[pulseaudio-discuss] [PATCH 3/4] pactl: Separate stat and info actions

[Maarten Bosmans]

2011/8/4 Colin Guthrie <gmane at colin.guthr.ie>:
> 'Twas brillig, and David Henningsson at 04/08/11 10:40 did gyre and gimble:
>> Sorry for the four months late reply, but now as I'm testing the new
>> PulseAudio builds I realize I have scripts depending on "pactl stat"
>> returning e g the default sink (and just yesterday I told somebody else
>> to do the same!). Also, "pactl info" is not documented in the man page
>> for pactl.
>>
>> For me the best thing would be to let backwards compatibility win this
>> time and revert this patch. What do other people think?
>
> In retrospect, yes, that makes sense. Backwards compat is probably key here.
>
> I think David suggested adding an "info" and a "mem" command and let
> "stat" stay the same? Is that sensible?

Well, the idea of the two separate commands 'stat' and 'info' was to
make those two commands equivalent to
pa_context_stat/pa_context_get_server_info and
PA_COMMAND_STAT/PA_COMMAND_GET_SERVER_INFO. That's much cleaner to
anyone familiar with the API/native protocol. The output of
server_info should never have been in pactl stat anyway.

So how much is backwards compatibility worth to us? I'd say using
scripts/grep to parse pactl output is already quite fragile IMHO. Now
(pre 1.0) seems like the right time to 'fix' pactl and use the most
logical command names.

I'll make sure the manpages get updated and the release notes say
something about the possible breakage. Would that be enough to keep
the change in 1.0?

Maarten