"make autoloads" fails

"make autoloads" fails

[Eli Zaretskii]

Here's the error message:

  $ make autoloads
  EMACSLOADPATH=/d/gnu/bzr/emacs/trunk/lisp LC_ALL=C /d/gnu/bzr/emacs/trunk/src/emacs -batch --no-site-file --no-site-lisp -l autoload \
     --eval "(setq generate-autoload-cookie \";;;###cal-autoload\")" \
     --eval "(setq generated-autoload-file (unmsys--file-name \"/d/gnu/bzr/emacs/trunk/lisp/calendar/cal-loaddefs.el\"))" \
     --eval "(setq make-backup-files nil)" \
     -f batch-update-autoloads /d/gnu/bzr/emacs/trunk/lisp/calendar
  Wrong type argument: number-or-marker-p, nil

It looks like it's related to the latest changes in coding.c, because
the backtrace is this (see frame #7):

  #0  wrong_type_argument (predicate=53917842, value=53864474) at data.c:188
  #1  0x011603eb in arith_driver (code=Asub, nargs=2, args=0x82e2a4)
      at data.c:2597
  #2  0x011608ae in Fminus (nargs=2, args=0x82e2a4) at data.c:2745
  #3  0x011c00f1 in exec_byte_code (bytestr=19877665, vector=19877685,
      maxdepth=36, args_template=53864474, nargs=0, args=0x0) at bytecode.c:1406
  #4  0x011790d0 in funcall_lambda (fun=19877605, nargs=5, arg_vector=0x82e494)
      at eval.c:3024
  #5  0x01178766 in Ffuncall (nargs=6, args=0x82e490) at eval.c:2839
  #6  0x01178128 in call5 (fn=56400666, arg1=57411777, arg2=53864474,
      arg3=55667906, arg4=53864474, arg5=57411793) at eval.c:2657
  #7  0x0112813f in Fchoose_write_coding_system (start=57411777, end=53864474,
      filename=57411793, append=53864474, visit=53864474, lockname=57411793)
      at fileio.c:4697
  #8  0x01128548 in Fwrite_region (start=57411777, end=53864474,
      filename=57411793, append=53864474, visit=53864474, lockname=57411793,
      mustbenew=53864474) at fileio.c:4870
  #9  0x011786c6 in Ffuncall (nargs=4, args=0x82ed98) at eval.c:2816
  #10 0x011be963 in exec_byte_code (bytestr=57412897, vector=53824293,
      maxdepth=20, args_template=1028, nargs=1, args=0x82ef70) at bytecode.c:903
  #11 0x01178d25 in funcall_lambda (fun=53824317, nargs=1, arg_vector=0x82ef6c)
      at eval.c:2958
  #12 0x01178766 in Ffuncall (nargs=2, args=0x82ef68) at eval.c:2839
  #13 0x011be963 in exec_byte_code (bytestr=57410433, vector=58392469,
      maxdepth=12, args_template=0, nargs=0, args=0x82f140) at bytecode.c:903
  #14 0x01178d25 in funcall_lambda (fun=56405885, nargs=0, arg_vector=0x82f140)
      at eval.c:2958
  #15 0x01178766 in Ffuncall (nargs=1, args=0x82f13c) at eval.c:2839
  #16 0x011be963 in exec_byte_code (bytestr=54251329, vector=56710493,
      maxdepth=60, args_template=512, nargs=1, args=0x82f3e0) at bytecode.c:903
  #17 0x01178d25 in funcall_lambda (fun=55975853, nargs=1, arg_vector=0x82f3e0)
      at eval.c:2958
  #18 0x01178766 in Ffuncall (nargs=2, args=0x82f3dc) at eval.c:2839
  #19 0x0117764b in Fapply (nargs=2, args=0x82f3dc) at eval.c:2276
  #20 0x0117840d in Ffuncall (nargs=3, args=0x82f3d8) at eval.c:2771
  #21 0x011be963 in exec_byte_code (bytestr=54248241, vector=56913669,
      maxdepth=40, args_template=0, nargs=0, args=0x82f5f0) at bytecode.c:903
  #22 0x01178d25 in funcall_lambda (fun=56913805, nargs=0, arg_vector=0x82f5f0)
      at eval.c:2958
  #23 0x01178766 in Ffuncall (nargs=1, args=0x82f5ec) at eval.c:2839
  #24 0x011be963 in exec_byte_code (bytestr=19665353, vector=19665373,
      maxdepth=88, args_template=1028, nargs=1, args=0x82f7cc) at bytecode.c:903
  #25 0x01178d25 in funcall_lambda (fun=19665333, nargs=1, arg_vector=0x82f7c8)
      at eval.c:2958
  #26 0x01178766 in Ffuncall (nargs=2, args=0x82f7c4) at eval.c:2839
  #27 0x011be963 in exec_byte_code (bytestr=19652537, vector=19652557,
      maxdepth=68, args_template=0, nargs=0, args=0x82f9cc) at bytecode.c:903
  #28 0x01178d25 in funcall_lambda (fun=19652517, nargs=0, arg_vector=0x82f9cc)
      at eval.c:2958
  #29 0x01178766 in Ffuncall (nargs=1, args=0x82f9c8) at eval.c:2839
  #30 0x011be963 in exec_byte_code (bytestr=19651001, vector=19651021,
      maxdepth=32, args_template=0, nargs=0, args=0x82fb50) at bytecode.c:903
  #31 0x01178d25 in funcall_lambda (fun=19650981, nargs=0, arg_vector=0x82fb50)
      at eval.c:2958
  #32 0x01178a46 in apply_lambda (fun=19650981, args=53864474) at eval.c:2899
  #33 0x01177357 in eval_sub (form=55571486) at eval.c:2205
  #34 0x011769ad in Feval (form=55571486, lexical=53864474) at eval.c:1993
  #35 0x010e2b93 in top_level_2 () at keyboard.c:1173
  #36 0x011751f9 in internal_condition_case (bfun=0x10e2b77 <top_level_2>,
      handlers=53918890, hfun=0x10e2723 <cmd_error>) at eval.c:1289
  #37 0x010e2bc5 in top_level_1 (ignore=53864474) at keyboard.c:1181
  #38 0x01174c80 in internal_catch (tag=53908770, func=0x10e2b95 <top_level_1>,
      arg=53864474) at eval.c:1063
  #39 0x010e2b00 in command_loop () at keyboard.c:1142
  #40 0x010e22b3 in recursive_edit_1 () at keyboard.c:776
  #41 0x010e2463 in Frecursive_edit () at keyboard.c:840
  #42 0x010e06fe in main (argc=15, argv=0xa44288) at emacs.c:1550

  Lisp Backtrace:
  "select-safe-coding-system" (0x82e494)
  "write-region" (0x82ed9c)
  "autoload-ensure-default-file" (0x82ef6c)
  "autoload-find-generated-file" (0x82f140)
  "update-directory-autoloads" (0x82f3e0)
  "apply" (0x82f3dc)
  "batch-update-autoloads" (0x82f5f0)
  "command-line-1" (0x82f7c8)
  "command-line" (0x82f9cc)
  "normal-top-level" (0x82fb50)

TIA

Re: "make autoloads" fails

[Paul Eggert]

"make autoloads" works for me (Fedora 17, GCC 4.8.1, x86-64)
for the latest Emacs trunk (bzr 113225).  So I suspect the
problem is platform-dependent.

Re: "make autoloads" fails

[Juanma Barranquero]

On Sun, Jun 30, 2013 at 3:02 AM, Paul Eggert <eggert@cs.ucla.edu> wrote:

> "make autoloads" works for me (Fedora 17, GCC 4.8.1, x86-64)
> for the latest Emacs trunk (bzr 113225).  So I suspect the
> problem is platform-dependent.

It works for me on Windows 7 64-bit:

  $ make autoloads
  cd /c/Devel/emacs/repo/trunk/lisp && chmod +w ps-print.el
emulation/tpu-edt.el  mail/rmail.el  dired.el  ibuffer.el
htmlfontify.el  emacs-lisp/eieio.el
  cd /c/Devel/emacs/repo/trunk/lisp; subdirs=`find . -type d -print`;
for file in $subdirs; do  case $file in */.* | */.*/* | */=* |
*/obsolete | */term ) ;;  *) wins="$wins $file" ;;  esac;  done; \
  echo Directories: $wins; \
  EMACSLOADPATH=/c/Devel/emacs/repo/trunk/lisp LC_ALL=C
/c/Devel/emacs/repo/trunk/src/emacs -batch --no-site-file
--no-site-lisp -l autoload \
      --eval '(setq autoload-builtin-package-versions t)' \
      --eval '(setq generated-autoload-file (unmsys--file-name
"/c/Devel/emacs/repo/trunk/lisp/loaddefs.el"))' \
      -f batch-update-autoloads $wins
  Directories: . ./calc ./calendar ./cedet ./cedet/ede
./cedet/semantic ./cedet/semantic/analyze ./cedet/semantic/bovine
./cedet/semantic/decorate ./cedet/semantic/symref
./cedet/semantic/wisent ./cedet/srecode ./emacs-lisp ./emulation ./erc
./eshell ./gnus ./international ./language ./mail ./mh-e ./net ./nxml
./org ./play ./progmodes ./textmodes ./url ./vc
  Saving file c:/Devel/emacs/repo/trunk/lisp/loaddefs.el...
  Wrote c:/Devel/emacs/repo/trunk/lisp/loaddefs.el

Re: "make autoloads" fails

[Eli Zaretskii]

> From: Juanma Barranquero <lekktu@gmail.com>
> Date: Sun, 30 Jun 2013 03:35:21 +0200
> Cc: Kenichi Handa <handa@gnu.org>, Eli Zaretskii <eliz@gnu.org>,
> 	Emacs developers <emacs-devel@gnu.org>
> 
> On Sun, Jun 30, 2013 at 3:02 AM, Paul Eggert <eggert@cs.ucla.edu> wrote:
> 
> > "make autoloads" works for me (Fedora 17, GCC 4.8.1, x86-64)
> > for the latest Emacs trunk (bzr 113225).  So I suspect the
> > problem is platform-dependent.
> 
> It works for me on Windows 7 64-bit:

Including if you delete, say, cl-loaddefs.el?

Re: "make autoloads" fails

[Juanma Barranquero]

On Sun, Jun 30, 2013 at 4:43 AM, Eli Zaretskii <eliz@gnu.org> wrote:

> Including if you delete, say, cl-loaddefs.el?

Yes.

Juanma@ODIEONE /c/Devel/emacs/repo/trunk/lisp
$ rm emacs-lisp/cl-loaddefs*

Juanma@ODIEONE /c/Devel/emacs/repo/trunk/lisp
$ ls -l emacs-lisp/cl-load*
ls: emacs-lisp/cl-load*: No such file or directory

Juanma@ODIEONE /c/Devel/emacs/repo/trunk/lisp
$ touch emacs-lisp/cl-macs.el

Juanma@ODIEONE /c/Devel/emacs/repo/trunk/lisp
$ make autoloads
cd /c/Devel/emacs/repo/trunk/lisp && chmod +w ps-print.el
emulation/tpu-edt.el  mail/rmail.el  dired.el  ibuffer.el
htmlfontify.el  emacs-lisp/eieio.el
cd /c/Devel/emacs/repo/trunk/lisp; subdirs=`find . -type d -print`;
for file in $subdirs; do  case $file in */.* | */.*/* | */=* |
*/obsolete | */term ) ;;  *) wins="$wins $file" ;;  esac;  done; \
echo Directories: $wins; \
EMACSLOADPATH=/c/Devel/emacs/repo/trunk/lisp LC_ALL=C
/c/Devel/emacs/repo/trunk/src/emacs -batch --no-site-file
--no-site-lisp -l autoload \
    --eval '(setq autoload-builtin-package-versions t)' \
    --eval '(setq generated-autoload-file (unmsys--file-name
"/c/Devel/emacs/repo/trunk/lisp/loaddefs.el"))' \
    -f batch-update-autoloads $wins
Directories: . ./calc ./calendar ./cedet ./cedet/ede ./cedet/semantic
./cedet/semantic/analyze ./cedet/semantic/bovine
./cedet/semantic/decorate ./cedet/semantic/symref
./cedet/semantic/wisent ./cedet/srecode ./emacs-lisp ./emulation ./erc
./eshell ./gnus ./international ./language ./mail ./mh-e ./net ./nxml
./org ./play ./progmodes ./textmodes ./url ./vc
Making generated-autoload-file local to  *autoload-file* while let-bound!
Generating autoloads for emacs-lisp/cl-macs.el...
Wrote c:/Devel/emacs/repo/trunk/lisp/emacs-lisp/cl-loaddefs.el
Generating autoloads for emacs-lisp/cl-macs.el...done
Saving file c:/Devel/emacs/repo/trunk/lisp/loaddefs.el...
Wrote c:/Devel/emacs/repo/trunk/lisp/loaddefs.el
Saving file c:/Devel/emacs/repo/trunk/lisp/emacs-lisp/cl-loaddefs.el...
Wrote c:/Devel/emacs/repo/trunk/lisp/emacs-lisp/cl-loaddefs.el

Juanma@ODIEONE /c/Devel/emacs/repo/trunk/lisp
$

Re: "make autoloads" fails

[Eli Zaretskii]

> Date: Sat, 29 Jun 2013 18:02:56 -0700
> From: Paul Eggert <eggert@cs.ucla.edu>
> Cc: Kenichi Handa <handa@gnu.org>, emacs-devel@gnu.org
> 
> "make autoloads" works for me (Fedora 17, GCC 4.8.1, x86-64)
> for the latest Emacs trunk (bzr 113225).  So I suspect the
> problem is platform-dependent.

Maybe so, but then on what platform(s)?  See

  http://lists.gnu.org/archive/html/emacs-devel/2013-06/msg01329.html

Re: "make autoloads" fails

[Waiyian Chong]

[-- Attachment #1: Type: text/plain, Size: 1205 bytes --]

Eli Zaretskii <eliz@gnu.org> writes:

> Here's the error message:
>
>   $ make autoloads
>   EMACSLOADPATH=/d/gnu/bzr/emacs/trunk/lisp LC_ALL=C /d/gnu/bzr/emacs/trunk/src/emacs -batch --no-site-file --no-site-lisp -l autoload \
>      --eval "(setq generate-autoload-cookie \";;;###cal-autoload\")" \
>      --eval "(setq generated-autoload-file (unmsys--file-name \"/d/gnu/bzr/emacs/trunk/lisp/calendar/cal-loaddefs.el\"))" \
>      --eval "(setq make-backup-files nil)" \
>      -f batch-update-autoloads /d/gnu/bzr/emacs/trunk/lisp/calendar
>   Wrong type argument: number-or-marker-p, nil
...
>   Lisp Backtrace:
>   "select-safe-coding-system" (0x82e494)
>   "write-region" (0x82ed9c)
>   "autoload-ensure-default-file" (0x82ef6c)
>   "autoload-find-generated-file" (0x82f140)
>   "update-directory-autoloads" (0x82f3e0)
>   "apply" (0x82f3dc)
>   "batch-update-autoloads" (0x82f5f0)
>   "command-line-1" (0x82f7c8)
>   "command-line" (0x82f9cc)
>   "normal-top-level" (0x82fb50)

I encountered the same problem, and traced it back to the function
`autoload-ensure-default-file', which seems to call `write-region'
with wrong arguments.  Build succeeded after applying attached patch.


-- 
waiyian


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: patch --]
[-- Type: text/x-patch, Size: 560 bytes --]

diff --git a/lisp/emacs-lisp/autoload.el b/lisp/emacs-lisp/autoload.el
index 22713c6..d1722d0 100644
--- a/lisp/emacs-lisp/autoload.el
+++ b/lisp/emacs-lisp/autoload.el
@@ -354,7 +354,7 @@ not be relied upon."
 (defun autoload-ensure-default-file (file)
   "Make sure that the autoload file FILE exists and if not create it."
   (unless (file-exists-p file)
-    (write-region (autoload-rubric file) nil file))
+    (with-temp-file file (insert (autoload-rubric file))))
   file)
 
 (defun autoload-insert-section-header (outbuf autoloads load-name file time)

Re: "make autoloads" fails

[Eli Zaretskii]

> From: Waiyian Chong <waiyian.chong@gmail.com>
> Date: Sun, 30 Jun 2013 23:30:00 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Here's the error message:
> >
> >   $ make autoloads
> >   EMACSLOADPATH=/d/gnu/bzr/emacs/trunk/lisp LC_ALL=C /d/gnu/bzr/emacs/trunk/src/emacs -batch --no-site-file --no-site-lisp -l autoload \
> >      --eval "(setq generate-autoload-cookie \";;;###cal-autoload\")" \
> >      --eval "(setq generated-autoload-file (unmsys--file-name \"/d/gnu/bzr/emacs/trunk/lisp/calendar/cal-loaddefs.el\"))" \
> >      --eval "(setq make-backup-files nil)" \
> >      -f batch-update-autoloads /d/gnu/bzr/emacs/trunk/lisp/calendar
> >   Wrong type argument: number-or-marker-p, nil
> ...
> >   Lisp Backtrace:
> >   "select-safe-coding-system" (0x82e494)
> >   "write-region" (0x82ed9c)
> >   "autoload-ensure-default-file" (0x82ef6c)
> >   "autoload-find-generated-file" (0x82f140)
> >   "update-directory-autoloads" (0x82f3e0)
> >   "apply" (0x82f3dc)
> >   "batch-update-autoloads" (0x82f5f0)
> >   "command-line-1" (0x82f7c8)
> >   "command-line" (0x82f9cc)
> >   "normal-top-level" (0x82fb50)
> 
> I encountered the same problem, and traced it back to the function
> `autoload-ensure-default-file', which seems to call `write-region'
> with wrong arguments.  Build succeeded after applying attached patch.

Thanks.  However, there's no problem with that call to write-region:
it is perfectly valid to pass to it a string.  The root cause was in
select-safe-coding-system, which didn't handle correctly the case of a
string instead of a region of buffer positions.

Should be fixed now on the trunk.

Re: "make autoloads" fails

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 334 bytes --]


   > Should be fixed now on the trunk.



Sorry for this elementary question.
I obtained yesterday the trunk (with the bug)

bzr branch bzr://bzr.savannah.gnu.org/emacs/trunk 

In git to obtain the changes which were submitted to solve the bug would
be 

git checkout master


What is the corresponding bzr command?

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5556 bytes --]

Re: "make autoloads" fails

[Jambunathan K]

Uwe Brauer <oub@mat.ucm.es> writes:

>    > Should be fixed now on the trunk.
>
>
>
> Sorry for this elementary question.
> I obtained yesterday the trunk (with the bug)
>
> bzr branch bzr://bzr.savannah.gnu.org/emacs/trunk 
>
> In git to obtain the changes which were submitted to solve the bug would
> be 
>
> git checkout master
>
>
> What is the corresponding bzr command?

bzr update

Re: "make autoloads" fails

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 586 bytes --]

>> "Jambunathan" == Jambunathan K <Jambunathan> writes:

   > Uwe Brauer <oub@mat.ucm.es> writes:
   >> 
   >> 
   >> What is the corresponding bzr command?

   > bzr update

In the trunk directory I suppose?

bzr branch bzr://bzr.savannah.gnu.org/emacs/trunk 
created a directory called trunk with all the relevant files, now when I
do 

bzr update 


in the trunk directory I obtain 

Tree is up to date at revision 113223 of branch
/home/oub/ALLES/Add-Import/Soft/trunk

which seems to me the trunk of yesterday...
Also no report of any new downloaded files

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5556 bytes --]

Re: "make autoloads" fails

[Jambunathan K]

Uwe Brauer <oub@mat.ucm.es> writes:

>>> "Jambunathan" == Jambunathan K <Jambunathan> writes:
>
>    > Uwe Brauer <oub@mat.ucm.es> writes:
>    >> 
>    >> 
>    >> What is the corresponding bzr command?
>
>    > bzr update
>
> In the trunk directory I suppose?
>
> bzr branch bzr://bzr.savannah.gnu.org/emacs/trunk 
> created a directory called trunk with all the relevant files, now when I
> do 
>
> bzr update 
>
>
> in the trunk directory I obtain 
>
> Tree is up to date at revision 113223 of branch
> /home/oub/ALLES/Add-Import/Soft/trunk
>
> which seems to me the trunk of yesterday...
> Also no report of any new downloaded files

I followed the instructions at 
        
        http://www.emacswiki.org/emacs/BzrForEmacsDevs

May be you need to do a bzr bind?

Re: "make autoloads" fails

[Eli Zaretskii]

> From: Jambunathan K <kjambunathan@gmail.com>
> Date: Sun, 30 Jun 2013 23:35:50 +0530
> Cc: emacs-devel@gnu.org
> 
> > bzr update 
> >
> >
> > in the trunk directory I obtain 
> >
> > Tree is up to date at revision 113223 of branch
> > /home/oub/ALLES/Add-Import/Soft/trunk
> >
> > which seems to me the trunk of yesterday...
> > Also no report of any new downloaded files
> 
> I followed the instructions at 
>         
>         http://www.emacswiki.org/emacs/BzrForEmacsDevs
> 
> May be you need to do a bzr bind?

Yes, either "bzr bind", and then "bzr update" will do the trick, or
leave the branch un-bound and use "bzr pull" to re-sync with upstream.

Re: "make autoloads" fails

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 424 bytes --]

>> "Eli" == Eli Zaretskii <eliz@gnu.org> writes:
   >> 
   >> http://www.emacswiki.org/emacs/BzrForEmacsDevs
   >> 
   >> May be you need to do a bzr bind?

   > Yes, either "bzr bind", and then "bzr update" will do the trick, or

For some reason this does not work.
   > leave the branch un-bound and use "bzr pull" to re-sync with upstream.

Yes, this works, sync starts, lets see whether makes now succeeds. 

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5556 bytes --]

Re: "make autoloads" fails

[Eli Zaretskii]

> From: Uwe Brauer <oub@mat.ucm.es>
> Cc: Jambunathan K <kjambunathan@gmail.com>, emacs-devel@gnu.org
> Date: Sun, 30 Jun 2013 20:17:55 +0200
> 
>    > Yes, either "bzr bind", and then "bzr update" will do the trick, or
> 
> For some reason this does not work.

Probably my fault: the full command is

  bzr bind bzr://bzr.savannah.gnu.org/emacs/trunk

followed by "bzr update".

Re: "make autoloads" fails

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 1166 bytes --]

>> "Eli" == Eli Zaretskii <eliz@gnu.org> writes:

   >> From: Jambunathan K <kjambunathan@gmail.com>
   >> Date: Sun, 30 Jun 2013 23:35:50 +0530
   >> Cc: emacs-devel@gnu.org

Another annoyance. My actual emacs24 is in prefix=/opt/emacs24

However even choosing 

./configure --prefix=/opt/emacs24

Make 

I obtain the following error

Warning: arch-independent data dir
`/usr/local/share/emacs/24.3.50/etc/': No such file or directory 
Error: charsets directory not found:
/usr/local/share/emacs/24.3.50/etc/charsets
Emacs will not function correctly without the character map files.
Please check your installation!
make[2]: ***
[/home/oub/ALLES/Add-Import/Soft/emacs/trunk/lisp/calendar/cal-loaddefs.el]
Error 1 
make[2]: Leaving directory `/home/oub/ALLES/Add-Import/Soft/emacs/trunk/lisp'
make[1]: *** [/home/oub/ALLES/Add-Import/Soft/emacs/trunk/src/../lisp/loaddefs.el] Error 2
make[1]: Leaving directory `/home/oub/ALLES/Add-Import/Soft/emacs/trunk/src'
make: *** [src] Error 2


Why the hack is he looking in /usr/local/share/emacs/24.3.50

And by the way does he suppose there is already a working GNU emacs
installation?

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5556 bytes --]

Re: "make autoloads" fails

[Eli Zaretskii]

> From: Uwe Brauer <oub@mat.ucm.es>
> Cc: Jambunathan K <kjambunathan@gmail.com>, emacs-devel@gnu.org
> Date: Sun, 30 Jun 2013 20:29:34 +0200
> 
> Another annoyance. My actual emacs24 is in prefix=/opt/emacs24
> 
> However even choosing 
> 
> ./configure --prefix=/opt/emacs24
> 
> Make 
> 
> I obtain the following error
> 
> Warning: arch-independent data dir
> `/usr/local/share/emacs/24.3.50/etc/': No such file or directory 
> Error: charsets directory not found:

Let's slow down.  The commands you should invoke are, in order:

  ./autogen.sh
  ./configure --prefix=...
  make bootstrap

("make bootstrap" because you already tried to build and failed, so
something is not clean in that directory.)

Also, what do you mean by "my actual emacs24 is in
prefix=/opt/emacs24" -- is that where you want to install Emacs after
building, or is that where the sources are?

> Why the hack is he looking in /usr/local/share/emacs/24.3.50

Probably because you built it incorrectly.  Did you also invoke "make
install"?

> And by the way does he suppose there is already a working GNU emacs
> installation?

No.

Re: "make autoloads" fails

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 1407 bytes --]

>> "Eli" == Eli Zaretskii <eliz@gnu.org> writes:

   > Let's slow down.  The commands you should invoke are, in order:

   >   ./autogen.sh
   >   ./configure --prefix=...
   >   make bootstrap

   > ("make bootstrap" because you already tried to build and failed, so
   > something is not clean in that directory.)

Ok, that was the issue! Now it works.

   > Also, what do you mean by "my actual emacs24 is in
   > prefix=/opt/emacs24" -- is that where you want to install Emacs after
   > building, or is that where the sources are?

   >> Why the hack is he looking in /usr/local/share/emacs/24.3.50

   > Probably because you built it incorrectly.  
Ok, the message confused me I thought he was looking for an existing GNU
emacs 24 installation. On Kubuntu I have the official GNU emacs 23 and
Xemacs 21.4, + a compiled Xemacs 21.5.X which is in /usr/local

And for some bizarre reason I cannot have GNU emacs 24 there (etag
conflict, which most likely could be solved, but the simpler solution
was to have GNU emacs 24 in /opt/emacs24)

   > Did you also invoke "make
   > install"?

No.

So I already tried it out and the visual cursor works nicely. I will do
further tests in the coming days and then I report back. I am not sure
what to think about the text selection on a line which has r2l and l2r
text.


Thanks a zillion for this nice work!!!!

Uwe 


[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5556 bytes --]

"make autoloads" fails

[Jan Djärv]

Hello.

A clean checkout on OSX fails on make bootstrap:

cd ../lisp; make  autoloads EMACS=/Users/jhd/src/emacs/obj-constr/src/bootstrap-emacs
EMACSLOADPATH=/Users/jhd/src/emacs/constr/lisp LC_ALL=C /Users/jhd/src/emacs/obj-constr/src/bootstrap-emacs -batch --no-site-file --no-site-lisp -l autoload \
	   --eval "(setq generate-autoload-cookie \";;;###cal-autoload\")" \
	   --eval "(setq generated-autoload-file (unmsys--file-name \"/Users/jhd/src/emacs/constr/lisp/calendar/cal-loaddefs.el\"))" \
	   --eval "(setq make-backup-files nil)" \
	   -f batch-update-autoloads /Users/jhd/src/emacs/constr/lisp/calendar
Loading macroexp.elc...
Wrong type argument: number-or-marker-p, nil
make[3]: *** [/Users/jhd/src/emacs/constr/lisp/calendar/cal-loaddefs.el] Error 255
make[2]: *** [/Users/jhd/src/emacs/constr/src/../lisp/loaddefs.el] Error 2
make[1]: *** [src] Error 2
make: *** [bootstrap] Error 2

Is it possible that existing trees work fine, but new one fails?

	Jan D.

Re: "make autoloads" fails

[Eli Zaretskii]

> From: Jan Djärv <jan.h.d@swipnet.se>
> Date: Sun, 30 Jun 2013 10:21:09 +0200
> 
> A clean checkout on OSX fails on make bootstrap:
> 
> cd ../lisp; make  autoloads EMACS=/Users/jhd/src/emacs/obj-constr/src/bootstrap-emacs
> EMACSLOADPATH=/Users/jhd/src/emacs/constr/lisp LC_ALL=C /Users/jhd/src/emacs/obj-constr/src/bootstrap-emacs -batch --no-site-file --no-site-lisp -l autoload \
> 	   --eval "(setq generate-autoload-cookie \";;;###cal-autoload\")" \
> 	   --eval "(setq generated-autoload-file (unmsys--file-name \"/Users/jhd/src/emacs/constr/lisp/calendar/cal-loaddefs.el\"))" \
> 	   --eval "(setq make-backup-files nil)" \
> 	   -f batch-update-autoloads /Users/jhd/src/emacs/constr/lisp/calendar
> Loading macroexp.elc...
> Wrong type argument: number-or-marker-p, nil
> make[3]: *** [/Users/jhd/src/emacs/constr/lisp/calendar/cal-loaddefs.el] Error 255
> make[2]: *** [/Users/jhd/src/emacs/constr/src/../lisp/loaddefs.el] Error 2
> make[1]: *** [src] Error 2
> make: *** [bootstrap] Error 2
> 
> Is it possible that existing trees work fine, but new one fails?

Yes.  The bug was with passing a string (as opposed to region of
buffer positions) to select-safe-coding-system, which only happens
when the autoload files don't exist.

Re: "make autoloads" fails

[Angelo Graziosi]

Paul Eggert wrote:
> "make autoloads" works for me (Fedora 17, GCC 4.8.1, x86-64)
> for the latest Emacs trunk (bzr 113225).  So I suspect the
> problem is platform-dependent.

Jan Djärv wrote
> A clean checkout on OSX fails on make bootstrap:
[...]
> Is it possible that existing trees work fine, but new one fails?

On Cygwin r. 113227 fails to bootstrap:
[...]
Wrote /work/emacs/lisp/emacs-lisp/autoload.elc
make[3]: uscita dalla directory "/work/emacs/Work/lisp"
cd ../lisp; /usr/bin/make -w autoloads 
EMACS=/work/emacs/Work/src/bootstrap-emacs.exe
make[3]: ingresso nella directory "/work/emacs/Work/lisp"
EMACSLOADPATH=/work/emacs/lisp LC_ALL=C 
/work/emacs/Work/src/bootstrap-emacs.exe -batch --no-site-file 
--no-site-lisp -l autoload \
    --eval "(setq generate-autoload-cookie \";;;###cal-autoload\")" \
    --eval "(setq generated-autoload-file (unmsys--file-name 
\"/work/emacs/lisp/calendar/cal-loaddefs.el\"))" \
    --eval "(setq make-backup-files nil)" \
    -f batch-update-autoloads /work/emacs/lisp/calendar
Loading macroexp.elc...
Wrong type argument: number-or-marker-p, nil
Makefile:418: recipe for target 
`/work/emacs/lisp/calendar/cal-loaddefs.el' failed
make[3]: *** [/work/emacs/lisp/calendar/cal-loaddefs.el] Error 255
make[3]: uscita dalla directory "/work/emacs/Work/lisp"
Makefile:818: recipe for target `/work/emacs/src/../lisp/loaddefs.el' failed
make[2]: *** [/work/emacs/src/../lisp/loaddefs.el] Error 2
make[2]: uscita dalla directory "/work/emacs/Work/src"
Makefile:381: recipe for target `src' failed
make[1]: *** [src] Error 2
make[1]: uscita dalla directory "/work/emacs/Work"
Makefile:992: recipe for target `bootstrap' failed
make: *** [bootstrap] Error 2
build-emacs.sh: Bootstrap failure...

Notice that r113214 (two days ago) bootstraps fine.

I would suggest to bootstrap in a clean tree.

Usally I have a local copy of trunk BZR repo and to bootstrap I do

rsync -av --exclude=.bzr* --exclude=.g* --delete /work/emacs-trunk/ 
/work/emacs/

cd /work/emacs/
./autogen.sh
mkdir Work
cd Work
../configure ...
make [-j3] bootstrap
etc..

Ciao,
Angelo.