Izumi Tsutsui <tsutsui@ceres.dti.ne.jp> writes:

> th-nozaki@netwrk.co.jp wrote:
>
>> thanks for your log, this is Makefile's problem.
>> libc don't try open(2) /usr/lib/i18n/libUTF8.so.*  and
>> /usr/share/i18n/esdb/UTF8.esdb.
>> apparently, libc.so lacks -D18N_DYNAMIC flag during their own building.
>> 
>> if you don't wait until fix is committed to CVS,
>> (need some time to do it, i'm not bsd.lib.mk guru...)
>
> CPPPICFLAGS seems used only to compile .so from *.s files.
> I guess we should use CPICFLAGS instead even for cpp(1) flags
> on compiling .so files. (it already contains -DPIC)
> ---
> Izumi Tsutsui
>
>
> Index: Makefile
> ===================================================================
> RCS file: /cvsroot/src/lib/libc/Makefile,v
> retrieving revision 1.128
> diff -u -r1.128 Makefile
> --- Makefile	15 Mar 2007 00:59:32 -0000	1.128
> +++ Makefile	17 Mar 2007 18:15:22 -0000
> @@ -106,7 +106,7 @@
>  # multibyte for libc.so.  the quirk should be removed when we support
>  # dlopen() from within statically linked binaries.
>  .if ${MKPICLIB} != "no"
> -CPPPICFLAGS+=	-D_I18N_DYNAMIC
> +CPICFLAGS+=	-D_I18N_DYNAMIC
>  .elif ${MKPIC} != "no"
>  CPPFLAGS+=	-D_I18N_DYNAMIC
>  .endif

I reverted to 1.127 of src/lib/libc/Makefile, and redid libc
(clean/build/install) and now iconv(1) and gnome panel applets work.

The CSHLIBFLAGS was added by yamt@ in 2002, but matt@ recently changed
to PICFLAGS to avoid some other lossage, so reverting the change
doesn't seem wonderful.  Are you saying that the patch you posted
results in a working libc with iconv?  I don't understand the maze of
twisty preprocessor flags, all alike.