Hi Luiz,
> Coverage is disabled by default but it is automatically enabled by
> bootstrap-configure using --enable-coverage, it creates a new make target
> 'coverage' that can be run after make check or any tester that produces
> .gdca files.
>
> The build time using --enable-coverage is considerable bigger so if
> that is a problem just disable it, also this is the reason why it is not
> added directly under --enable-maintainer-mode so it is possible to
> turn off coverage without having to turn off maintainer mode.
>
> Since lcov apparently cannot detect changes to .gcda there is another
> make target 'clean-coverage' so after that make coverage can be
> run again.
> ---
> Makefile.am | 17 +++++++++++++++++
> acinclude.m4 | 8 ++++++++
> bootstrap-configure | 1 +
> configure.ac | 5 +++++
> 4 files changed, 31 insertions(+)
>
> diff --git a/Makefile.am b/Makefile.am
> index 15cc149..4b3351a 100644
> --- a/Makefile.am
> +++ b/Makefile.am
> @@ -331,5 +331,22 @@ lib/bluetooth/%.h: lib/%.h
> $(AM_V_at)$(MKDIR_P) lib/bluetooth
> $(AM_V_GEN)$(LN_S) -f "$(abs_top_builddir)"/$< $@
>
> +if COVERAGE
> +clean-coverage:
> + $(RM) -r coverage
> +
> +coverage:
> + $(AM_V_at)$(MKDIR_P) -p coverage
> + @lcov --compat-libtool --directory $(top_builddir) --capture \
> + --output-file coverage/check.info
> + @genhtml -o coverage/ coverage/check.info
I would actually prefer if we can have some sort AM_V_GEN thing working here. So that make V=1 will end up showing the exact command.
> +
> +clean-local: clean-coverage
> + @find $(top_builddir) -name "*.gcda" -exec $(RM) {} \;
> + @find $(top_builddir) -name "*.gcno" -exec $(RM) {} \;
> + @lcov --directory . --zerocounters
> + $(RM) -r lib/bluetooth
> +else
> clean-local:
> $(RM) -r lib/bluetooth
> +endif
This if for clean-local is not something I like that much. Can we not come up with a better way and leave clean-local as it is. Just make clean-coverage a null-op.
> diff --git a/acinclude.m4 b/acinclude.m4
> index 2065852..4bb8bde 100644
> --- a/acinclude.m4
> +++ b/acinclude.m4
> @@ -51,6 +51,14 @@ AC_DEFUN([MISC_FLAGS], [
> misc_ldflags="$misc_ldflags -pie"
> fi
> ])
> + AC_ARG_ENABLE(coverage, AC_HELP_STRING([--enable-coverage],
> + [enable coverage flags]), [
> + enable_coverage=${enableval}
> + if (test "${enable_coverage}" = "yes"); then
> + misc_cflags="$misc_cflags --coverage"
> + misc_ldflags="$misc_ldflags --coverage"
> + fi
> + ])
> AC_SUBST([MISC_CFLAGS], $misc_cflags)
> AC_SUBST([MISC_LDFLAGS], $misc_ldflags)
> ])
> diff --git a/bootstrap-configure b/bootstrap-configure
> index c7f08ed..35aa98b 100755
> --- a/bootstrap-configure
> +++ b/bootstrap-configure
> @@ -14,4 +14,5 @@ fi
> --enable-experimental \
> --enable-android \
> --enable-sixaxis \
> + --enable-coverage \
> --disable-datafiles $*
> diff --git a/configure.ac b/configure.ac
> index 18d0b55..5e684a3 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -252,4 +252,9 @@ AC_ARG_ENABLE(android, AC_HELP_STRING([--enable-android],
> [enable_android=${enableval}])
> AM_CONDITIONAL(ANDROID, test "${enable_android}" = "yes")
>
> +if (test "${enable_coverage}" = "yes"); then
> + AC_CHECK_PROG([LCOV], [lcov], [yes], AC_MSG_ERROR(lcov is required))
> +fi
> +AM_CONDITIONAL([COVERAGE], [test "${enable_coverage}" = "yes"])
> +
Still not 100% convinced that we really want this. I get the feeling that maintainer-mode would be just fine.
Regards
Marcel
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
