Theodore Y. Ts'o wrote in
|On Tue, Jun 23, 2020 at 12:43:01AM +0200, Steffen Nurpmeso wrote:
|> #?0|kent:steffen$ pkginfo -o `command -v col`
|> Package File
|> util-linux usr/bin/col
|> * This command is deprecated. The utility is in maintenance mode,
|> * meaning we keep them in source tree for backward compatibility
|> * only. Do not waste time making this command better, unless the
|> * fix is about security or other very critical issue.
|> * See Documentation/deprecated.txt for more information.
|I'll note that the Austin Group / The Open Group marked those commands
|as LEGACY in the Single Unix Specification V2 in 1997:
| The utilities in the table below are marked LEGACY. Various factors
| may have contributed to the decision to class a utility
| LEGACY. Application writers should not use functionality marked
| If a migration path exists, advice is given to application
| developers regarding alternative means of obtaining similar
| functionality. This information may be found in the APPLICATION
| USAGE sections on the relevant pages.
| No requirement beyond that which was in effect at the time that
| these utilities were marked LEGACY shall be applied to these
| calendar cancel cc col cpio cu dircmp dis egrep fgrep line lint lpstat
| mail pack pcat pg spell sum tar unpack uulog uuname uupick uuto
|The quote from the Single Unix Specification: "No requirement beyond
|that which was in effect at the time that these utilities were marked
|LEGACY shall be applied to these utilities." is not that different
|from "Do not waste time making this command better, unless the fix is
|about security or other very critical issue."
|And while I'm sure the Linux haters will be happy to try to blame
|Linux for the decision to declare pg, col, et.al as "legacy", in 1997
|The Open Group was hardly filled with Linux developers; '97 predates
|IBM and Oracle declaring their support for Linux (1998), the
|publication of "The Cathedral and the Bazaar (also 1998), the start of
|the GNOME project (1999), Red Hat Enterprise Linux (first released
|in 2002), and Ubuntu Linux (2004).
| - Ted
|P.S. Of course, the fact that the The Open Group tried to convince
|the world to stop using tar and cpio in favor to pax seems to be a
|strong indication that they forgot the lesson of King Canute. :-)
--End of <20200625015009.GB4655(a)mit.edu>
Ok. Yes, POSIX is a condensed pool of tradition, history and
knowledge. I do not play it well, let alone in its entirety.
STREAMS not at all. After the time when the "Unix community" came
together, more or less, as i understand it, in POSIX, to finally
find a common ground (again), the current "POSIX group" more or
less started collecting "behind the pack" what can be standardized
as a common foundation, to fill gaps, and to clarify ambiguities.
I have the impression that POSIX is referred to quite often ---
sometimes i have to search the net to find more or less nothing
good, stackoverflow, superuser, you-know (or maybe even not, having
grown in this environment and even at such an outstanding
position, and for such a long time), then POSIX shines through and
stands out from this mess. One of the main contributors stated to
me years ago something like "POSIX is an excellent foundation.."
The current Issue 7, 2018 edition of POSIX states on page 3481,
line 117589 ff.
The term ``legacy’’ was included in earlier versions of this
standard but is no longer used in the current version.
For col(1) that means, unfortunately, to abandon all hope. It
does not exist. (It never did in any POSIX document i have in
full text, not a moment i had an intent to look there.)
Actually i was astonished. You said IBM (i do not know Oracle)
declared support for Linux in 1998! I would have sworn that there
was a big billion dollar announcement in 2003 or around that time,
with "evangelists" appearing on eg Linux Magazine (of Germany,
what i know..). I always, foolishly, state this must have been
around 2005, the time when the wonderful Linux HOWTOs entered
their bit rot age.
Just yesterday, i use Linux for real since April 2019 (1999-2002
does not count, with all the out-of-the box SuSE, RedHat and
Halloween Linux, and Debian Woody) and take a week of learning
Linux better, ZFS on Monday, switch from proxy_arp=1 network for
my VMs and such to separate netns with
net<->host:veth-netns-veth:bridge<->vms++ on Tuesday (with almost
400 percent performance penalty, but: cool!), and since yesterday
i try for the first time to do my own kernel with initrd (only big
one for now, efi stub, builtin firmware, no modules except one
from staging which cannot be linked in etc.), in order to get an
allyesmod thing which i can use to plug into a box, lsmod>MODS,
and then do localyesconfig, you know. I hope i find some time.
However, just yesterday i wondered what would happen if suddenly
virt-manager etc. would stop working, how many people would know
how to continue? You really find nothing good on the web, and all
the billion dollar players and their forums etc, they are no help
at all. Luckily for ZFS there is the FreeBSD handbook (people who
care), which is a kind starter (or would be if you would really go
that route, i not for the moment, at least), and that i can apply
a bridge onto a veth on a wifi network device that allows no
bridge to be attached crossed my mind, finally, yay. For that
proxy_arp thing i could read iptables-tut and Linux Advanced
Routing & Traffic Control HOWTO, which says "Can be very useful
when building 'ip pseudo bridges'. Do take care that your netmasks
are very correct before enabling this! Also be aware that the
rp_filter, mentioned elsewhere, also operates on ARP queries!" How
could anyone hate people who write such HOWTOs?
But other than that, POSIX is also always worth looking at, right.
Quite off topic that all, though..
|Der Kragenbaer, The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)