On 2024/10/24 19:13, Klemens Nanni wrote:
> 24.10.2024 17:46, Stuart Henderson пишет:
> > On 2024/10/20 15:32, Klemens Nanni wrote:
> >>>> +
> >>>> +.if ${MODPY_TEST_LINK_SO:L} == "yes" && !empty(MODPY_TEST_LIBDIR)
> >>>> +MODPY_TEST_TARGET := for _dir in ${MODPY_TEST_LIBDIR:S,:, ,g}; do \
> >>>> +	if [ -e $$$${_dir} ]; then \
> >>>> +		cd $$$${_dir} && \
> >>>> +		find . -name '*.so' -type f \
> >>>> +			-exec ln -sf $$$${_dir}/{} ${WRKSRC}/{} \; ;\
> >>>> +	fi; done; ${MODPY_TEST_TARGET}
> >>>> +.endif
> >>>
> >>> Looks fine make-wise, but I lack python experience to judge the approach itself.
> >>>
> >>> I think you can do without double-escaping and prepending the target variable
> >>> by using your own and using it fist, this way it looks a clearer;  diff below.
> >>>
> >>>> +
> >>>> +
> >>>>  # dirty way to do it with no modifications in bsd.port.mk
> >>>>  .if empty(CONFIGURE_STYLE)
> >>>>  .  if !target(do-build)
> >>>
> >>> Since that is a default-off opt-in, I'm fine committing it and tweaking in-tree.
> >>>
> >>> Feedback or OKs for either of the two diffs?
> >>
> >> None so far,  we'll land this in a bit unless there's objection.
> > 
> > Can I have a diff rather than a "either of two diffs" to look at
> > please?
> 
> Sure, here's the same diff I sent out last time.
> 
> Index: lang/python/python.port.mk
> ===================================================================
> RCS file: /cvs/ports/lang/python/python.port.mk,v
> diff -u -p -r1.192 python.port.mk
> --- lang/python/python.port.mk	25 Jul 2024 15:42:03 -0000	1.192
> +++ lang/python/python.port.mk	5 Oct 2024 10:16:58 -0000
> @@ -303,6 +303,18 @@ MODPY_COMPILEALL = ${MODPY_BIN} -m compi
>  MODPY_COMPILEALL = ${MODPY_BIN} -m compileall -j ${MAKE_JOBS} -s ${WRKINST} -o 0 -o 1
>  .endif
>  
> +MODPY_TEST_TARGET =
> +MODPY_TEST_LINK_SO ?=	No
> +.if ${MODPY_TEST_LINK_SO:L} == "yes" && !empty(MODPY_TEST_LIBDIR)
> +MODPY_TEST_SO_CMD = for _dir in ${MODPY_TEST_LIBDIR:S,:, ,g}; do \
> +	if [ -e $${_dir} ]; then \
> +		cd $${_dir} && \

I think that would be better to use "if [ -d" rather than -e.

Otherwise OK sthen@