On Tue, Aug 13, 2019 at 01:37:14AM +0900, Masahiro Yamada wrote: > On Fri, Aug 9, 2019 at 2:15 PM Nathan Chancellor > > This is problematic because there is no way for CC, HOSTCC, and HOSTCXX > > to be passed to make within scripts/kconfig/merge_config.sh. > Is it so problematic? > If you start from make, CC=clang and HOSTCC=clang are propagated to sub-make > even via shell scripts such as merge_config.sh > Only the problem I see is the situation where > a user directly runs scripts/kconfig/merge_config.sh > without using make as a start-point. This is really a very common thing for testing infrastructure to do, it'll combine a base defconfig with a fragment enabling extra stuff either to directly cover that extra stuff or to ensure that configuration options needed for testsuites get turned on. > A user can wrap merge_config.sh with a simple Makefile > if they want to override CC, HOSTCC, etc. If we want to do that it seems sensible to provide that Makefile upstream so there's a standard thing, it might also help people notice that they need to do this and avoid getting surprised.
Attachment:
signature.asc
Description: PGP signature