IRC log of #maemo for Tuesday, 2014-07-29

*** DocScrutinizer05 has quit IRC00:00
*** DocScrutinizer05 has joined #maemo00:00
*** pdz has quit IRC00:04
sixwheeledbeastM4rtinK: :)00:04
*** shentey has quit IRC00:05
sixwheeledbeastre tracklogs I cannot reproduce the reported issue of not stopping a tracklog. However I did notice the stop button displays the banner "saving" even if the tracklog is off.00:05
M4rtinKsixwheeledbeast: BTW, any feedback on the track logging fix ? :)00:05
M4rtinKyeah, it should do that00:06
M4rtinKfor about 3 seconds00:06
M4rtinKbecause if you leave it running for a couple of hours00:06
M4rtinKthere will be a multi-MB CSV file that needs to be converted to GPX00:06
*** rm_work is now known as rm_work|away00:07
*** pdz has joined #maemo00:07
M4rtinKso while usually it is quite redundant, sometimes it is indeed "saving" for some time :)00:07
sixwheeledbeastWell I tried to use the tracklog for the first time this weekend out on the motorbike. However the device became unresponsive by the time I got to the petrol station 5 minutes away.00:07
M4rtinKweird00:08
sixwheeledbeastNot tried again since.00:08
M4rtinKI've use00:08
sixwheeledbeastBlack screen with backlight on only modRana was open. I did get a SMS in that time so maybe that caused some weirdness?.00:09
M4rtinKd modRana a bit on a field trip this Friday and it worked fine00:09
sixwheeledbeastI plan to track my route up Snowdon in a few weeks.00:10
sixwheeledbeastSo when I walk or ride I am testing it out.00:10
M4rtinKlets hope we can track it down before then :)00:10
M4rtinKI think there are 3 most probable possible causes:00:11
M4rtinK1) memory leek - it is possible modRana is adding to some data structure somewhere and not removing from it00:11
M4rtinK2) modRana triggering some race or other bug in Cairo/GTK/X00:12
*** flynx_ has joined #maemo00:12
M4rtinK(could b related to the SMS - might trigger a notification, which shows an overlay, etc.)00:13
M4rtinK3) that SGX driver bug resurfacing for some reason00:13
M4rtinKso I think that black screen could be either the N900 swapping like hell and being unresponsive or X/GPU driver crashing00:14
sixwheeledbeastM4rtinK: seemed like the SGX bug symptoms however I was away from a device I could SSH with.00:15
M4rtinKI'm thinking some sort of script that logs top or free memory could be useful to track it down00:16
sixwheeledbeastM4rtinK: could swap on uSD be a constant? Maybe reports are from people with uSD swaps?00:16
M4rtinKas for potential countermeasures I'm thinking about introducing performance tuning settings00:17
M4rtinKfor the tile cache size, download thread counts, etc.00:17
M4rtinKin hope that setting them to lower values it could be possible the work around the issue :)00:17
M4rtinKsixwheeledbeast: for the record I'm using the stock partitioning on my N90000:18
M4rtinKonly bigger change is CSSU + Thumb + 900 MHz overclock (since like ever)00:18
M4rtinKso yeah, I guess it could be influenced by it00:19
M4rtinKlike say io errors crashing the kernel :)00:19
M4rtinKactually I wonder what happens if you kill the block device under an active swap :)00:20
sixwheeledbeast:nod: well it is popular to use uSD swaps now. no overclocking here.00:20
M4rtinKwell, could improve the eMMC lifetime I guess00:20
M4rtinKI wonder about throughput though00:21
*** wotan147 has quit IRC00:21
M4rtinKuSD bandwidth generally sucks00:21
sixwheeledbeasteMMC lifetime and workaround for swap defragmentation00:21
M4rtinKoh, interesting :)00:22
sixwheeledbeast~flopswap00:22
infobotwell, flopswap is http://wiki.maemo.org/Flopswap00:22
sixwheeledbeast;)00:22
M4rtinKBTW, what about compressed swap ? :)00:22
M4rtinKusing something like LZOo/LZ4/snappy could work00:23
M4rtinKmight also help with the limited bandwidth00:25
sixwheeledbeastI dont believe there are any benefits, you have to uncompress it.00:25
M4rtinKsure, bu the benefit is that only compressed data will go over the presumably slow uSD bus00:28
M4rtinKall the compression would happen in RAM and LZO & co are _FAST_00:28
M4rtinKbut who knows, it might still be too much for the CPU of the N900 to provide any benefit00:30
M4rtinKon the other hand we have had some good experience with zRAM (previously called compcache)00:30
M4rtinKN9 is using it's predecessor and it is in mainline kernel since 3.1000:31
M4rtinKit is basically a fixed size compressed block device residing in RAM you can use to setup swap00:32
*** wotan147 has joined #maemo00:32
M4rtinKand it indeed seem to work quite nicely00:32
M4rtinKwe have enabled it on one of our virtualisation servers and the claim of the zRAM people that using it basically duplicates possible RAM usage before the real swap is hit seems to hold :)00:34
M4rtinKoh, looks like it is in kernel power already: http://talk.maemo.org/showthread.php?t=76112 :)00:35
*** florian has quit IRC00:40
sixwheeledbeastWell I believe Swappolube should be able to tune the I/O00:40
M4rtinKBTW, provided the swap on uSD can be disabled00:44
M4rtinKit might be good to try if you get more/less modRana freezes if it is ON or OFF00:44
*** wotan147 has quit IRC00:49
*** xes has joined #maemo00:55
*** fanf has joined #maemo00:58
fanf'lo ?00:59
*** fanf has left #maemo01:00
*** fanf has joined #maemo01:02
*** fanf has left #maemo01:03
*** rm_work|away is now known as rm_work01:06
*** xes_ has joined #maemo01:08
*** xes has quit IRC01:10
sixwheeledbeastwell yes it can be disabled, I'll check that out.01:12
*** xes_ is now known as xes01:13
ShadowJKThe current day deent uSDs have better iops performance than 2008-2009 era emmc01:21
ShadowJKThough of course, you can also buy worse performing cards.01:21
M4rtinKnew modRana is out! :D http://talk.maemo.org/showpost.php?p=1434142&postcount=161901:23
*** N-Mi has quit IRC01:38
*** jooles has joined #maemo01:47
joolesHey all. Does anyone have a copy of the latest rx-48 firmware image? All the sites that had it seem to have gone down and my N810 won't boot anymore :C01:48
ShadowJKdid you take out battery for 5 minutes?01:48
joolesnot for 5 minutes01:48
jooleseven so, it's in a mess and could do with being re-flashed01:49
ShadowJKrx-48, is that wimax model?01:50
joolesI succumbed to temptation like I always do, some months back, and tried to build a new kernel for it which never ends well01:50
joolesit is01:50
DocScrutinizer05err01:50
DocScrutinizer05~tabletsdev01:50
infobotextra, extra, read all about it, tabletsdev is http://tablets-dev.nokia.com/  http://wiki.maemo.org/Tabletsdev ,  http://tabletsdev.maemo.org (all defunct, thanks Nokia) or the nice site http://www.fladnag.net/downloads/telephone/n900/tools/, or http://www.mmnt.net/db/0/0/93.81.63.203/repositories/skeiron.org/skeiron.org/tablets-dev/maemo_dev_env_downloads01:50
joolesin order, down, 404 and has no images01:51
DocScrutinizer05~combined01:51
infobotextra, extra, read all about it, combined is the rootfs fiasco image of maemo. For N900 latest (PR1.3.1) see http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2011/9/--FID--A0A22MVWFVFAM/--LID--FiRe1317015685654/RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin, or http://galif.eu/nokia/01:51
DocScrutinizer05~emmc01:51
infobotemmc is, like, magnet:?xt=urn:btih:402fb5cc8a48ecbc18a77c9cf70d869a775bcf53&dn=RX-51%5F2009SE%5F10.2010.13-2.VANILLA%5FPR%5FEMMC%5FMR0%5FARM.bin or https://dl.dropboxusercontent.com/u/4105928/RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.torrent, or http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2010/5/--FID--A0A22YHFSICNA/--LID--FiRe1275051276916/AE98ED9D_RX-51_2009SE_10.2010.13-2.CENTRAL-EUROPE_PR_EMMC_MR0_ARM.bin01:51
DocScrutinizer05~flasher01:52
infobotsomebody said flasher was at http://www.jedge.com/n810/flasher/maemo_flasher-3.5_2.5.2.2.tar.gz (also .exe!), or http://www.chakra-project.org/ccr/packages.php?ID=5027 or generally http://www.google.com/search?q=maemo_flasher-3.5_2.5.2.2.tar.gz.   HARMattan(N9): https://aur.archlinux.org/packages/fl/flasher-harmattan/flasher-harmattan.tar.gz, or -- list of filenames/md5sums: http://pastebin.com/sYKdNJSH, or http://galif.eu/nokia/01:52
joolesDocScrutinizer05, Care to read what I wrote or not? I get that these questions get annoying but once all the sites went down I had no choice but to come and ask. I'm not an idiot but clearly you can't tell that RX-48 != RX-5101:53
DocScrutinizer05thanks for your kind way to tell me that you didn't notice http://www.jedge.com/n810/flasher/maemo_flasher-3.5_2.5.2.2.tar.gz01:54
joolesright, the flasher I could find01:54
joolesI'm looking for the firmware image01:54
DocScrutinizer05*sigh*01:54
joolesAs I also said earlier. Also spamming links at me is kinda rude01:54
*** N-Mi has joined #maemo01:55
DocScrutinizer05meh01:55
DocScrutinizer05good luck01:55
joolesWell thanks anyway. Sorry everyone for causing so much spam.01:57
*** jooles has left #maemo01:58
DocScrutinizer05friendly guy01:58
bencohwhat the ... hmm.02:02
*** M4rtinK has quit IRC02:03
nox-no n8xx firmware anymore?02:06
*** odin_ has joined #maemo02:10
*** antranigv has joined #maemo02:15
antranigvHay, I have N900, I changed the /bin/sh to /bin/bash in /etc/passwd file... I wanted to use Bash as default... so now I can't boot the device :( it's stuck at loading :( any help ??02:17
nox-~rescueos02:20
infobot[rescueos] http://n900.quitesimple.org/rescueOS/02:20
antranigvP.S. MicroUSB not working  :( ... I can fix it after I go back home (still outside of my country)02:22
nox-ow02:23
* nox- leaves you to DocScrutinizer05 maybe he has a better idea02:23
antranigvLet's hope and wait :))02:24
DocScrutinizer05there's no idea how to fix such sort of broken rootfs02:25
antranigvDocScrutinizer05: is flashing a good idea ? I can get the USB port fix02:25
DocScrutinizer05when you can fix the USB, then get rescueos and revert the passwd edit02:26
antranigvDocScrutinizer05 , nox- thanks !02:27
DocScrutinizer05nox-: how would I publish a torrent?02:27
DocScrutinizer05noob here02:28
xeshttp://www.btkitty.com/DcjJAQAgCAOwmcBSyjh47T-C5hnUNWOxIzmdvZc84MZ0QfilCJzxAA~S8ssyi1PLEotBgA.torrent.html02:30
nox-DocScrutinizer05, i havent tried that yet either...02:33
nox-oh xes found it it seems :)02:35
xesanyway md5 must be checked02:37
nox-*nod*02:38
bencohbroken usb + broken fs abroad ? sounds terrible :/02:39
*** dima5` has joined #maemo02:39
nox-indeed02:40
DocScrutinizer05xes: I found this, and it doesn't work for me. but that's maybe because I'm behind a NAT anyway. And actually I asked how to *publish* a torrent/magnet02:42
nox-downloading a torrent more or less needs portfowards yeah02:43
nox-(if you're behind nat)02:43
*** arcean_ has quit IRC02:52
DocScrutinizer05yeah, I know. I have to admit I simply don't understand that crap (not port forwarding since that's easy - but that torrent client)02:56
DocScrutinizer05it seems to miss any server it can coordinate with02:57
DocScrutinizer05neither sends nor receives a single byte02:57
DocScrutinizer05search is more or less broken too02:58
*** bef0rd has joined #maemo02:59
*** bef0rd has joined #maemo02:59
dos1there is no "server it can coordinate with" aside of trackers, but that's per torrent file (but DHT can be used instead anyway)03:02
*** goldkatze has quit IRC03:02
dos1generally you just create a .torrent file with a hash and some open tracker specified, then point your local client to it in a way so it treats your file as already downloaded, and then send .torrent file to someone else (or to some public index)03:03
dos1(or magnet link can be used instead of .torrent file)03:05
DocScrutinizer05no connection to any trackers here03:06
dos1http://www.mepis.org/docs/en/index.php?title=Creating_a_torrent03:06
*** RiD has quit IRC03:06
*** RiD has joined #maemo03:07
dos1trackers are per-file, so if you don't have any torrents down/uploading then that's expected :P03:07
DocScrutinizer05HAHAHA!!!  >>You must add at least one tracker: type or paste in the URL: http://linuxtracker.org:2710/announce and click the add button.<<03:07
DocScrutinizer05suuure, I know this by heart, and thus the shitty ktorrent doesn't even suggest it03:08
DocScrutinizer05and the magnet seems to linger there like it does nothing at all03:09
dos1dunno what you want to seed - if it's not some 100% legal linux iso, then something like http://openbittorrent.com/ will probably be a better choice :P03:09
DocScrutinizer05probably best choice is to forget about that torrent shit03:09
nox-torrents are faster for some people03:10
DocScrutinizer05meh, I don't care03:10
nox-:)03:10
TheOnlyJoey|deski normally download torrent with the speed of my internet connection, 50mbit as we speak03:11
DocScrutinizer05it's not like it gives me headache to NOT share some of the stuff I have here03:11
dos1it all depends03:11
DocScrutinizer05this ktorrent client however for sure gives me headache03:11
TheOnlyJoey|deskDocScrutinizer05, try Transmission, has served me well for years :)03:11
dos1ktorrent is a pretty good client imo - you would probably have similar problems with any client03:12
dos1unless it has some "friendly" trackers etc. so it autoprefills this stuff for you ;)03:13
DocScrutinizer05ok, so much for wasted time with torrents03:13
dos1anyway, that mepis wiki page has it described pretty well03:15
*** xes has quit IRC03:15
*** janemba has quit IRC03:18
nox-magnet:?xt=urn:btih:49f11696a576b26adc8254216728484b268554e3&dn=firmwares03:18
nox-doesnt seem to work at least in rtorrent03:18
*** janemba has joined #maemo03:19
nox-(from xes' link)03:19
*** antranigv has quit IRC03:20
DocScrutinizer05exactly03:23
*** eMHa_ has joined #maemo03:24
DocScrutinizer05and http://openbittorrent.com/  reports "Error: invalid data from tracker"03:24
nox-theres a thunder:// link too which seems to belong to https://en.wikipedia.org/wiki/Thunder_Download03:32
nox-"Although not widely reported, Xunlei client has been discovered spreading the malware Win32/KanKan to thousands of Chinese Windows PCs and Android devices.[5] In 2013 the company admitted its employees bundled malware into Xunlei client.[6]"03:32
nox-uh03:32
hurrian_such a surprise03:32
jon_yisn't xunlei also one of those exploiting torrent clients?03:35
jon_yuh, selfish client03:35
* nox- wouldnt know03:35
jon_ylast I read, it was banned from private trackers03:36
nox-well if it spreads malware who wants to use it anyway...03:36
*** jonwil has joined #maemo03:37
jon_ylusers want to, for faster speeeeed03:37
nox-any idea whats wrong with that magnet link?  or is it just the tracker being down?03:39
jon_yI think its just the tracker since you got the torrent already03:40
nox-there doesnt even seem to be a .torrent to download03:44
hurrian_man, and people wonder why nobody outside of china wants to use that crap03:44
*** dima5` has left #maemo03:46
*** robbiethe1st has joined #maemo03:49
*** flynx_ has quit IRC03:50
*** Humpelstilzchen has quit IRC04:10
*** Kabouik has quit IRC04:11
*** Humpelstilzchen has joined #maemo04:15
*** bef0rd has quit IRC04:20
*** janemba has quit IRC04:27
*** janemba has joined #maemo04:27
*** erlehmann has joined #maemo04:29
*** RiD has quit IRC04:39
*** RiD has joined #maemo04:39
DocScrutinizer05I gave up on torrent, that's silly. When you need sth, just ask me04:48
DocScrutinizer05I got an amazingly huge backup folder here04:48
*** janemba has quit IRC05:04
*** nox- has quit IRC05:04
*** janemba has joined #maemo05:05
*** maybeWTF has quit IRC05:34
*** maybeWTF has joined #maemo05:39
*** lxp1 has joined #maemo06:01
*** lxp has quit IRC06:04
*** dima5` has joined #maemo06:05
*** janemba has quit IRC06:11
*** janemba has joined #maemo06:12
*** RiD has quit IRC06:26
*** janemba has quit IRC06:38
*** janemba has joined #maemo06:40
*** Psi has quit IRC06:58
*** Psi has joined #maemo07:09
*** LauRoman has joined #maemo07:11
*** robbiethe1st has quit IRC07:25
*** dhbiker has joined #maemo07:52
*** erlehmann has quit IRC07:59
*** maybeHere has joined #maemo08:09
*** maybeArgh has joined #maemo08:09
*** maybeHere has quit IRC08:10
*** jormungandr has joined #maemo08:25
*** jonwil has quit IRC08:52
*** kolp has joined #maemo08:54
*** fuz_ has quit IRC08:57
*** fuz_ has joined #maemo09:02
*** DocScrutinizer05 has quit IRC09:14
*** DocScrutinizer05 has joined #maemo09:14
*** Gadgetoid has quit IRC09:16
*** Gadgetoid has joined #maemo09:17
*** mavhc has quit IRC09:22
*** mavhc has joined #maemo09:22
*** mavhc has quit IRC09:29
*** mavhc has joined #maemo09:32
*** Pilke has joined #maemo09:56
*** AD-N770 has joined #maemo09:59
*** tanty_off is now known as tanty10:00
*** norayr has joined #maemo10:02
*** LaoLang_cool has joined #maemo10:22
*** dhbiker has quit IRC10:42
*** rcg has joined #maemo10:43
*** protem has quit IRC10:48
*** goldkatze has joined #maemo10:48
*** janemba has quit IRC10:56
*** janemba has joined #maemo10:57
*** APic has quit IRC10:57
*** shamus has quit IRC10:59
*** shamus has joined #maemo10:59
*** geaaru has joined #maemo11:07
*** dhbiker has joined #maemo11:14
*** norayr has quit IRC11:16
*** LaoLang_cool has quit IRC11:24
*** tanty has quit IRC11:25
*** Kup has joined #maemo11:29
*** Pilke has quit IRC11:32
*** goldkatze has quit IRC11:44
*** tanty_off has joined #maemo11:48
*** tanty_off is now known as tanty11:48
*** geaaru has quit IRC11:53
*** dima5` has quit IRC11:55
*** robink has quit IRC11:58
*** robink has joined #maemo12:05
*** geaaru has joined #maemo12:06
*** Kabouik has joined #maemo12:13
*** Aspi has joined #maemo12:21
Aspithere are lots of entries of phones world wide promoting "UNLOCKED". what is this "unlocked" really, can a vendor lock a device, especially a N900 that will be flashed?12:23
kerioAspi: rapuyama don't care, rapuyama don't give a fuck12:25
kerioyou can flash as hard as you want, the n900 will stay oplocked/simlocked12:25
Aspiso it's done in some ROM?12:26
hurrian_there are locked N900s?12:29
*** mkaindl has joined #maemo12:30
Aspithat's what I don't know. I dunno if "unlocked" is Chinese marketing scam.12:31
hurrian_the Hong Kong refurbished N900s are unlocked, but are probably reballed with cigarette lighters12:31
*** eMHa_ has quit IRC12:33
Aspiso, I have the dilemma that if I buy second hand, previous owner might in turn have bought a HK :)12:33
hurrian_do you already have a N900?12:34
Aspino12:34
hurrian_reusing HK N900s isn't too bad if you just want spare parts12:34
hurrian_welp, that's pretty risky then12:34
fizzieThere's an "Enter code to remove operator lock" option in the title-bar-menu of the "Settings" application, which kind of sounds like they sold operator-locked N900s somewhere. Unless it's there just in case.12:34
*** ALo has quit IRC12:35
Aspiah, okay. I'm just gonna ask seller if I buy.12:35
Aspiare y'all excited about the Neo900? or will you live with the N900 to your death? :)12:36
*** DocScrutinizer05 is now known as DocScrutinizer0612:36
hurrian_my N900 was swiped, and all the N900s on the local equivalent of eBay are dead/sold as-is.12:37
*** DocScrutinizer06 is now known as DocScrutinizer0512:37
*** ALo has joined #maemo12:37
Aspiditto, so I look international12:37
hurrian_unfortunately, I won't be happy with just an OMAP3730, having been totally spoiled by how powerful octa-core balls out Android phones are.12:38
AspiI might take a chance on a HK refurb if I can't get something else, as they are not extremely expensive, and I might get lucky and get a reasonably working item (?)12:38
hurrian_I'll probably just run NeoFremantle on my N9, when it materializes.12:39
dos1[02:24:25] <DocScrutinizer05> and http://openbittorrent.com/  reports "Error: invalid data from tracker"12:40
dos1this url is for the browser, tracker url is on the page :)12:41
*** eMHa_ has joined #maemo12:58
*** arcean has joined #maemo12:58
*** LauRoman has quit IRC12:59
*** DocScrutinizer05 is now known as DocScrutinizer0613:08
*** DocScrutinizer06 is now known as DocScrutinizer0513:08
Aspikerio: after reading the logs about rapuyama O_o maybe I'll need a second thinkthrough about N90013:08
*** Aspi has left #maemo13:10
DocScrutinizer05hurrian_: good luck! I doubt anybody will backport the N9-BB5 cmt stack to fremantle13:16
DocScrutinizer05other than that, you're right that N9 has basically exactly same SoC and Storage/Memory chip like Neo90013:18
hurrian_isn't the BB5 comm rather similar?13:18
*** janemba has quit IRC13:21
*** janemba has joined #maemo13:22
DocScrutinizer05I don't think so. N9 BB5 is responsible for handling aegis certs/tokens and actually this is what forbids downgrading of HARM "firmware"13:22
hurrian_that's something interesting to test, unlzo'ing PR1.2 and running it on N9.13:24
hurrian_as far as I've read in the kernel hooks for aegis, I haven't seen any calls to BB5, but it might be in userland13:24
hurrian_besides, if BB5 doesn't get it's aegis fix, what's it gonna do, shut down the device? Nemo clearly works without it* (ofono as the RIL instead of Nokia's stuff, but hey, it works)13:26
*** dhbiker has quit IRC13:30
*** norayr has joined #maemo13:31
*** goldkatze has joined #maemo13:46
*** janemba has quit IRC13:46
*** janemba has joined #maemo13:47
*** janemba has quit IRC13:51
*** janemba has joined #maemo13:52
*** merlin1991 has quit IRC14:03
*** janemba has quit IRC14:06
*** merlin1991 has joined #maemo14:06
*** janemba has joined #maemo14:07
*** b1101 has quit IRC14:10
*** dhbiker has joined #maemo14:13
*** jas4711 has quit IRC14:14
*** Kup has left #maemo14:36
*** janemba has quit IRC14:44
*** janemba has joined #maemo14:44
*** janemba has quit IRC15:03
*** janemba has joined #maemo15:04
*** Ninja has quit IRC15:13
*** Wulfe has joined #maemo15:13
*** Wulfe has joined #maemo15:13
*** erlehmann has joined #maemo15:25
*** janemba has quit IRC15:33
*** janemba has joined #maemo15:34
*** b1101 has joined #maemo16:00
*** b1101 has joined #maemo16:00
*** flynx_ has joined #maemo16:04
*** janemba has quit IRC16:26
*** janemba has joined #maemo16:27
*** janemba has quit IRC16:37
*** jas4711 has joined #maemo16:37
*** janemba has joined #maemo16:38
*** SAiF1 has joined #maemo16:41
*** RiD has joined #maemo16:43
*** RedM has joined #maemo16:56
*** RedW has quit IRC16:56
*** jas4711 has quit IRC16:58
*** DrCode has quit IRC17:02
*** DrCode has joined #maemo17:03
*** konelix__ has joined #maemo17:32
*** konelix__ has quit IRC17:34
*** konelix__ has joined #maemo17:34
*** bluelupo has joined #maemo18:01
*** bluelupo_ has joined #maemo18:07
*** bluelupo_ has quit IRC18:15
*** bluelupo_ has joined #maemo18:22
*** bluelupo has quit IRC18:22
*** SAiF has joined #maemo18:24
*** SAiF1 has quit IRC18:26
*** bluelupo has joined #maemo18:49
*** ced117 has joined #maemo18:51
*** kerio has quit IRC18:54
*** kerio has joined #maemo18:54
*** rcg has quit IRC18:56
bencohhm, looking at scim_panel_gtk code ... no wonder scim eats battery ...19:00
bencoh    _check_exit_timeout = gtk_timeout_add (500, check_exit_timeout_cb, NULL);19:00
bencohcouldnt get uglier19:01
bencoh(checks every 500ms whether it should exit or not)19:03
*** N-Mi has quit IRC19:04
*** N-Mi has joined #maemo19:07
*** LjL has quit IRC19:08
*** AD-N770 has quit IRC19:11
*** heroux has quit IRC19:11
*** heroux has joined #maemo19:11
*** tanty is now known as tanty_off19:13
*** tanty_off is now known as tanty19:14
*** LjL has joined #maemo19:15
*** Kabouik has quit IRC19:41
*** kraft has quit IRC19:43
*** fuz_ has quit IRC19:44
*** fuz_ has joined #maemo19:49
*** Kabouik has joined #maemo19:57
*** tanty is now known as tanty_off19:58
*** LauRoman has joined #maemo20:00
*** ced117 has quit IRC20:12
*** Katajakasa has quit IRC20:15
*** infobot has quit IRC20:15
*** infobot has joined #maemo20:17
infobotDocScrutinizer: infobot joined!20:17
*** ChanServ sets mode: +v infobot20:17
*** RzR is now known as rZr20:31
DocScrutinizer05bencoh: every 500ms is almost OKish20:47
DocScrutinizer05also aiui it passes a callback to some gtk_timeout_add () function, which - depending on the way this function works - could cancel the previous callback registered. When it does, then this is an absolutely correct way to implement a 500ms timeout20:50
DocScrutinizer05however when it doesn't cancel the previous check_exit_timeout_cb registration then that's terribly broken indeed20:51
DocScrutinizer05I guess the real problem happens somewhere else, prolly in some busy loop that polls for input20:53
*** mbs1337 has joined #maemo21:01
*** mbs1337 has quit IRC21:03
*** mbs1337 has joined #maemo21:04
*** mbs1337 has quit IRC21:05
bencohDocScrutinizer05: well according to powertop it does fix the issue21:05
DocScrutinizer05what does fix which issue?21:06
bencohremoving this timer let the cpu go to c421:07
bencohlets*21:07
kerioisn't it c021:07
keriono wait, it goes the other way right21:07
bencohyeah the other way21:07
DocScrutinizer05well, randomly removing timers for sure fixes some timer-caused issues eventually, but it prolly will introduce some severe other breakage21:08
bencohit doesnt break a functionality as such (and newer versions of scim have this line commented as well) .... but it wont die on sigterm anymore21:10
bencohsigs are processed in another thread which updates a global var (ugly)21:11
DocScrutinizer05I cannot tell for sure what's the root problem from a single line of code (in this case), but I can *guess* that the code registers 100 timer callbacks per second and never cancels a single one of them. And when push comes to pull it will register a new timer callback event when the callback function of a previous timer event gets called21:11
bencohthis line gets executed once and the _cb doesnt seem to register a timer .... but meh :)21:14
DocScrutinizer05the line >>  _check_exit_timeout = gtk_timeout_add (500, check_exit_timeout_cb, NULL); << doesn't look like it could cancel a previously registered timer, since for that it would need the _check_exit_timeout handle to that previous instance as a parameter handed to the function21:14
*** sLumPia has joined #maemo21:15
*** geaaru has quit IRC21:16
bencohhttp://scim-for-maemo.googlecode.com/svn-history/r162/trunk/scim-1.4.7-maemo/extras/panel/scim_panel_gtk.cpp :)21:16
*** jas4711 has joined #maemo21:16
DocScrutinizer05>> gtk_timeout_add has been deprecated since version 2.4 and should not be used in newly-written code. Use g_timeout_add() instead.<<21:19
bencohit's a 5yo code ;)21:19
bencohbut yeah, looks like others patched it to use g_timeout_add() as well21:19
DocScrutinizer05http://www.gtk.org/tutorial1.2/gtk_tut-17.html#ss17.1  mhm21:20
DocScrutinizer05>>...can create a timeout function that will be called every "interval" milliseconds.<<21:20
bencohhmm actually they also removed this callback as well21:21
DocScrutinizer05so it's a recurring timer21:21
bencohyeah it is21:21
DocScrutinizer05that's pretty odd21:21
bencoh(that's actually what I was implying, sorry)21:21
*** florian has joined #maemo21:21
bencohit's an ugly recurring timer to check "should I die now ?"21:22
DocScrutinizer05The return value is an integer "tag" which may be used to stop the timeout by calling:21:22
DocScrutinizer05 void gtk_timeout_remove( gint tag );21:22
DocScrutinizer05braindamaged code21:22
bencohyup21:22
DocScrutinizer05HAHA >>You may also stop the timeout function by returning zero or FALSE from your callback function.<<21:22
bencohI might try to update scim to 1.4.14 (part of it has been rewritten and it should behave better)21:23
bencohbut I cant even build 1.4.7 in sb21:23
bencohI half-baked the tool from binary lib21:23
bencohI wonder how it got built at first21:24
*** at1as has joined #maemo21:24
DocScrutinizer05I'd not be surprised to find gtk do a busy loop for every instance of timer event registered via gtk_timeout_add()21:24
*** mkaindl has left #maemo21:25
*** jas4711 has quit IRC21:26
DocScrutinizer05I honestly wonder why this code needs such nonsense instead of doing it right in signalhandler(int sig)21:34
bencohthey eventually had it write21:37
bencohthe author was probably on heavy drug .... or just sleep deprived ;)21:37
bencohs/write/right/ : https://bugzilla.redhat.com/show_bug.cgi?id=44415021:38
bencohfunny thing is we got a borked version for maemo, and people started writing a kludgy menu applet "switcher" instead of patching/looking for a patch21:39
bencoh(this bug ticket predates n900)21:39
DocScrutinizer05 if (!_panel_agent->run ())21:39
DocScrutinizer05        std::cerr << "Failed to run Panel.\n";21:39
DocScrutinizer05    G_LOCK (_global_resource_lock);21:39
DocScrutinizer05    _should_exit = true;21:39
DocScrutinizer05WTF?21:39
bencohexactly21:39
bencohtold you it was ugly21:40
DocScrutinizer05this whole code has more cruft and commented-out lines of code than actually meaningful ones21:40
DocScrutinizer05and this coude looks to me pretty similar to a chinese PCB21:41
DocScrutinizer05code*21:41
bencoh:))21:42
bencohconsidering scim is mostly used for cjk input .... it probably is21:42
bencoh(or japanese code, maybe)21:42
DocScrutinizer05apply random changes until it starts to work. then cut out components until it stops working. Revert last component removal. Produce21:42
bencohhm no, definitely chinese, from the names21:43
*** sLumPia has quit IRC21:44
DocScrutinizer05God, I hate such uninspired code21:47
DocScrutinizer05reading it feels like drinking sour milk21:47
*** sq-one has joined #maemo21:48
DocScrutinizer05and thinking of that code maybe running on my machine gives me a feeling like in a restaurant where I know they cook with rotten milk21:48
DocScrutinizer05ugh, meeting today21:49
*** ciroip has quit IRC21:53
*** ciroip has joined #maemo21:55
bencohI'm pretty sure half (most ?) of desktop software looks or looked like that at some point21:58
bencohand maemo prolly isnt an exception21:58
*** b1101 has quit IRC22:01
DocScrutinizer05I'm quite sude most of KDE looks similar, yes22:03
DocScrutinizer05sure*22:03
DocScrutinizer05most (core) maemo code I've seen isn't all that bad22:04
bencohlook at hildon-desktop then :>22:04
DocScrutinizer05some stuff (hildon for example) could use some more love for parameter checking etc22:05
bencohor just the fact that you cant have desktop widgets without waking cpu22:05
DocScrutinizer05err wut?22:05
bencoh(I mean, waking cpu more than is needed)22:05
bencohI spotted a few loops/condition inversion in hildon-home22:06
bencoh(or hildon-desktop, I dont remember where this part of the code sits)22:06
bencohbasically it loops through widgets and checks whether display is *inside* the loop for each widget22:07
bencohand I suspect there is more22:07
*** totalizator has quit IRC22:08
DocScrutinizer05you mean hildon executing all widget code frequently to do display updates as long as desktop is visible? well, that's not exactly nice but sort of tolerable22:08
DocScrutinizer05the fact that a single segfaulting or freezing widget makes whole desktop go south is more severe22:09
bencohdepending on the widget it can get pretty heavy22:09
bencohtrue, that's quite bad as well, but this is more of a design thing22:10
DocScrutinizer05display update for all widgets is as well22:10
bencohon a different level/of a different nature22:11
DocScrutinizer05since the widgets are executed in hildon context, as simple function calls, there's no reasonable way for them to announce they don't need to update their display22:11
DocScrutinizer05a widget simply returning as fast as possuble when nothing to do is the most lightweight way to handle this in this architecture22:12
*** totalizator has joined #maemo22:13
DocScrutinizer05and anyway since the homescreen and thus the widgets are visible, there's a certain system load anyway, and particularly a certain energy consumption by backlight and mere display data bus. So a 10 wakeups per second for CPU don't really hurt22:13
bencohanyway ... now to build http://tomoe.sourceforge.jp against scim in sb22:13
bencohDocScrutinizer05: actually I was talking of display status check22:14
bencohs/of/about/22:14
infobotbencoh meant: DocScrutinizer05: actually I was talking about display status check22:14
DocScrutinizer05display status check?22:14
bencoh"is display on" "is desktop visible" "is device unlocked"22:14
bencohthat kind of things22:14
DocScrutinizer05that's supposed to get done in hildon22:14
DocScrutinizer05makes absolutely no sense to do it in every widget22:15
bencohthat's done in hildon but in reverse order/in a peculiar way22:15
bencohlemme find the source22:15
DocScrutinizer05when display of homescreen isn't visible, not a single widget is visible, so why should hildon run any wiget code in that sitation?22:16
DocScrutinizer05as soon as display status changes from invisible to visible, all widgets need to update their display content, and then every 50ms or whatever22:17
DocScrutinizer05as soon as status changes from visible to not visible (aka background, or locked), hildon should stop executing +any* widget code22:18
bencohhd_comp_mgr_update_applets_on_current_desktop_property22:18
DocScrutinizer05every widget in turn is supposed to check early if there's actually sth to update and return when e.g. last (seconds) digit of the clock didn't change22:19
bencohit goes through widgets twice and make Xlib calls whether display is on or not22:19
*** erlehmann has quit IRC22:19
bencohwhich roughly means going through X11 crazy codepath22:19
DocScrutinizer05I never checked that hildon code. I'm talking from mere best practice which I'd expect to see in hildon22:20
bencohwell ... I'd definitely love to see more best practice everywhere :)22:21
DocScrutinizer05and I'm pretty sure hildon does NOT do what you say it does, when screen is locked. Otherwise my battery wouldn't stay a 5h even22:21
bencohI didnt say it would really update widgets when locked, but it would do some X calls22:22
*** qwazix_ has quit IRC22:22
DocScrutinizer05X calls are as worse as display updates22:22
jaskai guess i need to look up the dual bl-5j hack, my mugen is getting shorter22:23
bencohDocScrutinizer05: well some widgets (hello qwb !) can do much worse ... but yeah, X calls arent free, that's my point ;)22:26
DocScrutinizer05my point is that I honestly doubt they happen while screen locked22:29
DocScrutinizer05if they actually do, then that's prolly a bug22:30
bencohI think it does at least for XDeleteProperty22:30
bencohthough I didnt rebuild hildon/mb to trace that22:30
bencohbut then again, powertop has different output with and without widgets (be it one simple widget)22:31
DocScrutinizer05that's bad22:32
DocScrutinizer05at least when you talk about locked screen22:32
bencohyup22:32
DocScrutinizer05actually with a locked screen hildon itself shouldn't run anymore22:33
DocScrutinizer05unless some idiotic widget scheduled a timer22:33
*** konelix__ has quit IRC22:34
DocScrutinizer05you know, one of the kind that sends a signal to wake up the process on expiry22:34
*** ult has joined #maemo22:35
DocScrutinizer05"the process" is hildon here, and of course it needs to run *all* widgets so the idiotic one has a chance to service the timer expiry22:35
bencohI think I tested with DCE widget at that time22:36
bencoh(which behaved better than qwb, but still)22:36
DocScrutinizer05whatever that is22:36
bencohdesktop cmd exec22:36
*** xes has joined #maemo22:36
bencohand I checked the source as well22:36
DocScrutinizer05umm well22:37
*** ult has left #maemo22:37
bencohI stopped using widgets at that point22:38
DocScrutinizer05I dunno what QBW does, but it prolly shouldn't run any timers. it has nothing that needs a timed execution22:38
bencohqwb has timed functionality (as well as dce)22:38
*** dhbiker has quit IRC22:38
*** bluelupo has quit IRC22:38
bencohbut qwb is heavier in everything22:38
DocScrutinizer05~qbw22:39
DocScrutinizer05~qwb22:39
*** ult has joined #maemo22:40
DocScrutinizer05it's probably an absolutely idiotic idea to implement something like a daemon that get executed 10 times per second in a widget22:41
DocScrutinizer05I can't think of any such function that would need that22:41
bencohit is, but ... looks like some people found it neat22:41
bencohI dont understand either :)22:41
*** at1as has quit IRC22:44
*** javispedro has joined #maemo22:46
*** goldkatze has quit IRC22:52
DocScrutinizer05I mean, sure I could integrate the fan control of my PC into xeyes.... BUT...22:53
DocScrutinizer05shouldn't a widget do *only* stuff that's actually visible on display? And particularly shouldn't it STOP doing this as soon as the display isn't visible?22:54
DocScrutinizer05a maemo widget isn't just an arbitrary X-app with a funny shaped window22:56
kerioDocScrutinizer05: X has a way to figure out whether or not you're drawn23:01
DocScrutinizer05kerio: so what?23:03
DocScrutinizer05you're aware that all widgets run as "hildon" ?23:03
kerioso even if it was an arbitrary X program, it doesn't really have an excuse to waste cpu23:03
bencoh:)23:04
DocScrutinizer05kerio: that's completely off the point we discussed, which is timer ALARM23:05
kerio"shouldn't a widget do *only* stuff that's actually visible on display? And particularly shouldn't it STOP doing this as soon as the display isn't visible?"23:05
kerioi expect every fucking program to do that, not only widgets23:05
DocScrutinizer05aha23:05
DocScrutinizer05kerio: that's completely off the point we discussed, which been daemons23:05
DocScrutinizer05and those for sure don't give an F that they are not visible usually23:06
DocScrutinizer05lemme put it simple: you MUST NOT make an X program out of crond23:08
*** xes_ has joined #maemo23:12
*** jas4711 has joined #maemo23:14
*** xes has quit IRC23:14
*** xes_ is now known as xes23:17
*** DrCode has quit IRC23:21
*** DrCode has joined #maemo23:22
*** alexript has joined #maemo23:24
*** alexript has quit IRC23:26
*** sq-one has quit IRC23:26
*** YnG has joined #maemo23:38
*** YnG has quit IRC23:40
*** louisdk has joined #maemo23:49
*** florian has quit IRC23:51
*** at1as has joined #maemo23:52

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