[haskell-mode] Work around coding cookie breakage

[Stephen J. Turnbull]

David Kastrup writes:

 > many people hating MULE with a bellicose vengeance switched to
 > XEmacs,

Interesting POV, although Erik Naggum has never once posted to our
channels that I know of. ;-)

Jokes aside, I never saw evidence of that.  We had a lot of USians
arguing against early removal of the no-Mule option, but most of them
were dyed-in-the-wool Lucid users, not reactionary ex-GNU Emacs users.
The big influx of developers/testers to XEmacs came before Emacs got
Mule.  They were generally coming *for* Mule, or for the open
development model.  They weren't coming to escape Mule.

And the to-XEmacs flow slowed as soon as Emacs 20 was released and the
developers' lists were opened, and maybe reversed as Mule stabilized
around 20.3 (at the very least we lost most of our Japanese developers
and never did get any Chinese or Korean developers, and of course the
Indic languages didn't have an option at XEmacs since we couldn't even
display them).

Of course we could both be right: people who were vocally against Mule
at Emacs switched to XEmacs (or vi?) and shut up.  They weren't
posting to our channels that I can recall, though.

 > If they don't get the choice of not using MULE at all, they'll
 > probably turn to the least painful MULE implementation that is
 > available.  And I don't think that, for a while to come, this would
 > mean XEmacs.

*chuckle*  You overestimate the difficulty of the problem.  And you
underestimate Ben Wing.  But the point is moot; I really don't think
Mule is going to drive adoption decisions.  It's going to be Unicode
support, and the question is whether Mule gets in the way of Unicode
or not.  Handa's record on that is two decades of foot-dragging.

Is emacs-unicode2 a genuine change of heart?  We'll see.  I have to
wonder though.  What Emacsen need for real Unicode support far more
than Unicode-in-the-buffer is bidi-in-the-display, and I never saw any
interest in that at all at the Mule Lab.  At the XEmacs/Mule
conference in 1998 or so, Eli Zaretskii and several other bidi
proponents were invited.  Those invitations went out about ten years
ago, but there's still no bidi support in any Emacs.

All of this is more relevant to Emacs than to XEmacs, though.  Unless
Ben Wing decides to start massive contribution again, we don't have
anybody we can count on for the code we need for bidi and Indic
support, and without Ben we can't easily borrow it from GNU either
because the Mule implementations are very divergent now.