On 10/24/2017 06:45 PM, Philippe Mathieu-Daudé wrote: > Hi Richard, > > On 10/21/2017 09:46 PM, Richard Henderson wrote: >> Do not require the submodule, but use it if present. Allow the >> command-line to override system or git submodule either way. >> >> Signed-off-by: Richard Henderson <richard.hender...@linaro.org> >> --- >> Makefile | 13 +++++++++++++ >> .gitmodules | 3 +++ >> capstone | 1 + >> configure | 60 >> +++++++++++++++++++++++++++++++++++++++++++++++++----------- >> 4 files changed, 66 insertions(+), 11 deletions(-) >> create mode 160000 capstone >> >> diff --git a/Makefile b/Makefile >> index 9372742f86..beecc85bee 100644 >> --- a/Makefile >> +++ b/Makefile >> @@ -383,6 +383,19 @@ subdir-dtc: .git-submodule-status dtc/libfdt dtc/tests >> dtc/%: .git-submodule-status >> mkdir -p $@ >> >> +# Overriding CFLAGS causes us to lose defines added in the sub-makefile. >> +# Not overriding CFLAGS leads to mis-matches between compilation modes. >> +# Therefore we replicate some of the logic in the sub-makefile. > > I'm having plenty of "missing-prototypes" warnings:
Yes, we use lots of -Wfoo that upstream Capstone does not. I do strip -Werror, so at least it builds. Are you suggesting that I drop most of our extra -Wfoo? I suppose that's reasonable. We don't want our developers worrying about warnings coming from upstream code. If in fact you believe that most of our developers won't just install libcapstone-dev and be done? r~
