Write a string in EBCDIC

[John Leslie]

I need to write a string to a file in EBCDIC.

Do I need to do it character by character using a translation table,
or is there a function to translate the whole string?

(I am aware that I can convert a whole file using Unix utilities, but
this file will have only a few header records in EBCDIC)

[SM Ryan]

jo********@mada safish.com (John Leslie) wrote:
# I need to write a string to a file in EBCDIC.
#
# Do I need to do it character by character using a translation table,
# or is there a function to translate the whole string?
#
# (I am aware that I can convert a whole file using Unix utilities, but
# this file will have only a few header records in EBCDIC)

http://www.room42.com/store/computer...ranslate.shtml

--
SM Ryan http://www.rawbw.com/~wyrmwif/
Don't say anything. Especially you.

[CBFalconer]

John Leslie wrote:

I need to write a string to a file in EBCDIC.

Do I need to do it character by character using a translation
table, or is there a function to translate the whole string?

If the system native char set is EBCDIC, just write the string.
The whole problem is system specific, not language specific, and
thus is off-topic in c.l.c. It requires knowing both the source
and destination char sets and their coding. I would create a
translation function, and customize it for each specific system.

Try a group that deals with your specific system.

--
Chuck F (cb********@yah oo.com) (cb********@wor ldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home .att.net> USE worldnet address!

[Lew Pitcher]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

SM Ryan wrote:
| jo********@mada safish.com (John Leslie) wrote:
| # I need to write a string to a file in EBCDIC.
| #
| # Do I need to do it character by character using a translation table,
| # or is there a function to translate the whole string?
| #
| # (I am aware that I can convert a whole file using Unix utilities, but
| # this file will have only a few header records in EBCDIC)
|
| http://www.room42.com/store/computer...ranslate.shtml

I'd be cautious in using that example. For one thing, it doesn't indicate
/which/ EBCDIC it is translating ASCII to, and for another, it performs at
least 256 improper ASCII-to-EBCDIC conversions, if not more.

For what it's worth, there are about 30 different EBCDIC variants, each with
differences in the order and codepoint of common characters, and the
assignment of codepoints to disparate characters. You might want to take a
look at the various characterset mappings at http://anubis.dkuug.dk/i18n/charmaps/


- --
Lew Pitcher

Master Codewright & JOAT-in-training | GPG public key available on request
Registered Linux User #112576 (http://counter.li.org/)
Slackware - Because I know what I'm doing.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFAy8VLagV FX4UWr64RAhUTAJ wKsjG10yqHt3ekE xcB3Mh8e0TdWgCg sxnz
c4MlgfDNzFr6oh/JVg3ju9I=
=+wfr
-----END PGP SIGNATURE-----

[Lew Pitcher]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Lew Pitcher wrote:
| SM Ryan wrote:
| | jo********@mada safish.com (John Leslie) wrote:
| | # I need to write a string to a file in EBCDIC.
| | #
| | # Do I need to do it character by character using a translation table,
| | # or is there a function to translate the whole string?
| | #
| | # (I am aware that I can convert a whole file using Unix utilities, but
| | # this file will have only a few header records in EBCDIC)
| |
| | http://www.room42.com/store/computer...ranslate.shtml
|
| I'd be cautious in using that example. For one thing, it doesn't indicate
| /which/ EBCDIC it is translating ASCII to, and for another, it performs at
| least 256 improper ASCII-to-EBCDIC conversions, if not more.

Correction - I meant "at least 128 improper ASCII-to-EBCDIC conversions"
- --
Lew Pitcher

Master Codewright & JOAT-in-training | GPG public key available on request
Registered Linux User #112576 (http://counter.li.org/)
Slackware - Because I know what I'm doing.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFAy8XnagV FX4UWr64RAvlCAJ 4tuCL5Rb8GWKYZ6 SmZgh1MQ2NsoACf Yhcz
hySLhMx4qKqqi+J jKfSeFnE=
=8/jR
-----END PGP SIGNATURE-----

[nospam]

CBFalconer <cb********@yah oo.com> wrote:

John Leslie wrote:

I need to write a string to a file in EBCDIC.

Do I need to do it character by character using a translation
table, or is there a function to translate the whole string?

If the system native char set is EBCDIC, just write the string.
The whole problem is system specific, not language specific, and
thus is off-topic in c.l.c. It requires knowing both the source
and destination char sets and their coding. I would create a
translation function, and customize it for each specific system.

Try a group that deals with your specific system.

Is it really off-topic? I'm not so sure. The problem can be solved
in a 100% portable way that is not system-specific at all:

void setup_xlate_tab le(void)
{
...

xlate['a'] = /* EBCDIC value for 'a' */;
xlate['b'] = /* EBCDIC value for 'b' */;
xlate['c'] = /* EBCDIC value for 'c' */;

...
}

Write a translate function and voila, you have a portable
native-charset-to-EBCDIC-charset conversion solution.

[Lew Pitcher]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

no****@nowhere. com wrote:

CBFalconer <cb********@yah oo.com> wrote:

John Leslie wrote:

I need to write a string to a file in EBCDIC.

Do I need to do it character by character using a translation
table, or is there a function to translate the whole string?

If the system native char set is EBCDIC, just write the string.
The whole problem is system specific, not language specific, and
thus is off-topic in c.l.c. It requires knowing both the source
and destination char sets and their coding. I would create a
translation function, and customize it for each specific system.

Try a group that deals with your specific system.

Is it really off-topic? I'm not so sure. The problem can be solved
in a 100% portable way that is not system-specific at all:

void setup_xlate_tab le(void)
{
...

xlate['a'] = /* EBCDIC value for 'a' */;
xlate['b'] = /* EBCDIC value for 'b' */;
xlate['c'] = /* EBCDIC value for 'c' */;

...
}

Write a translate function and voila, you have a portable
native-charset-to-EBCDIC-charset conversion solution.

1) /which/ EBCDIC?

2) How do you handle characters that exist in ASCII, but not in the target
EBCDIC? For instance, how do you translate 0x5d (ASCII ']') to EBCDIC-US?

3) How do you handle char values that are not ASCII to begin with? For
instance, how do you translate 0x9A (not an ASCII character) to EBCDIC-US?
- --
Lew Pitcher

Master Codewright & JOAT-in-training | GPG public key available on request
Registered Linux User #112576 (http://counter.li.org/)
Slackware - Because I know what I'm doing.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFAy9ZPagV FX4UWr64RAlp+AJ 459ha4S4Gyi46PK 9ZOeRsDsWdNgwCf UHN4
3joRaeW+GWPREx/MQDjVBkM=
=ImUU
-----END PGP SIGNATURE-----

[Ralmin]

"Lew Pitcher" <lp******@sympa tico.ca> wrote:

no****@nowhere. com wrote:

Is it really off-topic? I'm not so sure. The problem can be solved
in a 100% portable way that is not system-specific at all:

void setup_xlate_tab le(void)
{
...

xlate['a'] = /* EBCDIC value for 'a' */;
xlate['b'] = /* EBCDIC value for 'b' */;
xlate['c'] = /* EBCDIC value for 'c' */;

...
}

Write a translate function and voila, you have a portable
native-charset-to-EBCDIC-charset conversion solution.
1) /which/ EBCDIC?

Write a separate setup_xlate_tab le function for each EBCDIC you are
interested in.
2) How do you handle characters that exist in ASCII, but not
in the target EBCDIC? For instance, how do you translate
0x5d (ASCII ']') to EBCDIC-US?

3) How do you handle char values that are not ASCII to begin
with? For instance, how do you translate 0x9A (not an ASCII
character) to EBCDIC-US?

Simple answer: Replace it with a question mark, issue a warning.

Complex answer: Do whatever's appropriate for the particular type of file
you're translating.

--
Simon.

[Lew Pitcher]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ralmin wrote:

"Lew Pitcher" <lp******@sympa tico.ca> wrote:

no****@nowher e.com wrote:

Is it really off-topic? I'm not so sure. The problem can be solved
in a 100% portable way that is not system-specific at all:

void setup_xlate_tab le(void)
{
...

xlate['a'] = /* EBCDIC value for 'a' */;
xlate['b'] = /* EBCDIC value for 'b' */;
xlate['c'] = /* EBCDIC value for 'c' */;

...
}

Write a translate function and voila, you have a portable
native-charset-to-EBCDIC-charset conversion solution.
1) /which/ EBCDIC?

Write a separate setup_xlate_tab le function for each EBCDIC you are
interested in.

OK, so about thirty setup_xlate_tab le functions, and the /caller/ has to know
which EBCDIC to select. That means that the caller needs to incorporate the
infrastructure to determine which target EBCDIC is the one to use, and his
logic needs to encorporate some sort of user input query (cmdline options or
envvar checking or user-interactive options). From /my/ experience (and I've
had a lot of experience), it's not easy.

2) How do you handle characters that exist in ASCII, but not
in the target EBCDIC? For instance, how do you translate
0x5d (ASCII ']') to EBCDIC-US?

3) How do you handle char values that are not ASCII to begin
with? For instance, how do you translate 0x9A (not an ASCII
character) to EBCDIC-US?

Simple answer: Replace it with a question mark, issue a warning.

Bzzt. Wrong answer. Or, at least wrong replacement value. Tell me what the
correct replacement value should be. Also, issue a warning how, and to whom?

Speaking from experience, substitution is the 'correct' way to go, but also
the wrong way to go. Assume that the data the OP wants to translate looks like

#include <stdio.h>
int main(void)
{
char text[] = "2^3 = 8\n";
printf("%s",tex t);
return 0;
}

Will single character substitution result in an EBCDIC file that preserves the
meaning of the translated ASCII data? I can answer simply: no, it will not.
Square brackets do not exist in some EBCDIC variants, and exist with different
codepoints (varying in value from EBCDIC variant to variant) in others. The
caret symbol doesn't exist in most EBCDIC variants, and is typically
translated to the PL/1 'logical not' sign ( which sort of looks like: '-|' ).
Substituting '?' (or even the proper SUB character) for the untranslatable
characters would invalidate the meaning of the data being transferred. In this
case, with your suggested substitution, the data becomes

#include <stdio.h>
int main(void)
{
char text?? = "2?3 = 8\n";
printf("%s",tex t);
return 0;
}

Which results in a textual file that is /not/ equivalent to the original text
file.

Complex answer: Do whatever's appropriate for the particular type of file
you're translating.

Closer, but that's the whole point of this exercise. The OP wants to
"translate ASCII to EBCDIC"; that's the "particular type of file" he's
translating. What is appropriate for that?
- --
Lew Pitcher

Master Codewright & JOAT-in-training | GPG public key available on request
Registered Linux User #112576 (http://counter.li.org/)
Slackware - Because I know what I'm doing.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFAzJLxagV FX4UWr64RAmsbAJ 9d75Xv3Nsuepz8/FSgp0ITt5K3ygCg xCuD
etMA7J9MHahzBZ5 RySpiajo=
=lBrZ
-----END PGP SIGNATURE-----

[CBFalconer]

Lew Pitcher wrote:

Ralmin wrote:
.... snip ...
Which results in a textual file that is /not/ equivalent to the
original text file.

Complex answer: Do whatever's appropriate for the particular
type of file you're translating.

Closer, but that's the whole point of this exercise. The OP
wants to "translate ASCII to EBCDIC"; that's the "particular
type of file" he's translating. What is appropriate for that?

Seems to me that one rule that the tables should observe is
uniqueness. I.e. whether or not the translation makes sense, the
result should be translatable back into the original without
loss. In a way this is what happens when you pack a text file
into a zip file. The algorithm is lossless.

--
Chuck F (cb********@yah oo.com) (cb********@wor ldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home .att.net> USE worldnet address!