5:32 a.m.
Jim Meyering wrote:
Matthew Booth wrote:
> On 10/08/09 10:21, Jim Meyering wrote:
>> Matthew Booth wrote:
>>> On 10/08/09 09:57, Jim Meyering wrote:
>>>> Here are a few warning-removal changes for daemon/:
>> ...
>>> What warnings are you turning on? I normally compile with -Wall, and I
>>> don't get warning for unused function arguments:
>>
>> I started with these: -Wall -W -Wextra
>
> Now I see where you're coming from. ACK.
>
> Do you find it's generally useful to turn these on across the board?
Yes, Very.
Though I should qualify that.
It's a delicate balance.
The following is the configure.ac snippet that has evolved for coreutils:
The idea is to enable as all warnings, and to go through correcting
offending code or disabling individual warnings, as taste and time permit.
Eventually, you learn that certain classes of warnings are not worth
trying to avoid, because the cure is worse than the disease.
For others, the added value appears minimal, so not worth much time.
AC_ARG_ENABLE([gcc-warnings],
[AS_HELP_STRING([--enable-gcc-warnings],
[turn on lots of GCC warnings (for developers)])],
[case $enableval in
yes|no) ;;
*) AC_MSG_ERROR([bad value $enableval for gcc-warnings option]) ;;
esac
gl_gcc_warnings=$enableval],
[gl_gcc_warnings=no]
)
if test "$gl_gcc_warnings" = yes; then
gl_WARN_ADD([-Werror], [WERROR_CFLAGS])
AC_SUBST([WERROR_CFLAGS])
nw=
# This, $nw, is the list of warnings we disable.
nw="$nw -Wdeclaration-after-statement" # too useful to forbid
nw="$nw -Waggregate-return" # anachronistic
nw="$nw -Wlong-long" # C90 is anachronistic (lib/gethrxtime.h)
nw="$nw -Wc++-compat" # We don't care about C++ compilers
nw="$nw -Wundef" # Warns on '#if GNULIB_FOO' etc in
gnulib
nw="$nw -Wtraditional" # Warns on #elif which we use often
nw="$nw -Wcast-qual" # Too many warnings for now
nw="$nw -Wconversion" # Too many warnings for now
nw="$nw -Wsystem-headers" # Don't let system headers trigger
warnings
nw="$nw -Wsign-conversion" # Too many warnings for now
nw="$nw -Wtraditional-conversion" # Too many warnings for now
nw="$nw -Wunreachable-code" # Too many warnings for now
nw="$nw -Wpadded" # Our structs are not padded
nw="$nw -Wredundant-decls" # openat.h declares e.g., mkdirat
nw="$nw -Wlogical-op" # any use of fwrite provokes this
nw="$nw -Wformat-nonliteral" # who.c and pinky.c strftime uses
nw="$nw -Wvla" # warnings in gettext.h
nw="$nw -Wnested-externs" # use of XARGMATCH/verify_function__
nw="$nw -Wswitch-enum" # Too many warnings for now
nw="$nw -Wswitch-default" # Too many warnings for now
nw="$nw -Wstack-protector" # not worth working around
# things I might fix soon:
nw="$nw -Wfloat-equal" # sort.c, seq.c
nw="$nw -Wmissing-format-attribute" # copy.c
nw="$nw -Wunsafe-loop-optimizations" # a few src/*.c
nw="$nw -Winline" # system.h's
readdir_ignoring_dot_and_dotdot
nw="$nw -Wstrict-overflow" # expr.c, pr.c, tr.c, factor.c
# ?? -Wstrict-overflow
gl_MANYWARN_ALL_GCC([ws])
gl_MANYWARN_COMPLEMENT([ws], [$ws], [$nw])
for w in $ws; do
gl_WARN_ADD([$w])
done
gl_WARN_ADD([-Wno-missing-field-initializers]) # We need this one
gl_WARN_ADD([-Wno-sign-compare]) # Too many warnings for now
gl_WARN_ADD([-Wno-pointer-sign]) # Too many warnings for now
gl_WARN_ADD([-Wno-unused-parameter]) # Too many warnings for now
# In spite of excluding -Wlogical-op above, it is enabled, as of
# gcc 4.5.0 20090517, and it provokes warnings in cat.c, dd.c, truncate.c
gl_WARN_ADD([-Wno-logical-op])
gl_WARN_ADD([-fdiagnostics-show-option])
AC_SUBST([WARN_CFLAGS])
AC_DEFINE([lint], [1], [Define to 1 if the compiler is checking for lint.])
AC_DEFINE([_FORTIFY_SOURCE], [2],
[enable compile-time and run-time bounds-checking, and some warnings])
AC_DEFINE([GNULIB_PORTCHECK], [1], [enable some gnulib portability checks])
fi
I am hoping to find time to do this for libguestfs.