| *** Milhouse has quit IRC | 00:11 | |
| *** Milhouse has joined #maemo | 00:13 | |
| *** futpib has quit IRC | 00:17 | |
| *** _rd has quit IRC | 00:24 | |
| *** florian has quit IRC | 00:29 | |
| *** antranigv has left #maemo | 00:34 | |
| *** florian has joined #maemo | 01:07 | |
| *** Axel_H has joined #maemo | 01:18 | |
| *** Axel_H has joined #maemo | 01:18 | |
| *** protem has joined #maemo | 01:51 | |
| *** florian has quit IRC | 01:57 | |
| *** drawkula has quit IRC | 02:07 | |
| DocScrutinizer05 | Maxdamantus: actualy tracker, I guess | 02:10 |
|---|---|---|
| DocScrutinizer05 | that goes mad on too many files. Not the camera app itself | 02:11 |
| *** SpeedEvil has quit IRC | 02:13 | |
| *** SpeedEvil has joined #maemo | 02:14 | |
| *** MikaT_ has quit IRC | 02:19 | |
| *** aloril has quit IRC | 02:19 | |
| *** MikaT has joined #maemo | 02:20 | |
| *** aloril has joined #maemo | 02:21 | |
| *** drawkula has joined #maemo | 02:36 | |
| *** pentanol has joined #maemo | 02:37 | |
| *** vectis has quit IRC | 02:38 | |
| *** Venusaur has quit IRC | 02:47 | |
| *** trumee has quit IRC | 02:57 | |
| *** trumee has joined #maemo | 02:58 | |
| *** Venusaur has joined #maemo | 03:03 | |
| *** unclouded has quit IRC | 03:04 | |
| *** L29Ah has left #maemo | 03:05 | |
| *** L29Ah has joined #maemo | 03:06 | |
| *** unclouded has joined #maemo | 03:15 | |
| *** LauRoman has quit IRC | 03:22 | |
| *** beford has joined #maemo | 03:42 | |
| Maxdamantus | DocScrutinizer05: I think I replaced tracker with a script that does nothing. | 03:47 |
| Maxdamantus | oh, or not, but I don't think it does anything. | 03:47 |
| Maxdamantus | Meh, can't remember. | 03:48 |
| *** arossdotme has joined #maemo | 03:59 | |
| *** saponga has joined #maemo | 04:13 | |
| *** saponga has quit IRC | 04:20 | |
| *** eMHa__ has joined #maemo | 04:25 | |
| *** trumee has quit IRC | 04:25 | |
| *** trumee has joined #maemo | 04:28 | |
| *** eMHa_ has quit IRC | 04:28 | |
| *** beford has quit IRC | 04:54 | |
| *** Humpelst1lzchen has joined #maemo | 04:58 | |
| *** Humpelstilzchen has quit IRC | 04:59 | |
| *** lxp has quit IRC | 05:06 | |
| *** saponga has joined #maemo | 05:24 | |
| *** sparetire_ has quit IRC | 05:25 | |
| *** saponga has quit IRC | 05:44 | |
| *** Luke-Jr has quit IRC | 06:01 | |
| *** robbiethe1st has joined #maemo | 06:05 | |
| *** Roth has joined #maemo | 06:08 | |
| *** supergaban has joined #maemo | 06:12 | |
| *** supergaban has quit IRC | 06:16 | |
| *** L29Ah has left #maemo | 06:19 | |
| *** L29Ah has joined #maemo | 06:20 | |
| *** L29Ah has left #maemo | 06:20 | |
| *** beford has joined #maemo | 06:25 | |
| *** Smily has quit IRC | 06:44 | |
| *** Smily has joined #maemo | 06:44 | |
| *** Oksana has quit IRC | 07:04 | |
| *** Oksana_ has joined #maemo | 07:09 | |
| *** Oksana_ has quit IRC | 07:10 | |
| *** Oksana_ has joined #maemo | 07:16 | |
| *** Oksana_ is now known as Oksana | 07:22 | |
| *** robbiethe1st has quit IRC | 07:23 | |
| *** protem has quit IRC | 07:33 | |
| KotCzarny | ds3: there are cmdline exif renamers | 07:37 |
| *** Luke-Jr has joined #maemo | 07:45 | |
| *** _rd has joined #maemo | 07:48 | |
| *** delt has joined #maemo | 07:49 | |
| delt | Hello | 07:50 |
| delt | my n810 is completely dead all of a sudden... won't turn on. | 07:50 |
| KotCzarny | battery dead? | 07:50 |
| delt | tried pulling the battery and putting it back, no effect | 07:50 |
| KotCzarny | charge in other device? | 07:50 |
| delt | KotCzarny: nope... it's plugged in | 07:50 |
| delt | even when plugged in, i hold down the power button and it's just dead...won't start. | 07:51 |
| delt | is there any firmware trick i could use in a case like this? | 07:51 |
| KotCzarny | dont remember if n8x0 needs good battery to work, but with n900 you cant run without | 07:51 |
| delt | just plugged it into the usb port with the (weird) cable that came with it... still nothing. doesn't show up in 'lsusb' (running linux mint on the pc) | 07:53 |
| *** japa-fi has joined #maemo | 07:54 | |
| Maxdamantus | I don't think the N900 would show up in lsusb if it's unable to power on (dunno about N8x0) | 07:55 |
| delt | this is a n810 | 07:55 |
| delt | oh ok... sorry, misread -) | 07:55 |
| delt | yeah, normally if the device doesn't power up it won't show in the lsusb listing.... | 07:56 |
| delt | anyway. i just read something on the web about removing the battery for at least 15 minutes. Even though this one has been dead and ununsed for some time, i'll give it a try... | 07:57 |
| *** _rd has quit IRC | 07:59 | |
| delt | same page says to flash it with some new rom update.. how do i do that if it doesn't even power on?? | 07:59 |
| delt | anyway.. there's no trick like hold down a certain button while plugging it, or push power 5 times before, etc etc...? | 08:00 |
| Maxdamantus | with the N900 you just have to let it charge at a slower rate while it's off. | 08:05 |
| Maxdamantus | unless maybe you have a dumb charging source (indicated by the shorted data pins) | 08:05 |
| Maxdamantus | I'd try plugging it into a charger you know does that, if you have one. | 08:06 |
| Maxdamantus | (official Nokia chargers should, other chargers often won't) | 08:06 |
| delt | yeah, the charger i have for this is a foreign (i think european) plug (i'm in Canada) plugged into an adapter to make it fit into the normal wall plug.. | 08:07 |
| delt | how do i tell about the "shorted data pins" though? | 08:08 |
| delt | (preferably without cutting it to pieces) | 08:08 |
| *** Smily has quit IRC | 08:15 | |
| delt | plugged the charger into another socket.... nope. wont even flash a led or something, completely dead. | 08:16 |
| *** _rd has joined #maemo | 08:17 | |
| Maxdamantus | Have you tried just leavin it plugged in for a while? | 08:17 |
| delt | yeah, it was plugged in for months, and always worked. suddenly it just decided to die. | 08:18 |
| Maxdamantus | I mean over a shorter period, like half an hour or so. | 08:18 |
| Maxdamantus | and not trying to turn it on during that time. | 08:19 |
| delt | ok, i'll try that | 08:19 |
| delt | after being unplugged for how long? | 08:20 |
| Maxdamantus | Not being unplugged. | 08:20 |
| Maxdamantus | just leave it plugged in for half an hour, then turn it on. | 08:20 |
| inz | delt, as N810 does not do USB charging, the shorted data pins does not apply | 08:20 |
| Maxdamantus | Ah. | 08:21 |
| delt | inz: yeah, i remember something about that... | 08:21 |
| delt | Maxdamantus: it was plugged in for weeks/months. now what :D | 08:21 |
| delt | i could always see the green/blue-ish led in the top left corner pulsating/flashing, but it just stopped. went dead. | 08:23 |
| *** Roth has quit IRC | 08:24 | |
| Maxdamantus | The point with leaving it plugged in (at leat, if it charges from USB, and abides by the specificaotion as the N900 does) is it might otherwise use the small amount of charge it has to power on but not get to the point where it can draw more power (from USB, by negotiating with the host.) | 08:25 |
| *** _rd has quit IRC | 08:27 | |
| delt | heh... refigerator trick =) ---> http://techieworldz.blogspot.ca/2012/04/nokia-n810-dead-battery-issue.html | 08:27 |
| KotCzarny | i think that battery just died | 08:28 |
| KotCzarny | buy new/used one | 08:29 |
| KotCzarny | plugged in for months == 100% charge all the time == death for li-ions | 08:29 |
| delt | yeah, must be the case :/ | 08:29 |
| KotCzarny | luckily you can have cheap good replacement | 08:29 |
| KotCzarny | look for polarcell brand | 08:29 |
| delt | i will, when i have a few $$ for it -) | 08:32 |
| delt | ...but even plugged in, the n810 won't turn on without a battery? (i don't remember, haven't checked) | 08:33 |
| inz | I wonder how long my N810 will keep serving web pages, being constantly plugged in | 08:34 |
| DocScrutinizer05 | delt: forget about "shorted data pins", N810 carges from 1.5mm(?) barrel connector | 08:34 |
| DocScrutinizer05 | ooh inz beat me to it | 08:34 |
| delt | inz: in fact i do remember seeing a web server for it (: | 08:35 |
| inz | Doc, IIRC it's 2.0, way too thin to bear any kind of load anway :/ | 08:36 |
| DocScrutinizer05 | yep | 08:36 |
| DocScrutinizer05 | one bent, but worked. Then the PSU broke. Used the PSU from spare. Then the N810 suddenly didn't come up from one of its random reboots it did every 2 weeks to 3 months | 08:38 |
| DocScrutinizer05 | looks like corrupted filesystem, or defect MMC | 08:38 |
| DocScrutinizer05 | could try to reflash it, didn't get around to it | 08:39 |
| DocScrutinizer05 | well, it worked like 7 years 24/7/350 | 08:39 |
| DocScrutinizer05 | N810 charger is 100% software controlled, so it suffers massively from flatbat deadlock | 08:43 |
| DocScrutinizer05 | try charging battery in external charger | 08:44 |
| DocScrutinizer05 | harger here is internal charging circuit, not the wallcharger | 08:45 |
| *** ecc3g has quit IRC | 08:45 | |
| DocScrutinizer05 | the "walcharger" is a silly DC power supply | 08:45 |
| delt | <--- *nods* | 08:45 |
| KotCzarny | i think n8x0 can be charged via usb too | 08:46 |
| * DocScrutinizer05 soesn't even know the voltage | 08:46 | |
| DocScrutinizer05 | no | 08:46 |
| KotCzarny | i should check when i get home | 08:46 |
| KotCzarny | but i do remember using usb plug for charging | 08:47 |
| DocScrutinizer05 | check ~flashing, old version for N8x0. It clearly says "plug in to USB. THEN plug in power and power it up" | 08:48 |
| DocScrutinizer05 | N810 doesn't boot up and doesn't charge from USB connection | 08:49 |
| KotCzarny | you might be right | 08:50 |
| delt | nope.... zero sign of life on this thing. completely dead. | 08:50 |
| KotCzarny | which is a pity, because charges is 5V anyway | 08:51 |
| KotCzarny | AC-4U - Output 5v 890 mA | 08:51 |
| delt | fuck :( | 08:53 |
| delt | anyway... thanks for your help guys | 08:53 |
| *** delt has left #maemo | 08:53 | |
| *** kerio has quit IRC | 09:05 | |
| *** kerio has joined #maemo | 09:06 | |
| *** ecc3g has joined #maemo | 09:08 | |
| DocScrutinizer05 | technically it *might* charge from USB. at least it has all the wires needed: http://wstaw.org/m/2015/09/18/plasma-desktopoQ1947.png | 09:09 |
| DocScrutinizer05 | see hidden docs | 09:15 |
| DocScrutinizer05 | don't dare to post full URLs! | 09:15 |
| DocScrutinizer05 | or they will be gone the same moment | 09:16 |
| *** LauRoman has joined #maemo | 09:21 | |
| DocScrutinizer05 | however http://www.ti.com/product/tps65030 simply doesn't support chraging battery from VBUS | 09:41 |
| *** geaaru has joined #maemo | 09:56 | |
| *** kerio has quit IRC | 10:13 | |
| *** Ex-Opesa has quit IRC | 10:17 | |
| *** vectis has joined #maemo | 10:22 | |
| *** florian has joined #maemo | 10:28 | |
| *** pentanol has quit IRC | 10:32 | |
| *** kerio has joined #maemo | 10:37 | |
| *** kerio has quit IRC | 10:41 | |
| *** kerio has joined #maemo | 10:51 | |
| *** kerio has quit IRC | 10:51 | |
| *** vectis has quit IRC | 10:52 | |
| *** kerio has joined #maemo | 10:54 | |
| *** filosofem_ has quit IRC | 10:56 | |
| *** FlameReaper-PC has quit IRC | 11:24 | |
| *** sparetire_ has joined #maemo | 11:45 | |
| *** filosofem has joined #maemo | 12:01 | |
| *** LauRoman has quit IRC | 12:21 | |
| *** zGrr has joined #maemo | 12:34 | |
| *** _rd has joined #maemo | 12:38 | |
| DocScrutinizer05 | http://www.bildschirmarbeiter.com/content/images/picdump-15-09-18/picdump-15-09-18-001.png | 12:48 |
| *** _rd has quit IRC | 13:08 | |
| *** L29Ah has joined #maemo | 13:14 | |
| *** lxp has joined #maemo | 13:15 | |
| *** FlameReaper has joined #maemo | 13:19 | |
| Sicelo | naughty Germans! | 13:34 |
| *** _rd has joined #maemo | 14:11 | |
| *** _rd has quit IRC | 14:21 | |
| Vajb | action packed 9 hours. | 14:21 |
| *** ecc3g has quit IRC | 14:36 | |
| *** ecc3g has joined #maemo | 14:43 | |
| merlin1991 | Sicelo: well one in there is from austria ;) | 14:48 |
| KotCzarny | aren't austrians kind of germans? | 14:50 |
| Wizzup | KotCzarny: don't let them hear you say that ;) | 14:50 |
| KotCzarny | :) | 14:50 |
| Maxdamantus | I thought they were osteriches. | 14:51 |
| *** ssvb has quit IRC | 14:53 | |
| *** _rd has joined #maemo | 15:15 | |
| Sicelo | haha Maxdamantus. you're asking for trouble :D | 15:31 |
| *** xes has quit IRC | 15:42 | |
| *** darkschneider has quit IRC | 15:44 | |
| *** darkschneider has joined #maemo | 15:44 | |
| *** _rd has quit IRC | 15:48 | |
| *** beford has quit IRC | 16:02 | |
| *** ssvb has joined #maemo | 16:04 | |
| *** lobito has quit IRC | 16:07 | |
| *** lobito has joined #maemo | 16:07 | |
| *** xes has joined #maemo | 16:09 | |
| *** arossdotme has quit IRC | 16:21 | |
| *** shentey has joined #maemo | 16:28 | |
| *** arossdotme has joined #maemo | 16:32 | |
| *** shentey has quit IRC | 16:54 | |
| KotCzarny | ~lcd | 18:23 |
| infobot | methinks lcd is liquid crystal display. See http://www.earthlcd.com/ for inexpensive lcd hardware, and http://www.repairfaq.org/filipg/LINK/F_LCD_menu.html for good background information. | 18:23 |
| KotCzarny | any howto on screen replace for n900? | 18:23 |
| Sicelo | L1-2 manual? | 18:24 |
| KotCzarny | hum | 18:25 |
| L29Ah | infobot: hey can you google | 18:25 |
| L29Ah | infobot: google shit | 18:25 |
| L29Ah | i think there must be a video on youtube | 18:25 |
| *** Sicelo is now known as qhubekela | 18:27 | |
| *** qhubekela is now known as Sicelo | 18:28 | |
| KotCzarny | doh, some idiot screwed the screen with aluminum torx screw | 18:48 |
| Vajb | why did it matters? | 18:49 |
| Vajb | or u break the head? | 18:50 |
| L29Ah | b-but it's some milligrams lighter! | 18:51 |
| KotCzarny | it matters because: im trying to replace bad screen with a good screen, and i dont have torx+ screwdriver atm | 18:52 |
| KotCzarny | also, aluminum screws tend to flatout easily | 18:52 |
| *** florian has quit IRC | 18:54 | |
| Vajb | that is true | 18:55 |
| Vajb | hmm no way to use pliers to remove it? | 18:55 |
| KotCzarny | um | 19:00 |
| KotCzarny | will require breaking the plastics, luckily i managed to unscrew it | 19:00 |
| KotCzarny | Valkommen | 19:00 |
| KotCzarny | anyway, device booting | 19:01 |
| KotCzarny | lets see how much the seller f*cked me up | 19:01 |
| KotCzarny | 'good condition without any problems' turned out as: broken screen, empty battery and some sticky goo inside | 19:06 |
| *** shentey has joined #maemo | 19:08 | |
| KotCzarny | but at least it booted, lets see if i can get any discriminate pictures of the f*cker | 19:10 |
| buZz | why dont you have torx screwdrivers :? | 19:18 |
| KotCzarny | because i'm at gf's place | 19:18 |
| buZz | seems they are used in a shitton of mobile hw | 19:18 |
| KotCzarny | luckily in n900 they are only used for motherboard | 19:19 |
| buZz | my last gf had more tools than me :D | 19:19 |
| KotCzarny | :) | 19:19 |
| KotCzarny | umkay | 19:21 |
| KotCzarny | modem chip seems ok | 19:21 |
| KotCzarny | lets see the other things | 19:22 |
| buZz | can a n900 with broken baseband still work? | 19:22 |
| buZz | i desire a faster replacement for my n800, but refuse to have a GSM | 19:22 |
| KotCzarny | yes, i have one (the lcd donor) that has modem chip problem | 19:22 |
| KotCzarny | but it boots and wifi works | 19:22 |
| buZz | cool! | 19:22 |
| buZz | guess i should hunt one :) | 19:22 |
| KotCzarny | also, install tablet mode | 19:23 |
| buZz | maybe drill through modemchip just to be sure | 19:23 |
| KotCzarny | it turns off gsm chip i think | 19:23 |
| buZz | i analyzed my android with 'airplane mode' on, still does GSM comm | 19:23 |
| buZz | even without a SIM inside | 19:23 |
| KotCzarny | check your n900 | 19:23 |
| buZz | i have none | 19:23 |
| KotCzarny | o.o | 19:23 |
| buZz | just n800 since couple weeks | 19:23 |
| buZz | found in someone's estate after they died | 19:24 |
| KotCzarny | btw. do you listen to music? | 19:24 |
| ecc3g | ARRGGGH DIE ANDROID DIE... | 19:24 |
| buZz | lol | 19:24 |
| ecc3g | sorry | 19:24 |
| buZz | KotCzarny: yes :D | 19:24 |
| KotCzarny | buzz: | 19:24 |
| KotCzarny | ~oscp | 19:24 |
| infobot | rumour has it, oscp is http://talk.maemo.org/showthread.php?t=94590 | 19:24 |
| buZz | KotCzarny: ooo that was you :D | 19:24 |
| buZz | is it in extra repo? | 19:24 |
| KotCzarny | buzz: dont remember, should be | 19:24 |
| buZz | *open app manager* *wait until cpu drops off 100%* | 19:25 |
| buZz | :( | 19:25 |
| KotCzarny | http://repository.maemo.org/extras-devel/pool/diablo/non-free/o/oscp/ | 19:25 |
| KotCzarny | just grab the deb | 19:25 |
| KotCzarny | umkay, wifi works | 19:26 |
| KotCzarny | hooray! | 19:26 |
| buZz | \o/ | 19:26 |
| KotCzarny | so that leaves me with 3 fully working n900s and 1 with broken screen/modem chip | 19:27 |
| KotCzarny | still, not bad. will be good a source of spare parts :) | 19:27 |
| buZz | n900 is also 800x480, isnt it? | 19:27 |
| KotCzarny | yes | 19:27 |
| buZz | 0.9.7-27 ? | 19:27 |
| buZz | of oscp | 19:27 |
| buZz | thats in repo | 19:28 |
| buZz | ah yep, latest | 19:28 |
| KotCzarny | dont remember the dependencies, but should be strict system libs | 19:28 |
| buZz | we'll see :) | 19:29 |
| buZz | was using the built-in mediaplayer but thats really a HORRIBLE interface | 19:29 |
| buZz | such bad | 19:29 |
| KotCzarny | yes. you will love oscp then | 19:29 |
| KotCzarny | it has both, ncurses and remote interface | 19:29 |
| KotCzarny | and example pygtk interface as a clicky gui | 19:30 |
| *** _rd has joined #maemo | 19:30 | |
| buZz | remote interface has a hildon interface? | 19:30 |
| KotCzarny | no, network interface so you can control oscp-core from pc for example | 19:30 |
| buZz | hmm weird, installing from app manager doesnt work | 19:30 |
| KotCzarny | or from n900, or another n8x0 | 19:30 |
| buZz | i'll fix it later | 19:30 |
| KotCzarny | dpkg -i some.deb ? | 19:30 |
| Sicelo | and vote when done testing, buZz :p | 19:32 |
| *** vectis has joined #maemo | 19:32 | |
| buZz | vote? | 19:32 |
| KotCzarny | to move packages from extras-devel into extras | 19:33 |
| KotCzarny | if package is good and worthy | 19:33 |
| KotCzarny | (and if passes QA, but that's history) | 19:33 |
| *** _rd has quit IRC | 19:46 | |
| *** ced117 has joined #maemo | 19:48 | |
| *** ced117 has quit IRC | 19:49 | |
| KotCzarny | audio works. | 19:49 |
| *** zGrr has quit IRC | 19:54 | |
| *** xray256_ has joined #maemo | 19:55 | |
| *** Oksana has quit IRC | 19:57 | |
| *** xray256 has quit IRC | 19:57 | |
| *** Oksana has joined #maemo | 19:58 | |
| KotCzarny | hmm, packages for n900 should have additional field. battery-drain | 20:44 |
| Sicelo | haha. meaning? | 20:47 |
| KotCzarny | meaning min/max/avg mA for app | 20:47 |
| KotCzarny | measured for 24h | 20:47 |
| buZz | lol | 20:49 |
| buZz | you mean, for maemo | 20:49 |
| Sicelo | i don't know how useful that'd be... some applications, by their function, are already guranteed to take a lot of juice | 20:50 |
| Sicelo | think map applications | 20:50 |
| buZz | or a mp4 video player :P | 20:51 |
| buZz | or quake3 | 20:51 |
| KotCzarny | sicelo, but minimal would be for idle, maximum for max | 20:51 |
| KotCzarny | that would help weed out battery vampires | 20:51 |
| *** FlameReaper has quit IRC | 20:51 | |
| *** florian has joined #maemo | 20:55 | |
| *** futpib has joined #maemo | 21:00 | |
| *** geaaru has quit IRC | 21:04 | |
| *** L29Ah has left #maemo | 21:09 | |
| *** infobot has quit IRC | 21:10 | |
| *** infobot has joined #maemo | 21:11 | |
| *** L29Ah has joined #maemo | 21:14 | |
| *** L29Ah has left #maemo | 21:24 | |
| *** shentey has quit IRC | 21:24 | |
| DocScrutinizer05 | Sicelo: most are not, and that's why such value is mostly useless. Any decent app's power consumption will usually depend *massively* on user interaction with it. An app that's drawing power while in background is most likely doing sth bad | 21:25 |
| DocScrutinizer05 | for those apps that actually draw power even while not inter-*active* it might actually be useful to document *that* they do and how much power they draw | 21:26 |
| ds3 | how do you propose to do that? | 21:27 |
| KotCzarny | ibstall app, run for 24h, observe data | 21:28 |
| DocScrutinizer05 | however I don't see anybody equipped and willing to do such evaluation. after all you need to do an A/B test with exactly same conditions. IOW let same device rest idle for 24H two times, once with app running and once without. Do that 3 or 4 times, to rule out unrelated and statistical factors (bad weather, poor reception, device getting flooded by pings which you woun't even notice) | 21:29 |
| ds3 | but how do you know it is that app and not some other component? | 21:29 |
| ds3 | each time I had to look at something like this, the only 'reliable' way is to look at sleep stats for a process | 21:30 |
| DocScrutinizer05 | and of course you must make sure that no other apps (IRC!!!) are running during this week of testing on the device | 21:30 |
| KotCzarny | ds3: same default test system | 21:30 |
| KotCzarny | ie. one device just for that | 21:30 |
| ds3 | what would you conclude if an app sends out a dbus message then sleeps. the app strictly speaking doesn't draw much/an power but that dbus command can do some power guzzling thing | 21:31 |
| ds3 | is that the app that is charged with the power guzzling or not? | 21:31 |
| KotCzarny | ds3, worst offenders are those drawing power in idle | 21:31 |
| DocScrutinizer05 | then the question is: which apps would that be that are even _candidates_ for such tests? Obviously only logging apps and online apps qualify, since other apps shouldn't even stand the testing promotion due to eating resources for no obvious reason | 21:32 |
| ds3 | KotCzarny: in my example, the app is not drawing power when idle yet from a system point of view, the app started something that is drawing power | 21:33 |
| KotCzarny | ds3: thing is, if you install an app, then you see radical power draw, then it is this app related | 21:33 |
| KotCzarny | its not perfect, but could be a start | 21:34 |
| ds3 | got it, I see which direction you are approaching it from | 21:34 |
| ds3 | problem I see with that is it makes it harder to fix the problem (short of removing the app) | 21:35 |
| KotCzarny | but at least you will know why there is unexpected battery drain | 21:36 |
| ds3 | DocScrutinizer05: even off line games would be candidates | 21:36 |
| ds3 | seen stuff that redraws too often | 21:36 |
| DocScrutinizer05 | that's clearly a loser on testing | 21:36 |
| DocScrutinizer05 | see criteria list for apps in maemo-testing | 21:37 |
| ds3 | has the CSSU folks looked into sandboxing apps? | 21:38 |
| DocScrutinizer05 | *any* app in background MUST NOT (2119) draw any power (or resources except RAM in general) unless it does some updating on other stuff than a timer | 21:38 |
| DocScrutinizer05 | no, sandboxing apps is not a concept of maemo | 21:38 |
| ds3 | sand boxing in a different way then Android | 21:39 |
| ds3 | You can use containers to freeze apps or limit their swapping | 21:39 |
| ds3 | that way apps that does draw power can be frozen | 21:39 |
| KotCzarny | um, can linux do that? | 21:39 |
| DocScrutinizer05 | I thought about a gneric system sanity monitor that checks unusual resoure consumption of any kind | 21:40 |
| DocScrutinizer05 | and tags the possible causes | 21:40 |
| ds3 | (and things like the Browser (microB?) can be put on ice or limited in swapping so it won't render the N900 useless when a bad page loads) | 21:40 |
| ds3 | KotCzarny: Linux containers can do stuff like that | 21:40 |
| KotCzarny | ds3: containers as in virtual machines? or something else? | 21:41 |
| DocScrutinizer05 | err, for such stuff unix has ulimits and kill | 21:41 |
| ds3 | 2 things that kill me - Modest and MicroB | 21:41 |
| ds3 | KotCzarny: not as heavy weight as a full VM | 21:41 |
| DocScrutinizer05 | maemo even has OOM | 21:41 |
| DocScrutinizer05 | aka OHM | 21:41 |
| ds3 | those things kick in too late | 21:42 |
| DocScrutinizer05 | then reconfigure them | 21:42 |
| *** L29Ah has joined #maemo | 21:42 | |
| ds3 | sometimes I have 30seconds to pull up an old SMS.... if it is busy swapping, I have lost my 30seconds | 21:42 |
| *** L29Ah has left #maemo | 21:42 | |
| DocScrutinizer05 | ? | 21:43 |
| *** L29Ah has joined #maemo | 21:43 | |
| DocScrutinizer05 | when you have problems "pulling up old SMS" you probably should delete your eventsdb | 21:43 |
| ds3 | Modest/MicroB winds up using 95% of the memory so to do anything else you have to wait for them to page out | 21:43 |
| ds3 | the problem is I cannot get to conversations to try to pull it out | 21:43 |
| DocScrutinizer05 | yes, web browsers are a pest, or rather 'modern' webpages are (and those twits that code them). Anyway that seems to be fixable with a simple ulimit, no? | 21:45 |
| ds3 | Modest seems to have a memory leak (AFAICT)... combine that with a web page being open and I am in deep crap | 21:45 |
| ds3 | ulimit kills things | 21:45 |
| ds3 | i want to be able to change the swap policy on just the browser (which is what you can do with containers) | 21:45 |
| DocScrutinizer05 | how would that help? | 21:45 |
| KotCzarny | or configure browser to use smaller mem cache? | 21:45 |
| DocScrutinizer05 | we already have swap pressure 100 | 21:46 |
| ds3 | the browser can be forced to lagged more (more pre-emptive paging out) | 21:46 |
| ds3 | the other stuff has a better chance of not having to wait for free mem before it can page in/start | 21:46 |
| DocScrutinizer05 | and no, afaik a task doesn'T get killed when it runs into mem ulimit. It simply gets a 'out of memory' return code on malloc. At least that's what I'd expect to happen | 21:47 |
| ds3 | put it another way- the browser can act as if can use no more then 70% of real memory before it pages. (currently, any process can use 100%) | 21:47 |
| DocScrutinizer05 | ((more pre-emptive paging out)) again, we already have swappiness=100 | 21:47 |
| ds3 | and most programs kill themselves when malloc returns NULL | 21:48 |
| KotCzarny | so simply browser needs to know its limits | 21:48 |
| ds3 | and ulimit doesn't effect swapping either | 21:48 |
| KotCzarny | as i said, lower (or disable mem cache) | 21:48 |
| *** _rd has joined #maemo | 21:48 | |
| ds3 | KotCzarny: doesn't seem to be useful. tried it. | 21:48 |
| DocScrutinizer05 | again, we already have swappiness=100 | 21:48 |
| ds3 | DocScrutinizer05: that is for the whole system | 21:49 |
| ds3 | consider things - | 21:49 |
| DocScrutinizer05 | makimum paging-out pressure for all apps | 21:49 |
| ShadowJK | Any paging activity at all will cause enormous latency for the entire system anyway, even if it's just one process causing pageouts | 21:49 |
| ds3 | you open a "bad" web page. browser expands to fill all free RAM | 21:49 |
| ds3 | now you try to do something else - say start Conversations. | 21:49 |
| ds3 | since there is no free RAM, it has to page out enough to load in Conversation == Lag. | 21:50 |
| ds3 | what I am say is: put the brower in a container and restrict it to using 70% free RAM. | 21:50 |
| ds3 | now if I start Conversations, it will have RAM to start. | 21:50 |
| KotCzarny | or: write a browser/proxy that converts big images on the fly | 21:50 |
| ds3 | what the container does is make the process act as if it had less ram so it swaps out earlier | 21:51 |
| ds3 | I am willing to as far as to hook into the UI and put the browser process on ice when it looses focus | 21:51 |
| ds3 | (this has bad side effects like background loading of pages stop working) | 21:52 |
| KotCzarny | because 'bad site' is usually something with big gallery | 21:54 |
| KotCzarny | but, if 4000x3000 images got rescaled into 800x600 or something, and there are 8-40 images, it saves quite a big chunk of memory | 21:55 |
| ds3 | there are more bad stuff then that | 21:56 |
| KotCzarny | also, would require bad sites that just do width=somebignumber | 21:56 |
| DocScrutinizer05 | it has way more bad effects, like browser not _freeing_ any memory while frozen | 21:56 |
| ds3 | ideally the browser should cooperate. the containers approach is a big hammer | 21:57 |
| DocScrutinizer05 | exactly | 21:57 |
| DocScrutinizer05 | and browser backend is FOSS, so rewrite it | 21:57 |
| ds3 | i wonder if there is performance gains to swapping out to the uSD card instead | 21:57 |
| ds3 | <--- not an app guy | 21:57 |
| Wizzup | ds3: why not juse use a proper rbac + limits | 21:57 |
| Wizzup | I'm allergic to the word "container" | 21:58 |
| ds3 | Wizzup: rbac? | 21:58 |
| Wizzup | role based access control | 21:58 |
| ds3 | Wizzup: container in general or container as implemented in Linux? | 21:58 |
| Wizzup | container as in the popular term for container | 21:58 |
| ds3 | ah | 21:58 |
| KotCzarny | another thing, compressed swap | 21:58 |
| Wizzup | where many layers are virtualised, and people think they're safe that way | 21:58 |
| Wizzup | for example grsec's rbac is awesome | 21:58 |
| ds3 | the container stuff in Linux would give you a form of rbac | 21:59 |
| Wizzup | I don't think that's an rbac | 21:59 |
| Wizzup | just virtualisation | 21:59 |
| ds3 | if I had the time, I woudl shove all of android in a container ;) | 21:59 |
| Wizzup | selinux, apparmor and grsec's rbac are rbacs | 21:59 |
| DocScrutinizer05 | I guess the overhead would be higher than the gain | 21:59 |
| Wizzup | generally, rbacs do not have a large impact on speed | 21:59 |
| ds3 | Wizzup: I have a looser definition of 'role' | 21:59 |
| ds3 | selinux is too painful to configure in a useful way. lots of finegrain controls and lots of headaches | 22:00 |
| Wizzup | selinux is bloat, yes | 22:00 |
| ds3 | it comes down to is - the real world and the CS theories being taught is not quite compatible | 22:00 |
| Wizzup | I don't believe that shit | 22:01 |
| ds3 | but since new CS grads are cheap, we have lots of crappy SW | 22:01 |
| *** L29Ah has left #maemo | 22:01 | |
| ds3 | Wizzup: I mean writing a policy for selinux is very painful | 22:01 |
| Wizzup | grsec's rbac - https://en.wikibooks.org/wiki/Grsecurity/The_RBAC_System / | 22:01 |
| Wizzup | ds3: yes, I gave selinux as an example of rbac, not my suggested rbac | 22:01 |
| DocScrutinizer05 | why not just LD_PRELOAD-patch malloc (and whatever does handle page faults) for processes you want to make more "swappy" | 22:01 |
| Wizzup | ld preload is unsafe | 22:02 |
| *** ced117 has joined #maemo | 22:02 | |
| DocScrutinizer05 | unsafe | 22:02 |
| Wizzup | I mean, if you know what you're running, you can use ldp | 22:02 |
| DocScrutinizer05 | oh well | 22:02 |
| Wizzup | for all container purposes it is not | 22:02 |
| Wizzup | s/container/containing/ I guess | 22:02 |
| ds3 | ld preload fails under many conditions | 22:02 |
| DocScrutinizer05 | whatever, afk | 22:03 |
| Wizzup | ds3: I don't think that it makes sense to create virtual worlds where $app can do whatever it wants, and then say it's safe because likely the virtual world is isolated enough | 22:03 |
| DocScrutinizer05 | busy bringing 1GB RAM to that OS | 22:03 |
| Wizzup | DocScrutinizer05: yay! | 22:03 |
| ds3 | Mr Bringing a Bigger Hammer ;) | 22:03 |
| Wizzup | haha | 22:03 |
| ds3 | more RAM == more power usage ;) | 22:04 |
| Wizzup | Depends on the efficiency | 22:04 |
| KotCzarny | i wonder if n900's ram chip could be resoldered with something bigger | 22:04 |
| ds3 | unless someone wants to port the ram bank power management patches | 22:04 |
| ds3 | KotCzarny: yes. but the trick is finding one and getting it initialized correctly | 22:04 |
| KotCzarny | i remember old wrt54g had such feature if auto configuring bigger chip | 22:05 |
| DocScrutinizer05 | anyway swapping takes ages. When you want brwoser to start earlier with swapping, I'd actually prefer it getting killed on malloc right away | 22:05 |
| ds3 | the bootloader needs to shove in right timings or the memory controller won't be happy | 22:05 |
| KotCzarny | or swap out its mem cache in big contiguous chunks | 22:06 |
| DocScrutinizer05 | as soon as you bring it to background, it's supposed to swap out everything anyway | 22:06 |
| DocScrutinizer05 | ds3: we did | 22:07 |
| DocScrutinizer05 | and power consumption is not THAT bad | 22:07 |
| ds3 | DocScrutinizer05: is that source renewable or a lifetime buy? | 22:08 |
| DocScrutinizer05 | LPDDR is pretty humble in idle mode | 22:08 |
| DocScrutinizer05 | well, who knows | 22:08 |
| ds3 | I remember the guys doing that patch had numbers that indicated noticeable savings | 22:08 |
| DocScrutinizer05 | right now I could get another 2000 | 22:08 |
| ds3 | DocScrutinizer05: is that the end of it? (the other 2000) | 22:09 |
| DocScrutinizer05 | dunno | 22:09 |
| ds3 | if it is a bootloader compatible piece, I am tempted to rework an N900 | 22:09 |
| DocScrutinizer05 | I bet there are more | 22:09 |
| ds3 | the hardest part should be removing the cans | 22:09 |
| DocScrutinizer05 | no | 22:10 |
| ds3 | ? | 22:10 |
| DocScrutinizer05 | cans are easy. PoP is hard | 22:10 |
| ds3 | what's the harder problem besides re-POP'ing? | 22:10 |
| ds3 | POP is easy. done it on the Beagle already | 22:10 |
| ds3 | :D | 22:10 |
| DocScrutinizer05 | we too | 22:11 |
| ds3 | just can't do it too many times or PCB delaminates | 22:11 |
| KotCzarny | do i have to take out sim for reflash or it doesnt matter? | 22:11 |
| DocScrutinizer05 | I sent some 3 chips to a guy who planned to rework his N900. Never again heard of him | 22:11 |
| KotCzarny | ~flashing | 22:12 |
| infobot | rumour has it, maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh | 22:12 |
| ds3 | *shrug* | 22:12 |
| DocScrutinizer05 | we have 2 BB-xM with 1GB RAM | 22:13 |
| DocScrutinizer05 | and yes, bootloader was a mess to get right | 22:13 |
| ds3 | is that 1GB RAM or 1GB RAM + NAND? | 22:13 |
| DocScrutinizer05 | have fun doing that to NOLO | 22:13 |
| DocScrutinizer05 | 1GB + 512MB | 22:14 |
| ds3 | oops, thought you meant the chip is understood by NOLO already | 22:14 |
| ds3 | blah | 22:14 |
| ds3 | didn't the 3730 use the EMIF? | 22:14 |
| DocScrutinizer05 | NOLO is completely opaque | 22:14 |
| DocScrutinizer05 | EMIF? | 22:15 |
| ds3 | (vs the 3430 using the SDRC) | 22:15 |
| ds3 | the ram controller block | 22:15 |
| DocScrutinizer05 | I think there's no difference | 22:15 |
| ds3 | uboot handles both, IIRC | 22:15 |
| DocScrutinizer05 | yes, uBoot handles it, after quite some patching | 22:16 |
| ds3 | eh? it should be in one header file | 22:16 |
| DocScrutinizer05 | NOLO though... | 22:16 |
| DocScrutinizer05 | ask freemangordon | 22:17 |
| DocScrutinizer05 | I didn't follow what was the problem. Think we had 2 or 3 of them | 22:17 |
| DocScrutinizer05 | I sort of understood the problem back when I read the sourcecode. Now I can't recall anymore | 22:18 |
| DocScrutinizer05 | cheers | 22:18 |
| DocScrutinizer05 | o/ | 22:18 |
| ds3 | yep it is the same SDRC | 22:19 |
| *** robink_ is now known as robink | 22:20 | |
| *** L29Ah has joined #maemo | 22:32 | |
| KotCzarny | actually, n900 with 512-1024M of ram would rock | 22:35 |
| Wizzup | almost like the neo900 | 22:36 |
| Wizzup | oh wait :) | 22:36 |
| KotCzarny | but much cheaper ;) | 22:36 |
| *** L29Ah has left #maemo | 22:37 | |
| *** L29Ah has joined #maemo | 22:39 | |
| KotCzarny | flashing, always full of suspense | 22:45 |
| * Maxdamantus suspects the RAM issue on a browser can be solved simply enough by setting a ulimit on a multi-process browser. | 22:50 | |
| Maxdamantus | (one that handles subprocesses crashing, that is) | 22:51 |
| Maxdamantus | Chromium, Firefox with Electrolysis, .. | 22:52 |
| *** LauRoman has joined #maemo | 22:56 | |
| *** L29Ah has left #maemo | 23:04 | |
| *** L29Ah has joined #maemo | 23:06 | |
| *** L29Ah has left #maemo | 23:15 | |
| *** L29Ah has joined #maemo | 23:16 | |
| *** L29Ah has left #maemo | 23:24 | |
| *** _rd has quit IRC | 23:43 | |
| *** _rd has joined #maemo | 23:45 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!
