23.0.60; bootstrap broken by removal of cal-loaddefs.el

23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Tim Van Holder]

Ever since the removal of cal-loaddefs from CVS, I had gotten lisp
compilation errors during my more-or-less-daily rebuild. They seemed
harmless enough (especially since I don't actually use the calendar),
but this morning I decided to go for a full bootstrap to (hopefully)
make them go away.
Unfortunately, it looks like the bootstrap does not generate the file in
time for the full lisp compilations, so I received

Compiling /home/tim/gnu/src/emacs/lisp/./gnus/yenc.el
Wrote /home/tim/gnu/src/emacs/lisp/gnus/yenc.elc
Compiling /home/tim/gnu/src/emacs/lisp/./calendar/appt.el

In toplevel form:
../../../../src/emacs/lisp/calendar/appt.el:79:1:Error: Cannot open load file: cal-loaddefs
make[2]: *** [compile] Error 1
make[2]: Leaving directory `/home/tim/gnu/build/linux/emacs/lisp'
make[1]: *** [bootstrap-build] Error 2
make[1]: Leaving directory `/home/tim/gnu/build/linux/emacs'
make: *** [bootstrap] Error 2

It's worked around easily enough (I just fetched the last version from
CVS), but should probably be fixed.


In GNU Emacs 23.0.60.1 (i686-pc-linux-gnu, GTK+ Version 2.8.20)
 of 2008-03-14 on leeloo
Windowing system distributor `RealVNC Ltd', version 11.0.3370
configured using `configure  '--with-x-toolkit=gtk' '--with-x' '--with-freetype' '--with-xft' '--with-libotf''

Re: 23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Tim Van Holder]

On Fri, Mar 14, 2008 at 10:48 AM, Tim Van Holder
<tim.vanholder@gmail.com> wrote:
>
>  Ever since the removal of cal-loaddefs from CVS, I had gotten lisp
>  compilation errors during my more-or-less-daily rebuild. They seemed
>  harmless enough (especially since I don't actually use the calendar),
>  but this morning I decided to go for a full bootstrap to (hopefully)
>  make them go away.
>  Unfortunately, it looks like the bootstrap does not generate the file in
>  time for the full lisp compilations, so I received
>
>  Compiling /home/tim/gnu/src/emacs/lisp/./gnus/yenc.el
>  Wrote /home/tim/gnu/src/emacs/lisp/gnus/yenc.elc
>  Compiling /home/tim/gnu/src/emacs/lisp/./calendar/appt.el
>
>  In toplevel form:
>  ../../../../src/emacs/lisp/calendar/appt.el:79:1:Error: Cannot open load file: cal-loaddefs
>  make[2]: *** [compile] Error 1
>  make[2]: Leaving directory `/home/tim/gnu/build/linux/emacs/lisp'
>  make[1]: *** [bootstrap-build] Error 2
>  make[1]: Leaving directory `/home/tim/gnu/build/linux/emacs'
>  make: *** [bootstrap] Error 2
>
>  It's worked around easily enough (I just fetched the last version from
>  CVS), but should probably be fixed.
>
>
>  In GNU Emacs 23.0.60.1 (i686-pc-linux-gnu, GTK+ Version 2.8.20)
>   of 2008-03-14 on leeloo
>  Windowing system distributor `RealVNC Ltd', version 11.0.3370
>  configured using `configure  '--with-x-toolkit=gtk' '--with-x' '--with-freetype' '--with-xft' '--with-libotf''

Sorry for replying to myself here, but the same goes for the two other calendar
loaddefs (diary-loaddefs and hol-loaddefs). They need to exist for
bootstrap to work,
and the bootstrap process does NOT regenerate them at any point (nor does a
"make recompile updates" in lisp/, which I run as part of my daily
rebuild-from-CVS).
So it looks as though the Makefile.in in lisp/ has not been correctly
instrumented to
create them.

Re: 23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Andrew W. Nosenko]

On Fri, Mar 14, 2008 at 2:12 PM, Tim Van Holder <tim.vanholder@gmail.com> wrote:
>
> On Fri, Mar 14, 2008 at 10:48 AM, Tim Van Holder
>  <tim.vanholder@gmail.com> wrote:
>  >
>  >  Ever since the removal of cal-loaddefs from CVS, I had gotten lisp
>  >  compilation errors during my more-or-less-daily rebuild. They seemed
>  >  harmless enough (especially since I don't actually use the calendar),
>  >  but this morning I decided to go for a full bootstrap to (hopefully)
>  >  make them go away.
>  >  Unfortunately, it looks like the bootstrap does not generate the file in
>  >  time for the full lisp compilations, so I received
>  >
>  >  Compiling /home/tim/gnu/src/emacs/lisp/./gnus/yenc.el
>  >  Wrote /home/tim/gnu/src/emacs/lisp/gnus/yenc.elc
>  >  Compiling /home/tim/gnu/src/emacs/lisp/./calendar/appt.el
>  >
>  >  In toplevel form:
>  >  ../../../../src/emacs/lisp/calendar/appt.el:79:1:Error: Cannot open load file: cal-loaddefs
>  >  make[2]: *** [compile] Error 1
>  >  make[2]: Leaving directory `/home/tim/gnu/build/linux/emacs/lisp'
>  >  make[1]: *** [bootstrap-build] Error 2
>  >  make[1]: Leaving directory `/home/tim/gnu/build/linux/emacs'
>  >  make: *** [bootstrap] Error 2
>  >
>  >  It's worked around easily enough (I just fetched the last version from
>  >  CVS), but should probably be fixed.
>  >
>  >
>  >  In GNU Emacs 23.0.60.1 (i686-pc-linux-gnu, GTK+ Version 2.8.20)
>  >   of 2008-03-14 on leeloo
>  >  Windowing system distributor `RealVNC Ltd', version 11.0.3370
>  >  configured using `configure  '--with-x-toolkit=gtk' '--with-x' '--with-freetype' '--with-xft' '--with-libotf''
>
>  Sorry for replying to myself here, but the same goes for the two other calendar
>  loaddefs (diary-loaddefs and hol-loaddefs). They need to exist for
>  bootstrap to work,
>  and the bootstrap process does NOT regenerate them at any point (nor does a
>  "make recompile updates" in lisp/, which I run as part of my daily
>  rebuild-from-CVS).
>  So it looks as though the Makefile.in in lisp/ has not been correctly
>  instrumented to
>  create them.

You don't need to fetching now deleted *-loaddefs.
Try to regenerate lisp/Makefile (just by calling 'configure' or by
moving old lisp/Makefile awaiy followed by 'configure' call).

At least it worked for me.

-- 
Andrew W. Nosenko <andrew.w.nosenko@gmail.com>

Re: 23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Glenn Morris]

"Andrew W. Nosenko" wrote:

> Try to regenerate lisp/Makefile (just by calling 'configure' or by
> moving old lisp/Makefile awaiy followed by 'configure' call).

Is there still a problem after `make maintainer-clean'?

Re: 23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Andrew W. Nosenko]

On Fri, Mar 14, 2008 at 7:50 PM, Glenn Morris <rgm@gnu.org> wrote:
> "Andrew W. Nosenko" wrote:
>
>  > Try to regenerate lisp/Makefile (just by calling 'configure' or by
>  > moving old lisp/Makefile awaiy followed by 'configure' call).
>
>  Is there still a problem after `make maintainer-clean'?
>

If you asked about my case, then I didn't use 'maintainer-clean' but
emulated it by hands (as you can see).

If you asked about Tim Van Holder case, then I hope, Tim will answer by self.

-- 
Andrew W. Nosenko <andrew.w.nosenko@gmail.com>

Re: 23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Claus]

Hi,

On Fri, Mar 14, 2008 at 6:50 PM, Glenn Morris <rgm@gnu.org> wrote:
> "Andrew W. Nosenko" wrote:
>
>  > Try to regenerate lisp/Makefile (just by calling 'configure' or by
>  > moving old lisp/Makefile awaiy followed by 'configure' call).
>
>  Is there still a problem after `make maintainer-clean'?
>

There ain't no 'maintainer-clean' for me (MingGW Env. on Windows Vista):

cjk@cjk-mobile nt% make <TAB>
make target
addpm                 bootstrap-clean        clean-other-dirs-gmake
 ddeclient   install                   preprep          stamp_BLD
addsection            bootstrap-clean-gmake  clean-other-dirs-nmake
 doit        install-other-dirs-gmake  realclean        which-sh
all                   bootstrap-clean-nmake  cleanall
 force-info  install-other-dirs-nmake  recompile
all-other-dirs-gmake  bootstrap-gmake        cleanall-other-dirs-gmake
 info        maybe-bootstrap           recompile-gmake
all-other-dirs-nmake  bootstrap-nmake        cleanall-other-dirs-nmake
 info-gmake  maybe-bootstrap-CMD       recompile-nmake
bootstrap             clean                  cmdproxy
 info-nmake  maybe-bootstrap-SH        runemacs

I had the same problem as Andrew and Tim. Only manually deleting most
"untracked" CVS/GIT files (makefiles, generated lisp/**/*.el files,
etc.) solved the problem.


Claus

Re: 23.0.60; bootstrap broken by removal of cal-loaddefs.el

[Glenn Morris]

Claus wrote:

> There ain't no 'maintainer-clean' for me (MingGW Env. on Windows Vista):

<insert disparaging remark about Windows here>
I've no idea what the equivalent Windows make command would be.

I did forget to update the Windows makefile, but Eli (thanks) fixed
that recently.