IRC log of #maemo-ssu for Sunday, 2012-09-30

*** fredrinLap has quit IRC00:01
* DocScrutinizer05 wonders why fw190 vanished00:02
*** fredrinLap has joined #maemo-ssu00:03
DocScrutinizer05just when I was about to just install patterns-opensuse-lxde as that pulls in like 50 other pkgs as dependencies00:03
*** fredrinLap has quit IRC00:12
*** fredrinLap has joined #maemo-ssu00:14
*** Pali has quit IRC00:16
*** desaparecido has joined #maemo-ssu00:17
DocScrutinizer05yay, Donnie Darko in TV00:18
*** arcean has quit IRC00:18
DocScrutinizer05one of the stranger movies I ever saw00:18
*** fredrinLap has quit IRC00:19
*** fredrinLap has joined #maemo-ssu00:20
*** arcean has joined #maemo-ssu00:20
Skryexcellent movie00:22
*** fredrinLap has quit IRC00:24
*** fredrinLap has joined #maemo-ssu00:29
*** arcean has quit IRC00:34
*** Climbz has quit IRC00:37
*** fredrinLap has quit IRC00:39
*** Guest57704 has quit IRC00:42
*** fredrinLap has joined #maemo-ssu00:50
*** fredrinLap has quit IRC01:13
*** fredrinLap has joined #maemo-ssu01:40
*** fredrinLap has quit IRC01:46
*** fredrinLap has joined #maemo-ssu01:55
*** fredrinLap has quit IRC02:00
*** fredrinLap has joined #maemo-ssu02:15
*** fredrinLap has quit IRC02:20
*** arcean has joined #maemo-ssu02:32
*** fredrinLap has joined #maemo-ssu02:33
*** NIN101 has quit IRC02:34
*** fredrinLap has quit IRC02:40
*** desaparecido has quit IRC02:42
*** arcean has quit IRC02:53
*** _xnt14 has quit IRC03:10
*** _xnt14 has joined #maemo-ssu03:13
*** fredrinLap has joined #maemo-ssu03:26
*** fredrinLap has quit IRC03:43
*** M4rtinK has quit IRC03:51
*** fredrinLap has joined #maemo-ssu03:56
*** fredrinLap has quit IRC04:04
*** fredrinLap has joined #maemo-ssu04:09
*** fredrinLap has quit IRC04:15
*** fredrinLap has joined #maemo-ssu04:15
*** fredrinLap has quit IRC04:30
*** nox- has quit IRC04:31
*** fredrinLap has joined #maemo-ssu04:32
*** fredrinLap has quit IRC04:47
*** fredrinLap has joined #maemo-ssu04:50
*** fredrinLap has quit IRC05:09
*** amiconn has quit IRC05:15
*** amiconn_ has joined #maemo-ssu05:15
*** amiconn_ is now known as amiconn05:15
*** fredrinLap has joined #maemo-ssu05:37
*** DocScrutinizer05 has quit IRC06:03
*** DocScrutinizer05 has joined #maemo-ssu06:03
*** jon_y has joined #maemo-ssu06:49
DocScrutinizer05anybody ready for fun with sftp?08:20
DocScrutinizer05I have a rather strange bug in KDE08:20
DocScrutinizer05(maybe not limited to KDE)08:20
DocScrutinizer05you probably know ~/.ssh/id_dsa.pub08:21
DocScrutinizer05AIUI it's just needed on targets, not on origin of a ssh/sftp connection08:22
DocScrutinizer05even there it's called known_hosts08:22
DocScrutinizer05so obviously there's nothing wrong with `mv id_dsa.pub id_dsa.pub-away`08:23
DocScrutinizer05sftp://user@host:port   still works great, in both konqueror and dolphin08:24
DocScrutinizer05but!! only once, since then some sneaky magic created a new id_dsa.pub, alas a crippled truncated one, and next connection via sftp:// will fail due to that :-o08:25
DocScrutinizer05plus each time you open a normal ssh connection it spamms that trunkated key for a greeting08:26
DocScrutinizer05err, correction:08:28
DocScrutinizer05even there it's called authorized_keys08:28
*** Guest57704 has joined #maemo-ssu09:04
*** timo^ has joined #maemo-ssu10:08
*** amiconn has quit IRC10:33
*** amiconn has joined #maemo-ssu10:34
*** M13 has quit IRC10:40
*** timo^ has quit IRC10:55
*** fw190 has joined #maemo-ssu10:56
fw190DocScrutinizer05: I vanished yesterday as I was on my last legs10:56
fw190all went well10:57
fw190LXDE works ok10:57
fw190I installed patterns-opensuse-lxde and it pulled everything as it should10:57
fw190thanks for the tips yesterday10:58
keriofw190: how's lxde?11:00
kerioand have you realized that it's worse than xfce?11:00
fw190kerio: well it looks like lxde11:07
fw190newver used xfce11:07
kerioto be fair, the latest xfce is somewhat bad11:07
kerioit used to be good11:07
fw190everyone is saying that lxde is lighter11:07
fw190so I stick to it11:07
kerioxfce used to be the "light" DE11:08
keriothen it became heavier11:08
fw190but comparing with KDE I do not see much speed improvment11:08
keriowell, is KDE slow for you?11:08
fw190no11:08
kerio:)11:09
fw190I just have to find the settings which changes from one click to two11:10
fw190although I'm starting to get used to it11:10
fw190just found it11:20
fw190have to say it11:20
fw190this suse thingy looks better and better to me11:20
*** CodenameStrike has joined #maemo-ssu11:34
*** CodenameStrike has joined #maemo-ssu11:35
*** fw190 has quit IRC11:38
*** CodenameStrike has quit IRC11:57
*** NIN101 has joined #maemo-ssu11:59
*** M4rtinK has joined #maemo-ssu12:04
*** Pali has joined #maemo-ssu12:06
freemangordonmerlin1991: http://talk.maemo.org/showthread.php?p=1274218#post127421812:13
merlin1991freemangordon: I just saw it12:14
freemangordonyou'd better disable ipv6 DNS until everyone is on CSSU :D:D:D12:14
merlin1991ipv6 has been there all the time12:14
merlin1991I wonder why it fails now12:15
freemangordonhmm, then it is your ISP12:15
freemangordonlemme try on my device12:15
freemangordonwell, ping6 fails here as well, but I don't have problems with apt-get update12:18
merlin1991odd, http://centralops.net/co/Ping.aspx for example works12:19
freemangordonmerlin1991: in the meantime:12:20
freemangordonhttp://maemo.merlin1991.at/cssu/develdebs/gcc-4.7.2_fremantle_armel/scratchbox-toolchain-linaro-4.7-2012.09-fremantle-armv7a_1.0.17_i386.deb12:20
*** int_ua has joined #maemo-ssu12:20
*** CodenameStrike has joined #maemo-ssu12:21
merlin1991freemangordon: btw what do you get for ping6 ipv6.google.com    on your device?12:21
freemangordonnetwork unreachable12:21
merlin1991also you have 384 dl of the mp so far :D12:22
freemangordonwow12:22
int_uahey, I just got second N900 ^_^12:23
freemangordonhmm, not good, i'd better not screw with pulseaudio :D:D:D12:23
merlin1991int_ua: welcome to the club of madmen who own several n900s :D12:23
int_uabut haven't slept 28 hours)12:23
int_uasee you tomorrow :)12:23
merlin1991have a nice sleep :)12:24
int_uathanks :)12:24
freemangordonint_ua: night12:24
merlin1991freemangordon: do we have dig on the device?12:24
*** int_ua has quit IRC12:24
freemangordonnslookup?12:24
freemangordonnot sure for dig12:24
merlin1991dang dig is so much better :P12:25
* merlin1991 currently has no n900 nearby, all 4 of them are ~65 km away from him12:25
*** arcean has joined #maemo-ssu12:25
merlin1991hm I could drop the ipv6 dns entry on maemo.merlin1991.at12:26
merlin1991aka AAAA entry12:26
freemangordonmerlin1991: need to reboot my PC12:33
*** freemangordon has quit IRC12:34
merlin1991~lart rndc12:34
* infobot installs PocketPC on rndc's PDA12:34
*** freemangordon has joined #maemo-ssu12:35
freemangordonmerlin1991: no idea, but that should be something with bind (or whatever DNS server you use)12:37
merlin1991well I wonder why his/her n900 decided to query the dns for AAAA instead of A entries12:39
merlin1991ah wait12:39
merlin1991afaik I set up ipv6 dns entries on my side, but the guy who hosts the real dns servers had ipv6 disabled on his bind back then12:39
keriofreemangordon pls12:47
*** dhbiker has joined #maemo-ssu13:00
merlin1991gotta love my university, finally found my course which has a description page that is awesomely emtpy except an entry about some jabber chatroom, so I dl psi try to remember my jabber pw again, and finally logon, to find the chat room empty13:27
merlin1991yay13:27
kerioyo merlin199113:38
keriocan we update connui-home-cellular pls13:38
DocScrutinizer05yo dawg, heard you like info. So we put info about info into info of your course13:39
merlin1991kerio: eh is there anyhting new on gitorious?13:40
DocScrutinizer05yo dawg, heard you like chat... ;-P13:40
keriomerlin1991: the logging fix13:40
DocScrutinizer05is it complete13:40
DocScrutinizer05ß13:41
DocScrutinizer05?13:41
merlin1991ah yeah http://gitorious.org/community-ssu/operator-name-cbs-widget/commit/ff96ce3653b9951158badc150cb4c62aec28b1fc13:41
DocScrutinizer05still, is it *complete* now?13:41
DocScrutinizer05has anybody finally *tested* that ?13:42
keriotested what?13:42
keriothe whole thing or just that fix?13:42
* DocScrutinizer05 takes a huge sip of his coffe, and breathes deep13:43
DocScrutinizer05the fact that/if log files won't appear despite logging off, with random loglines in them13:43
merlin1991kreio wanna play guinea pig?13:44
merlin1991s/kreio/kerio/13:44
infobotmerlin1991 meant: kerio wanna play guinea pig?13:44
merlin1991should use tab instead of typing myself :P13:45
keriosure, why not13:45
* merlin1991 hits the reboot button13:45
DocScrutinizer05merlin1991: btw I yesterday spotted something I'd frown on as a pretty evil flaw in maemo boot UX13:46
DocScrutinizer05fsck gets started without *any* feedback to user about "prepare to watch your device acting_bricked"13:46
kerioDocScrutinizer05: to be fair it's ext3 fsck13:47
kerioit's journaled and stuff13:47
DocScrutinizer05well, my fsck.ext3 takes quite some time here13:47
DocScrutinizer05naw, fsck doesn't use journal13:47
DocScrutinizer05once you need or want fsck, you forget about journal.13:48
DocScrutinizer05journal is for not needing fsck at all13:48
DocScrutinizer05basically&afaik13:48
DocScrutinizer05anyway my fsck.ext3 here takes dozens of minutes13:49
DocScrutinizer05on PC13:49
* kerio tune2fses his mmcblk0p213:49
merlin1991hm regarding fucked up course pages, http://ti.tuwien.ac.at/rts/teaching/courses/osue/ is the best so far timeout when trying to access it13:49
merlin1991why the fuck can't they use the centralized information system, every single stupid course has a link to their "own" homepage that doesn't work most of the time13:50
DocScrutinizer05yeah, down4me213:51
DocScrutinizer05yo dawg...13:51
merlin1991and the best thing about it, the semester begins tomorrow, and guess what, the course in question has a reservation for a room on monday13:51
DocScrutinizer05errrwut?13:52
DocScrutinizer05sounds like it's a sane thing13:52
DocScrutinizer05surprisingly sane?13:52
keriotomorrow my courses begin too ^_____^13:52
merlin1991kerio: what do you study?13:53
keriomath13:53
merlin1991ah yeah /me forgot13:53
DocScrutinizer05merlin1991: call the Prof on Sunday noon ;-P "Hey Prof, where's the room?"13:53
DocScrutinizer05that explains some bits13:53
merlin1991call the prog "sup dwag, I heard you like questions in questions, so what's up with the page and which room would it show?" :D13:54
merlin1991s/prog/prof/13:54
infobotmerlin1991 meant: call the prof "sup dwag, I heard you like questions in questions, so what's up with the page and which room would it show?" :D13:54
DocScrutinizer05aaah, that makes more sense than prog ;-)13:55
merlin1991oh ffs automake madness in my sb again13:56
DocScrutinizer05merlin1991: less +/else /etc/event.d/rcS-late13:57
DocScrutinizer05waddooyathink? Add some LP5523 machine pattern for notifier LED?13:58
merlin1991hm what's the right way (tm) to edit symlinks in /etc/alternatives ?14:00
DocScrutinizer05err14:02
DocScrutinizer05usually postinst does that afaik14:03
DocScrutinizer05prolly there's even a scriptie dedicated to that daunting task14:03
merlin1991seems to be update-alternatives, but it keep claiming that there's no alternative for aclocal14:03
kerioreplacing rcS-late will fuck up soooooooooo many things14:05
DocScrutinizer05kerio: please elaborate14:06
keriowell, it's the place where a lot of system mods live14:06
DocScrutinizer05that's what patch is meant for14:06
keriopeople with weird swaps, people with strange partitioning...14:06
DocScrutinizer05and if we don't have patch, we still can use sed to tweak whatever sits in /etc/event.d/rcS-late.14:07
merlin1991kerio: http://cdnm.at/~christian/maemo/cssu/connui-home-cellular/14:07
DocScrutinizer05and if even that's not feasible or PC, we can wrap fsck.* by a wrapper taking care about proper user-notification14:07
merlin1991srly why don't people ever use dpkg-divert on packaged scripts they edit?14:08
keriomerlin1991: because they don't know how dpkg works14:08
merlin1991It's so stupid to go "uh I'm gonna edit this file here and don't think about anything but my immediate gain"14:09
DocScrutinizer05why don't we just use proper md5sum checks to detect edited config/scrript files and do like all good updaters: install xyfile.new instead of overwriting xyfile unconditionally?14:09
*** ZogG has quit IRC14:09
kerioDocScrutinizer05: guess what, apt actually does that14:09
kerioit asks the user14:09
kerio(for config files)14:09
kerio(for modified config files)14:10
DocScrutinizer05thought as much14:10
merlin19912 reasons, firstly there is no such support for scripts in dpkg14:10
merlin1991only for "configfiles" and stuff in /etc14:10
kerio"config file modified, install (N)ew, keep (O)ld, show (D)iff, ask (L)ater"14:10
merlin1991and secondly, because ham!14:10
DocScrutinizer05~fsck HAM14:10
infobote2fsck /dev/HAM : warning! filesystem contains dumbasses!14:11
keriomerlin1991: opername.log seems to not appear14:11
DocScrutinizer05kerio: good14:11
DocScrutinizer05now please also check if enabling logging will change that14:11
DocScrutinizer05also check if that silly timestamp got fixed14:12
merlin1991kerio: you got this fix aswell: http://www.gitorious.org/community-ssu/operator-name-cbs-widget/commit/95ca519c5b29dfb7a7692f19828781ddfcc29d6314:13
merlin1991DocScrutinizer05: which silly timestamp?14:13
DocScrutinizer05merlin1991: no comment from your side about general nastiness to start fsck.* and let user stare at moving dots for 'hours'?14:13
keriomerlin1991: the timestamp in the connui logs is 0112093014:14
merlin1991eh no n900 so no chance to comment till later this evening14:14
DocScrutinizer05merlin1991: the 1911214:14
DocScrutinizer05year14:14
DocScrutinizer05merlin1991: (comment) mompls, I pastebin the rcs-late for you14:15
* merlin1991 is about to fix the fsckd timestamp14:18
merlin1991what's the timestamp log convention?14:18
DocScrutinizer05http://maemo.cloud-7.de/maemo5/etc/events.d/rcS-late14:18
merlin1991oh nice, autofsck14:19
DocScrutinizer05yeah14:19
DocScrutinizer05without any notification to user14:19
DocScrutinizer05a bug in my book14:19
merlin1991kerio: did you enable the log and check it?14:22
keriono14:22
merlin1991judging from the source the timestamp should have a format of xx.xx.xx xx:xx:xx14:22
merlin1991though the order on the xx.xx.xx is day month year which imo should be the other way around :P14:23
merlin1991so we have 3 things for next -testing already, openssl dbg strip, connui-home-cellular logfix + namefix and camera-ui updates14:25
*** ZogG has joined #maemo-ssu14:27
merlin1991DocScrutinizer05/kerio please test something reagarding camera-ui, set the captured image preview to unlimited, take a pic and the delete it in the preview window14:27
merlin1991do you get the viewfinder back or a black screen?14:28
DocScrutinizer05"happy" that I still can test that ;-P14:28
DocScrutinizer05though, first of all, WTF do I configure the review timing?14:30
merlin1991there's a wrench symbol somewhere on the side of the screen iirc14:30
merlin1991or it's one of the buttons on the bottom lef14:31
merlin1991+t14:31
DocScrutinizer05alas the menu under wrench has no fsckng setting for that14:31
DocScrutinizer05got it14:31
DocScrutinizer052nd left on bottom left14:31
kerioi get the viewfinder back14:32
merlin1991good, then mag did compile that one14:33
merlin1991so the new stuff in camera ui would be launcher support and proper cam directory code14:37
DocScrutinizer05video is totally useless14:38
DocScrutinizer05slow as molasses14:38
DocScrutinizer05takes like 20s to start capture14:38
merlin1991might be faster with the coming update14:38
DocScrutinizer05but yes, viewfinder comes back14:38
DocScrutinizer05autofocus distance is totally wrong 90% of the times14:39
merlin1991well that's still the same as with the stock camera14:39
DocScrutinizer05hitting play button in preview opens mediaplayer. Deleting video file there and then closing mp returns me to preview where I could delete *again*14:40
DocScrutinizer05closing preview seems to freeze the whole system for several seconds, until viewfinder re-appears14:41
merlin1991don't hate so much on our beloved cam :D14:41
kerioDocScrutinizer05: camera-ui in cssu-thumb is much, much faster :314:45
DocScrutinizer05and the worse: it seems to ignore trigger button presses that are shorter than 0.5s14:45
DocScrutinizer05obviously trackerd all the time hogs CPU like mad14:45
DocScrutinizer05while it *should* *just* 'index' the _one_ media file that just got created/deleted, and then STFU again14:46
DocScrutinizer05(unconfirmed by any means like htop etc!)14:47
DocScrutinizer05(just cpu load widget constantly red)14:47
DocScrutinizer05honestly tracked needs to get suspended during realtime stuff like camera usage14:48
DocScrutinizer05when you short a rapid sequence, "rapid" like >1pic/s (if that's even possible), you for sure don't need any friggin trackerd hogging system resources for indexing that as well could be done later14:50
DocScrutinizer05same when you return from preview to viewfinder14:51
DocScrutinizer05as long as there's viewfinder visible, trackerd has to go hybernate mode14:51
DocScrutinizer05may just store up dbus msgs about media files to examine, for later processing14:52
DocScrutinizer05but to answer your question: yes, viewfinder comes back... eventually ;-P14:53
DocScrutinizer05UX however is like with a 1880's camera XP14:54
DocScrutinizer05"please don't move!" PFOOOOSH-FLASH <magnesia smoke goes up>14:55
DocScrutinizer05"please stay! I'm just inserting another plate"14:55
DocScrutinizer05on a "stock" PR1.3 (PK47 or sth, my t900 devel device), turnaround for the sequence >start video capture, stop video capture, review shows up, hit delete in preview, hit "ok", returns to viewfinder, start video capture< is like 5 times faster than on CSSU-T15:02
DocScrutinizer05also reacts reliably on short trigger button push15:02
DocScrutinizer05reliably and instantly15:02
DocScrutinizer05not with 2s delay like nicocam/cssu-t15:03
kerioDocScrutinizer05: cache?15:03
DocScrutinizer05honestly, this stuff isn't ready for prime time15:04
kerioit's pretty fast here, idk15:04
kerio...actually you could try it too, you have a n900 with cssu-thumb installed15:05
DocScrutinizer05I just tried on cssu-thumb by mistake, looked all the same like on cssu-t15:06
kerioDocScrutinizer05: is it updated?15:07
kerionicocam was thumbified recently15:07
DocScrutinizer05kerio: well, actually cssu-thumb is significantly faster than cssu-t, but still not as fast as stock15:08
DocScrutinizer05esp invocation of preview and return to viewfinder after hitting "ok" for delete is still sluggish, compared to stock15:09
DocScrutinizer05while reaction to trigger button is almost bearable15:09
DocScrutinizer05both for starting and stopping capture15:09
DocScrutinizer05but... could it be that charging is completely fsckdup in cssu-thumb?15:12
kerioD:15:14
keriowhy would it be?15:14
keriobme is the same15:14
DocScrutinizer05somehow device been stuck at "USB mode: just charging" in sysmenu. When I click it, the selector AssRage/PCShit pops up. Selecting AssRage gives me another notifier "not possible, storage already in use"15:15
DocScrutinizer05I plugged and unplugged USB 10 or 15 times before that, and it didn't even recognize15:16
keriowell, maybe the storage is in use15:16
DocScrutinizer05NB all the above been with USB *detached*15:16
keriohaha15:16
kerioneat15:16
kerio~n900-full-reset15:16
infobotmethinks n900-full-reset is when the user presses the PWRON (power-on) button for 8 seconds and removes the battery in the next 8 seconds, the TPS65950 enters NO SUPPLY state instead of BACKUP state, even if a valid backup battery is present. In such a situation, the backup domain registers are also reset, along with the VRRTC domain registers.15:16
keriowell, ke-recv isn't thumby15:17
DocScrutinizer05nah, now after that error-msg I quoted, it's back to normal. I inserted USB and it did "swooooosh" and notifier about charging came up15:17
DocScrutinizer05I guess something severely fsckd up in recent KP, regarding fast charger detect15:18
DocScrutinizer05and related stuff15:18
DocScrutinizer05h-e-n had to get over quite some headache due to this abomination of "layering violations" when reading /sys/*/*/charger nukes PHY settings and thus messes up musb-hdrc driver state15:19
DocScrutinizer05we had "some asshole is trying to probe for charger" klog in original h-e-n kernel. dunno what happened to it15:20
keriohah15:20
kerioplease, tell me that was the actual message15:20
DocScrutinizer05yes15:21
DocScrutinizer05verbatim15:21
kerio=D15:21
DocScrutinizer05should've told a story to devels messing with it (/me glares at pali)15:22
kerioPali: *glare*15:22
DocScrutinizer05you MUST NOT probe for D+- short during musb-hdrc has 'allocated' USB to do meaningful stuff like ENUM or data transfer!15:23
DocScrutinizer05~211915:23
infobotThe key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED",  "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.15:23
Palinew kernel-power has patch which disable that sysfs node charger when host mode is active15:23
DocScrutinizer05well, this probelm isn't hostmode specific15:24
Paliwhen musb go back to peripheral mode, then charger detection is enabled again15:24
DocScrutinizer05welcome Pali :-)15:24
DocScrutinizer05Pali: iirc musb-hdrc'.ko' initializes PHY as soon as somebody tries to do something over USB, like e.g. enable mass storage mode, or PC-Suite (or hostmode, yes). Any subsequent probing for D+- short nukes those PHY config settings and renders whole USB in an undefined defunct state15:27
DocScrutinizer05no matter if done via /sys readout, or via direct call to related musc-hdrc function15:28
DocScrutinizer05(aka kernel-internal)15:28
DocScrutinizer05the even more funny stuff about that: PHY gets backed up via main bat even during real device powerdown15:29
DocScrutinizer05so any such mess is relatively sticky, until musb-hdrc takes care to initialize it again the proper way15:29
DocScrutinizer05and!!! there's NO way to really reliably reset PHY, unlike twl403015:30
DocScrutinizer05only reliable reset is: remove main bat, short + and - battery contacts in device. wait 10 minutes (to be sure)15:31
DocScrutinizer05if PHY is misconfigured, even flatbatrecover may fail15:31
DocScrutinizer05since it depends on PHY detecting D+- short and signalling it to bq2415015:32
kerioDocScrutinizer05: why do you need to short + and -? :o15:32
DocScrutinizer05to discharge buffer capacitors15:32
DocScrutinizer05there are quite some huge cubes on Vbat15:32
DocScrutinizer05several 100 uF15:32
kerioDocScrutinizer05: does a working bupbat screw with this, too?15:34
kerioalso, do i have a problem if my usb works?15:34
DocScrutinizer05and there's no *reliable* way to discharge them in a short time, though trying power button during battery removed may help15:34
DocScrutinizer05no, as long as your usb and charging works, no problem15:34
keriohmm, does this also mean that in case of water damage, you're basically screwed?15:34
kerioif removing the battery still leaves quite a bit of charge15:35
DocScrutinizer05if however charging doesn't kick in, or device thinks it's connected to PC while in fact it's connected to fastcharger or to nothing at all, odds are this is cause by PHY15:35
DocScrutinizer05kerio: nope, since the reason for battery removal on water incidents is to stop electrolysis which means rapid discharge of buffer capacitors if they actually seen shorts by water15:36
keriooh, so it basically solves itself15:37
kerioi see15:37
DocScrutinizer05lemme put it this way: if water shorts your + and -, your capacitors are empty during few seconds15:37
DocScrutinizer05for buffering PHY 1707 on a intact device however they may last for hours15:38
keriogot it15:38
kerioto reset PHY, i must dunk my device in water15:38
DocScrutinizer05yeah, kinda15:39
DocScrutinizer05(disclaimer: all the above is not verified on a normal scientific level, but rather just off top of my head without properly checking each detail)15:40
DocScrutinizer05which *usually* is just good enough for any RL purpose ;-D15:40
kerioto be honest i don't remember the last time i connected my n900 to my pc when fremantle was booted15:41
kerioonly backupmenu/bootmenu15:41
DocScrutinizer05but don't call me out on it, when it turns out there is some incorrect or fuzzy bit in the whole story15:41
DocScrutinizer05tbh I *think* I've seen similar lockups on USB with stock kernel15:46
DocScrutinizer05there's a known race condition in kernel, with detecting Vbus vs detecting D+- short15:47
DocScrutinizer05the window for that race might have grown in KP15:47
DocScrutinizer05however so far on my daily iroN900 I'm used to an occasional USB charger detect miss, but that's just a simple ignore and gets fixed by un- and re-plugging fastcharger15:49
kerioyeah, i get that too15:49
DocScrutinizer05this time, on cssu-thumb not even 15 rounds did fix it15:49
kerioanyway, going back to the nicocam thing15:50
keriowhat's your installed version for camera-ui?15:50
DocScrutinizer05i'm done with testing. Noting to add so far15:50
DocScrutinizer05what could it be? latest T15:51
kerio1.1.29.2+0cssu14+thumb0 then?15:51
DocScrutinizer05as well as cssu-thumb as of when I installed thumb on that other device, a few weeks ago15:52
kerioi don't really remember *when* fmg thumbified camera-ui15:52
DocScrutinizer05nfc, I have not even a ssh open to the thumb thingie15:52
keriohe also added maemo-invoker support15:52
kerioand optification15:53
kerio(i think)15:53
DocScrutinizer05that might help a lot, with switching to and from mediaplayer15:53
keriohm, he didn't optify it15:53
keriobut it's maemo-invokered15:53
*** CodenameStrike has quit IRC15:53
DocScrutinizer05anyway, have to get some glimpse of TheBigBlueBox aka RL15:54
DocScrutinizer05o/15:54
keriowatch out, it's dangerous out there15:54
DocScrutinizer05yeah15:54
DocScrutinizer05I watched Monk, I heard the music15:54
DocScrutinizer05...cause it's jungle out there, it's jungle out there!15:55
DocScrutinizer05http://www.magistrix.de/lyrics/Randy Newman/Its-A-Jungle-Out-There-96950.html15:57
DocScrutinizer05http://www.youtube.com/watch?v=HIW69IRODxM15:58
PaliDocScrutinizer05, freemangordon, seems that I found problem17:22
freemangordonPali: where?17:22
Palisee log17:23
Paliproblem is that stupid battery status area plugin cannot handle hal changeset17:23
PaliI changed all calls libhal_changeset_set_property_ to libhal_device_set_property_string17:23
Paliand then status area battery plugin show message "battery is low"17:24
freemangordonPali: which log, there is a wall of text :)17:27
freemangordonbtw why did you change that?17:27
freemangordondo you remember me REing the original bme-addon17:27
Palifreemangordon, find your name in log17:27
Palifreemangordon, I changed it now (only for testing, not in git yet), because statusbar battery plugin not worked17:28
Paliwithout this change statusbar battery plugin did not show me message "charger disconnected" and message "battery is low"17:29
freemangordonaah, i see17:29
freemangordonPali: AFAIK bme status plugin waits for dbus calls, though I might be wrong17:30
freemangordondbus signals even17:30
PaliI'm sending same dbus signals as bme/hald-addon-bme17:30
freemangordonhmm17:30
freemangordondoes it work now?17:30
Paliseems that after I changed code to use libhal_device_set_property_string is working...17:31
DocScrutinizer05Pali: nice :-)17:31
Palinow I'm waiting until battery will be empty :D17:31
freemangordonyeah, great17:31
freemangordonhehe17:31
DocScrutinizer05we now just wanna know why the other stuff been used or if it really been used at all in original bme*, and why this doesn't work in your RE if it's actually different17:32
Palifreemangordon, in hald-addon-bme README.properties.gz is written that for information about "battery low" is need first to change battery.charge_level.capacity_state to "" and then to new real value17:33
DocScrutinizer05I mean, there's probably a difference in semantics between libhal_device_set_property_string and libhal_changeset_set_property_17:33
freemangordonDocScrutinizer05: I didn't do 100% RE replacement (as with VKB)17:34
freemangordonjust basic sceleton17:34
Paliso we need to decide when to send this "signal" battery is empty17:34
DocScrutinizer05aah, so that's been an educated guess17:34
Palioriginal hald-addon-bme sent it more times (you can see that phone show message battery is too low/empty more times...)17:35
DocScrutinizer05well, near miss fair enough then17:35
freemangordonPali: I think the signal is sent every time voltage goes onder the threshold17:35
freemangordons/onder/bellow/17:36
infobotfreemangordon meant: Pali: I think the signal is sent every time voltage goes bellow the threshold17:36
DocScrutinizer05nah, it also has a gate timer, to stop too frequent msgs/signals17:37
freemangordonthe other option is to send a signal on every percent down17:37
freemangordon(after we are in low zone)17:37
PaliDocScrutinizer05, now new hald-addon-bme sending empty state (which cause that dsme will turn device off) when EDV1 or EDVF flag in bq27200 is set17:37
Paliit is ok?17:37
DocScrutinizer05edv1 iirc, yes17:37
freemangordondo we have enough charge for graceful shutdown?17:38
DocScrutinizer05edv1 with some 90s delay, to give user a chance to close stuff17:38
DocScrutinizer05edvf immediately17:38
Palinow both edv1 or edvf immedialety turn device off17:39
freemangordonDocScrutinizer05: what do you think about: <freemangordon> the other option is to send a signal on every percent down17:39
freemangordonsounds like the best option to me17:40
Paliand low state is when bq27200 report capacity below 15%17:40
Pali(7% is for edv1)17:40
freemangordonwell, so we'll have 7 beeps before shutdown17:40
Palilow state = show notification17:40
DocScrutinizer05so: when edv1 got hit for >15s, you send a warning "going to power down!" and start timer for next msg that does real teardown. On either exiry of that timer, or on edvf you send "I'm shutting down NOW!" and do that - unless meanwhile charging kicked in or voltage is above edv1 again17:40
DocScrutinizer05freemangordon: I guess that's exactly what bme does17:41
DocScrutinizer05(per 1%)17:41
freemangordonDocScrutinizer05: only charger should preven shutdown once "shutting down" message appears17:42
freemangordonDocScrutinizer05: yeah, that is what i think too17:42
DocScrutinizer05nah, not really. Battery can recover pretty much from a brownout on heavy load17:42
freemangordonthen no "shutting down" message :)17:43
freemangordonand TBH I think it is better to shutdown with a slightly more charge, than to run battery flast17:43
DocScrutinizer05btw percentage of charge my be vastly unrelated to actual voltage of cell17:43
freemangordon*flat17:44
DocScrutinizer05may*17:44
freemangordonhmm, how precise is remaining percenatge if the battery is not calibrated?17:45
DocScrutinizer05I'll think about it during my brinnerunch17:45
DocScrutinizer05freemangordon: might be totally off17:45
freemangordonI mean after a couple of discharge cycles17:46
freemangordonso bq has had a chance to measure at least something17:46
DocScrutinizer05we had OM GTA02 shut down at ~60% real capacity remaining, due to bq27000 miscalibration and shutdown mechanism only looking for percent-charge rather than voltage17:46
DocScrutinizer05basing shutdown on EDV1/EDVF is sane. Waiting for 0% count down to 0% for warnings isn't17:48
freemangordonI don't think it is a good idea to force calibration only in order to have correct shutdown17:48
freemangordonAIUI that is what Pali does17:48
DocScrutinizer05I'll think about it during my brinnerunch17:48
DocScrutinizer05o/17:48
freemangordon~brinnerunch17:48
DocScrutinizer05dinner+brunch17:49
kerioi say we wait the 10 or so seconds on EDV1 so we can properly calibrate bq2720017:49
freemangordonyou mean auto-calibrate on first battery empty?17:49
DocScrutinizer05I say we only warn on EDV1, and shut down on EDVF17:50
Paliif edv1 is set, then learning cycle is complete17:50
freemangordonagree17:50
freemangordonPali: only if you have a full charge AFAIK17:50
freemangordon(at the start of discharge cycle)17:50
kerioPali: hm?17:50
Paliyes17:50
keriooh girhgt17:50
kerio*right17:50
DocScrutinizer05even a short intermediate charging stops learning cycle17:51
PaliI mean that it is ok to turn device off (if learning cycle was active, then after edv1 is complete)17:51
freemangordonwell, how is learning syscle activated?17:51
freemangordon*cycle17:51
DocScrutinizer05EDVF = End of Discharge Voltage Final. we should take it as that17:52
Palistart charging when edv1 is set and stop charging when battery is charged17:52
keriofreemangordon: VDQ set17:52
kerioaka fully charge17:52
Palithen vdq is set and when edv1 is set again, cycle is complete17:53
freemangordonyou mean there is no need SW to take care of it?17:53
freemangordonto set some bit or something?17:53
keriono17:53
DocScrutinizer05a valid learning cycle includes: charging to 100% -> VDQ:1. discharge without any intermediate charging or other extrema, until EDV1 is hit17:53
kerioit's just a bq27200 thing17:54
freemangordonand if learning cycle is interrupted, what happens with the calibration?17:54
*** krayonWork has quit IRC17:54
DocScrutinizer05untouched17:54
freemangordongreat17:54
freemangordonI don't see any problem then17:54
DocScrutinizer05VDQ:017:54
DocScrutinizer05there17:54
DocScrutinizer05's no problem17:54
keriothe problem is that you need to wait a handful of seconds below EDV1 voltage17:54
keriousually it's not doable with bme17:54
DocScrutinizer05just we shouldn't shut down on EDV117:55
freemangordonkerio: no bme there17:55
kerioindeed17:55
freemangordonDocScrutinizer05: :nod:17:55
freemangordonPali: ^^^17:55
DocScrutinizer05since that per definitionem wwastes 6% of bat capacity17:55
kerioshut down after 60-90s under EDV1 or when you reach EDVF17:55
freemangordonhmm , we'd better just ignore EDV117:56
DocScrutinizer05that's what I said17:56
freemangordonand shutdown on EDVF17:56
DocScrutinizer05no, not that :-)17:56
kerioDocScrutinizer05: we should hold enough charge to boot, though17:56
DocScrutinizer05we start warnings at EDV117:56
kerioboot to ACT_DEAD at least17:56
DocScrutinizer05kerio: that's granted17:56
keriogranted how?17:56
keriohow big is EDVF?17:57
DocScrutinizer05edvf still allows device to boot17:57
kerioi see17:57
DocScrutinizer05hw cutout threshold is waaay lower17:57
DocScrutinizer05anyway, cya. I'm going hypoglycemic17:58
kerioeat some food17:58
keriofood is good!17:58
keriohttp://youtu.be/g-uLTTT3BPc17:58
freemangordonPali: so, AIUI -> start battery low beeps on N% (10-15), beep on every % down, ignore EDV1 and shutdown on EDVF17:59
Paliok17:59
freemangordonBTW we may want to keep the current BME threshol for low warnings18:00
freemangordonor lower it just a bit (2-3 %)18:00
DocScrutinizer05well, increase warning freequency to a timed schedule on EDV1 since bq27200 switches from ~6% to 0% on EDV1 anyway18:00
freemangordonaah, i see18:01
DocScrutinizer05or maybe I recall wrong18:01
DocScrutinizer05read DS please18:01
freemangordonyeah, batt low every 30 seconds should be ok18:01
DocScrutinizer05or ~jr/bme18:01
freemangordon(once we are at EDV1)18:01
DocScrutinizer05there are some nice logs, with comments18:01
DocScrutinizer05~jrtools18:02
infoboti heard jrtools is http://wiki.maemo.org/User:Joerg_rw/tools18:02
DocScrutinizer05../bme*18:02
freemangordonPali: ^^^18:02
DocScrutinizer05http://wiki.maemo.org/User:Joerg_rw/jrbme/bme_discharge_learn_log18:02
DocScrutinizer05http://wiki.maemo.org/User:Joerg_rw/jrbme/bme_chargefloat2_log et al18:03
DocScrutinizer05study those18:03
kerio30 seconds? ew18:03
kerioafter i got the first warning, i *know* the battery is low18:03
DocScrutinizer05http://people.openmoko.org/joerg/battery/batlog.txt.odf18:08
DocScrutinizer05make as many as possible of those timing and level parameters configurable, via module parameters or whatever18:09
DocScrutinizer05so kerio can set that warning frequency to once per day ;-D18:10
kerioomfg i liek configurations18:10
DocScrutinizer05also timeout levels for those "90sec stable EDV1 "18:10
DocScrutinizer05cyacyacyacya18:11
kerioPali: eventually will you rewrite the statusmenu button to allow for detailed data from bq27k and to choose usb host mode?18:12
Palikerio, MAG already started rewriting it18:12
PaliI have pending merge request for it, MAG wrote me that he merge it after he commit some changes...18:13
kerio:D18:13
kerioPali: also, can you make your UI a bit more configurable so we can actually choose the usb speed and/or manually enumerate?18:14
Palikerio, I have no time for it now...18:14
DocScrutinizer05is there a way to detect user clicking away a one-line-notifier?18:14
kerioDocScrutinizer05: make it permanent :)18:14
DocScrutinizer05nah!!!18:15
DocScrutinizer05the idea is to open the permanent requester when user clicks on the bat-low warning18:15
DocScrutinizer05in that requester you can say "don't warn for this low-power-event again"18:16
merlin1991a click on notifier discards it18:16
DocScrutinizer05or set period of warning to something more pleasant for you18:16
DocScrutinizer05merlin1991: that's not what I asked18:16
DocScrutinizer05I know that18:16
merlin1991I'm saying to change that for 1 notifier would be a break of ui18:17
DocScrutinizer05I didn't suggest to change that18:17
keriojust detect whether the user dismissed it or it was just timeout'd18:18
DocScrutinizer05I asked "is it detectable that user did it (click it away)"18:18
merlin1991<DocScrutinizer05> the idea is to open the permanent requester when user clicks on the bat-low warning18:18
merlin1991thus you're suggesting to change the default behaviour even if not implemented in the notifier18:18
merlin1991which is inconsistent to the rest of the ui18:18
DocScrutinizer05hmm, i'll think about that. Maybe you got a pojnt, maybe I don't think it's a good point ;-)18:19
DocScrutinizer05or maybe I do18:19
keriowe could just make it a bigass system notification18:19
kerioone of the ones in the middle, that you *have* to dismiss18:19
* DocScrutinizer05 goes blackout18:20
*** int_ua has joined #maemo-ssu18:20
*** ZogG has quit IRC18:26
*** ZogG has joined #maemo-ssu18:29
*** arcean_ has joined #maemo-ssu18:37
*** arcean has quit IRC18:41
*** arcean_ has quit IRC19:02
DocScrutinizer51after dealing with pending blackout due to *my* charge reaching EDVF, I retought the notifier thingie and I guess merlin1991 is right. Anything related to powerflow-alarm config etc belongs to a requester popping up after clicking battery button in system menu19:19
DocScrutinizer51s/flow/-low/19:20
infobotDocScrutinizer51 meant: after dealing with pending blackout due to *my* charge reaching EDVF, I retought the notifier thingie and I guess merlin1991 is right. Anything related to power-low-alarm config etc belongs to a requester popping up after clicking battery button in system...19:20
kerioDocScrutinizer51: actually, config belongs in Settings19:21
keriothe battery button in the system menu is for live stuff19:21
DocScrutinizer51I however already wanna come up with another weird idea tweaking UI: the battery icon as such is BS for low battery19:21
DocScrutinizer51kerio: exactly what I meant19:22
DocScrutinizer51snoozing a bat-low-alarm *is* a live thing19:22
DocScrutinizer51bat icon is utter crap for detecting change of 1 row of pixels from green to red. I suggest to change the iconset and animation19:24
Palihald-addon-bme changes pushed to git on gitorious19:26
kerioD:19:28
keriodon't do this to me19:28
kerioi'm not at home tomorrow19:28
kerioi really want to try it, but i know i mustn't19:28
Palisee http://pastebin.com/mppAFAKd19:31
PaliEDVF was set when voltage was only 302119:32
Paliit is good idea to wait for EDVF?19:32
PaliDocScrutinizer51, see ^^^19:32
DocScrutinizer51according to bq27200 DS this is incorrect behaviour/voltage19:33
kerioD:19:33
DocScrutinizer51EDV1 set at 3239mV in that log19:38
DocScrutinizer51EDVF at 30nn mV is correect and sane19:38
DocScrutinizer51device is meant to shut down immeidiately though gracefully at tripping EDVF19:39
keriois "LOW" edvf?19:41
DocScrutinizer51I guess19:42
DocScrutinizer51not my version of script19:42
kerio~shadowjk19:42
*** timo^ has joined #maemo-ssu19:42
keriohrmpf19:42
DocScrutinizer51hrmpf?19:42
kerioi thought i added that factoid19:43
DocScrutinizer51you won't add a nick as factoid, will you?19:43
kerioof course19:43
kerio~pali19:43
infoboti heard pali is http://atrey.karlin.mff.cuni.cz/~pali/19:43
*** toxaris has joined #maemo-ssu19:44
*** toxaris has quit IRC19:44
DocScrutinizer51whatever it is you're after, maybe you find it next to19:44
DocScrutinizer51~jrtools19:44
infoboti guess jrtools is http://wiki.maemo.org/User:Joerg_rw/tools19:44
kerioi wanted the url to bq27200.sh19:45
DocScrutinizer51^^^^19:45
DocScrutinizer51../jrbme/19:45
keriooic19:45
DocScrutinizer51and of course at maemo.cloud-7.de19:48
DocScrutinizer51maemo/patches'n'tools/ or sth like that19:49
keriohehe, http://maemo.cloud-7.de/Aegis-kills-device.jpg19:49
*** ZogG has quit IRC20:14
*** arcean_ has joined #maemo-ssu20:16
*** arcean_ is now known as arcean20:29
*** Guest57704 has quit IRC20:30
*** Guest57704 has joined #maemo-ssu20:38
*** timo^ has quit IRC20:39
*** timo^ has joined #maemo-ssu20:55
*** fw190 has joined #maemo-ssu21:02
*** Pali has quit IRC21:06
*** fw190 has quit IRC21:15
*** int_ua has quit IRC21:23
*** timo^ has quit IRC21:29
*** timo^ has joined #maemo-ssu21:30
*** timo^ has quit IRC21:41
*** dafox has joined #maemo-ssu22:03
*** nox- has joined #maemo-ssu22:54
*** dhbiker has quit IRC23:23
*** M13 has joined #maemo-ssu23:29

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!