Daniel Chen <seven.steps at gmail.com> writes: > [I don't know whether this was moderated or whether you're a > subscriber, so I added you.] Reading through gmane. > On Mon, Sep 27, 2010 at 7:52 AM, David Kastrup <dak at gnu.org> wrote: >> ? ?amixer -D hw:DSP cget numid=23 >> >> now displays >> >> numid=23,iface=CARD,name='Precise Pointer' >> ?; type=BOOLEAN,access=rw------,values=1 >> ?: values=on >> >> where the default is 'off'. ?Which does not appear to disturb Jack. ?But >> pulseaudio most certainly does not seem to like it, with tsched=1 >> suffering much worse than tsched=0. >> >> Setting this to 'On' makes pulseaudio work worse with tsched=0, and >> apparently have no problems with tsched=1. > > There are two places where this control can be set (i.e., the > "default" changed), via alsactl's init db or in the driver itself. > Because Maverick is in deep freeze, and because Lucid is in an even > deeper freeze (having been released months ago), the best thing to do > at this point is to file a bug using "ubuntu-bug alsa-base" so that > the appropriate mixer information can be uploaded to Launchpad, and a > patch can be generated for one of the above contexts. Personally, the > former makes more sense and is less likely to cause regressions in the > short term. The main thing to test then becomes confirming that jackd > works correctly with that setting. After "sufficient" testing, the > driver could be updated. Well, I have no idea how to do this right. One obvious snag is that a setting like "Precise Pointer" is likely there for a reason. Why would anyone ask for "imprecise pointers" (the default) without there being an actual advantage? I have not yet found a rationale why this setting is there in the first place. I can't just believe there is a "shoot yourself in the foot" variable in a driver with a default of "yes". I also have no idea about Maverick at all, and how it would cater for the Hammerfall with regard to Pulseaudio, defaults and so on. Would it be considered useful if I made the jump to Maverick prerelease right now? Who will then be doing the testing for a putative backport? -- David Kastrup
