Re: address@hidden: Building Emacs CVS against DJGPP CVS]

[Kenichi Handa]

In article <address@hidden>, Richard Stallman <address@hidden> writes:
>     I think the basic problem is that just loading fontset.el
>     calls functions in fontset.c (set-fontset-set, etc).  That's
>     why bytecompiling lisp/term/x-win.el fails if it requires
>     fontset unconditionally.  Perhaps, fontset.el should just
>     define a data for the default fontset, and x-win.el should
>     setup the default fontset from that data.  Then, we can
>     compile x-win.el without fontset.c.

> That seems like a good idea.

I've just installed these changes in HEAD, and confirmed
that --without-x also works on GNU/Linux.

2003-01-17  Kenichi Handa  <address@hidden>

        * international/fontset.el: Don't setup the default fontset and
        font-encoding-alist at the top level.
        (setup-default-fontset): New function.
        (create-fontset-from-fontset-spec): Delete autoload cookie.

        * term/x-win.el: Require fontset unconditionally again.
        Call setup-default-fontset at the top level.

        * term/w32-win.el: Call setup-default-fontset.

        * term/mac-win.el: Require fontset and call setup-default-fontset.

I have not yet changed w32-win.el and mac-win.el to require
fontset unconditionally because I can't test them.  I just
added a call of setup-default-fontset if new-fontset is
bound.

Could maintainer of these files please test if removing the
check of "(if (fboundp 'new-fontset)" works well?

---
Ken'ichi HANDA
address@hidden