Latest master is broken

Latest master is broken

[Eli Zaretskii]

It produces a warning, and then aborts during loadup:

  In file included from fns.c:30:0:
  fns.c: In function 'make_hash_table':
  fns.c:3693:53: warning: implicit conversion from 'float' to 'double' to match other operand of binary expression [-Wdouble-promotion]
     eassert (0 < rehash_threshold && rehash_threshold <= 1.0);
						       ^
  lisp.h:196:27: note: in definition of macro 'eassert'
      (suppress_checking || (cond)     \
			     ^
  [...]
    CCLD     temacs.exe
  /bin/mkdir -p ../etc
  make -C ../lisp update-subdirs
  make[2]: Entering directory `/d/gnu/git/emacs/trunk/lisp'
  make[2]: Leaving directory `/d/gnu/git/emacs/trunk/lisp'
  ./temacs --batch  --load loadup bootstrap

  fns.c:3693: Emacs fatal error: assertion failed: 0 < rehash_threshold && rehash_threshold <= 1.0

Let me know if I can provide more information.

Re: Latest master is broken

[Kaushal Modi]

[-- Attachment #1: Type: text/plain, Size: 948 bytes --]

If it helps, I just did a dirty build of emacs (no make bootstrap) and it
built fine.

Emacs version: GNU Emacs 26.0.50 (build 33, x86_64-unknown-linux-gnu, GTK+
Version 2.24.23)
 of 2017-02-19, built using commit 143bc75c414434badcb324db056ad37c15893bd9.

./configure options:
  --with-modules --prefix=/home/kmodi/usr_local/apps/6/emacs/master
'--program-transform-name=s/^ctags$/ctags_emacs/' 'CPPFLAGS=-fgnu89-inline
-I/home/kmodi/usr_local/6/include -I/usr/include/freetype2 -I/usr/include'
'CFLAGS=-ggdb3 -O0' 'CXXFLAGS=-ggdb3 -O0'
'LDFLAGS=-L/home/kmodi/usr_local/6/lib -L/home/kmodi/usr_local/6/lib64
-ggdb3'

Features:
  XPM JPEG TIFF GIF PNG RSVG IMAGEMAGICK SOUND GPM DBUS GCONF GSETTINGS
NOTIFY ACL LIBSELINUX GNUTLS LIBXML2 FREETYPE LIBOTF XFT ZLIB
TOOLKIT_SCROLL_BARS GTK2 X11 MODULES


On Sun, Feb 19, 2017 at 11:12 AM Eli Zaretskii <eliz@gnu.org> wrote:

> It produces a warning, and then aborts during loadup:
>
> --

Kaushal Modi

[-- Attachment #2: Type: text/html, Size: 1477 bytes --]

Re: Latest master is broken

[Eli Zaretskii]

> From: Kaushal Modi <kaushal.modi@gmail.com>
> Date: Sun, 19 Feb 2017 16:29:55 +0000
> Cc: emacs-devel@gnu.org
> 
> If it helps, I just did a dirty build of emacs (no make bootstrap) and it built fine.

Sheer luck.

Re: Latest master is broken

[Eli Zaretskii]

> Date: Sun, 19 Feb 2017 18:13:01 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> It produces a warning, and then aborts during loadup:
> 
>   In file included from fns.c:30:0:
>   fns.c: In function 'make_hash_table':
>   fns.c:3693:53: warning: implicit conversion from 'float' to 'double' to match other operand of binary expression [-Wdouble-promotion]
>      eassert (0 < rehash_threshold && rehash_threshold <= 1.0);
> 						       ^
>   lisp.h:196:27: note: in definition of macro 'eassert'
>       (suppress_checking || (cond)     \
> 			     ^
>   [...]
>     CCLD     temacs.exe
>   /bin/mkdir -p ../etc
>   make -C ../lisp update-subdirs
>   make[2]: Entering directory `/d/gnu/git/emacs/trunk/lisp'
>   make[2]: Leaving directory `/d/gnu/git/emacs/trunk/lisp'
>   ./temacs --batch  --load loadup bootstrap
> 
>   fns.c:3693: Emacs fatal error: assertion failed: 0 < rehash_threshold && rehash_threshold <= 1.0

Fixed.

Re: Latest master is broken

[Stefan Monnier]

>> fns.c:3693:53: warning: implicit conversion from 'float' to 'double' to match other operand of binary expression [-Wdouble-promotion]

Should be harmless.

>> fns.c:3693: Emacs fatal error: assertion failed: 0 < rehash_threshold && rehash_threshold <= 1.0
> Fixed.

Thanks,


        Stefan

Re: Latest master is broken

[Paul Eggert]

Stefan Monnier wrote:
> Should be harmless.

Yes, although it was annoying that the change caused 
(hash-table-rehash-threshold (make-hash-table)) to yield 0.800000011920929 
instead of the correct value of 0.8, due to a rounding error caused by 
converting double to float and back again.

I just now fixed this by changing struct Lisp_Hash_Table to contain a double 
rather than a float. If it's a significant win to save 4 bytes in struct 
Lisp_Hash_Table by using a float rather than a double, then we can do that, 
albeit at the cost of complicating the code elsewhere.