ODP supports C and as such all normal C idioms apply. False is 0 and true is != false. I don't see a need to overspecify beyond that. For ODP routines that return an odp_bool_t they return 1 for true.
On Fri, Nov 28, 2014 at 8:25 AM, Ola Liljedahl <ola.liljed...@linaro.org> wrote: > In C conditional expressions (as used by if, for, while, ? ), any > non-zero value means true so yes you can pass "2" to indicate true. > Using "true" or "TRUE" enums or preprocessor symbols that evaluate to > "1" is just to enhance readability and understanding (at least that's > the hope but then we're having this conversation...). > > "#define TRUE 1" (or the corresponding enum definition) does not mean > that *only* "1" represents true. > > If the ODP specification mentions (specifies) that *only* 1 can be > used as true, then this is wrong in my opinion and should be changed. > > -- Ola > > > On 28 November 2014 at 14:13, Taras Kondratiuk > <taras.kondrat...@linaro.org> wrote: > > On 11/28/2014 02:42 PM, Bill Fischofer wrote: > >> > >> Not really. int has been used for booleans for some time and the 0 = > >> false, 1 = true convention is well established. The purpose of enums is > >> to support remapping. Under no circumstances would one expect to see > >> some platform define ODP_TRUE to be anything other than 1 or ODP_FALSE > >> to be anything other than 0. So in this case an enum is just adding > >> syntactic clutter for no benefit. > > > > > > In this case purpose of enums is not remapping, but value validation. > > > > In different parts of our specification we have true = 1 or true = !0. > > So it is not clear if '2' can be passed as 'true' to a function that had > > odp_bool_t argument. If true is exactly 1, then it would be more > > clear to see ODP_TRUE as input value. >
_______________________________________________ lng-odp mailing list lng-odp@lists.linaro.org http://lists.linaro.org/mailman/listinfo/lng-odp
