NetBSD 絹の日記

modena@makoto 14:57:59/130423(..wp-content/languages)% msgfmt -c -o ja.mo ja.po 
msgfmt: ja.po: headerfield `Last-Translator' missing in header
msgfmt: ja.po: headerfield `Language-Team' missing in header
ja.po:98: keyword "msgctxt" unknown
ja.po:98:8: parse error
ja.po:103: keyword "msgctxt" unknown
ja.po:103:8: parse error
ja.po:108: keyword "msgctxt" unknown
ja.po:108:8: parse error
....
....
msgfmt: too many errors, aborting
modena@makoto 14:58:03/130423(..wp-content/languages)%

I did not installed gettext-tools (aside from built-in ones). Now I have 0.18.2.1 and sucessfull with following interaction:

modena@makoto 15:09:10/130423(..wp-content/languages)% msgfmt -c -o ja.mo ja.po
ja.po:4: header field 'Last-Translator' missing in header
ja.po:4: header field 'Language-Team' missing in header
ja.po:4: warning: header field 'Language' missing in header

以前に Segmentation Fault する話を書いていて、どうも同じことを send-pr して下さった方がいらっしゃるようなのでもう一度見ている。不思議だが NetBSD/amd64 6.1_RC2 では lib-src/Makefile に \ がいくつも抜けているのか、と思われる症状があって、とりあえず、NetBSD/i386 5.2 で試して見ている。
一応再現する

LOADING its/kata...
LOADING its/hankaku...
LOADING its/zenkaku...
Loading japanese...
Pure bytes used: 372380
Dumping under names emacs and emacs-19.28.1
./temacs -batch -l
  /export/CHROOT/usr/pkgsrc/editors/mule/work/mule/src/../lisp/mule-inst.el
  /export/CHROOT/usr/pkgsrc/editors/mule/work/mule/src/../lisp/quail/[a-zA-Z]*.el
Warning: arch-dependent data dir (/usr/pkg/lib/mule/19.28/i386--netbsd/) does not exist.
Loading /export/CHROOT/usr/pkgsrc/editors/mule/work/mule/src/../lisp/mule-inst.el...
./emacs -batch -l
.  /export/CHROOT/usr/pkgsrc/editors/mule/work/mule/src/../lisp/mule-diag.elc
.  -f dump-charset
gmake[1]: *** [../etc/charset] Segmentation fault (core dumped)
gmake[1]: Leaving directory `/export/CHROOT/usr/pkgsrc/editors/mule/work/mule/src'
gmake: *** [src] Error 2
*** Error code 2

Reading symbols from /usr/libexec/ld.elf_so...done.
Loaded symbols for /usr/libexec/ld.elf_so
Core was generated by `emacs'.
Program terminated with signal 11, Segmentation fault.
#0  0xbb483167 in kill () from /usr/lib/libc.so.12
(gdb) bt
#0  0xbb483167 in kill () from /usr/lib/libc.so.12
#1  0x08083f7d in fatal_error_signal (sig=11) at emacs.c:225
#2  <signal handler called>
#3  0x080b5f8c in make_uninit_string (length=44) at alloc.c:943
#4  0x080b60c9 in make_string (contents=0xbf7ff175 
    "TIMEFMT=%U %S %*E %P %X+%Dk %I+%Oio %Fpf+%Ww", length=44) at alloc.c:922
#5  0x080b6110 in build_string (str=0xbf7ff175 
    "TIMEFMT=%U %S %*E %P %X+%Dk %I+%Oio %Fpf+%Ww") at alloc.c:931
#6  0x080da8ab in set_process_environment () at callproc.c:1229
#7  0x0808388b in main (argc=6, argv=0xbf7fdef4, envp=Cannot access memory at address 0x8
) at emacs.c:722
(gdb)

mule/src/alloc.c:

    934 Lisp_Object
    935 make_uninit_string (length)
    936      Lisp_Object_Int length;
    937 { 
    938   register Lisp_Object val;
    939   register Lisp_Object_Int fullsize = STRING_FULLSIZE (length);
    940   
    941   if (length < 0) abort ();
    942   
    943   if (fullsize <= STRING_BLOCK_SIZE - current_string_block->pos)
    944     /* This string can fit in the current string block */
    945     {

以前の
patches/patch-src_alloc_c
がちょっと気になる。

＠ PKG_OPTIONS.mule= -wnn4:

以前に、FreeWnn を変更したら、だめになったのでは、という話があったので、一応

PKG_OPTIONS.mule= -wnn4

こうしても同じことになる。断定出来るかどうかは不明ではあるが、関係ないと思う。

2013年04月23日(火) 旧暦 [n年日記] [更新:"2013/04/24 21:56:01"]

#1 [pkgsrc] gettext-tools

#2 [pkgsrc] editors/mule

＠ PKG_OPTIONS.mule= -wnn4: