On 2016/01/13 10:15, Daniel Bolgheroni wrote:
> On Tue, Jan 12, 2016 at 03:05:05PM +0000, Brandon Mercer wrote:
> > I'm somehow missing the initial email that Daniel wrote that proposes
> > building libstdc++ as part of arm-none-eabi, however, that bit me while
> > working on some stuff last week. I think the toolchain is not very useful
> > without it.
>
> OK, I generated a diff to apply to devel/arm-none-eabi.
>
> I removed the python files conflicts with lang/gcc/4.9 by adding these
> to ${PREFIX}/arm-none-eabi/share instead of ${PREFIX}/share.
>
> Running portcheck gives the usual
>
> Python module without compiled version, consider using ${MODPY_BIN} ...
>
> but I didn't add python. lang/gcc/4.9 also has .py files but doesn't add
> it either.
>
> patch needs -p0 to apply correctly.
>
> > I don't think that a FLAVOR is the right way to go, in this case I feel it
> > should be built by default.
>
> Yes.
>
> Any other suggestions?
The idea seems reasonable but I don't know these ports well.
I would suggest setting some variable like GCC_V=4.9.3 and using
SUBST_VARS+=GCC_V and substituting it in the PLIST files, so that it's
easier to see what is happening with any future update diffs.
Bump REVISION in everything you touch. I noticed it's missing in
newlib/Makefile but I haven't checked the other files.
> -@comment @man man/man7/fsf-funding.7
> -@comment @man man/man7/gfdl.7
> -@comment @man man/man7/gpl.7
> +@man man/man7/fsf-funding.7
> +@man man/man7/gfdl.7
> +@man man/man7/gpl.7
remove this chunk, it conflicts with lang/gcc.
