On 19.04.2017 09:36, Manuel Amador (Rudd-O) wrote:
> On 04/19/2017 02:36 AM, Jungsup Lee wrote:
>> Sink's "media.name" property is mostly NULL, therefore
>> using sink's name instead of sink's "media.name" property.
>> ---
>> src/modules/module-allow-passthrough.c | 5 +----
>> 1 file changed, 1 insertion(+), 4 deletions(-)
>>
>> diff --git a/src/modules/module-allow-passthrough.c b/src/modules/module-allow-passthrough.c
>> index 31ff270..856c6c6 100644
>> --- a/src/modules/module-allow-passthrough.c
>> +++ b/src/modules/module-allow-passthrough.c
>> @@ -59,7 +59,6 @@ static pa_sink *ensure_null_sink_for_sink(struct userdata *u, pa_sink *s, pa_cor
>> pa_module *m;
>> pa_sink *sink;
>> uint32_t idx;
>> - const char *name;
>>
>> sink = pa_hashmap_get(u->null_sinks, s);
>> if (sink != NULL) {
>> @@ -67,10 +66,8 @@ static pa_sink *ensure_null_sink_for_sink(struct userdata *u, pa_sink *s, pa_cor
>> return sink;
>> }
>>
>> - name = pa_proplist_gets(s->proplist, PA_PROP_MEDIA_NAME);
>> -
>> t = pa_sprintf_malloc("sink_name=allow_passthrough_null_%s sink_properties='device.description=\"%s\"'",
>> - name ? name : "", _("Dummy Output"));
>> + s->name, _("Dummy Output"));
>> m = pa_module_load(c, "module-null-sink", t);
>> pa_xfree(t);
>>
> Nono, please, do not remove this option. Sometimes we have programs
> (like mixing boards) that load null sinks and *need* to name them so the
> end user knows to distinguish them from actual null sinks that get
> loaded by default when there's no output. Do not remove features just
> because *you* do not see the use case.
>
I don't really understand this comment. There is no option removed, the
null-sink for module-allow-passthrough is just named differently with the
sink name as ending instead of the media name. Where is the problem
with that?
