Re: mktexpk: non-POSIX compliant use of "tail"

[Paul Eggert]

address@hidden (Karl Berry) writes:

> What harm can there be in *GNU* tail always accepting tail -10?

I am becoming more inclined to agree with you.  I don't know about
Jim, though.

Part of the motivation for removing support for usages not allowed by
POSIX was conformance.  But part of it, to be honest, was the hope of
simplifying the maintenance burden.  The obsolete usage was buggy and
was never documented properly, and I thought it easier to drop it than
to fix and document it.  But simplification hasn't happened as much as
I'd like.

Assuming this POSIX compromise becomes official, I'm becoming more
inclined to add back support for the more commonly-used obsolete
usages that do not conflict with the revised POSIX.  This would
include "tail -1".  It would not include usages that conflict with
POSIX (e.g., "tail -c file" as an alias for "tail -c 10 file", or
"touch 04030201 file" as an alias for "touch -t 04030201 file").
However, I doubt whether many people rely on these conflicting usages
nowadays and you have to draw the line somewhere.

Alas, doing this correctly will require some thought -- it's not just
a simple switch-flip, since the new default behavior would be neither
the 1992 nor the 2001 behavior -- so it's not likely to happen this
week.

As I mentioned, I was concerned about the hassle of maintaining the
user documentation, as describing all the obsolete usages would bloat
the documentation and place an extra burden on all users.  So, if we
do add back support for these obsolete usages, I'd rather not document
them in detail.  I don't want to burden new users with this ancient
backwards-compatibility gorp.