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

RE: [Tiny-cobol-users] Some fixes for indexed file...

[Rildo P. <rpr...@ac...>]

Hi Andrew,

I have got your mail from our archive, for I'm still without pop3 :(

Well, if the "key" clause is needed, then Glen's code is buggy. It tried
to open the indexed file without it. Please see his test suite.  He will
only come back saturday.
If you agree on that, I can enforce this condition in the compilation,
generating a warning/error.

regards,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

[Tiny-cobol-users] About making the compiler output C.

[Rildo P. <rpr...@ac...>]

Hi,

Andrew have already said everything. We prefer to have something now, even
in assembly language, but that is usable for many Cobol developers, than
later.  When we finish this, we will start a redesign of the compiler for
generating C. In fact, I have already started this, but I prefer to
concentrate my bullets in delivering the current compiler.

The redesign is needed because of many constraints I found neccessary
when making the original compiler and now it's difficult to clean up
everything. But even by patching this ugly code, we are learning much to
make a cleaner and simple design at the next phase.

Of course, we have many plans for the months even years to come. We plan
embedded sql, a gui interface, and more. And we want C output, naturally.

best wishes,
Rildo

P.S. - My isp is doing a bad work again. I'm without pop3 connection, so
please don't send mails directly to my mailbox. I expect tomorrow it will
be back.  Sorry for any unanswered messages.

 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

Re: [Tiny-cobol-users] Some fixes for indexed file and copy statement.

[Andrew C. <an...@an...>]

On Wed, 15 Dec 1999, Rildo Pragana wrote:

> Other problem is when compiling an indexed file without any key
> description.  The compiler previously dumped core.  Now it generates a
> NULL descriptor pointer (so Andrew, take care of it in the open procedure,
> please).

As far as I know an Indexed File MUST have a key and so the Compiler
should generate an Error.
My Cobol manual States.

After you have specified INDEXED for a file organization, you MUST name a
field that appears in the file record as the key (primary key) to the
file.

Regards
Andrew
-----------------------------------------------------------------------------

Andrew Cameron
Internet : an...@an...
X.400    : C=ZA G=Andrew S=Cameron Admd=TELKOM400

----------------------------------------------------------------------------

[Tiny-cobol-users] Alternate Keys,

[Andrew C. <an...@an...>]

Hi,

I have written the code for READ PREVIOUS and READ NEXT for Alternate
Keys.

I have tested READ PREVIOUS see t04 and it works.

Read Next also works but there is a bug in the compiler that needs to be
fixed before you can use it.

Rildo:- Please investigate why the compiler returns "sort_return" instead
of "cob_read_next" as this causes it to call the wrong routines. I think
the problem is in htcobol.y where the read syntax is defined.

 if ($1==1)
	gen_return($2,$4)
  
Regards
Andrew Cameron

-----------------------------------------------------------------------------

Andrew Cameron
Internet : an...@an...
X.400    : C=ZA G=Andrew S=Cameron Admd=TELKOM400

----------------------------------------------------------------------------

[Tiny-cobol-users] Some fixes for indexed file and copy statement.

[Rildo P. <rpr...@ac...>]

Hi,

As Glen reported, the copy statement have a problem when the last
character in file name is the period (end of statement). I have change a
little the scanner to fix that. The period at the end is still required
(or there is no way to know the copy ended).  The drawback is that we cant
have a file name ending with a period...

Other problem is when compiling an indexed file without any key
description.  The compiler previously dumped core.  Now it generates a
NULL descriptor pointer (so Andrew, take care of it in the open procedure,
please).

So, both bugs are fixed.
I'm still waiting for bug reports on the "inspect" statement (replacing
clause).

regards,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

RE: [Tiny-cobol-users] Re: Tiny Cobol

[Rolle, T. <Ted...@ug...>]

FWIW: The current effort focuses on implementing an old version of COBOL.
As a practicing COBOL programmer I believe that our effort would be better
received if we target COBOL-3 syntax.

One of the first things we do with legacy code is to upgrade it to COBOL-3;
implementing a compiler that doesn't support the new syntax is -- IMNSHO --
a waste of time.

By the time we get around to implementing COBOL-3, the 'suits' will already
have a taste in their mouths for our compiler.  We put the "backward" in
"backward compatible".

[Tiny-cobol-users] Alternate Keys

[Andrew C. <an...@an...>]

Hi,

I have written the basic code for read to handle alternate keys.

So far OPEN, CLOSE, READ KEY IS, WRITE all work for Alternate KEYS.

You can have as many Alternate keys as you like but each key will slow
down the overall speed that your programs will run at.

Still to be coded READ NEXT, READ PREVIOUS, START, DELETE, REWRITE as well
as handling error conditions for the code that currently works.

Have a look at t04 for some examples.

Glen please add some programs to your test suite for alternate keys.

Regards
Andrew Cameron

-----------------------------------------------------------------------------

Andrew Cameron
Internet : an...@an...
X.400    : C=ZA G=Andrew S=Cameron Admd=TELKOM400

----------------------------------------------------------------------------

[Tiny-cobol-users] Re: Tiny Cobol

[Andrew C. <an...@an...>]

Hi,

Yes we would like to ultimatly see this as a front end for gcc.

Rildo is currently driving the direction we take.

The reason that we have started off on the current route is that Rildo
released a version of his cobol compiler that he had written for handheld
devices. This compiler generated normal Intel Assembler. We have used this
compiler as a starting point to attempt to get a Compiler for cobol up and
running as fast as possible. All we really need to do currently is finish
off the library routines and add the syntax that the original compiler did
not support.

The reason that we are in a hurry is that there have been many requests
for a usable GPL Cobol Compiler on the Linux/GNU OS. So Far there have
been attempts like the current GNU Cobol effort but they are taking far
too long. I think that there are many MainFrame shops that would LOVE to
have a GPL Cobol Compiler for Linux as it will help them Port many of
their Systems to linux.

To Port this current compiler to Sparc etc will not be difficult but as
you say it will be much easier once we have converted this compiler to a
front end for GCC

Rildo would you care to comment on this and map out the future direction
as you see it.

regards
Andrew


On Wed, 15 Dec 1999, Richard Stallman wrote:

> Date: Wed, 15 Dec 1999 13:19:42 -0500
> From: Richard Stallman <rm...@gn...>
> To: an...@an...
> Cc: tin...@so...
> Subject: Re: Tiny Cobol
> 
>     The current system creates Gnu Assembler which is then assembled and
>     linked with the library routines. A later version may output C code.
> 
> How do you port it to new machines?
> 
> Have you considered making this a front-end for GCC?
> Then it would automatically support all the machines
> that GCC supports, and you would have much less work to do.
> 

-----------------------------------------------------------------------------

Andrew Cameron
Internet : an...@an...
X.400    : C=ZA G=Andrew S=Cameron Admd=TELKOM400

----------------------------------------------------------------------------

[Tiny-cobol-users] Re: Tiny Cobol

[Richard S. <rm...@gn...>]

    The current system creates Gnu Assembler which is then assembled and
    linked with the library routines. A later version may output C code.

How do you port it to new machines?

Have you considered making this a front-end for GCC?
Then it would automatically support all the machines
that GCC supports, and you would have much less work to do.

Re: [Tiny-cobol-users] Indexed IO (problem?)

[Andrew C. <an...@an...>]

Hi Glen,

This is currently working. It seems that you may have found a bug.

Please e-mail me the program and I will see if I can find the problem.

Regards
Andrew

On Tue, 14 Dec 1999, Glen Colbert wrote:

> Date: Tue, 14 Dec 1999 17:23:50 -0700
> From: Glen Colbert <gco...@US...>
> Reply-To: tin...@so...
> To: tin...@so...
> Subject: [Tiny-cobol-users] Indexed IO (problem?)
> 
> Andrew,
> 
>   I tried writing a test program where the ORGANIZATION IS INDEXED and the
> ACCESS IS SEQUENTIAL.  I realize that we are early on in the I-O routines,
> but it crashed ugly (core dump during compile).  Do you want me to open a
> bug report, or is this just an oversight?
> 
> GLen
> 
> 
> _______________________________________________
> Tiny-cobol-users mailing list
> Tin...@li...
> http://lists.sourceforge.net/mailman/listinfo/tiny-cobol-users
> 

-----------------------------------------------------------------------------

Andrew Cameron
Internet : an...@an...
X.400    : C=ZA G=Andrew S=Cameron Admd=TELKOM400

----------------------------------------------------------------------------

Re: [Tiny-cobol-users] (no subject)

[Fred M. <fr...@mo...>]

Boris Kortiak wrote:

> I think BCD stands for binary coded decimal and the IBM implementation produces a binary number aligned to some boundary (byte, word, &c).  It is not a packed format.

Sorry, it _is_ a packed format and not a binary format. The only difference in the PICTURE string is the presence or absence of the sign indicator S. With the S it's a
packed decimal field, without it's a BCD field. Alignment is, as with packed decimal fields, at byte boundaries.

Alignment is just another story. Nearly all types of COBOL fields are aligned on byte boundaries. The SYNCHRONIZED clause for binary fields indicates alignment on other
boundaries. Binary integer fields of 16 bit are aligned on halfword boundaries, bynari integer fields of 32 and 64 bits on word boundaries. A word here means a 32 bit
word. Floating points of 32 bits are aligned on word boundary, floating points of 64 bit are aligned on double word boundary.

>
>
> As an aside BCD is part of EBCDIC: Extended Binary Coded ...

That is Extended Binary Coded Decimal Interchange Code.

Regards,

Fred Mobach

Re: [Tiny-cobol-users] COPY problem

[Fred M. <fr...@mo...>]

Glen Colbert wrote:

> I tried to include file definitions for some of the test programs using the
> COPY statement.  In the test.code, copy works with a rather complex COPY
> REPLACING.  However, I was just using COPY filename.cpy
>
> Is there something else I need to include to delimit the copy statement?
>
> SOURCE:
>        FILE-CONTROL.
>         SELECT TEST-FILE
>             ASSIGN TO "./idx_test.dat"
>             ORGANIZATION IS INDEXED
>             ACCESS MODE  IS SEQUENTIAL
>             RECORD KEY   IS TR-KEY
>             FILE STATUS  IS TR-FS.
>
>        COPY testse.cpy
>
>        DATA DIVISION.
>        FILE SECTION.
>        FD  TEST-FILE

Like my wife always said, I'm thinking as a COBOL compiler. As all
COBOL statements the COPY statement has to be terminated by a period and a
space.

Regards,

Fred Mobach

Re: [Tiny-cobol-users] I'll be back

[Rildo P. <rpr...@ac...>]

Hi Glen,

On Tue, 14 Dec 1999, Glen Colbert wrote:

> I get to go th Atlanta and flog a software vendor for the rest of the week.
> I'll see you all on Saturday!

Have a nice trip :)  I c you saturday.

best wishes,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

Re: [Tiny-cobol-users] COPY problem

[Rildo P. <rpr...@ac...>]

Hi Glen,

On Tue, 14 Dec 1999, Glen Colbert wrote:

>        COPY testse.cpy
> 

Please put a period after the file name.
And tell me the rules for it (I can't guess, sorry).

regards,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776

[Tiny-cobol-users] I'll be back

[Glen C. <gco...@US...>]

I get to go th Atlanta and flog a software vendor for the rest of the week.
I'll see you all on Saturday!

Glen

RE: [Tiny-cobol-users] (no subject)

[Glen C. <gco...@US...>]

> -----Original Message-----
> From: tin...@so...
> [mailto:tin...@so...]On Behalf Of Fred Mobach
> Sent: Tuesday, December 14, 1999 10:00 AM
> The packed decimal format, which is still used on mainframes because it
> saves memory and instruction cycles. It's not so fast as Integer binary
> arithmetic but faster than Display numeric arithmetic. So programmers
> should choose binary fields over packed decimal fields. But because of

Fred is right on regarding usage.  I did consider (for a while) that we may
want to store numbers in the FPU's register format.  Silly me.  An
interesting point, however, is that Cobol has never been a great
computational language.  Almost all math is simple arithmatic (add subtract
multiply divide) on integers with implied decimal points (strange integers?)

For purposes of data interchange with other Cobol systems, COMP-3s tend to
follow the IBM format.  I think we should consider this approach more for
the reasons of data interchangability than for performance.

> the very old Cobol standard in those days they didn't : the binary
> integer fields didn't have the possibility of the assumed decimal point,
> which is really needed in business practice.

This concept of implied decimal points in what are truly integers is central
to business programming (there are no fractional cents).  Business programs
RARELY round results.  The business default (regardless of language) is to
truncate.

Glen

RE: [Tiny-cobol-users] Suggestion for change of test suite.

[Glen C. <gco...@US...>]

It was a little more difficult that I expected, but the cobol_test.pl script
now builds the compiler from the PWD directory, instead of from the PATH.
cobpp and the other executables (gcc, as) still need to be on the path
however.

Glen



> -----Original Message-----
> From: tin...@so...
> [mailto:tin...@so...]On Behalf Of Rildo
> Pragana
> Sent: Tuesday, December 14, 1999 3:57 AM
> To: tin...@ma...
> Subject: [Tiny-cobol-users] Suggestion for change of test suite.
>
>
> Hi Glen,
>
> I suggest you make a small change in our test suite, so we don't need to
> manually setup the path for our files. Please make the
>
> $CCX=gcc;$LD=gcc;$ASM=as;$COB=cob;$COBPP=cobpp
>
> point to each argument in our directory tree, as everything is available
> there, now that your test suite is inside our directories.  This way will
> be easier for all developers just run cobol_test.pl.
>
> Example: COB=../compiler/htcobol
>
> regards,
>  ----------------------------------------------------------------
>  Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
>  P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
>  Brazil 54792-990              +55-81-459-1776
>
>
> _______________________________________________
> Tiny-cobol-users mailing list
> Tin...@li...
> http://lists.sourceforge.net/mailman/listinfo/tiny-cobol-users
>

RE: [Tiny-cobol-users] Fixed bug with "perform".

[Glen C. <gco...@US...>]

Looks good.  Once again, you are a programming god!

Thanks Rildo

> -----Original Message-----
> From: tin...@so...
> [mailto:tin...@so...]On Behalf Of Rildo
> Pragana
> Sent: Tuesday, December 14, 1999 3:50 AM
> To: tin...@ma...
> Subject: [Tiny-cobol-users] Fixed bug with "perform".
>
>
> Hi Glen,
>
> Thanx for the report. I hope there are no other bugs with it.
> BTW, we don't need to link lib/mccntrl.s, for there will no exit_procedure
> anymore.
>
> regards,
>  ----------------------------------------------------------------
>  Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
>  P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
>  Brazil 54792-990              +55-81-459-1776
>
>
> _______________________________________________
> Tiny-cobol-users mailing list
> Tin...@li...
> http://lists.sourceforge.net/mailman/listinfo/tiny-cobol-users
>

[Tiny-cobol-users] COPY problem

[Glen C. <gco...@US...>]

I tried to include file definitions for some of the test programs using the
COPY statement.  In the test.code, copy works with a rather complex COPY
REPLACING.  However, I was just using COPY filename.cpy

Is there something else I need to include to delimit the copy statement?

Glen

SOURCE:
       FILE-CONTROL.
        SELECT TEST-FILE
            ASSIGN TO "./idx_test.dat"
            ORGANIZATION IS INDEXED
            ACCESS MODE  IS SEQUENTIAL
            RECORD KEY   IS TR-KEY
            FILE STATUS  IS TR-FS.

       COPY testse.cpy

       DATA DIVISION.
       FILE SECTION.
       FD  TEST-FILE

COMPILER LISTING:
SCAN(  17):-> COPYFILE_ST
SCAN(  17): <IDSTRING>: testse.cpy
SCAN(  17):<-> previous_state>


SCAN(  19): <reserved>: DATA
*!ERROR!* 256 : line     19, symbol DATA: syntax error ***
*!ERROR!* 256 : line     19, symbol DATA: COPY statement not complete ***
SCAN(  19): <reserved>: DIVISION
SCAN(  19):<EOS>:

SCAN(  20): <reserved>: FILE
*!ERROR!* 256 : line     20, symbol FILE: I-O SECTION format wrong ***
SCAN(  20): <reserved>: SECTION
SC

[Tiny-cobol-users] Indexed IO (problem?)

[Glen C. <gco...@US...>]

Andrew,

  I tried writing a test program where the ORGANIZATION IS INDEXED and the
ACCESS IS SEQUENTIAL.  I realize that we are early on in the I-O routines,
but it crashed ugly (core dump during compile).  Do you want me to open a
bug report, or is this just an oversight?

GLen

RE: [Tiny-cobol-users] Coding Standard.

[Glen C. <gco...@US...>]

I tried -gnu -bap and REALLY liked it.  However, I can find a way to live
with whatever we do.

Glen

> -----Original Message-----
> From: tin...@so...
> [mailto:tin...@so...]On Behalf Of Andrew
> Cameron
> Sent: Monday, December 13, 1999 10:41 AM
> To: tin...@so...
> Subject: [Tiny-cobol-users] Coding Standard.
>
>
> Hi,
>
> I would like to suggest that all C code be beautified by using the GNU
> indent program with the gnu coding standard.
>
> Any Comments.
>
> Regards
> Andrew
>

Re: [Tiny-cobol-users] (no subject)

[Rildo P. <rpr...@ac...>]

Hi Boris,

On Tue, 14 Dec 1999, Boris Kortiak wrote:

>  I think BCD stands for binary coded decimal and the IBM implementation
>produces a binary number aligned to some boundary (byte, word, &c).  It
>is not a packed format.

I'm sorry to say, but it's packed, yes. You can place 2 digits/byte
because each digit only uses 4 bits to be represented and in a byte there
are 2 of such half-bytes or nibbles (as they are called).  This is comp-3
usage for cobol numbers.

There are even arithmetic units (in hardware design, my speciality) done
with bcd arithmetic. BTW, even the Intel processors have some instructions
to make easier operate with bcd numbers (DAA, the decimal adjust
accumulator, is one of such instructions. It adds 6 to each digit that
needs correction, forcing a carry on non-decimal digits).

> As an aside BCD is part of EBCDIC: Extended Binary Coded ...

EBCDIC is just a replacement for ASCII, as IBM does everything in it's own
way.  (at least it used to have such "personality" when it was the
monopolist).

> Of course I could be all wet.

Don't worry. I have made much larger mistakes with cobol code :)
Without resentments...

best regards,
 ----------------------------------------------------------------
 Rildo Pragana                 FPGA/uControllers * Linux * tcl/tk
 P.O. Box 721 Camaragibe PE    http://members.xoom.com/rpragana
 Brazil 54792-990              +55-81-459-1776