Hi Enrico, On Tue, Nov 05, 2013 at 05:16:45PM +0100, Enrico Scholz wrote: > > > Laurentiu Palcu > <laurentiu.palcu-ral2jqcrhueavxtiumw...@public.gmane.org> writes: > > > Flex needs m4 to run (see below) and, since the create_wrapper > > introduces a bash dependency, export the right m4 path when /etc/profile > > is sourced by the shell. > > > > Snippet from the flex documentation: > > "The macro processor m4 must be installed wherever flex is installed. > > <...> > > m4 is only required at the time you run flex." > > Content at profile.d/ won't be evaluated when program is started without > a shell (e.g. by 'execlp("flex", "flex", ...)'). > > > > [YOCTO #5329] > > This should be solved by defining a correct path for the m4 binary at > build time. > > EXTRA_OECONF += " ac_cv_path_M4=${M4}" > EXTRA_OEMAKE += " m4=${M4_class-native}" This is, indeed, a better solution for the target issue. But, I don't think we can remove the wrapper scripts for native/nativesdk packages (as you suggested in bugzilla). And the reason for that are those guys using sstate from build machines. They'll end up with flex searching for m4 in a path belonging to the build machine.
I'll prepare a v2. Laurentiu > > is probably a better solution. > > > > Enrico > _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core