On Fri, 2013-09-20 at 17:55 +0530, Arun Raghavan wrote: > On Fri, 2013-09-20 at 15:13 +0300, Tanu Kaskinen wrote: > > On Fri, 2013-09-20 at 17:27 +0530, Arun Raghavan wrote: > > > --- > > > src/pulsecore/hashmap.h | 4 ++++ > > > 1 file changed, 4 insertions(+) > > > > > > diff --git a/src/pulsecore/hashmap.h b/src/pulsecore/hashmap.h > > > index ae030ed..e42732a 100644 > > > --- a/src/pulsecore/hashmap.h > > > +++ b/src/pulsecore/hashmap.h > > > @@ -85,6 +85,10 @@ void* pa_hashmap_last(pa_hashmap *h); > > > #define PA_HASHMAP_FOREACH(e, h, state) \ > > > for ((state) = NULL, (e) = pa_hashmap_iterate((h), &(state), NULL); (e); (e) = pa_hashmap_iterate((h), &(state), NULL)) > > > > > > +/* A macro to ease itration through all key, value pairs */ > > > +#define PA_HASHMAP_FOREACH_KV(k, e, h, state) \ > > > + for ((state) = NULL, (e) = pa_hashmap_iterate((h), &(state), (const void **) &(k)); (e); (e) = pa_hashmap_iterate((h), &(state), (const void **) &(k))) > > > > Are the (const void **) casts really needed? > > Yes, the compiler complains about casts from (const type **) to (const > void **). Ok. Silly compiler. I noticed that this casting is not done anywhere in the current code base, but on closer look it turned out that in every case the original type is const void ** anyway. -- Tanu
