On Mon, 29 Sep 2014 09:35:36 -0400, John Gilmore wrote:
>
>The notion that some code points in a character set---SBCS, DBCS, or
>MBCS---are "characters" and others are not is, however, a supremely
>silly one.
>
At best, it's an utter impediment to bijective (round-trip) conversion.
Fortunately, however, z/OS iconv does not enforce that notion:
13 *-* 'iconv -f IBM-1047 -t UTF-8 <a1 >a2'
>>> "iconv -f IBM-1047 -t UTF-8 <a1 >a2"
14 *-* 'iconv -t IBM-1047 -f UTF-8 <a2 >a3'
>>> "iconv -t IBM-1047 -f UTF-8 <a2 >a3"
15 *-* 'ls -alrt'
>>> "ls -alrt"
total 48
drwxr-sr-x 67 user group 383 Oct 2 16:45 ..
-rwxr-xr-x 1 user group 328 Oct 2 16:56 ctest
-rw-r--r-- 1 user group 826 Oct 2 16:57 report
-rw-r--r-- 1 user group 256 Oct 2 16:57 a3
-rw-r--r-- 1 user group 384 Oct 2 16:57 a2
-rw-r--r-- 1 user group 256 Oct 2 16:57 a1
drwxr-sr-x 2 user group 7 Oct 2 16:57 .
16 *-* 'cksum a*'
>>> "cksum a*"
1313719201 256 a1
716397259 384 a2
1313719201 256 a3
17 *-* 'diff -su a1 a3'
>>> "diff -su a1 a3"
Files a1 and a3 are identical
total 48
4724741 drwxr-sr-x 67 user group 383 Oct 2 16:45 ..
10695729 -rwxr-xr-x 1 user group 328 Oct 2 16:56 ctest
10695735 -rw-r--r-- 1 user group 826 Oct 2 16:57 report
10695734 -rw-r--r-- 1 user group 256 Oct 2 16:57 a3
10695733 -rw-r--r-- 1 user group 384 Oct 2 16:57 a2
10695730 -rw-r--r-- 1 user group 256 Oct 2 16:57 a1
10695728 drwxr-sr-x 2 user group 7 Oct 2 16:57 .
The z/OS LF<->NL flipflop remains a nonconformance to standard
and an obstacle to portability.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
