Re: MH-E is broken

[Bill Wohler]

Miles Bader <address@hidden> writes:

>    Compiling /usr/local/src/emacs/master/lisp/mh-e/mh-alias.el
>
>    In toplevel form:
>    ../../../../src/emacs/master/lisp/mh-e/mh-alias.el:33:1:Error: Invalid 
> function: mh-require-cl
>    make[2]: *** [/usr/local/src/emacs/master/lisp/mh-e/mh-alias.elc] エラー 1
>
> That's after "make bootstrap".  It fails similarly (in other files
> sometimes) after other types of make too.

Glenn Morris <address@hidden> writes:
> I think its loaddef related. mh-e has always used "require" rather
> than "load" to load its loaddefs, to the extent of having a special
> Makefile rule in Emacs 22 to generate them.

Sorry folks, I've been buried under hundreds of unread messages and
didn't get to this message until now.

There have been few changes to MH-E since Emacs 23 was released, so an
incompatibility must have been introduced recently which broke
existing MH-E code.

I'd like to thank Stefan and Juanma for fixing MH-E, as well as Stefan
for replacing all the set-buffers with with-current-buffer--that was
on my list of things to do.

Glenn, you aren't suggesting that the use of require and the Makefile
rule are obsolete, are you? MH-E uses require to avoid reloading the
same file repeatedly and uses the same Makefile semantics as Emacs
uses to build loaddefs.el to build the mh-loaddefs.el file.

-- 
Bill Wohler <address@hidden>  http://www.newt.com/wohler/  GnuPG ID:610BD9AD