Re: [Jack-Devel] jack1 version jack2 public headers comparison

Paul Davis <[hidden]> writes:

> On Tue, Feb 28, 2012 at 9:26 AM, Nedko Arnaudov <[hidden]> wrote:
>
>> jack_get_version_string() is also required for reporting jack server
>> version over dbus. I want this information to be present somewhere in
>> gladish (the GUI frontend bundled in the ladish tarball).
>
> this just defers the question: what is the purpose of presenting
> anyone with this version number?

Diagnostics. Why its reported to stdout early in jackd code? Its the
same.

>>>> jack_port_type_id_t jack_port_type_id (const jack_port_t *port) JACK_OPTIONAL_WEAK_EXPORT;
>>>
>>> needed for?
>>>
>>>> typedef uint32_t jack_port_type_id_t;
>>>
>>> use?
>>
>> jackdbus uses it to report [integer] port types over dbus.
>
> why integer and not string as within the JACK API itself?

I cant recall why we decided like this with iuso back then. I can guess
that one reason could be that comparing integers is faster. Also its
more bandwidth efficient. There are jackdbus users that work with huge
number of ports. And it has to stay like this in dbus patchbay interface
for compatibility reasons. If we are going to remove integer port types
From the C API, then jackdbus will have to either do string compares on
each port creation event or we should guarantee that jack_port_type()
returns unique string pointers that can be compared as integers.

-- 
Nedko Arnaudov <GnuPG KeyID: 5D1B58ED>

1330441030.3140_0.ltw:2,a <874nubdn3s.fsf at arnaudov dot name>