Emacs fails to bootstrap, gnus / mh-e related

Emacs fails to bootstrap, gnus / mh-e related

[Miles Bader]

After doing `make bootstrap':

   ... lots of lines omitted ...
   Compiling /home/m/miles/src/emacs/devo/lisp/./gnus/gnus-kill.el
   Wrote /home/m/miles/src/emacs/devo/lisp/gnus/gnus-kill.elc
   Compiling /home/m/miles/src/emacs/devo/lisp/./gnus/gnus-logic.el
   Wrote /home/m/miles/src/emacs/devo/lisp/gnus/gnus-logic.elc
   Compiling /home/m/miles/src/emacs/devo/lisp/./gnus/gnus-mh.el

   In toplevel form:
   ../../../../src/emacs/devo/lisp/gnus/gnus-mh.el:38:1:Error: Symbol's function definition is void: mh-require-cl
   make[2]: *** [compile] Error 1
   make[2]: Leaving directory `/srv/data/home/m/miles/build/emacs/devo/lisp'
   make[1]: *** [bootstrap-build] Error 2
   make[1]: Leaving directory `/srv/data/home/m/miles/build/emacs/devo'
   make: *** [bootstrap] Error 2

[Using a fresh update from CVS]

Neither gnus-mh.el nor mh-e seems to have been touched in quite a while;
can other people bootstrap successfully?

-miles
-- 
80% of success is just showing up.  --Woody Allen

Re: Emacs fails to bootstrap, gnus / mh-e related

[Miles Bader]

... and it also fails when compiling mh-e itself, so I guess it's
probably a mh-e error, nothing to do with gnus:

   ... lots omitted ...
   Compiling /home/m/miles/src/emacs/devo/lisp/./mh-e/mh-alias.el

   In toplevel form:
   ../../../../src/emacs/devo/lisp/mh-e/mh-alias.el:33:1:Error: Symbol's function definition is void: mh-require-cl
   make[2]: *** [compile] Error 1
   make[2]: Leaving directory `/srv/data/home/m/miles/build/emacs/devo/lisp'
   make[1]: *** [bootstrap-build] Error 2
   make[1]: Leaving directory `/srv/data/home/m/miles/build/emacs/devo'
   make: *** [bootfast] Error 2

-miles
-- 
((lambda (x) (list x x)) (lambda (x) (list x x)))

Re: Emacs fails to bootstrap, gnus / mh-e related

[Andreas Schwab]

Miles Bader <miles@gnu.org> writes:

> After doing `make bootstrap':
>
>    ... lots of lines omitted ...
>    Compiling /home/m/miles/src/emacs/devo/lisp/./gnus/gnus-kill.el
>    Wrote /home/m/miles/src/emacs/devo/lisp/gnus/gnus-kill.elc
>    Compiling /home/m/miles/src/emacs/devo/lisp/./gnus/gnus-logic.el
>    Wrote /home/m/miles/src/emacs/devo/lisp/gnus/gnus-logic.elc
>    Compiling /home/m/miles/src/emacs/devo/lisp/./gnus/gnus-mh.el
>
>    In toplevel form:
>    ../../../../src/emacs/devo/lisp/gnus/gnus-mh.el:38:1:Error: Symbol's function definition is void: mh-require-cl

Do you have mh-e/mh-loaddefs.el?

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Emacs fails to bootstrap, gnus / mh-e related

[Miles Bader]

Andreas Schwab <schwab@suse.de> writes:
>>    In toplevel form:
>>    ../../../../src/emacs/devo/lisp/gnus/gnus-mh.el:38:1:Error: Symbol's function definition is void: mh-require-cl
>
> Do you have mh-e/mh-loaddefs.el?

I have this:

   ;;; mh-loaddefs.el --- automatically extracted autoloads

   ;; Copyright (C) 2003, 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
   ;; Author: Bill Wohler <wohler@newt.com>
   ;; Keywords: mail
   ;;; Commentary:
   ;;; Change Log:
   ;;; Code:

   (provide 'mh-loaddefs)
   ;; Local Variables:
   ;; version-control: never
   ;; no-byte-compile: t
   ;; no-update-autoloads: t
   ;; End:
   ;;; mh-loaddefs.el ends here


I gather it's supposed to contain something more?  What is supposed
to update it?

One thing that might be relevant:  My original command used make's -j4
option (for parallel makes):  "make -j4 bootstrap".  When that failed, I
redid it using a simple "make bootstrap", but perhaps if this file was
somehow screwed up by the make -j4, and isn't being properly regenerated by
"make bootstrap", that could be why it failed the 2nd time?

Thanks,

-Miles

-- 
Somebody has to do something, and it's just incredibly pathetic that it
has to be us.  -- Jerry Garcia

Re: Emacs fails to bootstrap, gnus / mh-e related

[Andreas Schwab]

Miles Bader <miles@gnu.org> writes:

> I gather it's supposed to contain something more?

Yes, all mh-e autoloads.

> What is supposed to update it?

make mh-autoloads in lisp should update it.  But this will only work if you
already have a working emacs, of course.

> One thing that might be relevant:  My original command used make's -j4
> option (for parallel makes):  "make -j4 bootstrap".  When that failed, I
> redid it using a simple "make bootstrap", but perhaps if this file was
> somehow screwed up by the make -j4, and isn't being properly regenerated by
> "make bootstrap", that could be why it failed the 2nd time?

Quite possible.  Looks like the bootstrap emacs lacks all mh-e autoloads.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Emacs fails to bootstrap, gnus / mh-e related

[Bill Wohler]

Miles,

Sorry for the long hiatus, but work (kepler.nasa.gov) has been keeping
me way busy.

Miles Bader <miles@gnu.org> writes:

> Andreas Schwab <schwab@suse.de> writes:
>>>    In toplevel form:
>>>    ../../../../src/emacs/devo/lisp/gnus/gnus-mh.el:38:1:Error: Symbol's function definition is void: mh-require-cl
>>
>> Do you have mh-e/mh-loaddefs.el?
>
> I have this:
>
>    ;;; mh-loaddefs.el --- automatically extracted autoloads
>
>    ;; Copyright (C) 2003, 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
>    ;; Author: Bill Wohler <wohler@newt.com>
>    ;; Keywords: mail
>    ;;; Commentary:
>    ;;; Change Log:
>    ;;; Code:
>
>    (provide 'mh-loaddefs)
>    ;; Local Variables:
>    ;; version-control: never
>    ;; no-byte-compile: t
>    ;; no-update-autoloads: t
>    ;; End:
>    ;;; mh-loaddefs.el ends here
>
>
> I gather it's supposed to contain something more?  What is supposed
> to update it?

Yes, as Andreas said, it should contain all of the autoloads.

Look for mh-autoloads in lisp/Makefile to see how it should be
created. In particular, the missing bits are created with the
following command:

        $(EMACS) $(EMACSOPT) \
           -l autoload \
           --eval "(setq generate-autoload-cookie \";;;###mh-autoload\")" \
           --eval "(setq generated-autoload-file \"$(lisp)/mh-e/mh-loaddefs.el\")" \
           --eval "(setq make-backup-files nil)" \
           -f batch-update-autoloads $(lisp)/mh-e

I pretty much copied the command that builds loaddefs.el.

> One thing that might be relevant:  My original command used make's -j4
> option (for parallel makes):  "make -j4 bootstrap".  When that failed, I
> redid it using a simple "make bootstrap", but perhaps if this file was
> somehow screwed up by the make -j4, and isn't being properly regenerated by
> "make bootstrap", that could be why it failed the 2nd time?

The mh-autoloads depends on $(lisp)/mh-e/mh-loaddefs.el, so if that
file existed, it wouldn't be rebuilt. If you remove it, and run "make
mh-autoloads" in the lisp directory, you should be set.

I just did an update after eons, and a make bootstrap (after a make
maintainer-clean) and the build proceeded without error.

Note that as you observed, we haven't touched MH-E (except for a few
minor bug fixes) since Emacs was frozen, nearly a year ago. I guess the
Makefiles aren't conducive to parallelism.

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