Monday, 2012-06-25

*** RhymeswA has joined #mer00:04
*** BeholdMyGlory has quit IRC00:07
*** Behold has joined #mer00:07
*** himamura has joined #mer00:23
*** M4rtinK has joined #mer00:58
*** M4rtinK has quit IRC01:12
*** harha has joined #mer01:16
*** tilgovi has joined #mer01:19
*** Openfree` has joined #mer01:25
*** phdeswer has quit IRC01:26
*** tilgovi has quit IRC01:33
*** swer has joined #mer01:36
*** slx has quit IRC01:40
*** Milhouse has quit IRC01:50
*** himamura_ has joined #mer02:01
*** himamura has quit IRC02:04
*** harha has quit IRC02:07
*** dijenerate has joined #mer02:39
*** shrikrishna has joined #mer02:56
*** Openfree` has quit IRC03:03
*** Estel_ has joined #mer03:06
*** Openfree` has joined #mer03:11
*** Estel_ has quit IRC03:30
*** himamura_ has quit IRC03:34
*** himamura has joined #mer03:35
*** Estel_ has joined #mer03:43
*** clopez has quit IRC03:43
*** furikku has joined #mer03:57
iekkumorning04:00
dm8tbrmoaning04:16
*** gimli_ has joined #mer04:23
*** harha has joined #mer04:39
*** harha has joined #mer04:40
*** rdqfdx has joined #mer04:40
Sage_morning04:49
*** harha has quit IRC04:50
iekkuhi sage04:52
Sage_lbt: when is the Mer release going to happen btw?04:53
Sage_https://bugs.nemomobile.org/show_bug.cgi?id=202 is something that would be nice to have fixed in next release. Need more info though.04:54
StskeepsSage_: nemo images needed to verify quality05:12
Stskeepsand i've been away too05:12
Sage_Stskeeps: well that feature isn't used in nemo images yet thus nobody has noted it earlier AFAIK.05:13
Stskeepsi didnt see bug yet05:15
Stskeepsbut we need those imges05:15
Sage_I meant that even if I do images we dont' see the bug in the image as the calendar isn't in nemo repos but only in their devel tree.05:17
Sage_the calendar that shows the bug that is05:17
Sage_I'll do images today when I get "few" things sorted out. :P05:17
*** lamikr has joined #mer05:20
Stskeepsok05:20
* Stskeeps just woke up and is extremely trashed05:21
iekkui hope i wake at some point05:22
* Sage_ feels better when hearing that he is not the only one :)05:22
*** harha has joined #mer05:23
iekkui had alarm set at 7:00. woke up about 5 times for wife's alarm and 6:10 he was yelling downstairs that he can't find he's wallet...05:24
w00tStskeeps: got my message?05:41
*** pomm is now known as borax05:41
Stskeepsw00t: whatmessageneedcoffee05:41
E-PMorning05:42
dm8tbrStskeeps: how was your flight? :)05:42
Stskeepsmorn E-P :)05:42
Stskeepsdm8tbr: bumpy, strong tail wind05:42
dm8tbric, did that make the kids shut up? :)05:42
Stskeepsit certainly added for a nice noise filter05:43
dm8tbr:]05:43
w00t06/24@21:23:15 <#mer/w00t> Stskeeps: i fixed the y-flip bug, i have no idea how05:43
w00t06/24@21:23:20 <#mer/w00t> Stskeeps: also, we require no qml patch05:43
w00t06/24@21:23:49 <#mer/w00t> Stskeeps: install qt5-qtqml-import-window205:43
Stskeepsw00t: ok, that solves it for you?05:43
Stskeepsthe installing05:44
w00thad 4 minimer demos running fine05:44
w00tin qml-compositor05:44
Stskeepsok05:45
Stskeepswithout qml patching?05:45
Stskeepsor with05:45
w00twithout05:45
w00tthe patching was caused by the missing package05:45
w00ti was just an idiot and didn't realise the error05:46
w00t:P05:46
Stskeepsokay, so a bare qt-qtwayland package, works with qt5-qtqml-import-window2 ?05:47
w00tyes05:48
w00ti guess if you package qml example seperate, then you could add a dependency and it'd work05:48
Stskeepsok05:48
Stskeepsk05:48
Stskeepsgood05:52
Stskeepsi wonder why that is though05:53
w00twhy it is? because the QML uses that import05:54
Stskeepsok05:54
Stskeepsi really wonder at times if rpm should autogenerate qml dependencies05:55
Stskeepser, rpmbuild05:55
* w00t packs the laptop up05:56
w00tdownstairs for breakfast, and then airport time05:57
w00to/05:57
Stskeepsenjoy05:58
*** Siosm has joined #mer06:04
*** pohly has joined #mer06:04
*** harha has quit IRC06:11
*** niqt has joined #mer06:23
*** beford has quit IRC06:26
*** harha has joined #mer06:28
*** gimli_ has quit IRC06:33
*** vilpan has joined #mer06:34
*** harha has quit IRC06:39
*** rcg has joined #mer06:49
*** Venemo_N9 has joined #mer06:49
Venemo_N9good morning :)06:50
iekkumorning06:50
E-Pmorning06:51
*** sonach has joined #mer06:51
rcgmorning06:52
*** _ol has quit IRC06:54
*** eebrah has joined #mer06:56
*** sonach_ has joined #mer06:58
*** sonach has left #mer07:00
*** sonach_ is now known as sonach07:00
*** niqt has quit IRC07:02
*** yunta has joined #mer07:04
*** Venemo_N9 has quit IRC07:09
*** stefan_schmidt_w has joined #mer07:12
*** phaeron_n9 has joined #mer07:13
*** harbaum has joined #mer07:15
*** harha has joined #mer07:16
*** harha has quit IRC07:17
*** otep has quit IRC07:31
*** lamikr has quit IRC07:32
*** otep has joined #mer07:33
rcgSage_: I filed a new bug for mer. Should I close the initial bug for nemo as "resolved" or as "triagedupstream"?07:34
rcgI also added a somewhat revised version of the patch07:35
Sage_resolve that bug somehow :)07:36
rcgright, done :)07:37
*** fvennetier has quit IRC07:37
*** harha has joined #mer07:40
*** harha has quit IRC07:46
*** ssirkia has joined #mer07:55
*** harha has joined #mer08:02
*** slaine has joined #mer08:05
Stskeepsmorn slaine o/08:05
slaineMorning Stskeeps08:08
lbtmorning all08:09
rcgmorning lbt08:09
Stskeepsmorn lbt08:09
rcgSage_: http://pastebin.com/bhLAv9qK08:09
* Stskeeps is finally back at his desk, missed it very much.08:09
rcgwb Stskeeps :)08:10
*** drussell has quit IRC08:10
iekkuStskeeps, hope you don't need to travel soon again08:10
sonachlbt: morning! I am re-running your scripts :)08:11
rcgSage_: i think that's what you meant with having the return in the cpp file? I also think that this gives the meaning away much better08:11
lbtsonach: good :)08:11
*** shrikrishna has quit IRC08:11
lbtrcg: I see #38708:11
rcglbt: yes08:11
lbtrcg: this isn't a regression is it?08:15
lbtjust a recently spotted bug?08:15
Stskeepsrcg: this looks like an awesome opportunity to learn how to submit changes to mer :)08:15
lbthttps://wiki.merproject.org/wiki/Contribution_in_detail08:16
rcglbt: dunno if it is a regression. i just stumbled over it while hacking on qmlcalendar and trying to get persistency to work08:16
lbtStskeeps: mainly was wondering about doing another snapshot with it - but think not08:16
Stskeepsno08:16
rcgStskeeps: hehe, ic :)08:17
Stskeepsnot a release blocker in my book and we're already delayed08:17
lbtStskeeps: agree, a regression would have been08:17
Stskeepsrcg: it's quite simple really, git clone, change patch, submit..08:17
*** andre__ has joined #mer08:19
rcgStskeeps: yeah, i had a quick look at the link lbt sent08:19
rcgunfortunately, right now i really need to get my ass up and go to work08:19
rcg:/08:19
Stskeepshehe08:19
lbtStskeeps: fyi I need to patch : rpm, build and osc to get quickbuild working08:19
Stskeepslbt: OK08:20
Stskeepsno objections from me08:20
rcggot some freaking important stuff i should have been working on much much more08:20
lbtrpmbuild runs %clean when passed -bb08:20
lbthttp://www.rpm.org/max-rpm-snapshot/rpmbuild.8.html08:20
rcgStskeeps: i might come back to you tonight when, hopefully, i finished the other work :)08:20
Stskeepslbt: well, my initial quickbuild was just on sources side08:21
lbtStskeeps: so are you OK if I patch our rpmbuild to not do that08:21
Stskeepslbt: that alone is a speedup08:21
lbtand instead to behave as documented08:21
Stskeepslbt: in case you're doing --skipprep or whatever08:21
Stskeepswell08:21
Stskeeps.. okay08:21
Stskeepscheck the man page in rpm first08:21
Stskeepsnot max-rpm-snapshot08:21
lbtI did08:21
lbtsame text08:21
Stskeepsafter that, it's OK08:21
lbtalso - it's actually quite unpleasant to reproduce an 'as shipped' source tree08:22
Stskeepshmm?08:22
lbteg for rpm08:22
lbtpackage has 14 patches08:23
Stskeepsyes08:23
Stskeepsi've looked into rpm 4.10 as well08:23
Stskeepssome can be cut08:23
lbt9 is missing08:23
lbt3 are not applied08:23
Stskeeps:nod:08:23
Stskeepswe should do some automated work to clean up our packages08:23
lbtit's just .... ugly :)08:23
lbtyes08:23
lbtOK - coffee now08:24
Stskeepsindeed08:24
*** Tm_T has quit IRC08:25
*** u1106-office has joined #mer08:28
*** cxl000 has joined #mer08:30
*** HazardousWaster has quit IRC08:41
*** HazardousRooster has joined #mer08:41
*** M4rtinK has joined #mer08:43
*** notmart has joined #mer08:48
*** ZiQiangHuan has joined #mer08:54
sonachlbt: how to change 'OBSRSYNC_INT="obsrun@obsfe.example.com"'?08:59
sonachlbt: if my OBSFE=192.168.188.143, is it OBSRSYNC_INT="obsrun@192.168.188.143"?09:00
*** jukkaeklund has joined #mer09:00
lbtyes, that should work09:00
*** harha has quit IRC09:05
sonachok,09:09
*** Attie has joined #mer09:16
*** sonach has left #mer09:24
*** rcg has quit IRC09:31
*** himamura has quit IRC09:35
*** harha has joined #mer09:37
*** Openfree` has quit IRC09:38
*** phaeron has joined #mer09:38
*** phaeron_n9 has quit IRC09:41
*** Openfree` has joined #mer09:43
*** niqt has joined #mer09:43
*** rcg has joined #mer09:52
*** clopez has joined #mer09:54
*** shrikrishna has joined #mer10:30
*** Behold has quit IRC10:35
*** BeholdMyGlory has joined #mer10:36
*** wmarone has quit IRC10:41
*** wmarone_ has joined #mer10:41
*** Behold has joined #mer10:45
*** BeholdMyGlory has quit IRC10:45
*** BeholdMyGlory has joined #mer10:53
*** Behold has quit IRC10:54
*** ZiQiangHuan has quit IRC10:55
*** ZiQiangHuan_ has joined #mer10:55
*** ZiQiangHuan_ is now known as ZiQiangHuan10:55
*** shrikrishna has quit IRC10:56
*** shrikrishna has joined #mer10:56
iekkumer bug triage starting soon10:59
*** swisstomcat has joined #mer11:08
swisstomcathiya11:08
Stskeeps'lo swisstomcat :) back from a lot of travel?11:09
swisstomcathey :)11:09
swisstomcatyeah11:09
Stskeepssame here.. first day back after 3 weeks on the move, so tired11:09
swisstomcati'll be in madrid soon11:09
swisstomcatfor spainjs11:09
*** rcg has quit IRC11:10
swisstomcati met with scott jenson when i was in palo alto11:11
Stskeepsyeah, i had a talk with him on the phone a week or two ago11:11
swisstomcattalking about ubiquitous computing and project grande11:11
Stskeepswent quite well, after we sorted out technical difficulties of intercontinential voice communication11:12
swisstomcatvery interesting stuff11:12
swisstomcati hope we can somehow push this forward11:12
lbtswisstomcat: hey11:13
swisstomcathi lbt11:13
Stskeepsyes - i think it might be worthwhile to establish some kind of mailing list for sharing ideas -- because alone we can't lift this stuff but even a small group of people can be doing a lot11:13
swisstomcatyep11:13
swisstomcati also met someone at mobilism who seemed very interested (scott spoke there)11:13
lbtswisstomcat: never heard back from Steve Lemke :(11:14
swisstomcatoh :(11:14
Stskeepsswisstomcat: we've been hacking a bit on sw side this weekend, we now have a nice stack with qt5, qml compositor and llvmpipe in a virtualbox, makes for great prototyping of next-gen systems11:14
swisstomcatcool, i'd be interested in testing that (not that i know much about qt5/qml)11:15
Stskeepsoh, and wayland11:15
lbtswisstomcat: maybe you can prod him :)11:15
Stskeepsas you can do anything you'd like with windows, shade them differently, etc11:15
swisstomcatlbt: i think they're still working on OE11:15
lbtah ... we'll wait for their eyes to bleed some more then :)11:16
swisstomcatlbt: i could remind steve11:16
lbt*nod*11:17
*** rcg has joined #mer11:17
alteregolbt: see the wiki migration went well :)11:18
Stskeepsimages/ ?11:19
lbtalterego: yes thanks - and you should be able to carry on looking after it too11:19
alteregoI can't seem to SSH in to the vm11:19
lbthmm11:19
*** lamikr has joined #mer11:19
lbtlet me finish bug triage11:20
alteregoYeah, not urgent.11:20
alteregoyes11:20
alterego~yet11:20
*** jluisn has joined #mer11:23
*** lizardo has joined #mer11:32
*** slaine has quit IRC11:32
*** ZiQiangHuan has quit IRC11:40
*** lizardo has quit IRC11:48
*** furikku has quit IRC11:50
*** furikku has joined #mer11:50
* alterego wonders if he can fit a gpg key on an NFC tag11:51
*** niqt has quit IRC11:51
Stskeepspublic or private?11:53
alteregoWell, public should fit11:55
dm8tbralterego: the hash should fit easily11:55
alteregoAnd private I think would be a bit of a security issue :)11:55
lbtindeed11:55
dm8tbrthere are small tags, but there are nice 1k tags too. 4k prolly too11:55
alteregoI've got 1k tags11:55
* dm8tbr always wanted to fit a vcard in there11:56
kimjuare those bits or bytes?11:56
alteregobytes11:56
Stskeepsalterego: perhaps better to fit in a keyserver ID?11:56
dm8tbrbytes11:56
dm8tbryes, hash plus ID, that's what I thought11:56
alteregoStskeeps: yeah, and as far as vcard, you could fit a small vcard on there. But maybe having some kind of link to an open id profile or something.11:56
alteregoIt's a shame the NFC stack in the N9 is closed.11:57
*** shrikrishna has quit IRC11:59
Stskeepsw00t: confirmed, -window2 did the trick12:04
w00tStskeeps: good12:05
Stskeepsno qml mods, etc12:05
phaeronStskeeps: you  mentioned you had testrunner-lite working ?12:18
Stskeepsphaeron: sure12:18
Stskeepsphaeron: http://releases.merproject.org/~carsten/mer-core-raspberrypi.ks12:19
Stskeepsand testrunner-lite and -ui in sdk12:19
phaeronok12:20
lbtalterego: what's the ssh issue?12:22
alteregolbt: just times out :/12:22
alteregoOh ffs, I'm using the wrong port again ..12:22
alteregoWhat's the port number?12:23
alterego2221?12:23
lbthmm12:24
alteregoHmm, that's not having it either ..12:24
lbt222212:24
alteregoThere we go :)12:25
alteregoCool,12:26
lbtI'll remove the old session12:26
lbt2003     30627 30626  0 Feb16 pts/2    00:00:00 -bash12:26
lbt:)12:26
alteregoI have an old session?12:28
lbttwo ... Feb 16 and Mar 2612:28
lbtkilled now12:28
alteregoInteresting, wonder why they weren't being reaped.12:28
alteregoNot got screen on there.12:29
lbtnot sure - I assumed they were in a screen elsewhere12:30
alteregoNope12:30
lbtmaybe ssh isn't doing keepalives and so doesn't notice dead connections12:30
lbtanyhow ... back to food and then more fastbuild12:32
lbtoh alterego I installed some Debian extensions too - Cite and a few others - maybe see if they're operational and we can put up a page12:32
lbtCategoryTree Footnote RSSReader and LDAP of course12:33
alteregoCool12:36
alteregoWell, I'm sorta busy this week, but I'll hopefully have some time in the evenings.12:36
*** fvennetier has joined #mer12:47
*** eebrah has quit IRC12:50
*** niqt has joined #mer12:51
*** eebrah has joined #mer12:52
*** eebrah is now known as Guest1557312:52
alteregoWell, just "Registered my interest" in a raspberry pi. Could have sworn I pre-ordered last year but meh.12:55
*** niqt has quit IRC12:57
*** lizardo has joined #mer12:58
*** niqt has joined #mer13:00
*** DocScrutinizer05 has quit IRC13:01
*** DocScrutinizer06 has joined #mer13:01
*** DocScrutinizer has quit IRC13:01
*** DocScrutinizer has joined #mer13:02
*** niqt has quit IRC13:07
*** harha has quit IRC13:08
*** DocScrutinizer06 has quit IRC13:09
*** jukkaeklund has quit IRC13:09
*** DocScrutinizer05 has joined #mer13:09
*** niqt has joined #mer13:09
*** Guest15573 has quit IRC13:10
*** DocScrutinizer05 has quit IRC13:10
*** DocScrutinizer05 has joined #mer13:10
*** Aurium has joined #mer13:12
*** Behold has joined #mer13:19
*** sonach has joined #mer13:20
*** sonach has left #mer13:22
*** BeholdMyGlory has quit IRC13:23
*** niqt has quit IRC13:30
*** niqt has joined #mer13:31
*** M4rtinK has quit IRC13:37
*** niqt has quit IRC13:38
*** niqt has joined #mer13:39
*** nj3ma has joined #mer13:41
*** niqt has quit IRC13:48
*** niqt has joined #mer13:50
*** niqt has quit IRC13:57
*** niqt has joined #mer13:59
*** niqt has quit IRC14:05
*** niqt has joined #mer14:19
*** himamura has joined #mer14:21
*** shadeslayer has quit IRC14:25
*** shadeslayer has joined #mer14:27
*** trbs has joined #mer14:29
*** niqt has quit IRC14:35
*** Anssi138 has quit IRC14:37
*** BeholdMyGlory has joined #mer14:43
*** Behold has quit IRC14:43
Stskeepsw00t: regarding mouse cursor, my virtualbox has an option for absolute pointer device14:44
w00tStskeeps: meaning?14:44
Stskeepsmeaning 'act like a touchscreen'14:44
w00thmm14:45
w00tstill not quite sure what you mean in practice, does that help the bad sensitivity problem or?14:45
Stskeepsyes, you'd get touch instead of relative mouse movement14:45
w00thmm14:45
w00tinteresting maybe14:46
*** shrikrishna has joined #mer14:52
*** Behold has joined #mer14:56
*** jluisn has quit IRC14:59
*** BeholdMyGlory has quit IRC15:00
*** talavis has quit IRC15:01
Stskeepswell, i don't know what we did, but nemo on newest prerelease feels faster15:02
Stskeepsi wonder if this is related to the zlib-SSSE3 optimizations15:02
w00twhat hw?15:02
*** ol has joined #mer15:03
Stskeepsexopc15:03
w00tcool15:03
*** ol is now known as Guest2447015:03
*** Guest24470 is now known as _ol15:03
w00ttime for me to go see if i have a gate, i suppose..15:03
w00tlaptop is just about dead15:03
*** the-gibson has quit IRC15:08
*** the-gibson has joined #mer15:09
*** the-gibson has quit IRC15:09
*** the-gibson has joined #mer15:09
*** phaeron has quit IRC15:12
*** Openfree` has quit IRC15:12
*** Pantti has quit IRC15:18
Stskeepstimoph: http://talk.maemo.org/showpost.php?p=1227178&postcount=7515:18
*** harbaum has quit IRC15:24
*** beford has joined #mer15:25
*** stefan_schmidt_w has quit IRC15:25
*** M4rtinK has joined #mer15:28
*** Behold has quit IRC15:31
*** BeholdMyGlory has joined #mer15:34
*** Pantti has joined #mer15:38
*** smoku has quit IRC15:41
*** NIN101 has joined #mer15:54
*** u1106-office has left #mer16:00
*** vilpan has quit IRC16:26
*** jluisn has joined #mer16:46
*** gimli has joined #mer16:58
*** ssirkia has quit IRC16:58
*** CosmoHill has joined #mer17:05
*** tsdedst has joined #mer17:07
*** nsuffys has joined #mer17:07
*** rcg has quit IRC17:07
*** zeq1 has joined #mer17:15
*** phaeron has joined #mer17:26
*** tilgovi has joined #mer17:32
CosmoHillwhat dynamic DNS providers do you guys use?17:36
*** lizardo has quit IRC17:45
*** lizardo has joined #mer17:49
*** tsdedst has quit IRC17:52
*** tsdedst has joined #mer17:52
*** tilgovi_ has joined #mer17:52
*** tilgovi has quit IRC17:53
zeq1Hi guys, over a #maemo we've been fixing Fennec not to use QtSensors for orientation information.  I was just wondering if you still use this broken polling API, or if it's being fixed/replace with an event/notification based API?17:58
*** tsdedst has quit IRC18:10
*** tilgovi_ is now known as tilgovi18:10
*** tilgovi_ has joined #mer18:15
*** tilgovi has quit IRC18:15
*** gabrbedd has joined #mer18:23
*** Siosm has quit IRC18:26
*** Alison_Chaiken has quit IRC18:26
*** furikku has quit IRC18:38
Stskeepszeq1: depends, send a patch?18:43
*** tilgovi_ has quit IRC18:44
*** freemangordon has joined #mer18:47
zeq1Stskeeps: I could take a stab at it, I was just wondering if you were using those libqtm APIs, and if so for what purposes.  Any new API has to meet existing requirements where they make sense, or else code like in Fennec needs to be modified to use a more suitable API.18:49
Stskeepszeq1: it might be better to look at qt5's ways of doing it18:49
Stskeepsas qtm is part of that18:49
Stskeepsqt4.* isn't likely to change APIs18:50
Stskeepsalso, for DocS, i can actually read #maemo irc logs too you know18:50
Stskeeps:P18:50
Stskeepszeq1: could you investigate android APIs as well in this respect?18:51
zeq1I did mention we were talking about it :)18:51
Ronksuevenings18:51
Stskeepsbut seriously though: the usual gripe is about a certain driver in a hardware adaptation, which is external to the mer project. the qt mobility api's may be sub-optimal18:52
Stskeepsand qt4.* apis aren't likely to be better18:52
zeq1I'll take a look at what's in qt5/android18:53
Stskeepsi can't help but wonder if it was at some point evaluated to be "not that significant an issue" though - i know the people who do wakeup checks and power management on n90018:54
Stskeepswell, before they lost their jobs18:54
*** beford has quit IRC18:55
Stskeepsand i doubt they'd oversee a wakeup source like that if it was significant18:56
DocScrutinizer05yeah, but when N900=maemo5, then that's the only architecture that uses a semi-decent IRQ driven approach18:56
DocScrutinizer05based on a mameo-specific hw driver18:56
freemangordonStskeeps: AFAIK Qt on maemo5 uses mce for orientation reading, could be wrong though18:56
zeq1probably hard to ask them now. certainly in principle you' want to wake up as little as possible regardless of adaptation18:56
*** cristi has joined #mer18:56
Stskeepslet's just say that beyond community there isn't a terribly much going to happen on the n900 kernel, or for that matter n950/n918:56
DocScrutinizer05the issue is embedded domain18:57
DocScrutinizer05not that it would hurt to have a decent arch on desktop platforms18:58
Stskeepszeq1: agreed18:58
yuntazeq1: just use ipheartbeat (if it's actually open) to synchronize polling. impact will be very minimal then.18:59
Stskeepsyunta: iphb library is but i think the kernel side wasn't really accepted upstream, i think..18:59
Stskeepsbut i might be wrong18:59
DocScrutinizer05when I hear heartbeat, i already frown18:59
yuntahehe, stupid little thing, but it's all we had in n90019:00
Stskeepsyunta: well, idea is pretty good19:00
DocScrutinizer05we got NO_HZ on what? kernel version 2.4.somebeers?19:01
yuntageneral idea yes, but their implementation missed the point completely, at least for high level services19:01
Stskeepsyunta: for sure19:01
freemangordondo we have hartbeat when the device is offline?19:01
Stskeepsit was vastly underdocumented19:01
DocScrutinizer05idea been a idle platform is actually iudle19:01
DocScrutinizer05you generally don't want *any* timers running on an idle platform, just to do some polling19:02
yuntawe still have heartbeat from battery and other sources (timer in syshb) when offline19:02
yuntayou know, as in : battery *HAS* to be checked every 12s or whatever19:03
DocScrutinizer05and if polling isn't avoidable, then the timer needs to be inside or close to the sensor that gets polled19:03
* Stskeeps has to get some sleep19:03
DocScrutinizer05battery *has* to be checked? just WHY?19:03
yuntadepends on battery implementation I guess19:04
yuntaspeaking about n900 and n9 - it's about safety19:04
DocScrutinizer05I for the life of mine couldn't think of any reason to check battery every 12 seconds19:05
*** icota has joined #mer19:05
DocScrutinizer05no, yunta. THIS is not related to safety19:05
DocScrutinizer05on N9* we got a proper battery management chip that needs no attention whatsoever during discharge19:06
DocScrutinizer05(end *very* little attention during charging)19:06
yuntahmmm19:06
yuntaI think 60s was normal discharge polling time19:07
freemangordon:nod:19:07
DocScrutinizer05to display the battery icon with proper amount of bars inside, yes19:07
DocScrutinizer05could be completely stopped when screen is locked so icon not showing19:08
DocScrutinizer05no sound rationale to *poll*19:08
DocScrutinizer05the story is always the same: "look, we're polling as there's already something else that does"19:09
freemangordonDocScrutinizer05: it is stopped19:09
freemangordon(when the screen is off)19:09
freemangordonlibbmeipc listens to mce and stops the timer when needed19:10
*** harbaum has joined #mer19:12
DocScrutinizer05it's a golden commandment: never use polling on embedded when there's *any* other way19:13
DocScrutinizer05and alas e.g. for acellerometer, avoiding polling and still allowing instant reaction on a movement of device (even during 'idle') means we can't use the API and concept based on polling, as the two concepts conflict19:15
yuntasure, if you want to make things perfect19:16
DocScrutinizer05for battery it's maybe no problem to poll as long as you need the actual values, and stop polling as soon as e.g. screen gets locked. Since battery usually can't create any unexpected unforseeable events19:17
yuntain practice if you need like 0.01s for you poll, you can poll every 0.5s and you won't notice a difference in power consumption with screen-on19:18
*** zeq has joined #mer19:18
yuntajust because cpu will use 5mA more, and screen uses 180mA anyway19:18
DocScrutinizer05for accel however we got a nice chip with proper IRQ and a whole zoo of comparators and filters19:18
DocScrutinizer05yunta: that's all pretty fine _as long as screen on_19:19
yuntayeah19:19
DocScrutinizer05if however you one day get the funny idea to maybe ring an alarm when somebody touches your device (movement detection), you'll start to curse the day you decided polling is the _simpler_ way to go for you as a developer19:20
*** zeq has quit IRC19:21
*** phdeswer has joined #mer19:21
*** zeq1 is now known as zeq19:21
aard_DocScrutinizer05: reminds me, there's some car whatever utilities app in the store, with one function 'take pictures on movement'. my wife installed that, and the app didn't shut down properly. took a while to figure out where the hundreds of pictures on her phone came from :)19:22
DocScrutinizer05and btw, polling ever 0.5s seems intolerably long a delay when we expect device to react promptly on moves19:22
*** M4rtinK has quit IRC19:23
DocScrutinizer05au contraire we already need a sample period of at least 150..200ms to filter out sporadic bouncing19:26
*** rcg1 has joined #mer19:27
yuntastill acceptable in screen-on, but I totally support your view19:27
DocScrutinizer05during that sample period we should scan the orientation way more often than a mere 2Hz19:27
*** lamikr has quit IRC19:28
*** Eismann has quit IRC19:31
*** Eismann has joined #mer19:32
*** clopez has quit IRC19:36
DocScrutinizer05and once you accept that a sensor data _maybe_ is also needed when actually nothing else in active on device (situation when an app is waiting for an orientation change IRQ/signal, while device is 'idle'), you suddenly realize that 1320mAh / n << 1320mAh / (n + 5mA)19:39
DocScrutinizer05s/<</>>/19:40
DocScrutinizer05where n is for example ~6mA for an idle N90019:40
*** jluisn has quit IRC19:43
*** pohly has quit IRC19:44
*** swisstomcat has left #mer19:50
*** CosmoHill has quit IRC19:51
*** mardy has quit IRC19:52
*** nj3ma has quit IRC19:52
*** mardy has joined #mer19:53
*** arcean has joined #mer19:55
lbtDocScrutinizer05: what happens when the device is idle in your pocket?20:02
DocScrutinizer05lbt: shall I guess what's the point you have in mind?20:03
lbtyou're talking about event vs poll20:04
DocScrutinizer05yep20:04
lbtwhen the device is in my pocket and I'm moving then it will be continuall emitting events20:04
DocScrutinizer05not more frequently than when you're polling it at a sane rate20:05
*** harbaum has quit IRC20:05
lbtwhy?20:05
lbtare events rate limited?20:05
DocScrutinizer05and even if it would, you're still free to power down the whole chip20:05
lbtI was actually thinking that rate limiting events would help20:05
lbtbut if they already are then ... no need20:05
DocScrutinizer05or simply wait with servicing the IRQ and re-loading it20:06
lbtand it may have to wakeup too much to simply process the event to rate limit it20:06
DocScrutinizer05nope20:06
DocScrutinizer05there's no second IRQ until you serviced the first one20:06
lbtOK, so rather than poll20:07
lbtyou say "won't accept another for Xms"20:07
lbtby sleep X; reload IRQ20:07
DocScrutinizer05and after all it's up to you (or your middleware) to set the thresholds in lis302 so that you only get events you're interested in20:08
lbtand yes, X could be parameterised for appropriate power usage20:08
lbtthanks - just getting my head round that aspect - I hadn't appreciated how the event stream could be managed20:09
lbtfor those sensors - how much noise is there? ie would a stationary device send a stream of interrupts due to minor fluctuations?20:10
DocScrutinizer05look, you got 2 apps that are intersted in orientation (at large). A wants to wake up on hard bouncing, B wants to know when you turn device upside down. Both need to register to a service that finds common denominator of the requests and adjusts lis302 accordinly20:10
DocScrutinizer05if nobody is interested in any orientation event at all, the driver should just power down the whole chip20:11
lbtof course20:11
DocScrutinizer05the rest is up to "you"20:11
DocScrutinizer05(noise) you adjust thresholds in chip20:12
lbtright - was just thinking I'd need to understand the chip capabilities20:12
DocScrutinizer05you got iirc 2 comparators, a set of highpass filters (to detect changes aka motion), and an adder to monitor 3 axes with just one comparator20:13
DocScrutinizer05you (again iirc) could set up all these blocks so one comparator gets highpass-filtered events of arbitrary set of axis, and would fire when change in orientation is >x/timeunit. A second comparator could monitor sum of all three axis raw unfiltered values, and fire when it gets <0.01g (means freefall). You even have two IRQ output pins, though I'm not sure both are connected on N90020:16
*** harha has joined #mer20:16
DocScrutinizer05all parameters of comparators and filter are configurable20:17
lbt*nod* ... was just googling a bit20:17
DocScrutinizer05and yes, there's also a mode where IRQ fires on a timepace, like 100/s, for *real* polling20:17
DocScrutinizer05of course if your config says "fire IRQ1 when X >0.8G" and you simply reset the IRQ without changing the setup nor the orientation, then the IRQ gets triggered again instantly20:19
DocScrutinizer05less of a problem for highpass filtered stuff though20:19
DocScrutinizer05also please take all that with a grain of salt, I last time looked into lis302 datasheet like 6 months ago20:20
lbtwhich is why you'd look at the desired responsiveness when sleeping before reseting IRQ20:21
lbt*nod* ... conceptually I'm sure it's sound :)20:21
*** harha has quit IRC20:21
DocScrutinizer05yep20:21
DocScrutinizer05kept it in a dark corner of my mind for >2 years20:21
DocScrutinizer05it's "ripe"20:22
lbtis it just accel btw? or orientation too?20:22
*** tsdedst has joined #mer20:22
DocScrutinizer05only orientation20:22
DocScrutinizer05meants which size and direction has the g vector20:22
DocScrutinizer05it measures acceleration20:23
lbtah, OK20:23
DocScrutinizer05just a small mass on 3 gauges20:24
lbtyep - my mental model was wrong20:24
*** jstaniek has joined #mer20:28
DocScrutinizer05lbt: that for it basically cant detect rotations, at least when roatation axis is || g-vector20:30
lbtthanks ... now, must not get *too* distracted :)20:32
DocScrutinizer05:nod:20:32
*** harha has joined #mer20:33
*** nsuffys has quit IRC20:34
*** tsdedst has quit IRC20:50
*** jluisn has joined #mer20:52
*** lizardo has quit IRC20:59
*** NIN101 has quit IRC21:01
*** notmart has quit IRC21:01
*** M4rtinK has joined #mer21:15
*** shrikrishna has quit IRC21:15
*** shrikrishna has joined #mer21:17
*** wicket64 has joined #mer21:22
*** ALoGeNo has quit IRC21:23
*** jluisn has quit IRC21:23
*** tilgovi has joined #mer21:24
*** gabrbedd has quit IRC21:30
*** zeq has quit IRC21:30
*** CosmoHill has joined #mer21:32
*** gabrbedd has joined #mer21:35
*** Alison_Chaiken has joined #mer21:36
*** harha has quit IRC21:36
*** ALoGeNo has joined #mer21:38
*** ALoGeNo has joined #mer21:38
*** zeq has joined #mer21:40
*** arcean has quit IRC21:40
*** harha has joined #mer21:41
*** Behold has joined #mer21:42
*** CosmoHill has quit IRC21:42
*** BeholdMyGlory has quit IRC21:45
*** harha has quit IRC21:45
*** sigkill_ has joined #mer21:47
*** s1gk1ll has quit IRC21:50
*** gimli has quit IRC21:52
*** BeholdMyGlory has joined #mer21:56
*** tsdedst has joined #mer21:56
*** freemangordon has left #mer21:56
*** BeholdMy- has joined #mer21:58
*** Behold has quit IRC21:59
*** BeholdMyGlory has quit IRC22:00
*** BeholdMy- has quit IRC22:02
*** BeholdMy- has joined #mer22:02
*** BeholdMy- is now known as BeholdMyGlory22:02
*** Behold has joined #mer22:04
*** BeholdMyGlory has quit IRC22:07
*** tsdedst has quit IRC22:10
*** cxl000 has quit IRC22:17
*** yunta has quit IRC22:17
*** himamura has quit IRC22:19
*** himamura has joined #mer22:19
*** tilgovi_ has joined #mer22:19
*** tilgovi has quit IRC22:20
*** himamura has quit IRC22:23
*** tsdedst has joined #mer22:24
*** niqt has joined #mer22:26
*** rcg1 has quit IRC22:37
*** niqt has quit IRC22:43
*** Estel_ has quit IRC22:47
*** Estel_ has joined #mer22:47
*** cristi has quit IRC22:54
*** Aurium has quit IRC22:58
*** rdqfdx has quit IRC23:17
*** trbs has quit IRC23:18
*** rzr has joined #mer23:18
*** tilgovi_ has quit IRC23:23
*** jstaniek has quit IRC23:25
*** DocScrutinizer05 has quit IRC23:30
*** DocScrutinizer05 has joined #mer23:30
*** jussi has quit IRC23:30
*** jussi has joined #mer23:30
*** wicket64 has quit IRC23:33
*** icota has quit IRC23:33
*** sethstorm has quit IRC23:34
*** sethstorm has joined #mer23:35
*** decibyte has quit IRC23:40
*** w00t has quit IRC23:40
*** w00t has joined #mer23:40
*** w00t has quit IRC23:40
*** w00t has joined #mer23:40
*** decibyte has joined #mer23:41
*** kimitake_idle has quit IRC23:42
*** shrikrishna has quit IRC23:44
*** kimitake_idle has joined #mer23:44
*** clopez has joined #mer23:46

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!