Error while bootstrapping GNU Emacs 22.1.50

[Peter Dyballa-2]

Hello!

Compilation ends with:

In end of data:
dcl-mode.el:2223:1:Warning: the function `imenu-default-create-index-
function'
     is not known to be defined.
Wrote /Users/pete/Quellen/Emacs_CVS/emacs/lisp/progmodes/dcl-mode.elc
Compiling /Users/pete/Quellen/Emacs_CVS/emacs/lisp/./progmodes/delphi.el
Wrote /Users/pete/Quellen/Emacs_CVS/emacs/lisp/progmodes/delphi.elc
Compiling /Users/pete/Quellen/Emacs_CVS/emacs/lisp/./progmodes/ebnf-
abn.el
Fatal error (4)make[2]: *** [compile] Error 1
make[2]: Target `bootstrap' not remade because of errors.
make[1]: *** [bootstrap-build] Error 2
make: *** [bootstrap] Error 2


In GNU Emacs 22.1.50.1 (powerpc-apple-darwin8.10.0, X toolkit, Xaw3d  
scroll bars)
  of 2007-07-14 on localhost
Windowing system distributor `The XFree86 Project, Inc', version  
11.0.40400000
configured using `configure  '--without-sound' '--without-pop' '--
with-xpm' '--with-jpeg' '--with-tiff' '--with-gif' '--with-png' '--
with-x-toolkit=athena' '--enable-locallisppath=/Library/Application  
Support/Emacs/calendar22:/Library/Application Support/Emacs/caml:/
Library/Application Support/Emacs:/sw/share/emacs21/site-lisp/elib'  
'CPPFLAGS=-no-cpp-precomp -I/usr/include/openssl -I/sw/include/
pango-1.0 -I/sw/lib/freetype219/include -I/sw/lib/freetype219/include/
freetype2 -I/sw/lib/fontconfig2/include -I/sw/include/libpng12 -I/usr/
local/include -I/sw/include' 'CXXFLAGS=-no-cpp-precomp -I/usr/include/
openssl -I/sw/include/pango-1.0 -I/sw/lib/freetype219/include -I/sw/
lib/freetype219/include/freetype2 -I/sw/lib/fontconfig2/include -I/sw/
include/libpng12 -I/usr/local/include -I/sw/include -idirafter /usr/
X11R6/include' 'LDFLAGS=-dead_strip -bind_at_load -L/sw/lib/
freetype219/lib -L/sw/lib/fontconfig2/lib -L/sw/lib/ncurses -L/usr/
local/lib -L/sw/lib' 'CFLAGS=-pipe -fPIC -mcpu=7450 -mtune=7450 -fast  
-mpim-altivec -ftree-vectorize -foptimize-register-move -freorder-
blocks -freorder-blocks-and-partition -fthread-jumps -fpeephole -fno-
crossjumping''

Important settings:
   value of $LC_ALL: nil
   value of $LC_COLLATE: nil
   value of $LC_CTYPE: de_DE.UTF-8
   value of $LC_MESSAGES: nil
   value of $LC_MONETARY: nil
   value of $LC_NUMERIC: nil
   value of $LC_TIME: nil
   value of $LANG: de_DE.UTF-8
   locale-coding-system: utf-8
   default-enable-multibyte-characters: t

Major mode: Compilation

Minor modes in effect:
   TeX-PDF-mode: t
   shell-dirtrack-mode: t
   show-paren-mode: t
   display-time-mode: t
   desktop-save-mode: t
   tooltip-mode: t
   mouse-wheel-mode: t
   menu-bar-mode: t
   file-name-shadow-mode: t
   global-font-lock-mode: t
   font-lock-mode: t
   blink-cursor-mode: t
   unify-8859-on-encoding-mode: t
   utf-translate-cjk-mode: t
   auto-compression-mode: t
   column-number-mode: t
   line-number-mode: t
   transient-mark-mode: t

--
Greetings

   Pete

The mathematician who pursues his studies without clear views of this  
matter, must often have the uncomfortable feeling that his paper and  
pencil surpass him in intelligence.
                                            Ernst Mach




_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug

[Eli Zaretskii]

> From: Peter Dyballa <Peter_Dyballa@...>
> Date: Sat, 28 Jul 2007 15:27:23 +0200
>
> Hello!
>
> Compilation ends with:
>
> In end of data:
> dcl-mode.el:2223:1:Warning: the function `imenu-default-create-index-
> function'
>      is not known to be defined.
> Wrote /Users/pete/Quellen/Emacs_CVS/emacs/lisp/progmodes/dcl-mode.elc
> Compiling /Users/pete/Quellen/Emacs_CVS/emacs/lisp/./progmodes/delphi.el
> Wrote /Users/pete/Quellen/Emacs_CVS/emacs/lisp/progmodes/delphi.elc
> Compiling /Users/pete/Quellen/Emacs_CVS/emacs/lisp/./progmodes/ebnf-
> abn.el
> Fatal error (4)make[2]: *** [compile] Error 1
> make[2]: Target `bootstrap' not remade because of errors.
> make[1]: *** [bootstrap-build] Error 2
> make: *** [bootstrap] Error 2

I cannot reproduce this with today's CVS, on:

   Linux fencepost 2.6.16.29-xen #1 SMP Wed Dec 6 07:32:36 EST 2006 x86_64 GNU/Linux

"Fatal error (4)" seems to say it died with SIGILL.  Could it be some
hardware problem?


_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug

[Peter Dyballa-2]

Am 28.07.2007 um 17:14 schrieb Eli Zaretskii:

> I cannot reproduce this with today's CVS, on:
>
>    Linux fencepost 2.6.16.29-xen #1 SMP Wed Dec 6 07:32:36 EST 2006  
> x86_64 GNU/Linux


I still can!

Without a 'make clean' or 'make distclean' it leads to exactly the  
same result.

>
> "Fatal error (4)" seems to say it died with SIGILL.  Could it be some
> hardware problem?

Of course ... I'll try once more, after distclean and configure, and  
then I'll try to check my hardware.

SIGILL stands for "illegal instruction." Is it so absurd to assume  
that some changed part of GNU Emacs, byte code compiler, is faulty?

--
Greetings

   Pete

Experience is what you get when you don't get what you want.




_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug

[Peter Dyballa-2]

Am 28.07.2007 um 17:14 schrieb Eli Zaretskii:

> "Fatal error (4)" seems to say it died with SIGILL.  Could it be some
> hardware problem?

The Mac OS X CrashReporter now has data from three crashes. All  
happen with the same sequence of calls, but at different RAM  
addresses. An excerpt:

        Command: bootstrap-emacs
        Path:    ../src/bootstrap-emacs
        Parent:  sh [12079]
       
        Version: ??? (???)
       
        PID:    13265
        Thread: 0
       
        Exception:  EXC_BAD_INSTRUCTION (0x0002)
        Code[0]:    0x00000002
        Code[1]:    0x001485d8
       
       
        Thread 0 Crashed:
        0   bootstrap-emacs 0x001485d8 Fsqrt + 100 (floatfns.c:561)
        1   bootstrap-emacs 0x0014491c Feval + 1380 (eval.c:2351)
        2   bootstrap-emacs 0x00144af0 Feval + 1848 (eval.c:2316)
        3   bootstrap-emacs 0x00140ac0 Fprogn + 92 (eval.c:448)
        4   bootstrap-emacs 0x0014097c Fif + 116 (eval.c:398)
        5   bootstrap-emacs 0x00144a44 Feval + 1676 (eval.c:2298)
        6   bootstrap-emacs 0x0014491c Feval + 1380 (eval.c:2351)
        7   bootstrap-emacs 0x00140ac0 Fprogn + 92 (eval.c:448)
        8   bootstrap-emacs 0x00140a44 Fcond + 176 (eval.c:433)
        9   bootstrap-emacs 0x00144a44 Feval + 1676 (eval.c:2298)
        10  bootstrap-emacs 0x00144af0 Feval + 1848 (eval.c:2316)
        11  bootstrap-emacs 0x00140ae0 Fprogn + 124 (eval.c:448)
        12  bootstrap-emacs 0x0014600c funcall_lambda + 752 (eval.c:3200)
        13  bootstrap-emacs 0x001457a8 Ffuncall + 708 (eval.c:3089)
        14  bootstrap-emacs 0x00144ad4 Feval + 1820 (eval.c:2324)
        15  bootstrap-emacs 0x00140ac0 Fprogn + 92 (eval.c:448)
        16  bootstrap-emacs 0x0014097c Fif + 116 (eval.c:398)
        17  bootstrap-emacs 0x00144a44 Feval + 1676 (eval.c:2298)
        18  bootstrap-emacs 0x0014488c Feval + 1236 (eval.c:2409)
        19  bootstrap-emacs 0x00140ae0 Fprogn + 124 (eval.c:448)
        20  bootstrap-emacs 0x00141f30 Flet + 544 (eval.c:1065)
        21  bootstrap-emacs 0x00144a44 Feval + 1676 (eval.c:2298)
        22  bootstrap-emacs 0x00140ac0 Fprogn + 92 (eval.c:448)
        23  bootstrap-emacs 0x0014600c funcall_lambda + 752 (eval.c:3200)
        24  bootstrap-emacs 0x00145c38 apply_lambda + 200 (eval.c:3131)
        25  bootstrap-emacs 0x0014475c Feval + 932 (eval.c:2418)

No other Fsqrt before the crash.


I'll try GCC 4.2 in the morning, then hardware test.

--
Greetings

   Pete

The light at the end of the tunnel has been turned off due to budget  
cuts.




_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug

[Eli Zaretskii]

> Cc: emacs-pretest-bug@...
> From: Peter Dyballa <Peter_Dyballa@...>
> Date: Sat, 28 Jul 2007 23:22:33 +0200
>
> SIGILL stands for "illegal instruction." Is it so absurd to assume  
> that some changed part of GNU Emacs, byte code compiler, is faulty?

No, it's not absurd.  But given that it worked to me, I'd say a
hardware problem is definitely a possibility one should consider.


_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug

[Eli Zaretskii]

> Cc: emacs-pretest-bug@...
> From: Peter Dyballa <Peter_Dyballa@...>
> Date: Sun, 29 Jul 2007 00:40:24 +0200
>
> 0   bootstrap-emacs 0x001485d8 Fsqrt + 100 (floatfns.c:561)

That's the line where `sqrt' is called.

> I'll try GCC 4.2 in the morning, then hardware test.

Yes, a good plan, IMO.  Since `sqrt' is inlined by GCC, it could well
be a compiler bug.  I believe there's a switch to disable this
inlining, so perhaps that's also something to try.


_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug

[Peter Dyballa-2]

Am 29.07.2007 um 05:11 schrieb Eli Zaretskii:

>> I'll try GCC 4.2 in the morning, then hardware test.
>
> Yes, a good plan, IMO.  Since `sqrt' is inlined by GCC, it could well
> be a compiler bug.  I believe there's a switch to disable this
> inlining, so perhaps that's also something to try.

In order to compile GNU Emacs with GCC 4.2 I had to patch configure  
(and configure.in) to not set CPP to "${CC-cc} -E -no-cpp-precomp" (-
no-cpp-precomp is an Apple addition that is not right now  
implemented, or will never be?) and to patch src/s/darwin.h to not  
define C_SWITCH_SYSTEM to '-fpascal-strings -DMAC_OSX' (-fpascal-
strings is another Apple add-on). It compiled OK three times with  
different levels of optimisation. Then I was encouraged to retry with  
GCC 4.0 – and it compiled, as well! I made changes to the compiler  
switches and to the level of optimisation – it still compiled OK.

I cannot recall how the faulty compilation was configured (the  
config.log was not saved), I just remember that there was something  
phony in the *compilation* buffer due to re-configuration, but this  
would have to mean that config.status was patched somehow during CVS  
update or such ... (too bad that machine is not built into Tiger!)


Launching new Emacs as src/emacs I get a warning:

        Warning (initialization): Building Emacs overflowed pure space.  
(See the node Pure Storage in the Lisp manual for details.)

Setting SYSTEM_PURESIZE_EXTRA to 1000 solved this issue.


Another strange effect becomes visible:

        Warning: desktop file appears to be in use by PID 26912. Using it  
may cause conflicts.  Use it anyway? (y or n)

or such appears in echo area for the tenth of a second, or less, and  
is then replaced with a

        Loading cal-bahai...done

so I can't answer the question. Using -q or -Q at emacs invocation  
time can't work, because then no desktop file would be used. In  
"normal" use of GNU Emacs, i.e. launching it off /usr/local/bin, the  
init files are respected and no religious calendar is loaded ...

--
Greetings

   Pete

The most exciting phrase to hear in science, the one that heralds new  
discoveries, is not "Eureka!" (I found it!) but "That's funny..."
                                       Isaac Asimov




_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@...
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug