[06:13] <hikiko> hi
[07:48] <pitti> bonjour tout le monde !
[07:51] <flocculant> hey pitti  - what's up with your blog then - HaCkeD By SA3D HaCk3D ... http://www.piware.de/2016/12/last-day-at-canonical/
[07:51] <flocculant> just appeared on planet
[07:53] <didrocks> bonjour pitti ! Le FOSDEM était bien ?
[07:56] <pitti> flocculant: yeah, just saw a warning in my inbox, I guess some server admin will be in order :-(
[07:57] <flocculant> yup ;)
[07:57] <pitti> didrocks: oui, merci ! J'ai rentré très tard, je suis un peu fatigue
[07:57] <pitti> mais c'était grand à nouveau, et ... my talk got quite some nice feedback
[07:58] <didrocks> pitti: oh, talked about cockpit or systemd?
[07:58] <didrocks> talk*
[07:58] <didrocks> ( :( for your server, sorry)
[07:59] <flocculant> morning didrocks :)
[07:59] <pitti> well, I have backups of the mysql db, and will of course change passwords --  but I am running the current wordpress version, not sure how to defend against it
[08:00] <didrocks> hey flocculant ;)
[08:00] <didrocks> oh, it's not even an ancient wordpress version? Urgh, unnice
[08:00] <didrocks> (I did move to static site generation for other reasons, but this is one of them)
[08:02] <seb128> good morning desktopers
[08:02] <seb128> hey pitti flocculant, re didrocks
[08:02] <didrocks> re seb128 !
[08:04] <pitti> bonjour seb128 ! c'était bien à te revoir *accolade*
[08:04] <flocculant> hi seb128
[08:04] <didrocks> pitti: what was your talk about? I'll jump on the video :)
[08:05] <seb128> pitti, salut, oui, dommage que je ne pouvais pas rester samedi soir j'aurais aimé manger avec toi ... c'était une bonne soirée ?
[08:05] <pitti> didrocks: video isn't yet available; it was about how ubuntu does CI, I'm sure you know all about it already :)
[08:06] <didrocks> oh *that* thing :)
[08:06] <pitti> seb128: oui, on a pris le diner avec huit gens dans le centre de ville
[09:00] <willcooke> morning all
[09:03] <Laney> morning!
[09:03] <willcooke> hi Laney, are you really here?
[09:03] <Laney> erm
[09:03] <Laney> physically
[09:03] <willcooke> You're on a swap day though?
[09:04] <pitti> hey Laney -- made it back in one piece?
[09:04] <Laney> willcooke: just checking on some bits, probably fade away in a while
[09:04] <Laney> hey pitti
[09:04] <willcooke> Laney, cool
[09:04] <Laney> there was some delays at midi
[09:05] <Laney> the announcements were blaming british border control /o\
[09:05] <Laney> so our eurostar was ~1h delayed, got back at like midnight
[09:05] <Laney> how about you?
[09:09] <seb128> hey willcooke Laney
[09:09] <Laney> hey seb128
[09:09] <Laney> get back ok?
[09:10] <seb128> yes mostly, had to change train and wait for a connection for like 30min but that was announced, they are work in the line to amsterdam scheduled over saturday/sunday
[09:10] <seb128> they are->there was
[09:11] <seb128> Laney, did the 1h hour delay cascade in making you miss your train in London, etc and end up making you be like 3 hours later than scheduled?
[09:11] <Laney> there was another one soon after we arrived
[09:12] <Laney> so not too much additional delay
[09:12] <seb128> cool
[09:12] <Laney> like 20 people were running from the eurostar
[09:12] <Laney> it was great
[09:12] <seb128> how was the rest of fosdem? fun saturday evening? did you manage to get on bed before midnight? ;-)
[09:13] <Laney> https://www.youtube.com/watch?v=TLbWBlB2aWA
[09:14] <Laney> yeah it was good
[09:14] <Laney> we went to the "usual" thai place
[09:15] <seb128> GNOME beer?
[09:15] <Laney> & sunday was fun, some nice talks & met more people that I didn't even manage to see
[09:15] <Laney> including doko, sladen, etc
[09:16] <Laney> didn't go there in the end, got quite late
[09:16] <Laney> desr_t did though
[09:17] <didrocks> hey willcooke, sad to hear you had delays Laney :(
[09:17] <Laney> hey didrocksssss
[09:20] <Laney> python asyncio is weird
[09:20] <Laney> Exception ignored in: <object repr() failed>
[09:20] <Laney> Traceback (most recent call last): File "/usr/lib/python3.5/asyncio/tasks.py", line 85, in __del__
[09:20] <Laney> AttributeError: 'NoneType' object has no attribute '_PENDING'
[09:26] <pitti> Laney: oh, did you see that talk yesterday, too?
[09:27] <pitti> I found it very interesting, particularly as I know how the usual callback-style "reverse code order" looks like in things like udisks
[09:29] <Laney> pitti: no - I didn't actually know there was one on this stuff
[09:29] <Laney> I must be playing with it independently
[09:29] <Laney> do you have a link?
[09:30] <pitti> Laney: https://fosdem.org/2017/schedule/event/python_coroutines/
[09:30] <pitti> Laney: no video yet
[09:30] <pitti> Laney: hah, what a coincidence then
[09:31] <Laney> awesome!
[09:32] <Laney> apart from this exception my thing works, but I'm sure it's not idiomatic
[09:48] <davmor2> Morning all
[09:49] <davmor2> Laney: good weekend?
[09:58] <Laney> hey davmor2
[09:59] <Laney> very good thanks, nice to see everyone
[09:59] <Laney> you should come one time
[10:01] <seb128> hey davmor2
[10:02] <davmor2> Laney: pfff I'd wind up breaking it you know me :)
[10:02] <Laney> you might break yourself :-o
[12:24] <davmor2> Laney: only if I go rockclimbing apparently ;)
[14:46] <ricotz> Sweet5hark, could you create series specific ppa for 5.3?
[15:32] <flexiondotorg> Morning
[15:32] <flexiondotorg> Had some connectivity issues earlier and my IRC bouncer ate itself.
[15:36] <seb128> hey flexiondotorg
[15:36] <flexiondotorg> seb128 o/
[16:04] <qengho> flexiondotorg: Arm support in Chromium is a bit wonky because of our constraint of Hard Float. Do you think the R F would be interested in helping sort out the run-time conditional Neon support? Upstream assumes ARM v7 == has Neon unconditionally, which I have had to hack out in many places.
[16:06] <qengho> And run-time conditional scares me a bit because I have heard of hardware that switches between CPUs with different capabilities and instruction-sets to save power.
[16:07] <flexiondotorg> qengho Thay might be.
[16:07] <flexiondotorg> Just being able to tell them this is much futher forward than they are right now.
[16:07] <flexiondotorg> They have no upstream insight.
[16:07] <flexiondotorg> And thank for taking a look :-)
[16:07] <flexiondotorg> Indeed.
[16:07] <qengho> flexiondotorg: My pleasure. It's pretty good code so far.
[16:08] <flexiondotorg> I've built other stuff for R F and had disable Neon and (I think) VFP.
[16:08] <flexiondotorg> qengho Does it need spliting up?
[16:09] <flexiondotorg> For example memcpy() optimisations in one patch and mmal in other etc?
[16:09] <qengho> flexiondotorg: Yes, it will. I suggested quilt patchsets in email, if they want guidance on how.
[16:10] <flexiondotorg> I think they would appreciate the help.
[16:10] <flexiondotorg> Are you able to help (sponsor in some fashion) get the patches accepted?
[16:10] <qengho> flexiondotorg: Yes.
[16:11] <flexiondotorg> Excellent. I thought that was the case :-)
[16:11] <qengho> flexiondotorg: I don't have a magic wand, but I know how.
[16:11] <flexiondotorg> Cool.
[16:11] <flexiondotorg> And I take it monolithic patches won't fly?
[16:12] <qengho> Yeah, the explanation and justification for making code larger and likely more buggy has to be pretty convincing, with no surprises or extraneous code.
[16:14] <qengho> Small patches that do one coherent thing are best.
[16:15] <qengho> Maybe not all will go in. One poison pill might ruin the whole patch. Best to be severable.
[16:19] <flexiondotorg> I've replied to your mail. Do you have any objection with being put in contact with the engineer at R F to get them up to speed on preparing quilt patch sets?
[16:19] <qengho> Not at all.
[16:19] <flexiondotorg> Cheers.
[16:55] <dobey> flexiondotorg: hi! how does MATE start indicators? does it use upstart/systemd, or is it relying on xdg autostart facilities?
[17:28] <flexiondotorg> dobey Via mate-indicator-applet which is a mate-panel applet.
[17:29] <flexiondotorg> Indicators are actually disabled by default.
[17:29] <flexiondotorg> MATE Tweak has an option to enable them, which just add mate-indicator-applet to the panel and reloads the panel.
[17:29] <flexiondotorg> qengho I've got engineer details from R F.
[17:30] <dobey> flexiondotorg: ok, so getting rid of the old xdg autostart in indicators shouldn't break mate image (i see some indicators are seeded there)?
[17:31] <flexiondotorg> dobey I'll have to refresh my memory to be sure, it's been a while.
[17:31] <flexiondotorg> I'll double check tonight.
[17:32] <flexiondotorg> But my initial feel this will actually benefit MATE :-)
[17:33] <flexiondotorg> Which indicators are you going to change BTW?
[17:33] <dobey> flexiondotorg: ok, that'd be great, thanks. i'm doing a quick sweep through indicators to try and get rid of some of this legacy stuff
[17:33] <flexiondotorg> I've got a test VM I can test with later.
[17:34] <dobey> flexiondotorg: all of them, but session, sound, power, messages, and indicator-application seem to be seeded in ubuntu-mate.
[17:34] <flexiondotorg> This is what Ubuntu MATE currently has in /etc/xdg/autostart
[17:35] <flexiondotorg> indicator-application, indicator-messages, indicator-sound
[17:35] <flocculant> dobey: I saw you mentioon xfce about this a few days ago - I can't answer the in's and out's I'm afraid - but I do know we had a bug a while back in the zesty cycle that might shed light > bug 1648889
[17:36] <flocculant> and copy what flexiondotorg said re /etc/xdg/autostart
[17:37] <dobey> flocculant: ah. that seems to suggest that xnox added a systemd target for xfce4 indicators.
[17:38] <dobey> and looks like it was using upstart before
[17:38] <xnox> flocculant, probably mate-indicator-applet needs a tweak similar to xfce4
[17:38] <xnox> indeed, no more upstart for indicators in zesty
[17:39] <dobey> xnox: do you know if any of the seeds are using xdg autostart for indicators? some are in kubuntu, lubuntu, budgie, and gnome, too
[17:39] <flexiondotorg> dobey I njust move indicator-* out of /etc/xdg/autostart and added to mate-indicator-applet.
[17:40] <flexiondotorg> Log out, log in and all indicators are present and correct.
[17:40] <dobey> flexiondotorg: ok great. thanks
[17:40] <flexiondotorg> So non-issue for MATE.
[17:41] <flocculant> dobey: hope that helps :)
[17:41] <flocculant> xnox: thanks
[17:41] <dobey> yep, thanks :)
[17:50] <flocculant> don't break it or I will whine :p
[17:55] <jbicha> dobey: ubuntu-gnome doesn't use indicators, libappindicator3-1 recommends indicator-application which is why it ends up in the -gnome seed
[17:56] <jbicha> for budgie, ask fossfreedom (he's in #ubuntu-devel )
[18:49] <jc44> Hi - John Cox here, flexiondotorg suggested that I try to talk to qengho about my Chromium patches here
[18:49] <qengho> jc44: Welcome.
[18:50] <qengho> jc44: I'm reading a single large patch now, so I know what's in it and what to suggest.
[18:51] <jc44> Good Oh - I was just about to ask that
[18:52] <jc44> Much of the patch is interrelated, but it does split into a few pieces
[18:52] <jc44> There's a bit that does decode & resize using the Pi h/w
[18:53] <jc44> There's a slew of patches that attempt to get the ARGB image created by the h/w to display without copying it
[18:54] <jc44> There's a patch to use shm to copy the image to the screen rather than putimage
[18:54] <qengho> jc44: Awesome!
[18:54] <qengho> All great.
[18:54] <jc44> And there are a few patches to get armv6 compiles to work
[18:54] <jc44> (and aauto select armv7 where appropriate)
[18:55] <jc44> and a patch or two so I get version info on my patch out of the image so I know what someone is running
[18:56] <jc44> None of these are currently controled by gn flags as I really wasn't expecting upstream to want them
[18:57] <qengho> Cool. Upstream has different people in charge of distinct directories of the source tree. One person is unlikely to know enough about other parts enough to say "this looks good enough". So, patches that are small as possible have a chance of success.
[18:58] <qengho> jc44: I can help with GN.
[18:59] <jc44> 1st off - do you have any clue if upstream would want he Pi-only h/w decode?  If not then much of teh reset becomes academic
[18:59] <jc44> That was meant to be "the rest"
[19:01] <qengho> jc44: I'm pretty sure they'd take it. RPi is not obscure. Half the people in that project own one or two Pis, I'm sure.
[19:04] <qengho> jc44: You'll also have to sign their agreement. http://dev.chromium.org/developers/contributing-code#TOC-Legal-stuff
[19:05] <jc44> There are a couple of bits that are more hacky than other bits - the decode itself is (I think) quite presentable, but the side channel I use for telling it what image size is wanted is really quite nasty...
[19:08] <jc44> I probably can't sign the agreement myself - I'm working for RPi so I imagine that they have to be the ones to do that
[19:10] <qengho> jc44: Right.
[19:14] <jc44> qengho: Any clues on how to start this? If you've got my patch in front of you I can help your understanding of the pieces
[19:16] <qengho> jc44: Well, first thing I'm going to do is try to split it into discrete pieces. Let's talk about those pieces in email.
[19:17] <jc44> Sure - I'm going to have to go home quite soon anyway (in UK)
[19:18] <qengho> jc44: Okay.
[19:50] <flexiondotorg> jc44 Hi
[19:50] <flexiondotorg> I was in a meeting.
[19:53] <flexiondotorg> qengho Thanks for that ^ :-)
[20:00] <willcooke> morning robert_ancell
[20:01] <robert_ancell> willcooke, hi
[20:14] <flexiondotorg> robert_ancell o/
[20:14] <flexiondotorg> willcooke You still alive?
[20:14] <willcooke> flexiondotorg, otp
[20:19] <willcooke> flexiondotorg, done
[20:19] <robert_ancell> flexiondotorg, hi!
[20:19] <flexiondotorg> willcooke I was just asking after your well being :-)
[20:20] <willcooke> flexiondotorg, that's very nice :)  I am indeed alive, feeling better actually
[20:20] <willcooke> (that this morning)
[20:20] <willcooke> *than
[20:20] <flexiondotorg> Good :-)
[20:30] <willcooke> right ho, night all
[20:51] <flocculant> robert_ancell: hi there - have you had chance to look further at the locking issue? bug 1656399
[20:53] <robert_ancell> flocculant, I haven't had time to look into more detail, but it's on my list of things to fix before 17.04
[20:54] <robert_ancell> Unless someone else wants to fix it :)
[20:54] <flocculant> robert_ancell: ok and yea I understand - only wish I could :)