SCO UNIX (3.2.2) 'ls' is broken... what a shame

[Larry Philps]

pizzi at esacs.UUCP (Riccardo Pizzi) writes:
>
>We installed recently the new SCO UNIX release, i.e. 3.2.2.
>
>Please, look at the output of the 'lc' command on my home directory:
>
>AAA            house.c        mbox           remind22.5     sps.2
>active.uunet   hv at uwasa.fi    mte215.zip.1   s.c            sps.3
>alt.groups     hv at uwasa.fi.h  mte215.zip.2   sb.1           sucker
>anonftp.sites  i2u.info.h     nws            sb.2           TTT
>batchout.c     i2u.info.i     periodic       sb.3           T1500.uunet
>casa           maild          prog           sb.5           uunet.games
>ciuccia        maild.c        remind22.1     sb.6           uunet.unix
>conq4          Makefila       remind22.2     sb.7           uunet.x
>ftp-server.h   makefile       remind22.3     sig
>gifhelp        Makefile       remind22.4     sps.1
>
>In the ASCII character set, I have always thought the capital (upper case)
>letters should come *before* the lower case letters, but this seems not true
>for SCO...
>If we look at the following files: maild.c, Makefile, makefile e Makefila,
>we would expect to see the following:
>
>	   Makefila
>	   Makefile
>	   maild.c
>	   makefile
>
>now, take a look and see how SCO sorted the files...
>The first problem that comes to my mind is that the old habit of giving
>important files an uppercase name (or capitalized, as you prefer) does not
>work here... what a shame!
>I thought the bug could depend on the internationalization stuff, but never

Bingo.

>had the time to dig into it to debug the problem (I don't like SCO UNIX
>at all, agreeing with most developers out there so I don't want to do it, too).
>
>BTW: today I found that the numbers always come *after* all other letters:

This is not a bug, rather a feature :-).  Under the ANSI and XPG3
internationalization (I18N for short), practically everything is
variable.  There are languages out there in which words are sorted by the
*first vowel in a word* rather than the first letter.  Some letters have
to be treated as 2, some strings have to be treated as a single letter.
It goes on and on.  The whole thing is incredibly complicated.

Anyway, what you have done is set your default environment to
english_us.8859, rather than english_us.ascii.  Thus you have asked for
ISO 8859 sorting rules rather than ascii sorting rules, and are getting
case independent sorting among other things.  I was also confused when
this happened the first time.

You can do lots of things to solve this (as is typical in I18N)

1) Edit /etc/default/lang, and change the 8859 to ascii.  This changes
   the default behaviour for the *entire* system.

2) Set the environment variable LANG from your shell

	% setenv LANG english_us.ascii

   to change just the behaviour you see, but leave everything else alone.

3) Set the environment variable LC_COLLATE (it controls sorting criteria)
   from your shell

	% setenv LC_COLLATE english_us.ascii

   Then your environment will use 8859 rules except for sorting, which
   will change to the ascii conventions. 

Have fun.

---

I am not an I18N expert, but have actually written code during a project
to use this stuff.  I found myself confused for months about the *right*
way to do things.

If anyone else out there has done any I18N programming (under SCO or any
other system) and has some *pearls of wisdom* (complaints?) to impart, I
would be happy to listen to them.  I don't promise to be able to solve,
fix or change anything, but I18N is virtually certain to affect my life
as a programmer more and more as time goes on, and I would not mind getting
a bit of a head start.

---
Larry Philps,	 SCO Canada, Inc (Formerly: HCR Corporation)
Postman:  130 Bloor St. West, 10th floor, Toronto, Ontario.  M5S 1N5
InterNet: larryp at sco.COM  or larryp%scocan at uunet.uu.net
UUCP:	 {uunet,utcsri,sco}!scocan!larryp
Phone:	 (416) 922-1937
Fax:	 (416) 922-8397