=== asac_ is now known as asac [08:16] hi [09:11] fta: whats the status on fonts? [09:11] is that fixed? [10:00] asac, i doubt it [12:02] fta: if you give me bug id i would like to reraise this ;) [12:16] asac, bug 305394, mozilla bug 458612 [12:16] Launchpad bug 305394 in fontconfig "No subpixel smoothing" [Medium,Confirmed] https://launchpad.net/bugs/305394 [12:16] Mozilla bug 458612 in GFX: Thebes "(Ubuntu) system fontconfig settings override GNOME font rendering screen settings" [Normal,Resolved: invalid] http://bugzilla.mozilla.org/show_bug.cgi?id=458612 [14:59] mconnor: awake/on? [15:25] asac, so, the font bug? [15:38] fta2: i targetted it as RC for jaunty now ... will try to raise that tomorrow when i have a minute again [16:11] asac: for some value of awake, yes [16:15] mconnor: access to email? [16:15] ;) [16:16] asac: always [16:16] (for better or for worse) [16:16] mconnor: i talked to our usability folks a bout the complains you raised and asked them to give me the real concept they want to go for ... i will send it to you [16:16] mconnor: can you take a quick look and tell what you think? [16:16] sure [16:18] mconnor: ok sent [16:19] asac: so, why the hell is this private? [16:19] mconnor: what do you mean? [16:19] This is a private draft. When complete, we will publish it and propose it [16:19] for integration into the Gnome Human Interface Guidelines and KDE4 Human [16:19] Interface Guidelines. [16:20] hmm [16:20] seems pretty Microsoft-ish [16:20] hmmm .... i asked them and they said it wasnt really private [16:20] let me check [16:20] please do [16:21] because I'd like to stick this on a wiki and link people to it, or something [16:22] I mean, a lotof this is pure assertion so far [16:22] I get their logic, I just think it's wrong [16:23] I mean, if I have a transient notification telling me stuff's done, then I know it's done, and I don't have to switch windows/desktops/whatever to find out that info [16:24] if I have to switch to the window to discover that yes, in fact, the download is finished, and not that it had an error or some other reason that window needs attention [16:24] hmm [16:24] then I've just been more forcefully switched out of my context [16:25] they say its not confidential ... its private in the sense that it hasnt been published yet. i will ask about wiki stuff now [16:25] I mean, the principle was "if I miss the notification, it's not a big deal" [16:25] if we use a WM call to get attention, then I will need to click on that window sooner or later to find out why it's asking for attention [16:26] that's making that notification _more_ imperative, no? [16:27] asac: even better would be a channel where I could talk to them about this [16:28] asac: I mean, it's a fundamental disagreement of what's more intrusive [16:29] it really seems like they're trying to reinvent the wheel here, but their rationale is weak [16:29] IMO [16:35] mconnor: ok ... we are working on getting this public to get input from multiple sides. i will paste the url to you once its somewhere [16:35] asac: ok, thanks [16:35] mconnor: do you have user-experience folks at mozilla that also could take a look? [16:35] yeah, we have at least five across various groups [16:36] as in, people who actually do UX full time [16:39] not counting the people who've done other design stuff as a secondary role [16:39] mconnor: sigh ... so it will still take a week; they seem to have promissed to key-community contributors to get their feedback first as it seems :/ [16:40] ah, politics [16:40] yes ... i think some community contributors dont like parts of it ;) [16:40] Open Source: You're Doing It Wrong [16:40] so maybe in a week it will look different ;) [16:41] Rheeeeeeeeet! [16:41] Rheeeeeeeeeeet! [16:41] mconnor: i agree ... but well. [16:41] <[reed]> it's mpt! [16:41] hi [reed], hi mconnor [16:41] <[reed]> mconnor: this is Linux... it's "free software" instead, didn'tyaknow? [16:41] mpt: are you behind this well-meaning travesty of wheel reinvention? [16:41] <[reed]> :p [16:41] mconnor, I have my fingers in every spoke [16:42] It's quite painful [16:42] I can imagine [16:42] and it's not even public yet! [16:42] yes [16:42] We are trying to minimize the whine factor [16:43] my flippant recommendation is "just duplicate Growl" fwiw ;) [16:43] and we think we can do that best by getting Top Men and Top Women to comment on it for a week before making it more public [16:43] You being among the Top Men, of course [16:44] King ;) [16:44] haha [16:45] mpt: so, my basic assertion is that, as I said above, Growl gets this right. There's no queue, there's no permanence. You get information that may/may not be something you want to act on (to use the examples from the email, an IM in the background, or wireless disconnection, or something [16:46] by the very act of showing me information, you're presenting me with a choice [16:46] but a choice I am equipped to make, because I have the basic context already [16:47] if you use a WM_NOTIFY or whatever, all I know is that my taskbar is saying "click on this window" and I have a harder choice [16:48] and it's a bigger interrupt than just glancing at the corner of my window [16:48] mconnor, I use Growl on my Mac, and it's nice most of the time [16:48] also, in most WMs I've used, it's a permanent state, so it will continue to demand attention [16:48] mconnor, but when it goes wrong (i.e. puts up a bubble where I was about to click something), it goes wrong badly [16:49] I end up having to click three times instead of once [16:49] I think that's fixable [16:49] by not showing notifications when the mouse is there [16:51] the problem is, I think that the usability win from being able to click the notification and getting immediately to the context I care about is [16:51] much greater than the annoyance in accidentally clicking [16:55] mpt: basically, the "one click to get to my downloads, regardless of what I'm doing" bit is pretty good, and better than "switch to app, open correct window" [16:55] So the "usability win" = the greater context provided by a bubble vs. a taskbar button [16:55] and the direct win in being able to jump directly there [16:55] mental model time: [16:55] That's no different, clicking on the taskbar button will take you directly there too [16:56] on platforms with taskbar button per window, yeah [16:56] not Win 7 or OS X, of course [16:56] yes [16:57] but the question is whether I want to do that [16:57] OS X doesn't have a notification bubble system built in in the first place :-) [16:57] yeah, instead they flash the dock, and you have to break your context even more to find out what it's bouncing the damn dock icon for [16:57] (despite Growl deliberately being licensed so that Apple could incorporate it if they wanted) [16:57] yes [16:58] that's my problem, I don't want a permanent notification that is basically saying "You should click me" [16:58] I want to say "oh hi, your stuff is done" in a transient way [16:59] if you want to go to that window, you can click the notify button [16:59] if you don't, you ignore it, and you aren't left with something annoying you, or having to break stride to find out why Firefox is bugging you [17:00] One issue with that is, it imposes a rather short and rather arbitrary time limit [17:00] it does, yes [17:00] If you click the box within X seconds you're fine, but if you reach it in X.01 seconds, sorry, you're clicking something else [17:01] well, I like growl's fade thing there [17:01] it'll unfade if you focus it during the slow fade out [17:01] and remain as long as you have it hovered with the mouse [17:02] which makes it fairly hard to accidentally click on whatever's behind that area, unless you're trying for FPS headshot quick twitch clicking [17:03] but that's the thing, it's not for something critical, it's primarily informational [17:03] you can still take action on that information without the callback, but it's just easier this way [17:04] and we only use these notifications as a "by the way, X, if you want to do somthing with it" [17:04] One of the problems we're trying to curtail is that application developers *have* been using it for critical things [17:05] e.g. http://cyberelk.net/tim/2009/02/03/screenshots-system-config-printer-11/ [17:05] hmm [17:06] that does seem rather odd... [17:06] Golden Hammer much? [17:07] I mean, the "printer added" one, minus the widgets, seems fine [17:07] I like that on Windows, "driver found and installed for your thing, kthxbye" [17:07] but, uh, auth required? [17:07] wtf [17:08] There's an explanation in the last comment (probably fixable another way, but anyway) [17:08] My favorite other example is http://xgoat.com/wp/wp-content/oops.png [17:08] ow [17:08] OW [17:09] so this is the clueless developer replacment for modal dialogs, or something? [17:09] pretty much [17:09] * mconnor sighs [17:09] Just like modal dialogs, but with added yellow! [17:10] The battery in your computer may have been recalled by %s and you may be at risk. For more information visit the %s battery recall website. ( Visit recall website ) ( Do not show me this again ) [17:10] these, yes, you need to use the WM notify bit [17:10] Actually "Do not show me this again" is probably the single most common button in the notification bubbles we've reviewed [17:10] because the user _has_ to do something with it [17:10] that, to me, is the key here [17:11] none of the stuff we use notifications are situations where we _require_ interaction [17:11] "stuff is done" "things are available to be updated, but that'll happen on next start anyway" [17:12] which is why we're not more aggressive [17:13] Ubuntu patches out the "things are available to be updated" case, so I haven't been able to see it yet [17:14] What happens when you click on it? [17:14] ubuntu disables the app updater, but not the extensions case [17:14] asac showed me that extensions are handled by the big wizardy window at launch [17:15] in the add-ons case, it opens the add-ons manager and focuses the updates tab [17:15] yeah, that window needs to die, sigh [17:16] that's if you don't respond to the notification, of course [17:16] if you do, you won't see it on the restart [17:16] iirc [17:16] how frequently do you check for extension updates if you didnt find any on startup? [17:16] we don't check on startup [17:16] we check daily [17:17] and we apply updates on next startup [17:17] same with software update for non distro packages [17:17] So the parallel of our Downloads proposal would be for the Add-Ons window to open, focus the Updates tab, and request attention [17:17] right [17:18] but that, to me, is more demanding and insistent [17:18] I'm going to have to switch to that window at some point to find out what Firefox wants [17:19] true [17:20] I'm very sorry about this, but there's a team expedition to a drowned spaceship and we're leaving shortly [17:20] and since that can also mean "download failed to complete" I need to click on it every time [17:20] I'll be glad to continue this discussion tomorrow if you like [17:20] dude, spaceship > UX [17:20] (disclaimer: not *really* a spaceship) [17:20] where are you based anyway? [17:20] NZ still? [17:20] Based in London, but working in Berlin this week [17:20] hmm [17:20] coming to FOSDEM perchance? [17:23] possibly, now that you suggest it [17:23] anyway, must dash now [17:23] ok [17:23] ttfn [20:43] fta: (RE: identi.ca notice) yeah, driver issue, may be fixed in 1.0.19 (but not likely) [20:43] fta: have you tried using tsched=0 ? [20:43] no, where should i put that? [20:44] /etc/pulse/default.pa : load-module module-hal-detect tsched=0 [20:44] should be line 61 [20:44] i just lost sound while playing openarena (while waiting for something to complete), yet p-a is still alive [20:44] the pa daemon being alive is not a good indicator of anything, really [20:45] unfortunately, one can't query internal sink/source state in 0.9.14 [20:45] at least, it's not an abort() that makes all other apps freeze or die [20:46] fta: Any progress on o-k? :p [20:46] piratenaapje, none [20:46] e.g., it will still be alive via `pgrep' but won't actually have grabbed anything in /dev/snd/* via libasound2 [20:46] Didn't feed a MOTU to take it over? [20:46] yeah, at least there's a null sink always loaded now [20:46] euh find [20:47] piratenaapje, been busy with work [20:47] ah alright [20:54] hi all. Just saw the last changelog entry ("we're using nspluginwrapper on amd64") of the flashplugin-nonfree package in jaunty and was wondering ...; Any plans to include the 64-bit plugin provided by Adobe into jaunty? [20:55] petski: if and only if it's released by Adobe as stable [20:56] their redistribution terms prohibit anything not stable [20:56] any place/PPA where I can already test the 64-bit version of lashplugin-nonfree [20:56] sure, a number of community member's ppas have varying packages [20:56] most of them don't bother to track the changes in jaunty [20:56] members' [21:01] Kees Cook (hehe, not "just a member"): Support alpha native 64bit flash plugin [21:01] he made that change in 10.0.15.3ubuntu1 [21:01] it was reverted in 10.0.15.3ubuntu2 [21:02] The 64-bit Flash Player 10 alpha refresh for Linux was released on December 16, 2008. .... Kees package was born on Mon, 17 Nov 2008 08:44:55 -0800 [21:04] ...and? [21:05] Kees' package doesn't include Adobe's "refresh" [21:07] i'm sure you can find other packages that have updated sha1sums/md5sums [21:07] barring that, you could just download from labs.a.c directly and slap the shared lib into ~/.mozilla/plugins/ [21:08] was just planning to do so, thank for the help [21:34] E: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there was actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the ALSA developers. We were woken up with POLLOUT set -- however a subsequent snd_pcm_avail_update() returned 0. [21:34] dtchen, ^^ [21:35] fta: ok, good; not a pa issue then - much as i suspected [21:35] sigh, so many deep stack magic tricks required [21:36] i just unpaused rhythmbox [21:37] how come i never see these bugs with conexant hw? ;) [21:37] 00:1b.0 Audio device: Intel Corporation 82801G (ICH7 Family) High Definition Audio Controller (rev 01) [21:37] Subsystem: Giga-byte Technology Device a002 [21:38] hoorah for realtek :/ [21:39] to be honest, realtek hda hw is actually more featureful than conexant hda hw but clearly is ... suboptimal in the driver [21:42] dtchen, it's a quite cheap gigabyte motherboard, i didn't care much when i bought it, i just wanted something with 2 eth ports, a socket for a dual core 2, 4 mem slots, and any 5.1+ integrated sound chip [21:44] i suppose we could try adjusting bdl_pos_adj for you [21:47] fta: do you still have output from your alsa-info.sh pastebinned? [21:48] hm, i can dig into my logs [21:52] dtchen, here is a fresh one: http://paste.ubuntu.com/113793/ [21:52] Pulseaudio: [21:52] Installed - Yes (/usr/bin/pulseaudio) [21:52] Running - No [21:52] i doubt it :) [21:53] yeah, i'll push an update to travis/takashi [21:54] quite a few of those checks are inane [21:55] see, code like this just warms my heart :/ [21:55] pos_adj = (pos_adj * runtime->rate + 47999) / 48000; [21:58] ok, please try adjusting [21:58] bdl_pos_adj [21:59] where? [21:59] intel hda defaults to 1; you need to increase it by powers of 2 until the symptom is relieved [21:59] one method is to kill all apps using audio, then rmmod snd-hda-intel and modprobe snd-hda-intel bdl_pos_adj=2 [22:00] you could also place an options line in /etc/modprobe.d/ [22:00] e.g., options snd-hda-intel bdl_pos_adj=2 [22:00] effected on modprobe [22:00] any value beyond 2**5 is a real problem, though [22:09] ok, will try but i don't want to kill or reboot now, i'm working on something. I will let you know. [23:38] mozilla bug 461277 [23:38] Mozilla bug 461277 in Embedding: GTK Widget "use G_TYPE instead of deprecated GTK_ macros in gtkmozembed.h" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=461277 [23:55] asac, [reed]: the patch in mozilla bug 461277 got a r+ but is not committed, yet, someone else committed a part (1/3) of it so it's weird now, http://hg.mozilla.org/mozilla-central/diff/ce7f39495675/embedding/browser/gtk/src/gtkmozembed.h [23:55] Mozilla bug 461277 in Embedding: GTK Widget "use G_TYPE instead of deprecated GTK_ macros in gtkmozembed.h" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=461277 [23:56] asac, i mean, your patch is now 1/3 obsolete [23:56] mozilla bug 474116 [23:56] Mozilla bug 474116 in Widget: Gtk "Clean up deprecated GTK symbols" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=474116 [23:58] too bad for your sr+