Re: Removing MULTI_KBOARD

[Dan Nicolaescu <dann@ics.uci.edu>]

Eli Zaretskii <eliz@gnu.org> writes:

  > > From: Dan Nicolaescu <dann@ics.uci.edu>
  > > Cc: "Juanma Barranquero" <lekktu@gmail.com>, eliz@gnu.org, emacs-devel@gnu.org
  > > Date: Fri, 01 Aug 2008 23:55:57 -0700
  > > 
  > > The code was turned off by default, unused and unnecessary, and that has
  > > been explained on the list almost a year ago.  If someone has a problem
  > > with that, it would be much more productive to explain why, but not just
  > > hand waving, with some real experience/examples.  Otherwise please let
  > > bygones be bygones, it's a terrible waste of time.
  > 
  > I did explain why I have a problem with this: it breaks the MSDOS
  > port.

You have handwaved that.  Can you please actually explain why you think that?
I have explained it is not a problem.

Given that you really insist on this, maybe it would be a good idea to
for have a look at the whole picture:

1. The MSDOS port is broken for other reasons, it has been like that for a
year.  Nobody has complained.
2. The MSDOS port was on a widely publicized list of platforms that we
indented to remove.  Nobody requested it to be kept, so RMS approved its
removal.
3. The MSDOS port might never be revived, you have never committed to do
so, the only statements you have made was "I might do it".

All working platforms use MULTI_KBOARD by default, removing the extra
code makes everything cleaner and easier to read, it was removed because
it has been getting in my way when trying to chase an infrequently
occurring event/input bug.

  > Now, since removing MULTI_KBOARD obviously was not discussed in
  > advance, 

This implies that there would be any need for such a discussion.  There
isn't, the code is better now without the #ifdefs.

  > I'm asking Yidong and Stefan to decide which of the following
  > is preferable:
  > 
  >   . Revert Dan's changes that removed MULTI_KBOARD.

Please do not do such a thing, this would be a very bad case of priority
inversion: forcing to keep code just because it might (although it is
doubtful) make it easier to revive a port to a very old, proprietary
platform with NO ACTIVE USERS, otherwise broken and unmaintained for one
year, hence a low priority task for a GNU project.  The code is an
inconvenience and requires more work when doing maintenance and
debugging maintenance on GNU/Linux, i.e a high priority task for GNU.

If the MSDOS port ever gets fixed (which, again, is not sure to happen),
it can be fixed from a CVS checkout prior to the removal of the
MULTI_KBOARD code.  If it turns out that the MULTI_KBOARD code is
needed, it can be restored at that time.  I stand by my claim that such
a thing would not be necessary.

  >   . Replace those #ifdef's that are needed by the MSDOS port (I don't
  >     know whether this means all of them or not) with #ifdef MSDOS.

That is even worse, the unnecessary code is still in the way of all
the GNU/Linux maintainers.