thanks, I took another look already and came up with nearly the same
diff and committed. (I didn't bump REVISION because it doesn't change
anything for archs which are able to build it).

On 2025/09/14 18:04, SASANO Takayoshi wrote:
> Hello,
> 
> > oh, it fails on i386...
> 
> To avoid the problem, SIMD needs to be disabled except amd64 and aarch64.
> The configuration word looks different from previous version.
> 
> Here is new diff, ok?
> 
> Index: Makefile
> ===================================================================
> RCS file: /cvs/ports/textproc/libunicode/Makefile,v
> diff -u -p -r1.4 Makefile
> --- Makefile	11 Sep 2025 12:46:46 -0000	1.4
> +++ Makefile	14 Sep 2025 08:58:24 -0000
> @@ -3,6 +3,7 @@ COMMENT =	Unicode support to C++ with si
>  GH_ACCOUNT =	contour-terminal
>  GH_PROJECT =	libunicode
>  GH_TAGNAME =	v0.6.0
> +REVISION =	0
>  
>  SHARED_LIBS  =	unicode		0.0 # 0.0
>  SHARED_LIBS +=	unicode_loader	0.0 # 0.0
> @@ -34,7 +35,7 @@ CONFIGURE_ARGS =	-DLIBUNICODE_TESTING=OF
>  			-DLIBUNICODE_UCD_DIR=${LOCALBASE}/share/unicode/ucd
>  
>  .if ${MACHINE_ARCH} != "amd64" && ${MACHINE_ARCH} != "aarch64"
> -CONFIGURE_ARGS +=	-DLIBUNICODE_USE_INTRINSICS=OFF
> +CONFIGURE_ARGS +=	-DLIBUNICODE_SIMD_IMPLEMENTATION=none
>  .endif
>  
>  .include <bsd.port.mk>
> 
> -- 
> SASANO Takayoshi (JG1UAA) <uaa@mx5.nisiq.net>
>