Jonathan Perkin wrote:
> * On 2006-03-14 at 14:04 GMT, Roland Illig wrote:
> 
> 
>>http://www.netbsd.org/Documentation/pkgsrc/platforms.html, 3.2.7.4
> 
> 
> Ick, we can't seriously be recommending that as a legitimate fix?

I think we can and should do that. The source code of /bin/ksh belongs 
to the things that are not published as part of OpenSolaris by Sun 
Microsystems. There must be some strong reason behind. :)

The segmentation fault occurs in the function arg_build, which is called 
by sh_exec. There does not seem to be related source code in the 
publicly accessible web.

(dbx) run
Running: ksh /home/j2001/1illig/pkg/bin/libtool --tag=CC --mode=link gcc 
-o libgettextpo.la -rpath /home/j2001/1illig/pkg/lib -version-info 1:0:1 
-rpath /home/j2001/1illig/pkg/lib libgettextsrc.la
(process id 26527)
Reading de_DE.ISO8859-1.so.2
signal SEGV (no mapping at the fault address) in (unknown) at 0x18480
0x00018480: __fsr_init_value+0x18480:   ldsb     [%i0 + 8], %g2

(dbx) where
   [1] 0x18480(0x2f, 0xffffffff, 0x53f40, 0x0, 0xb3880, 0xff1c2c04), at 
0x18480
   [2] arg_build(0x0, 0x0, 0x2f, 0x0, 0xffbedc98, 0x2fa04), at 0x1837c
   [3] sh_exec(0x4, 0x53ec0, 0x0, 0x0, 0x826a8, 0x53e38), at 0x3255c
   [4] sh_exec(0x4, 0x53ec0, 0x0, 0x0, 0x82688, 0x53e38), at 0x33c3c
   [5] sh_exec(0x4, 0x53ec0, 0x0, 0x0, 0x82760, 0x53e38), at 0x3369c
   [6] sh_exec(0x58b80, 0x5b258, 0x0, 0x0, 0x82554, 0x53e38), at 0x33998
   [7] sh_exec(0x4, 0x53ec0, 0x0, 0x0, 0x82818, 0x53e38), at 0x3367c
   [8] sh_exec(0xb26c4, 0x5b0a0, 0x0, 0x0, 0x72d90, 0x53e38), at 0x33998
   [9] sh_exec(0x4, 0x53ec0, 0x0, 0x0, 0xa530c, 0x53e38), at 0x3367c
   [10] sh_exec(0x66970, 0x53ec0, 0x66960, 0x0, 0x589e8, 0x53e38), at 
0x33d90
   [11] sh_exec(0x4, 0x53ec0, 0x0, 0x0, 0xb2624, 0x53e38), at 0x3367c
   [12] sh_exec(0x4, 0x53ec0, 0x20, 0x0, 0x57e1c, 0x53e38), at 0x33c54
   [13] 0x2a35c(0x57e1c, 0x52800, 0x200, 0x53ec0, 0x53ec4, 0x53ec0), at 
0x2a35c
   [14] main(0x57e12, 0xffbef4e9, 0x53d04, 0x53f4c, 0x53cf0, 0x53ec4), 
at 0x29b90

Roland