Re: Testing new abbrev tables in elisp

["Andreas Röhler" <andreas.roehler@online.de>]

Am Sonntag, 4. November 2007 03:42 schrieb Stephen J. Turnbull:
> Richard Stallman writes:
>  >     In fact (Richard, would you please confirm?) it may be a good idea
>  > to use the Lisp implementation as a base to avoid legal issues if it
>  > looks "too much like" XEmacs code (the problem is that AFAIK you have
>  > looked at the XEmacs code, so couldn't swear that it's not an
>  > unintentional copy of someone else's code).
>  >
>  > That is valid in general, but is it an issue here?  Our old C code was
>  > written by me, mostly.
>
> Well, of course any code in Emacs has all the necessary papers.
>
> But as I understand it, Andreas wants to code in C so he can borrow
> techniques he saw in the XEmacs code.  It's fairly likely that code
> varies significantly from your code.
>

The head of abbrev.c shows remarks concerning authors
and displays GPL, everything looks fine for me. Maybe
exists a precise reason not to use that code for GNU Emacs?

XEmacs don't rely on single word abbrevs to be
expanded, but takes several words too. That's of
interest with NLP/translations and it's coded in C.

As it's at stake to look back or forward, speed
difference between Lisp and C execution might be
significant. At least that idea was in my head...

Thanks all

Andreas Röhler