Problems building current CVS trunk

Problems building current CVS trunk

[Eli Zaretskii]

With today's CVS, I get a failure compiling mh-show.el:

    Compiling d:/gnu/emacs/lisp/mh-e/mh-show.el...
    No MH variant found on the system

    In toplevel form:
    mh-e/mh-show.el:196:32:Error: No setf-method known for mh-buffer-data

Also, running "cvs up -d" after building Emacs shows that "make
autoloads" causes two files look modified as compared to the
repository:

    cvs update: Updating lisp/calc
    cvs update: Updating lisp/calendar
    cvs update: Updating lisp/emacs-lisp
    M lisp/emacs-lisp/cl-loaddefs.el
    cvs update: Updating lisp/emulation
    M lisp/emulation/tpu-edt.el
    cvs update: Updating lisp/erc
    cvs update: Updating lisp/eshell

I thought these problems were already solved in the past, was I
dreaming?

Re: Problems building current CVS trunk

[Glenn Morris]

Eli Zaretskii wrote:

> With today's CVS, I get a failure compiling mh-show.el:
>
>     Compiling d:/gnu/emacs/lisp/mh-e/mh-show.el...
>     No MH variant found on the system
>
>     In toplevel form:
>     mh-e/mh-show.el:196:32:Error: No setf-method known for mh-buffer-data

Works for me. No changes to mh-e in over a month. Suggest
bootstrapping.

> Also, running "cvs up -d" after building Emacs shows that "make
> autoloads" causes two files look modified as compared to the
> repository:

[cl-loaddefs.el and tpu-edt.el]

> I thought these problems were already solved in the past, was I
> dreaming?

You're probably thinking of:

http://lists.gnu.org/archive/html/emacs-devel/2007-07/msg00336.html

I guess that whenever a file referred to in an autoload changes, no
matter how trivially (eg a comment), the file that contains the
generated autoloads will need regenerating due to a changed md5sum.