IRC log of #maemo-ssu for Saturday, 2012-04-07

*** Abissom has joined #maemo-ssu00:07
*** Sicelo has quit IRC00:08
*** Abissom is now known as Sicelo00:10
*** NIN101 has quit IRC01:21
*** Pali has quit IRC01:55
DocScrutinizerhah, nice01:59
*** trbs has quit IRC03:00
*** LaoLang_cool has joined #maemo-ssu03:15
*** arcean has quit IRC03:29
*** Estel_ has joined #maemo-ssu03:57
*** nox- has quit IRC04:03
*** M4rtinK has quit IRC04:23
*** LaoLang_cool has quit IRC05:22
*** amiconn has quit IRC05:45
*** amiconn_ has joined #maemo-ssu05:45
*** amiconn_ is now known as amiconn05:46
*** Estel_ has quit IRC06:19
*** infobot has quit IRC07:12
*** infobot has joined #maemo-ssu07:13
*** ChanServ sets mode: +v infobot07:13
*** Pali has joined #maemo-ssu09:04
*** psycho_oreos has quit IRC09:04
*** amiconn has quit IRC09:24
*** amiconn has joined #maemo-ssu09:25
*** Jade- has joined #maemo-ssu09:53
*** Jade- has joined #maemo-ssu09:53
*** Jade has quit IRC09:54
*** Jade- is now known as JAde09:54
*** JAde is now known as Jade09:54
Lava_Crofthooray, work on tracker10:01
*** Jade has quit IRC10:13
*** LaoLang_cool has joined #maemo-ssu10:54
*** NIN101 has joined #maemo-ssu11:36
*** LaoLang_cool has quit IRC12:18
Paliwhen will be new version of CSSU T?13:08
Paliwith patched ke-recv?13:08
*** M4rtinK has joined #maemo-ssu13:34
Palimerlin1991, ^^^13:46
DocScrutinizerdudes, I have a syslog for you. from when I did a call, inbound. FRIGGIN TRUCKLOAD OF MESS!14:03
DocScrutinizerwe should tackle that14:03
DocScrutinizeralso the annoying loglines like >>14:04
DocScrutinizerApr  7 12:23:55 IroN900 cellular: csd/libsms[792]: ISI_SMS .937683> ind_reg_status(): Net registration (ind) status:1 rc:014:04
DocScrutinizerApr  7 12:23:55 IroN900 cellular: csd/libsms[792]: ISI_SMS .938934> set_timeout(): Timeout 1200 s event type:-114:04
DocScrutinizerApr  7 12:24:08 IroN900 cellular: csd/libsms[792]: ISI_SMS .854400> incoming_cell_broadcast(): Incoming cell broadcast14:04
DocScrutinizerand >>14:05
DocScrutinizerApr  7 10:23:20 IroN900 cellular: csd/libsms[792]: ISI_SMS .492645> ind_reg_status(): Net registration (ind) status:1 rc:014:05
DocScrutinizerApr  7 10:23:20 IroN900 cellular: csd/libsms[792]: ISI_SMS .493896> set_timeout(): Timeout 1235 s event type:-114:05
DocScrutinizeroops14:05
DocScrutinizerlike >>14:06
DocScrutinizerApr  6 19:25:06 IroN900 wlancond[1616]: High signal14:06
DocScrutinizerApr  6 19:25:06 IroN900 kernel: [107646.301696] wlan0: roaming signal from driver, sending HIGHSIGNAL14:06
DocScrutinizerApr  6 19:25:46 IroN900 wlancond[1616]: Low signal14:06
DocScrutinizerApr  6 19:25:46 IroN900 kernel: [107686.238952] wlan0: roaming signal from driver, sending LOWSIGNAL14:06
DocScrutinizerI know I could log those lines to /dev/null, when my syslog.conf was correct14:07
DocScrutinizerbut I'd prefer to nuke tose spam logs out of kernel/$whatever14:07
DocScrutinizer.14:11
DocScrutinizer.14:11
DocScrutinizerIroN900:~# grep "wlan0: roaming signal" /var/log/syslog|wc -l14:11
DocScrutinizer266414:11
DocScrutinizerIroN900:~# grep "cellular: csd/libsms" /var/log/syslog|wc -l14:11
DocScrutinizer1546914:11
DocScrutinizerIroN900:~# wc -l /var/log/syslog14:11
DocScrutinizer57125 /var/log/syslog14:11
DocScrutinizer:-/14:11
DocScrutinizeralmost 30% are friggin "cellular: csd/libsms" lines14:12
DocScrutinizeractually the "wlan0: roaming signal" are * 2, so another 10%14:13
DocScrutinizerI'm not entirely clear yet about "GLIB DEBUG default"14:14
DocScrutinizermaybe I should adapt my loglevel somewhere?14:15
*** LaoLang_cool has joined #maemo-ssu14:33
*** DocScrutinizer has quit IRC14:55
*** DocScrutinizer has joined #maemo-ssu14:55
*** sam23m has joined #maemo-ssu15:22
sam23mpraying for testing update!15:23
*** arcean has joined #maemo-ssu15:36
sam23mi wish the n900's camera would take pictures faster15:38
*** Free-MG has joined #maemo-ssu16:22
*** arcean has quit IRC16:23
*** arcean has joined #maemo-ssu16:24
*** Sicelo has quit IRC16:28
*** LaoLang_cool has quit IRC16:32
*** NIN102 has joined #maemo-ssu16:50
*** NIN101 has quit IRC16:50
merlin1991Pali: it's getting built17:07
*** Sicelo has joined #maemo-ssu17:15
*** Sicelo has quit IRC17:20
*** Sicelo has joined #maemo-ssu17:20
Palimerlin1991, ok write me when packages will be in cssu T repository17:37
*** javispedro has joined #maemo-ssu17:42
*** sam23m has left #maemo-ssu18:16
DocScrutinizerwhat's that stuff with slow nicocam?18:18
*** trumee has quit IRC18:19
merlin1991DocScrutinizer: 2 mins too late18:19
DocScrutinizerallegedly taking ages to store pic18:19
merlin1991I never experienced slow nikocam18:20
merlin1991maybe he/she experienced swapping hell?18:20
DocScrutinizerI as well faul to see a story that explains it18:20
DocScrutinizerpossibly18:20
DocScrutinizerfail*18:20
DocScrutinizeror friggin tracker kicking in, turning system into a mess18:21
DocScrutinizer(btw sth that happens also when I accept a call - recall starts a new file, tracker rears its ugly head :-S )18:22
DocScrutinizerunbearable18:22
merlin1991why don't you add the folder recall uses to the tracker ignore config?18:23
merlin1991I mean it isn't trackers fault that it's not configured to ignore that set of files18:24
*** trumee has joined #maemo-ssu18:28
*** Free-MG has quit IRC18:33
*** mase76 has joined #maemo-ssu18:34
MohammadAGaccepting calls is a nightmare on Fremantle18:43
DocScrutinizerI don't actually want it to ignore the files, as I eventually want to playback then in mediaplayer18:44
DocScrutinizerjust friggin tracker could shutup until call finished18:45
RST38hOnly if your device has been up for a while :)18:45
merlin1991well recaller could send tracker the dbus command to pause18:46
merlin1991and continue when it's done18:46
merlin1991so the file gets picked up18:46
DocScrutinizernice idea18:46
RST38hDocScrutinizer: A sensible (albeit somewhat awkward) solution would be to 1) permanently shut doen background tracker process and 2) add a desktop or status menu option saying "Reindex media"18:46
merlin1991i think nikocam has that code somewhere18:46
merlin1991RST38h: I'd rather add 2 desktop icons for paus and unpause tracker18:47
RST38hPausing tracker takes time18:47
merlin1991and just let it run when you feel like18:47
DocScrutinizerpausing tracker no option when inbound pending18:47
RST38hmerlin: or one status item - "Tracker: Do | Shutup"18:47
merlin1991yeah18:47
merlin1991so have tracker not running in general, but once a day tell it now you can :D18:47
DocScrutinizeryeah18:48
DocScrutinizera possible solution18:48
RST38hnot once a day18:48
RST38hbecause it does not know when you are adding media and want it shown18:48
RST38hso it has to be a button saying "run tracker once"18:48
DocScrutinizerI could trigger it from dbus scripting when call finished18:48
RST38hplus an on/off switch if you wish so18:48
MohammadAGor do it the Android way18:49
DocScrutinizerkill -sigstop * ?18:49
MohammadAGrun indexer only in certain conditions, otherwise add data manually :P18:49
MohammadAGindexer runs when you disconnect mass storage, or insert an SD Card18:49
MohammadAGotherwise (when downloading), the browser adds the file to the db18:50
MohammadAGrather than having the indexer running18:50
MohammadAGof course this needs major changes in code18:50
DocScrutinizererrr18:50
MohammadAGkill -sigstop is not the android nor iOS way18:50
MohammadAGandroid allows app to run in the background18:51
DocScrutinizerfor that amount of changes I'd rather go for nuking tracker off the globe and replace it in apps like mediaplayer with a dirt straight fs access18:51
RST38hplease note that tracker is *already* configurable almost to this degree via config file18:51
RST38hso it may not be such a huge fix18:51
MohammadAGiOS gives apps some time to save state and quit18:51
merlin1991http://paste.debian.net/162447/18:51
MohammadAGiOS 4.0 mutlitasking simply added the overridable function viewDisappeared() or whatever18:51
MohammadAGsaving state and quitting is a nice idea18:52
javispedroit is not18:52
MohammadAGwhen you have a 6GHz CPU that can launch apps quick enough for the user not to notice it18:52
javispedroback to the PalmOS days we are.18:52
MohammadAGwait for the sentence to finish javispedro18:52
MohammadAGon iOS you see a splash screen for quite some time18:53
javispedroah, sorry :)18:53
DocScrutinizeranywqay:18:53
MohammadAGlike Settings on Maemo 518:53
DocScrutinizerApr  6 19:33:48 IroN900 tracker-indexer[4619]: GLIB WARNING ** Tracker - Unable to retrieve info from file (Error stating file '/home/user/MyDocs/calls/CurrentlyRecording.wav': No such file or directory)18:53
MohammadAGlol18:53
MohammadAGmerlin1991, 2220 tests passed18:54
merlin1991DocScrutinizer: does it stash does away later on?18:54
merlin1991s/h does/h those/18:54
infobotmerlin1991 meant: DocScrutinizer: does it stash those away later on?18:54
DocScrutinizerobviously18:54
MohammadAGfuck openssl18:54
merlin1991well then just get it to write somewhere else that curently file18:54
merlin1991tell tracker to ignore the fsckd folder18:54
MohammadAGbuilding Qt takes time but it's fun to watch18:54
MohammadAGtracker works on Harmattan18:54
MohammadAGwhy? cause all the internals were changed18:55
merlin1991and then tracker happily indexes your "encoded, renamed whatever" file that shows up somewhere else later18:55
DocScrutinizerwell, not somewhere else, just in this very dir18:55
MohammadAGand probably cause of the 768MB increase in RAM18:55
DocScrutinizeronly under a different name of file18:55
merlin1991I don't think you can get tracker to ignore a single file in a dir18:55
DocScrutinizerindeed18:55
merlin1991so recaller has to write the temp file into some ignored folder and later on copy/move into indexed folder18:56
javispedrohey, I couldn't get fremantle tracker to ignore anything until that patch I made :)18:56
DocScrutinizerthe issue however is that it indexes at all during call18:56
merlin1991javispedro: ignore patch?18:56
DocScrutinizerjavispedro: that too, indeed18:56
merlin1991on your page I only noticed mediaart stuff and removebale device fixes18:56
javispedroyeah, ignore patch18:56
javispedrohttp://depot.javispedro.com/nit/tracker/realignore.patch18:57
DocScrutinizerI noticed that flaw in tracker some 2 years ago, I think18:57
javispedrooh, but this was for external mountpoints18:57
DocScrutinizeralways thought it's just me18:57
javispedrothe problem is that tracker adds certain mountpoints by default to its "watch" list18:58
merlin1991silly tracker18:58
DocScrutinizeryep18:58
DocScrutinizeridiot prog18:58
javispedroand FORGETs to compare those autoadded mountpoints against the exclude list18:58
javispedrowell, as MohammadAG, it got some work on Harm, this bug and a few others have been fixed18:58
javispedro*said18:58
merlin1991javispedro: those 3 patches from you will end in the next cssu-testing :D18:59
MohammadAGdoes the N900 have tracker-utils preinstallde?18:59
DocScrutinizernot afaik18:59
DocScrutinizernot sure18:59
merlin1991MohammadAG: openssl finsihed18:59
MohammadAGfuuuuuuuuuu18:59
DocScrutinizerdang, I seem to be glued to this chair :-S18:59
merlin1991DocScrutinizer: err what?19:00
DocScrutinizerhotel california syndrome19:00
MohammadAGwhat about libtrackerclient019:00
MohammadAGYou can checkout any time you like...19:00
DocScrutinizerbut you can never leave19:00
* MohammadAG bangs head while listening to solo in head19:00
MohammadAGmerlin1991, openssl finished here19:01
MohammadAGmuhahaha19:01
merlin1991do you still want a tarball or using your build?19:01
MohammadAGI'll use my build :p19:02
merlin1991all those cycles19:02
MohammadAGlibssl0.9.8 and openssl right?19:02
merlin1991wasted for nothing19:02
merlin1991MohammadAG: I have a dump of the mp depends here19:03
merlin1991comparing :D19:03
MohammadAGcheck tracker while you're at it :p19:03
DocScrutinizerbtw I've seen odd thing: apps were rotated to portrait in taskswitcher while they rotated to landscape in fullscreen19:03
MohammadAGand check if libcrypto0.9.8-udeb is installed19:03
merlin1991 nope libcrypto0.9.8-ude is not in the depends19:04
MohammadAGtracker-utils?19:04
DocScrutinizerI'd guess when an app is allergic to portrait, it will barf up even when rendering for thumbnail portrait19:04
merlin1991yes19:04
merlin1991libtrackerlient0 aswell19:04
DocScrutinizerso I dare calling this behaviour a bug19:04
merlin1991DocScrutinizer: hm landscape only applications don't rotate in the switcher afaik19:05
DocScrutinizerapp that shows the behaviour (though not suffering from it): extended call log19:05
MohammadAGwe shouldn't depend on19:06
MohammadAG maemo-applet-tvout - TV-Out control panel applet19:06
MohammadAG maemo-tvout-control - Set the aspect ratio and scaling of the N900's TV out19:06
MohammadAG status-menu-applet-tvout - TV-Out status menu applet19:06
MohammadAGanymore right?19:06
merlin1991sec19:06
MohammadAGig waut19:06
MohammadAGoh wait19:06
MohammadAGig waut looks like german btw, probably means something too19:06
MohammadAGshould depend on maemo-applet-tvout19:06
merlin1991nah :D19:07
MohammadAGyes19:07
MohammadAGit's a stock replaced control panel applet19:07
merlin1991Pali: tv-out which packages do we have to depend on ?19:07
DocScrutinizerI got protrait desktop, I open extcalllog via desktop shortcut, it displays in landscape, even when device helt portrait and kbd closed. I go to taskswitcher: extcalllog renders portrait thumbnail19:07
merlin1991DocScrutinizer: what's a portrait thumbnail in you book?19:08
DocScrutinizerheld*19:08
Palistatus-menu-applet-tvout is needed for all Tvout packages19:08
merlin1991because I open quassel2go (landscape only qt app) go to switcher, hold device in portrait and nothing happens19:08
DocScrutinizerobviously a thumbnail that's not only rotated but also rearranged19:08
MohammadAGso basically19:08
MohammadAGall 3 packages?19:08
Paliand maemo-applet-tvout is control panel plugin19:08
merlin1991only the header of the thumbnail with the application names swapps19:08
MohammadAGand the backend19:08
Palimaemo-applet-tvout is needed too19:09
Paliand maemo-tvout-control is optional19:09
DocScrutinizeror in short: a thumbnail of a portrait screen19:09
merlin1991DocScrutinizer: screenshot that please19:09
merlin1991I'm still confused19:09
Palicssu meta package must depends on: maemo-applet-tvout and status-menu-applet-tvout19:09
DocScrutinizerhmm, lemme see what I can do and where to upload, now that OVI us dead19:09
Paliboth packages should not be in user/section19:10
MohammadAGso basically, nothing changed19:10
MohammadAGexcept that we need yet another package, libtvctl19:10
merlin1991yep19:10
Palimaemo-tvout-control is optional and this only should be user/ section19:10
MohammadAGhttp://pastebin.com/hk5qc3RN merlin199119:10
DocScrutinizerHAHAHA, screenshot of taskswitcher in portrait, with closed kbd. YEAH!19:10
merlin1991DocScrutinizer: ssh in19:11
merlin1991gst-launch19:11
DocScrutinizerin always19:11
merlin1991be a hero19:11
DocScrutinizerI'm19:11
merlin1991looking the commandline up currently19:11
MohammadAGdon't slide it all the way19:11
MohammadAGyou can do it19:11
MohammadAGthe engineer does it the hardware way19:11
MohammadAGdisassemble the device19:11
merlin1991I remember posting it on nokia developer some time ago19:11
DocScrutinizerwhat's the key shortcut?19:11
MohammadAGbreak the switch detecting slides19:11
MohammadAGCTRL+SHIFT+P19:11
MohammadAGor just cat /dev/fb0 > rawimage then convert that to png19:12
DocScrutinizeryeah, I can do this ;-P19:12
DocScrutinizer(sh-ctl-p)19:12
merlin1991or let me find the fsck commandline19:12
merlin1991here we go: http://www.developer.nokia.com/Community/Wiki/Capture_movie_directly_from_N9/N950_screen19:13
merlin1991gst-launch-0.10 ximagesrc num-buffers=1 ! ffmpegcolorspace ! pngenc ! filesink location=screenshot.png19:13
merlin1991i think it's gst-launch without the -0.10 on maemo19:13
MohammadAGcat /dev/fb0 > rawimage ffmpeg -vcodec rawvideo -f rawvideo -pix_fmt rgb32 -s 800x480 -i rawimage -f image2 -vcodec png screenshot.png19:13
RST38hDear gods, please, port SiB to Harmattan. Sorry19:13
MohammadAGforgot a ; somewhere19:13
MohammadAGmerlin1991, check the debian/control file19:14
MohammadAGpastebin link19:14
DocScrutinizeroops friggin crap, wasn't extcalllog19:15
merlin1991MohammadAG: libxml2 missing19:18
MohammadAGmerlin1991, added it now19:18
DocScrutinizerbut Fahrplan19:18
MohammadAGmerlin1991, is -utils in the mp?19:18
MohammadAGlibxml2-utils19:18
merlin1991nope19:18
MohammadAGmerlin1991, not mine, PR1.3's19:18
merlin1991yeah19:18
merlin1991that's what I checked :D19:18
* merlin1991 is a wizard19:19
DocScrutinizermerlin1991: install Fahrplan, start up in portrait mode, it forces to landscape19:19
MohammadAGwe're done here19:19
DocScrutinizermerlin1991: then while still holding device in portrait, open taskswitcher19:19
DocScrutinizermerlin1991: watch Fahrplan thumbnail rerendering and rearranging whle screen19:20
MohammadAGUploading19:20
*** psycho_oreos has joined #maemo-ssu19:20
merlin1991DocScrutinizer: nfc what you do19:20
merlin1991but fahrplan has a portrait mode here19:20
DocScrutinizerwell, it doesn't here19:20
merlin1991wtf19:20
merlin1991afaik I don't have forced enabled19:20
merlin1991version?19:20
MohammadAG<3 1MBps upload speed19:20
merlin1991I have 0.0.2919:21
DocScrutinizerand that's the way I like it and it's been until recently some crap nuked my UI-may-rotate gconf key19:21
DocScrutinizerno matter why, an app that's in ls in fullscreen mustn't switch to portrait in appswitcher *only*19:22
DocScrutinizerI wouldn't deny it's a bug in Fahrplan though19:22
DocScrutinizerNFC19:22
DocScrutinizeras evidently a lot of other apps (xterm etc) do NOT rotate to pt in appswitcher19:23
merlin1991even with ui_can_rotate false it rotates19:23
merlin1991do you by any chance have to portrait lock enabled?19:24
merlin1991or parts of it :D19:24
DocScrutinizercould you paste the gconftool line here to set ui_can_rotate false, please?19:24
merlin1991gconftool-2 -s /apps/osso/hildon-desktop/ui_can_rotate -t bool false19:24
DocScrutinizerhanks19:24
merlin1991found the cornercase you had19:26
merlin1991ui_can_rotate to true19:26
DocScrutinizererr, is gconf user specific?19:26
merlin1991but orientation_lock true19:26
merlin1991dunno19:26
merlin1991I always ran that stuff as user though19:26
DocScrutinizershould it take effect immediately?19:27
merlin1991yes19:27
DocScrutinizeroh yes, it does19:27
merlin1991the state you want is ui_can_rotate false19:27
DocScrutinizer:-D19:27
merlin1991and portrait_lock false19:27
merlin1991but you're right about the taskswitcher19:27
DocScrutinizerno more portrait appswitcher, no more way to see the "bug"19:27
merlin1991well fahrplan app itself supports portrait19:28
merlin1991but you managed to have the ui rotation enabled whilst the other 50% of the lock (apps) were engaged19:28
merlin1991so it got forced back to landscape but the taskswitcher could still rotate happily19:28
DocScrutinizermhm, I think the separation of that is the root bug19:28
merlin1991I think the lock is the root bug ;)19:29
*** freemangordon has joined #maemo-ssu19:29
DocScrutinizernah, portrait is the root of all bugs19:29
merlin1991hehe19:29
DocScrutinizerwill you open a ticket against $whatever regarding it?19:30
merlin1991well it was actually right in what it did19:30
merlin1991it acted accordingly to it's 2 config parameters19:30
DocScrutinizerthe result however is wrong19:31
DocScrutinizerso the definition of the semantics of those 2 parameters is flawed19:31
merlin1991but we've got the general "portrait lock comes from deeper layers of hell" ticket anyway19:31
merlin1991which we hopefully can start to tackle soon19:31
merlin1991MohammadAG: is the release uploaded?19:32
merlin1991I can't test till tomorrow evening (only brought my -stable n900 with me)19:33
freemangordonguys, any idea where the hell is jonwil?19:33
merlin1991nfc19:33
DocScrutinizer\o/ no more weird rotation when holding device a bit out of the flat19:33
*** trbs has joined #maemo-ssu19:33
merlin1991btw anybody in here who has experience with acl on debian (or another distro)?19:34
DocScrutinizereven if I'd appreciate portrait, I'd bitch about gmeter orientation detection algo even more19:34
DocScrutinizerit's unbearable19:34
freemangordonanyone willing to help RE libcellularsettings.so, i.e. "phone" control panel applet. I have some progress, but it would take me probably a month or so and I have to suspend all other projects :(19:35
DocScrutinizerI.E the threshold and hysteresis is way too low19:35
merlin1991yeah I agree abotu the threshold19:36
merlin1991sometimes I get rotation with the slightes change of physical rotation state19:36
DocScrutinizeryep, that's hysteresis though19:36
freemangordonhmm, lets try on TMO19:37
DocScrutinizeraka "on at 9, off at 4"19:37
merlin1991freemangordon: good luck with that ;)19:37
merlin1991at least closing an application isn't tied to the same system19:37
freemangordon:) thanks. damn library is 40k of object code :(19:37
merlin1991the am I going to close this now game on harm is so annoying19:38
DocScrutinizer??19:38
merlin1991harm closes applications with a swipe from the physical "up"19:39
DocScrutinizeraaah19:39
DocScrutinizeryeah19:39
merlin1991especially funky if the device is lying flat on the desk19:39
DocScrutinizerhehe19:39
merlin1991hm something went wrong with the ads on the page I'm currently visiting19:40
DocScrutinizerthe virus failed and splattered screen with its blood?19:40
merlin1991firstly it's about "related video" even though I'm reading some text19:41
merlin1991and secondly its "Getting Started in PowerPoint 2010"  and "How to Disable AutoComplete in Internet Explorer"19:41
merlin1991(page content is about acl)19:41
DocScrutinizerhihi19:42
* merlin1991 wonders how those fit19:42
DocScrutinizerprobably your IP inherited from a winlose noob19:42
merlin1991yeah dhcp from the isp here19:43
merlin1991actually was G-lined on qnet from one of the ips I got :D19:43
DocScrutinizerhaha, yeah that happens19:44
DocScrutinizersome foolish nets set IP glines19:44
DocScrutinizerbtw what's a gline?19:44
MohammadAGmerlin1991, uploaded19:44
* DocScrutinizer only knows k-lines19:44
merlin1991is that what freenode calls a ban?19:45
DocScrutinizeryup, server wide ban19:45
DocScrutinizernetwork wide19:45
DocScrutinizerhmm19:46
merlin1991on quakenet it's g-lined19:46
DocScrutinizer~kline19:46
DocScrutinizer~k-line19:46
*** ChanServ sets mode: +o merlin199119:46
*** merlin1991 changes topic to "Maemo Community Seamless Software Update "CSSU" channel, http://wiki.maemo.org/Community_SSU | Known bugs: http://j.mp/communityssu-bugs | Channel logs: http://mg.pov.lt/maemo-ssu-irclog/ | Sources: http://gitorious.org/community-ssu/ | Latest version (testing): 21.2011.38-1Tmaemo3; (stable): 21.2011.38-1Smaemo3 | meeting: 04.04 19:00 UTC"19:46
*** merlin1991 sets mode: -o merlin199119:47
merlin1991time to write a changelog19:47
merlin1991aint that fun :/19:47
merlin1991Pali: ke-recv is in the repos now19:48
Paliok19:48
*** NIN102 has quit IRC19:49
*** NIN101 has joined #maemo-ssu19:56
merlin1991changelog for current cssu release http://wiki.maemo.org/Community_SSU/Changelog#T-maemo320:00
*** mase_76 has joined #maemo-ssu20:05
merlin1991hm bugs.maemo.org is slow20:07
DocScrutinizerI thought it's dead ;-P20:07
DocScrutinizerbut no, that's been harm bt20:07
*** mase76 has quit IRC20:08
RaimuHmm, let's fire up the ol' appman, then.20:15
Raimu...and tell it to refresh after it's been churning 'round for a couple of minutes already. I love this app.20:15
merlin1991damn going over the buglist I realized that there are more bugs filed for the stuff fixed in this version20:17
merlin1991I think I added them all now20:19
merlin1991why o why did I start the garage page20:20
merlin1991one more thing to manage20:20
RaimuTracker-doc conflict... hmms, let's see what's up with that.20:21
*** mase_76 has quit IRC20:21
merlin1991Raimu: what's the exact package name of the conflict?20:26
NIN101I wonder whether it is safe to update from 21.2011.38-1Tmaemo1.1 haha :P20:26
merlin1991NIN101: updates from any cssu version up should be save20:26
* merlin1991 waits for the first entry in the -testing thread on tmo :D20:30
NIN101hmm. but I first have to make sure that /etc/event.d/rcS-late and /etc/init.d/rcS weren't changed20:31
merlin1991wth have you set up on your n900?20:34
merlin1991and afaik we don't touch them20:34
NIN101e. g. encrypted /home/ - needed some hacks in those bootscripts.20:34
merlin1991I guess you could use dpkg-divert to keep your changes save20:38
merlin1991srly, nobody who knows a thing or two about acls?20:40
NIN101just ask... :P20:42
merlin1991well i need to have 1 directory where a bunch of users (in a group) can drop files which 1 other user can modify at will20:54
DocScrutinizerhmm, creating something means modifying it20:58
DocScrutinizerthus your plan won't be that easy, by concept20:58
NIN101but the users in the group should only drop them, not modify? If so, then it will be very difficult.20:59
DocScrutinizeryou could use inotify-monitor to move files or change their perms once they got created20:59
DocScrutinizerthis way only the creator will have a valid write filehandle21:00
DocScrutinizerof course there's still a race, a window of "vulnerabiity"21:00
merlin1991tbh I don't care that much about denying modify21:01
merlin1991my bigger problem is to autoallow the other user to have rights21:01
NIN101setfacl -m user:name:rwx folder or whatever?21:02
merlin1991well I have nfc about acl, only thought that it probably is the sanest way to approach this21:02
DocScrutinizeryou want a "local superuser"?21:03
DocScrutinizerperfectly feasible with ACL21:03
*** sicelo_ has joined #maemo-ssu21:35
merlin1991DocScrutinizer: yes, and that leaves me with the original question, how :D21:42
*** trumee has quit IRC21:43
*** NIN101 has quit IRC21:48
*** NIN101 has joined #maemo-ssu21:55
arceanmerlin1991: Pali: freemangordon: we have a problem with situations like this: "1. Neither MyDocs, nor external card got mounted on startup"22:00
arceanfor e.g. http://talk.maemo.org/showpost.php?p=1189418&postcount=49422:00
freemangordonarcean, I saw that, but NFC what is that supposed to mean22:02
Paliarcean, bug when something points to /lib/init/upstart-job means problem with debhelper22:03
arceanlooks like emmc and memory card is not mounted on startup22:03
Palithis is big in creating package when calling dh_init*22:03
freemangordonPali, you know that bug?22:03
Paliyes22:04
Paliseems like new version of debhelper or dpkg22:04
Paliif you use regular scratchbox fremantle dpkg, this wont happen22:04
Pali(dpkg and debhleper)22:04
freemangordonscratchbox?22:04
freemangordonwhat has scratchbox to do with that?22:05
Paliscratchbox creating package22:05
freemangordonI mean those reports are for real devices22:05
freemangordonyou mean that the packages are prepared outside scratchbox?22:05
merlin1991Pali: so you mean with a newer debhelper the build goes bad?22:05
Palibut 412b said that ke-recv point to upstart-job22:06
arceanso what can we do, to fix the problem?22:06
freemangordonnew update?22:06
Palitry recompile package in scratchbox22:06
freemangordonmerlin1991, who prepared those packages?22:06
Paliand check if /etc/init.d/ke-recv point to upstart-job or not22:06
merlin1991freemangordon: mag22:07
Paliupstart-job is used on ubuntu22:07
Palimaemo still using /etc/event.d22:07
freemangordonMohammadAG ping22:07
merlin1991freemangordon: mag is gone for today22:07
freemangordongreat22:07
merlin1991well I'm still here ;)22:07
freemangordonmerlin1991, do you have scratchbox installed?22:08
merlin1991freemangordon: what a question :D22:08
arceannice, there's at least 5 people with bricked Maemo :D22:08
freemangordonwell, Paliu is still using mad AFAIK :D22:08
Palimadde has also fremantle compatible dpkg & debhelper22:08
merlin1991arcean: that's why it's called -testing and has a reflash warning on the wiki page :D22:08
freemangordonmerlin1991, make a new update ASAP22:09
arceanmerlin1991: yeah :D22:09
merlin1991freemangordon: sure, but I'll need the info as to why it went bad22:09
merlin1991Pali: which version of debhelper is bad?22:10
Paliall which is not in freamntle22:10
freemangordonbecause MAG has Harmattan SDK, that is why22:10
Paliso this can be reason22:10
merlin1991freemangordon: I have both aswell22:10
Paliharmattan dpkg & debhelper22:10
freemangordonAnd I am getting sick of that socalled maintanersip22:10
RaimuThis is hilarious to read as I see the CSSU installation crawl increase in yellow length. :p22:10
PaliI have qemu image for SDK fremantle22:10
RaimuOn my n900.22:10
PaliI can try to start it up and try to compile ke-recv22:11
merlin1991just wait22:11
Palibut I need to get version22:11
* merlin1991 builds ke-recv22:11
Paliok22:11
freemangordonmerlin1991, I have NFC how are those packages compiled, but it seems it is not done as it should22:11
merlin1991freemangordon: the problem is that you have so much influence on how a package is built from the buildpc22:12
freemangordonand lets fix this stuff before more devices are bricked. Will investigate afterwards, when mag is here22:12
freemangordonmerlin1991, disagree22:12
RaimuAnd reboot... let's hope I'm not bricked now. :-)22:12
freemangordonyou shold have a scratchbox environment which is used ONLY for CSSU22:12
freemangordonyou have AFAIK, ain't?22:13
merlin1991yeah I have22:13
freemangordonsee22:13
freemangordonare you sure MAG has one?22:13
merlin1991I know for sure that mags sb is kinda fucked up22:13
freemangordon:nod:22:13
merlin1991for example I never had those modest build bs22:13
freemangordonmerlin1991, make a post on TMO please, there is a little chance to save some innocent souls22:15
RaimuNo bricks here. Whew.22:15
merlin1991Raimu: got MyDocs mounted?22:15
merlin1991btw Raimu this isn't exactly bricking, only 90% of your datastorage is going to be missing22:16
Raimu:D22:16
Paliok, I see that problem is really with that upstart22:16
Palike-recv will not start22:16
Raimumerlin1991: I've got MyDocs mounted.22:16
freemangordoncould it be started manually?22:16
Palino22:16
merlin1991wtf22:16
Palibecause script is deleted22:16
Paliscript should be in /etc/init.d/ke-recv22:16
Palibut debhelper deleted it and created some upstart link22:17
freemangordonok, do we have terminal in that situation?22:17
Paliyou can downgrade22:17
freemangordonyeah, that was my idea22:18
Palicopy startup script22:18
Paliupgrade22:18
Paliand then add back that script22:18
freemangordonPali, no need for that, just downgrade until next update is ready22:18
Paliok22:18
Paliand when will be next update?22:18
freemangordonPali, do you know the exact command?22:19
freemangordonask merlin1991 :D22:19
merlin1991Pali: when I've built it :D22:19
Pali/etc/init.d/ke-recv start22:19
merlin1991Pali: how can I check if my build is proper?22:19
freemangordonno, for downgrade22:19
Palimerlin1991, update somewhere package22:19
Pali*upload22:19
merlin1991apt-get install ke-recv=MEMOVERSION22:19
Palior see if you have correct script /etc/init.d/ke-recv22:20
Paliin builded deb package22:20
merlin1991Pali: http://cdnm.at/~christian/maemo/ke-recv_3.19-14_armel.deb22:21
Paliseems is ok22:22
merlin1991hm but mags package has a /etc/init.d/ke-recv aswell22:22
Palireally?22:23
Palidiff it22:23
merlin1991yes22:23
merlin1991http://repository.maemo.org/community-testing/pool/fremantle/free/k/ke-recv/ke-recv_3.19-14_armel.deb22:23
Palino22:23
Palithat package has invalid symlink22:23
Palithat package is really bad22:24
Paliit show me: lrwxrwxrwx 1 pali pali 21 Apr  7 16:20 etc/init.d/ke-recv -> /lib/init/upstart-job22:24
merlin1991hm funky diff http://paste.ubuntu.com/919373/22:25
Palimerlin1991, there is really no file etc/init.d/ke-recv22:25
Paliit is only symlink!!!22:25
*** nox- has joined #maemo-ssu22:25
merlin1991hm why does ark resolve it to file then?22:25
merlin1991weird22:25
Palibecause you have ubuntu?22:26
merlin1991probably22:26
merlin1991anyway does my build look ok?22:26
PaliI wrote that that upstart is used on ubuntu22:26
Paliyes22:26
Paliseems your build is ok22:26
Palioh no22:26
Palithere is other problem22:26
Paliit is not easy to upgrade from MAG bad version to your new fixed22:27
merlin1991Pali: I'll upload with higher version22:27
Palithis is not enought22:27
merlin1991what else is problematic?22:27
Paliconffine22:27
Pali*confile22:27
Paliwhen you try to upgrade with dpkg -i <new version>22:28
*** Sicelo has quit IRC22:28
*** sicelo_ is now known as Sicelo22:28
Pali... old invalid link will not be replaced /etc/init.d/ke-recv22:28
Paliso you need to first uninstall ke-recv and then install new version again22:28
merlin1991hm?22:28
RaimuWhy not?22:28
freemangordonPali, why is that?22:29
merlin1991how can mags packge replace but a higher version not?22:29
Palifreemangordon, try it22:29
Palimerlin1991, try too22:29
merlin1991not -testing device with me22:29
Palifirst install bad mag version22:29
freemangordonI can't, my device is running ke-recv build by me22:29
freemangordonaah22:29
freemangordonyeah22:30
Palifreemangordon, install MAG version and then install fixed merlin version22:30
*** trumee has joined #maemo-ssu22:30
PaliI tried this and bad link /etc/init.d/ke-recv was still here22:30
freemangordonPali, don't have time to do that now, sorry. Do you know the reason22:30
freemangordon?22:30
Paliyes22:30
* merlin1991 tries22:30
Paliconffiles and dpkg22:30
merlin1991Pali what is the wrong link and what would the right one be?22:30
Palifile /etc/init.d/ke-recv must be shell script22:31
Palibut MAG version has symlink with that filename22:31
freemangordonwhatever it means :D. anyway, seems there is enough expertise without me bringing noise.22:31
Paliand when you try to upgrade file will not be replaced with symlink22:31
Palimerlin1991, at least release new version quickly22:32
merlin1991why does the symlink from mag kill that though?22:32
merlin1991Pali:  on my way22:32
Palibecause debian conffiles and dpkg behaviour22:33
merlin1991SILLY22:33
Palithere could be hope: HAM using apt-worker and it does not used dpkg but libapt22:33
Paliso maybe HAM will update it22:33
Palifix for that problem could be: delete /etc/init.d/ke-recv and reinstall new version of ke-recv22:34
Paliwhen file /etc/init.d/ke-recv will not exists, dpkg will unpack correct version from deb package22:34
Pali$ sudo rm -rf /etc/init.d/ke-recv && sudo apt-get install --reinstall ke-recv22:35
Palithis could fix this problem ^^^22:35
DocScrutinizerhaha, let's start the fun then22:39
merlin1991okay fixed package with increased version number is up22:39
* merlin1991 adjusts the metapackge22:40
DocScrutinizermerlin1991: no, this is not the way to *test* packages. it's not called devel after all, it's TESTING, implying it's already considered working, and that implies it got tested thoroughly by developer and a few guinea pigs22:40
merlin1991DocScrutinizer: yes yes yes22:40
merlin1991well it was22:40
merlin1991it was a build error that made it all bad22:41
merlin1991also no time, gotta prevent bricks ;)22:41
DocScrutinizerindeed22:41
merlin1991okay fixed file should be uploaded22:45
merlin1991yes it's there22:45
merlin1991okay now that we have a proper package out there, why did this upstart bs happen?22:49
merlin1991or rather, can we prevent that on source level?22:50
freemangordonmerlin1991, there is no way to prevent the MAINTAINER from doing BS to the packages22:53
merlin1991yes22:53
merlin1991though can we prevent debhelper from doing bs?22:53
Paliyes, if we use correct debhelper :D22:54
freemangordonmerlin1991, well, then we should fix all the control files22:55
freemangordonto depend on debheler version in SB22:55
freemangordons/debheler/debhelper/22:55
infobotfreemangordon meant: to depend on debhelper version in SB22:55
DocScrutinizerfreemangordon: :-D22:56
DocScrutinizerthumbs up22:56
DocScrutinizerplus I'd like to suggest to find a way to give a release a final check before enabling it to general public22:57
freemangordonbut that does not answer the question WHY THE HELL MAG UPLOADS STUFF IN THE REPOS WITHOUT BEING TESTED ON HIS DEVICE?22:57
DocScrutinizerand have advisory instructions from patch devels on how to test the possible areas of impact for their patch22:58
merlin1991DocScrutinizer: plus thing is getting handled22:58
merlin1991aka cssu-devel22:58
merlin1991aka I need them damn acl done so it can go live22:59
freemangordonmerlin1991, but then the maintainer should not build the stuff which is already in -devel repo22:59
merlin1991yea22:59
DocScrutinizerwell, I hope the man pages will be clear on ACL22:59
freemangordonBTW that is why i uploaded ke-recv on your server22:59
merlin1991DocScrutinizer: man pages on acl just intrduce a state of "brain is full of fuck"23:00
merlin1991s/intrduce/introduce/23:00
infobotmerlin1991 meant: DocScrutinizer: man pages on acl just introduce a state of "brain is full of fuck"23:00
DocScrutinizer man setfacl23:00
freemangordonmerlin1991, so THAT build to be used. BTW do we know if there are other packages fcked up?23:01
merlin1991freemangordon: we never know that23:01
DocScrutinizercould we find a way to know?23:01
merlin1991ofc, upgrade and check if everything still works :D23:01
freemangordonIf I were you, I will rebuild all the stuff from the last update and will make a new update23:02
DocScrutinizer*sigh*23:02
freemangordonexcluding ke-recv of course23:02
freemangordonsee, it just does not make sense to have bug-reporst for nonexisting bugs, it is the same situation as with modest.23:03
freemangordonit took 4 months to find that actually MAG's SB is fcked up.23:04
Palican we use autobuilder on garage?23:04
freemangordonno23:04
freemangordonno way23:04
Paliwhy?23:04
freemangordonQt23:04
*** javispedro has quit IRC23:04
freemangordonplu autobuilder will put the stuff in extras-devel :p23:05
freemangordon*plus23:05
merlin1991freemangordon: there actually is a builder specific for communit and community-testing23:05
freemangordonat least if we don't feed him with KP :D23:05
freemangordonmerlin1991, do we have access to it?23:06
merlin1991but it has an fsckd setup that is partly copied over from the -devel builder23:06
merlin1991which leads to most things not building23:06
merlin1991when I started -stable I've spent about a month untill I gave up and just compiled in a dedicated sb target23:06
merlin1991only that this sb target by now is polluted by -devel stuff, so for next stable I'll create yet another new sb target23:08
merlin1991fun times ahead23:08
merlin1991err not -devel but cssu-testing23:08
freemangordonso, effectively we don't have autobuilder and should do the job by hand23:09
merlin1991yep23:10
merlin1991though I've played with the idea to hack together an outobuilder on my server :)23:10
merlin1991but that is for the far future :D23:10
merlin1991lol @ outobuilder23:10
merlin1991see, that's what the acl manpages did to me!23:11
freemangordon:D23:11
merlin1991I really wonder though how we can prevent debhelper from doing silly ubuntu specific things23:12
freemangordonI was near the same state a few hours ago, while REing libcellularsettings23:12
freemangordonby using the EXACT version in control file23:13
merlin1991sure that's one way23:13
merlin1991though afaik newer debhelper somehow coexists with older one or something23:13
* merlin1991 could be spreading fud though, not sure23:14
merlin1991freemangordon: one day I'll know how to RE stuff aswell, for now I know way too much about various build systems source->binary :D23:14
merlin1991the most interesting thing I learned today though is, that ark when it finds a symlink inside a tar also presents you your local file23:17
merlin1991kinda wtf23:17
DocScrutinizermerlin1991: restore backup?23:17
merlin1991of my brain or what?23:18
merlin1991sry, you got me lost there23:18
DocScrutinizer<merlin1991> only that this sb target by now is polluted by -devel stuff, so for next stable I'll create yet another new sb target23:19
merlin1991yeah there is one thing about backups, you gotta do ones to be able to restore them :D23:21
merlin1991but creating a new sb target is quite fast, untar rootstrap, install maemo-whatever-mp, add cssu repo, install qt, hussa I'm done :)23:21
merlin1991hm I've got my acl as far as that the user in question can edit the files because he's got rw23:35
merlin1991but he can't move / delete the file23:35
merlin1991ah wait23:35
merlin1991it works23:36
merlin1991I tried to move the file within the folder, but whilst the user was allowed todo anyhting with the file, he couldn't write to the folder23:36
merlin1991:P23:36
freemangordonso, we have a working -devel repo?23:37
merlin1991about 5 commands away of going live23:37
merlin1991and ofc I'll have to email you guys the instructions :D23:37
merlin1991it beats me though why the acl tool is called setfacl (set file access list) even though it works on direcotries aswell23:38
freemangordonunix?23:38
Siceloeverything is a file :P23:39
freemangordoneverything is a file :p23:39
freemangordon:D:D:D23:39
merlin1991fu :D23:39
merlin1991shit23:40
DocScrutinizerLOL23:40
freemangordonPali, isn't it time to put KP50 in testing?23:41
Paliyes23:42
PaliI will promote it now23:42
freemangordongreat23:43
Palidone23:51
merlin1991freemangordon: http://maemo.merlin1991.at/cssu/community-devel/23:51
freemangordonmerlin1991, what versions are the packages there?23:52
merlin1991current -devel23:52
merlin1991err -testing23:52
freemangordonok23:52
* merlin1991 should not operate as root in such a state23:52
freemangordongreat23:52
freemangordonrm -rf / ?23:52
freemangordon:D23:52
merlin1991I just have to remember the pass I set for the reprepro user then I can start the handler for the incoming queue and you can start breaking it :D23:53
merlin1991hm I should write the passwords down one day23:54
freemangordonPali, KP49 is ready for extras, KP50 is still not in -testing. Maybe we should promote KP49 in -extras first :(23:54
merlin1991freemangordon: you need votes for that not promotion23:54
Paliok, I will try23:54
PaliI think we have needed votes23:55
freemangordonyeah, I can vote too if neede :D23:55
merlin1991okay the incoming system should be alive aswell23:56
merlin1991freemangordon: got some package ready you could upload as a test?23:56
freemangordonok, i have a new version of libconnui-dev23:56
merlin1991hehe the only package that isn't even in the repo23:57
freemangordonlet me change debian/changelog23:57
merlin1991but should be fine aswell23:57
freemangordonwhy it is not in repo?23:57
merlin1991because it is not part of the user end23:58
RaimuI find it hilarious how my n900 now goes into offline mode and probably reboots updating T3 to T3.123:58
merlin1991and possibly can wrek havok on devices actually23:58
merlin1991it's a re package for a library that exists otherwise on the device23:58
merlin1991coming from another package ondevice23:58
freemangordonit is -dev package, mp should not depend on it, but it should be available in repo23:59
merlin1991well it is on gitorious23:59
freemangordonhowever23:59

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