On Sun, Nov 24, 2013 at 09:58:34PM +0100, Stefan Weil wrote: > Am 24.11.2013 17:48, schrieb Michael S. Tsirkin: > > On Sun, Nov 24, 2013 at 05:39:58PM +0200, Michael S. Tsirkin wrote: > >> On Sun, Nov 24, 2013 at 02:54:51PM +0100, Stefan Weil wrote: > >>> Am 24.11.2013 14:42, schrieb Michael S. Tsirkin: > >>>> On Sun, Nov 24, 2013 at 03:04:04PM +0200, Marcel Apfelbaum wrote: > >>>>> On Sun, 2013-11-24 at 11:52 +0200, Michael S. Tsirkin wrote: > >>>>>> Useful for platforms with a broken IASL. > >>>>>> > >>>>>> Signed-off-by: Michael S. Tsirkin <m...@redhat.com> > >>>>>> --- > >>>>>> configure | 3 ++- > >>>>>> 1 file changed, 2 insertions(+), 1 deletion(-) > >>>>>> > >>>>>> diff --git a/configure b/configure > >>>>>> index 0592ba7..d0a0abe 100755 > >>>>>> --- a/configure > >>>>>> +++ b/configure > >>>>>> @@ -1080,6 +1080,7 @@ echo " --source-path=PATH path of source > >>>>>> code [$source_path]" > >>>>>> echo " --cross-prefix=PREFIX use PREFIX for compile tools > >>>>>> [$cross_prefix]" > >>>>>> echo " --cc=CC use C compiler CC [$cc]" > >>>>>> echo " --iasl=IASL use ACPI compiler IASL [$iasl]" > >>>>>> +echo " iasl='' (empty string) disables ACPI > >>>>>> compiler" > >>>>>> echo " --host-cc=CC use C compiler CC [$host_cc] for > >>>>>> code run at" > >>>>>> echo " build time" > >>>>>> echo " --cxx=CXX use C++ compiler CXX [$cxx]" > >>>>>> @@ -4269,7 +4270,7 @@ else > >>>>>> fi > >>>>>> echo "PYTHON=$python" >> $config_host_mak > >>>>>> echo "CC=$cc" >> $config_host_mak > >>>>>> -if $iasl -h > /dev/null 2>&1; then > >>>>>> +if test "$iasl" && $iasl -h > /dev/null 2>&1; then > >>>>>> echo "IASL=$iasl" >> $config_host_mak > >>>>>> fi > >>>>>> echo "CC_I386=$cc_i386" >> $config_host_mak > >>>>> I tried to disable iasl without this patch and it worked for me... > >>>>> I used: > >>>>> - [Qemu-devel] [PATCH] configure: make --iasl option actually work > >>>>> - ./configure --iasl="" (or --iasl=) > >>>>> and it worked (no sign of IASL in config-host.mak) > >>>>> > >>>>> I think this is because the prev test "$iasl -h > /dev/null 2>&1" > >>>>> fails of course. > >>>> By luck, yes. > >>>> > >>>> So this is just a documentation patch: cleaner to document the > >>>> interface explicitly. > >>>> I'll make this clearer in the commit log. > >>>> > >>> --iasl=false works without further modifications (because test "false > >>> -h" works and returns false). It also looks more natural than --iasl=. > >>> > >>> Cheers, Stefan > >> It seems that some people try --iasl= as the more natural way > >> to do this. It's not documented that a failing iasl will > >> cause a fall-back and I'd rather document an explicit option > >> than making it fail. > > I'm also curious why would you say =false is natural: > > it's a string option not a boolean one. > > > > 'false' is a string here, namely the name of the executable which is > normally found at /bin/false. > It takes any number of parameters and always returns 1 (which is boolean > false > in shell conventions). > > You could also write --iasl=/bin/false. > > I suggest this kind of patch: > > echo " --iasl=IASL use ACPI compiler IASL [$iasl]" > +echo " iasl=false disables the ACPI compiler"
It's still not a natural interface. It works by chance because we run iasl and test the return code, but we don't have to. For example, a reasonable implementation might produce an error if the user-specified iasl fails, breaking this hack. Interfaces should not follow implementation. If you don't want iasl a reasonable syntax is '' or --disable-iasl. > echo " --host-cc=CC use C compiler CC [$host_cc] for code run > at" > > > Stefan