On Thu, Oct 03, 2013 at 05:01:33PM +0200, David Coppa wrote:
> On Thu, 26 Sep 2013, David Coppa wrote:
>
> > On Thu, Sep 26, 2013 at 2:30 PM, Remi Pointel <remi.poin...@xiri.fr> wrote:
> > > On Thu, 26 Sep 2013 14:19:28 +0200
> > > David Coppa <dco...@gmail.com> wrote:
> > >> On Thu, 26 Sep 2013, David Coppa wrote:
> > >>
> > >> >
> > >> > Why not getting rid of MODPY_BADEGGS and doing this instead:
> > >> >
> > >> > This will assure the egg-info directory is always created with the
> > >> > right permissions...
> > >> >
> > >> > Comments?
> > >>
> > >> Oops!
> > >> Now with everything in its right place:
> > >>
> > >> Index: python.port.mk
> > >> ===================================================================
> > >> RCS file: /cvs/ports/lang/python/python.port.mk,v
> > >> retrieving revision 1.72
> > >> diff -u -p -u -p -r1.72 python.port.mk
> > >> --- python.port.mk 24 Sep 2013 16:55:02 -0000 1.72
> > >> +++ python.port.mk 26 Sep 2013 12:19:08 -0000
> > >> @@ -91,6 +91,7 @@ MODPY_SETUPUTILS = Yes
> > >> # The setuptools uses test target
> > >> TEST_TARGET ?= test
> > >> _MODPY_USERBASE =
> > >> +_MODPY_PRE_BUILD_STEPS += ;${MODPY_CMD} egg_info
> > >> .else
> > >> # Try to detect the case where a port will build regardless of
> > >> setuptools
> > >> # but the final plist will be different if it's present.
> > >> @@ -117,13 +118,6 @@ MODPY_BIN = ${LOCALBASE}/bin/python${MO
> > >> MODPY_INCDIR =
> > >> ${LOCALBASE}/include/python${MODPY_VERSION}${MODPY_LIB_SUFFIX}
> > >> MODPY_LIBDIR = ${LOCALBASE}/lib/python${MODPY_VERSION}
> > >> MODPY_SITEPKG = ${MODPY_LIBDIR}/site-packages
> > >> -
> > >> -.if defined(MODPY_BADEGGS)
> > >> -. for egg in ${MODPY_BADEGGS}
> > >> -_MODPY_PRE_BUILD_STEPS += ;mkdir -p ${WRKBUILD}/${egg}.egg-info
> > >> -. endfor
> > >> -.endif
> > >> -
> > >>
> > >> # usually setup.py but Setup.py can be found too
> > >> MODPY_SETUP ?= setup.py
> > >
> > > Could be useful to test this diff in a bulk, no?
> >
> > Running against all the ports using py-setuptools should be sufficient.
> >
> > > Remi.
>
> As seen on icb, not all the setuptools based stuff support this
> command. Some ports fail with:
>
> error: invalid command 'egg_info'
>
> New revision of the diff, hope I got it right this time:
>
> Index: python.port.mk
> ===================================================================
> RCS file: /cvs/ports/lang/python/python.port.mk,v
> retrieving revision 1.72
> diff -u -p -u -p -r1.72 python.port.mk
> --- python.port.mk 24 Sep 2013 16:55:02 -0000 1.72
> +++ python.port.mk 3 Oct 2013 15:01:04 -0000
> @@ -91,6 +91,7 @@ MODPY_SETUPUTILS = Yes
> # The setuptools uses test target
> TEST_TARGET ?= test
> _MODPY_USERBASE =
> +_MODPY_PRE_BUILD_STEPS += ;${MODPY_CMD} --help-commands | grep -qw egg_info
> && ${MODPY_CMD} egg_info
> .else
> # Try to detect the case where a port will build regardless of setuptools
> # but the final plist will be different if it's present.
> @@ -117,13 +118,6 @@ MODPY_BIN = ${LOCALBASE}/bin/python${MO
> MODPY_INCDIR =
> ${LOCALBASE}/include/python${MODPY_VERSION}${MODPY_LIB_SUFFIX}
> MODPY_LIBDIR = ${LOCALBASE}/lib/python${MODPY_VERSION}
> MODPY_SITEPKG = ${MODPY_LIBDIR}/site-packages
> -
> -.if defined(MODPY_BADEGGS)
> -. for egg in ${MODPY_BADEGGS}
> -_MODPY_PRE_BUILD_STEPS += ;mkdir -p ${WRKBUILD}/${egg}.egg-info
> -. endfor
> -.endif
> -
>
> # usually setup.py but Setup.py can be found too
> MODPY_SETUP ?= setup.py
>
That works for the three ports broken by the previous diff (terminator,
py-cef and py-greenlet) but that looks like a huge hack to me.
Landry
