"OBATA Akio" <obache%netbsd.org@localhost> writes: Really the only issues are the sense of naming - does the name imply that being cross-buildable is special, or that it's normal. the default - if there is no data, should one try to build cross, or not try I think in the glorious cross future, when our comrades have made all upstream distributions cross friendly, cross will be normal, and packages that aren't cross buildable will be viewed as deficient, just like upstream packages that fail with -j are considered broken. So I prefer a variable that indicates "cross is broken", defaulting to no. > I wonder it will be usable with bulk build. > If USE_CROSS_COMPILE_SUPPORTED=yes, and cross-build with fast machines. > If USE_CROSS_COMPILE_SUPPORTED=no, and build with native hosts. It may be easier and cleaner to try cross, and put those jobs on a queue to be done native, than to try to mark all packages. It would also be cool to compare native and cross versions of the same package to see if they are the same
Attachment:
pgpwmxAwa0JVS.pgp
Description: PGP signature