[vlc-devel] [PATCH v2 06/17] contrib: main: ensure NM is defined in makefile

[Alexandre Janniaux]

Ok,

I'll fix with the -gcc variant then, as the case of apple
platforms is already workarounded in buildscripts anyway
and minGW seems to have both variant anyway in the build
containers.

Steve, maybe you have more insight regarding whether mingw
in msys/cygwin have those tools prefixed with $(HOST)-gcc- ?

Regards,
--
Alexandre Janniaux
Videolabs

On Thu, Apr 30, 2020 at 03:24:34PM +0300, Rémi Denis-Courmont wrote:
> I have no idea what you are on about, sorry.
>
> Le 30 avril 2020 12:07:38 GMT+03:00, Alexandre Janniaux <ajanni at videolabs.io> a écrit :
> >Hi,
> >
> >This is right indeed for binutils, but I also have
> >
> >x86_64-apple-darwin-ar / nm / ranlib... without the -gcc
> >suffix (since, maybe, it is coming with clang and not gcc).
> >
> >Likewise, another flaw of this patch is that macosx doesn't
> >have prefixed tools so you should actually use ar/nm directly
> >when cross compiling to apple devices.
> >
> >I was wondering if we should add the `xcrun --find` here as
> >it is still needed in the buildscript to run the configure
> >step, and we would probably want to use =? syntax which
> >disallow the same hack as I wrote for symbols.mk.
> >
> >In the mean time, passing --nm=$(NM) for ffmpeg is needed but
> >currently not really correct since NM is not even defined by
> >the default makefile rules (make -p), even if it doesn't
> >really have any consequences for our builds.
> >
> >I wanted to have fallback first, but I didn't know how to
> >write the detection in a valid way for all platforms,
> >especially Windows ones. This would probably allow fixing
> >those issues in a smarter way.
> >
> >Regards,
> >--
> >Alexandre Janniaux
> >Videolabs
> >
> >On Thu, Apr 30, 2020 at 11:44:26AM +0300, Rémi Denis-Courmont wrote:
> >> Ok but correct value for GCC is gcc-nm (also gcc-ranlib and gcc-ar so
> >you could argue that it's a preexisting condition).
> >>
> >> Le 29 avril 2020 23:40:49 GMT+03:00, Alexandre Janniaux
> ><ajanni at videolabs.io> a écrit :
> >> >ffmpeg rules.make is passing the variable to configure, and nm can
> >be
> >> >present in different flavour, for example in linux/macosx
> >> >cross-compilation environments.
> >> >---
> >> > contrib/src/main.mak | 2 ++
> >> > 1 file changed, 2 insertions(+)
> >> >
> >> >diff --git a/contrib/src/main.mak b/contrib/src/main.mak
> >> >index 0a731f11dcc..534f1d11349 100644
> >> >--- a/contrib/src/main.mak
> >> >+++ b/contrib/src/main.mak
> >> >@@ -70,6 +70,7 @@ endif
> >> > ifneq ($(findstring $(origin AR),undefined default),)
> >> > AR := ar
> >> > endif
> >> >+NM ?= nm
> >> > RANLIB ?= ranlib
> >> > STRIP ?= strip
> >> > WIDL ?= widl
> >> >@@ -88,6 +89,7 @@ endif
> >> > ifneq ($(findstring $(origin AR),undefined default),)
> >> > AR := $(HOST)-ar
> >> > endif
> >> >+NM ?= $(HOST)-nm
> >> > RANLIB ?= $(HOST)-ranlib
> >> > STRIP ?= $(HOST)-strip
> >> > WIDL ?= $(HOST)-widl
> >> >--
> >> >2.26.2
> >> >
> >> >_______________________________________________
> >> >vlc-devel mailing list
> >> >To unsubscribe or modify your subscription options:
> >> >https://mailman.videolan.org/listinfo/vlc-devel
> >>
> >> --
> >> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez
> >excuser ma brièveté.
> >
> >> _______________________________________________
> >> vlc-devel mailing list
> >> To unsubscribe or modify your subscription options:
> >> https://mailman.videolan.org/listinfo/vlc-devel
> >_______________________________________________
> >vlc-devel mailing list
> >To unsubscribe or modify your subscription options:
> >https://mailman.videolan.org/listinfo/vlc-devel
>
> --
> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser ma brièveté.

> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel