On 2007-11-14, Adam Dunkels <[email protected]> wrote:
>>>> I see that the gcc "pointers to labels" implementation is what
>>>> is used by default (which removes the restriction on using
>>>> switch() statements in a protothread).
>>>
>>> I think it actually defaults to the switch()-based
>>> implementation.
>>
>> You're right. I misread the conditional in the include file.
>>
>> Defining LC_INCLUDE to be "lc-addrlabels.h" will use the
>> label-pointer version. Have you thought about that being the
>> default when gcc is used?
>
> Hadn't thought about that before... Is it possibly to reliably
> detect GCC? I heard about non-gcc compilers that #define
> __GCC__ to make more code compile.
ITYM mean __GNUC__. If other compilers define that and don't
implement the same feature set then they deserve to break. :)
Something like this would default to label-pointers for gcc but
still allow it to be overridden for platforms that define
__GNUC__ but don't implement label-pointers:
#if defined(LC_INCLUDE)
# include LC_INCLUDE
#else
# if defined(__GNUC__)
# include "lc-addrlabels.h"
# else
# include "lc-switch.h"
# endif
#endif
But, it's not a big deal to have to specify that you want to
use label-pointers in the Makefile.
--
Grant Edwards grante Yow! I'm receiving a coded
at message from EUBIE BLAKE!!
visi.com
