IRC log of #maemo-ssu for Tuesday, 2012-04-10

*** lizardo has quit IRC00:03
*** macmaN has quit IRC00:16
*** int_ua has joined #maemo-ssu00:19
*** Guest50125 has quit IRC00:22
*** macmaN has joined #maemo-ssu00:38
*** int_ua has quit IRC00:57
*** BCMM has quit IRC01:25
*** Jade has quit IRC01:35
*** Kaptenen has quit IRC01:40
merlin1991arcean: ping01:45
arceanmerlin1991: pong01:45
merlin1991do you knwo what the edit_mode_portrait gconf key does in hildon-desktop?01:46
arceanhmm, i'm not sure if it's not related to application list/launcher01:47
arceanmerlin1991: edit_mode_portrait is set when user enters desktop edit in portrait mode01:50
merlin1991hm I wonder why it's even neccesary01:51
merlin1991hm the only other place where I find it is in hildon-home in hd-backgrounds.c01:52
merlin1991I guess it's to tell h-h that user edits in portrait and not in landscape :)01:53
arceanyes01:58
arceanhmm, I'll try to change it tomorrow01:58
arceanso we will have one gconf key less :)01:59
freemangordonmerlin1991, (and others) http://maemo.merlin1991.at/cssu/community-devel/pool/free/o/osso-gnome-vfs2/01:59
merlin1991and what's it got?01:59
freemangordonnow sdcard partitions have correct labels02:00
merlin1991because on gitorious there still is only the stock maemo code ,)02:00
merlin1991I think you're a git push short02:00
freemangordonwait a minute :D02:00
freemangordonmerlin1991, ready02:02
freemangordoncommits pushed02:02
merlin1991btw could you look into something for me?02:03
freemangordonwhich one?02:03
merlin1991runing blkid as root shows you some info about all partitions in the system02:03
merlin1991but when I reformatted my sd card to only have one partition I still have ghost entries in the output from blkid02:03
merlin1991even after a reboot02:03
freemangordonyou want me to reformat my sdcard?02:04
merlin1991well simply adding a partition in the padding on the end02:04
merlin1991running blkid and removing it again02:04
merlin1991should be sufficient to reproduce :)02:04
merlin1991you'll have the partition that is long gone in the list even if you reboot your n90002:05
merlin1991and imo that is a bug :D02:05
freemangordonhmm, won't work, my sdcard has 3 partitions and there is no free space (afaik). but I can try on my development device02:06
merlin1991btw in case you wonder, blkid is part of e2fsprogs02:06
freemangordonblkid reads partition table, ain't?02:06
merlin1991nfc02:07
freemangordonhmm, very strange, that is the last line here - /dev/ramzswap0: TYPE="swap"02:07
freemangordonso I could just rmmod ramzswap02:07
freemangordonmerlin1991, yep, there is a bug, blkid still reports /dev/ramzswap0 az swap partition, even after there is NO /dev/ramzswap0 :D02:09
freemangordons/az/as/02:09
infobotfreemangordon meant: merlin1991, yep, there is a bug, blkid still reports /dev/ramzswap0 as swap partition, even after there is NO /dev/ramzswap0 :D02:09
merlin1991and if you reboot it will still report it :D02:10
freemangordonnice02:10
freemangordondooes your card have several partitions?02:10
merlin1991I went from 3 to 1 partition02:10
merlin1991but it still remembers partition 2 and 302:11
freemangordondamn, you don't classify as a guine pig then :(02:11
freemangordon*guinea02:11
merlin1991freemangordon: I can play guinea pig no problem02:11
merlin1991that sd card holds no valuable information :D02:11
freemangordonfor the new gnome-vfs02:11
freemangordonooh, another bug, ke-recv cannot rename ext3 partitions02:12
freemangordonand if you rename fat partition, you need to reboot so new name to appear in filemanager02:12
freemangordonactually 2 bugs02:12
freemangordonmerlin1991, then install new gnome-vfs from the repo , re-partition your sd-card to have 2 partitions and play a bit with ossofilemanager02:13
freemangordon1st partition vfat02:14
freemangordon2nd - initially no fstype, after that swap, and after that ext302:14
freemangordonI am curious if my changes would deal with that as intended02:15
merlin1991damn now I'll have to play with fsckd fdisk02:15
freemangordonfdisk?02:15
freemangordonos sfdisk02:15
freemangordon*or02:15
freemangordonaah, and initially do not set labels on partitions02:16
freemangordonthen set a label on ext302:16
freemangordonand then on vfat02:16
merlin1991hm nfc how to set lables actually :D02:17
freemangordonis going to have something for dinner, will be back in a couple of minutes02:17
merlin1991I've been using gui for paritioning all the time xD02:17
freemangordonconnect via usb and use your pc02:17
merlin1991runing windows currently :/02:18
freemangordonyou can use e2fstune to set up a label for ext302:18
freemangordonand n900 itself for vfat, but that will need a reboot (see the bugs above)02:19
freemangordonI meant the rename function in filemanager02:19
*** gri has quit IRC03:11
*** LaoLang_cool has joined #maemo-ssu03:25
*** gri has joined #maemo-ssu03:34
*** NIN101 has quit IRC03:34
*** gri has quit IRC03:39
*** NIN101 has joined #maemo-ssu03:39
*** NIN101 has quit IRC03:40
*** gri has joined #maemo-ssu03:41
*** LaoLang_cool has quit IRC03:43
*** arcean has quit IRC03:46
*** LaoLang_cool has joined #maemo-ssu03:46
*** M4rtinK has quit IRC04:29
*** nox- has quit IRC04:41
*** M4rtinK has joined #maemo-ssu04:49
*** javispedro has quit IRC04:53
*** M4rtinK has quit IRC04:56
*** int_ua has joined #maemo-ssu05:29
*** int_ua has quit IRC05:30
*** amiconn_ has joined #maemo-ssu05:43
*** amiconn has quit IRC05:43
*** amiconn_ is now known as amiconn05:43
*** LaoLang_cool has quit IRC06:08
*** merlin_1991 has joined #maemo-ssu07:54
*** merlin1991 has quit IRC07:55
*** Jade has joined #maemo-ssu08:33
DocScrutinizer<merlin1991> (you'll have the partition that is long gone in the list even if you reboot your n900)  WTF?!08:54
DocScrutinizermerlin_1991: arcean: could we go thru *all* gconfkeys eventually? I feel like quite a lot of mess there, maybe we can do a major refactoring08:56
DocScrutinizer(re partition table, labels) this is a problem that exists for all flavours of linux/unix. Actually you'll find yourself fixing upstream kernel stuff09:02
DocScrutinizerunix never planned to have hot swappable drives, and there are *still* quite a number of problems atound all that, which usually get solved the "windows way" aka "mandatory reboot"09:03
DocScrutinizeractually a really removable drive will see complete new initialization on plugin, afaik. For changes on fundamental drive structures/properties like partition table there's not yet a proper foolproof concept to do that re-initialization without such unplug&replug or reboot, afaik09:06
DocScrutinizermerlin_1991: NFC though how an obsolete drive-info could survive reboot. *might* be related to friggin partition table concept, which iirc has two redundant ways to describe drive and partitions, and not all tools work on both concurrently to keep them in sync09:09
DocScrutinizerfirst test to dissect the problem: create a uSD that shows the problem, see if the problem sticks to the uSD when you insert the uSD to a "new" device09:11
freemangordonDocScrutinizer, no need for SD card, a simple compcache drive is enough09:12
DocScrutinizer??09:12
freemangordons/drive/device/09:12
infobotfreemangordon meant: DocScrutinizer, no need for SD card, a simple compcache device is enough09:12
DocScrutinizerhow you gonna insert that to a "new" N900?09:13
freemangordonmodins?09:13
freemangordoninsmod, sorry, still having my coffe09:13
freemangordon:D09:13
DocScrutinizernfc what you're talking about09:13
freemangordonby doing inmod ramzswap.ko, you are creating /dev/ramzswap009:14
DocScrutinizerthat's completely unrelated and irrelevant here09:14
freemangordonand that is listed by blkid09:14
freemangordonsee backscroll09:14
DocScrutinizerwhat in >>first test to dissect the problem: create a uSD that shows the problem, see if the problem sticks to the uSD when you insert the uSD to a "new" device<< makes you think this could be done with ramzswap?09:15
DocScrutinizerthere's a couple of potential causes around the topic at large - kernel's device tables not really being flushable is just one, properties/structure of partition-table of one drive is another09:18
freemangordonDocScrutinizer, if you are talking about blkid bug, it could be triggered by using a virtual device, no need for a real SD card partition09:18
DocScrutinizerfor dissecting the "sticky across boot" problem which is the most severe one in my book, we want to tackle partition table the way I suggested above first09:20
freemangordonand compcache/ramzswap is such a device, blkid reports /dev/ramzswap0: TYPE="swap" long afater there is no such device in /dev09:20
freemangordonok09:20
freemangordonbut don't really see the difference09:21
DocScrutinizerafter solving this problem, we check what persisted and tackle those09:21
freemangordonanyway, I need more coffee09:21
DocScrutinizerthe non-flushable device tables in kernel are a known problem, since ages09:21
DocScrutinizerbut they won't explain why problem is sticky across boot09:22
* DocScrutinizer also runs for another coffee09:22
DocScrutinizerthough daywork waiting for me, need to wake up first09:23
DocScrutinizeralso all please make damn sure you check back on stock fremantle with stock kernel and manual (u)mounting, to get a clear idea where from the problems stem that we're chasing09:25
DocScrutinizerI anticipate most of them appear on stock as well09:26
DocScrutinizerfor problems clearly introduced by new ke-recv we know whom "to blame" and fix. For problems that are also in stock kernel and system even, we have to ponder where and how, or if at all to fix them. Might be mission impossible if the problem even exists in upstream 3.x kernel still09:30
DocScrutinizerfreemangordon: actually it doesn't matter at all if you use ramzswap partition or whatever else to create the problem, the point is we create the problem on a partition table we can insert to a new system that never before seen this problem, so we find out if the problem sits (and persists across boots) in partition table itself or rather somewhere in the rootfs09:39
DocScrutinizeror (worst case) in both together, in an either AND or OR way09:41
DocScrutinizer1) create problem  2a) "swap rootfs" to a "clean" one (by moving uSD to new system, or by restoing a clean rootfs backup to the affected device) 2b) on device with freshly created problem swap MBR/part-tbl of the affected storage to a known good version, via dd; reboot 3) evaluate your findings to conclude what's the nature of the issue and where to investigate further09:46
DocScrutinizermy suggestion above been about 2a)09:47
DocScrutinizerclearly done the most easy way by swapping uSD to a "new" device09:47
*** DocScrutinizer has quit IRC09:56
*** DocScrutinizer has joined #maemo-ssu09:56
*** dafox has joined #maemo-ssu10:05
*** dafox has quit IRC10:08
*** dafox has joined #maemo-ssu10:09
*** dafox has quit IRC10:11
*** dafox has joined #maemo-ssu10:21
*** LaoLang_cool has joined #maemo-ssu10:24
*** Jade has quit IRC10:31
*** dafox has quit IRC10:35
*** dafox has joined #maemo-ssu10:36
DocScrutinizerman addpart, man delpart10:40
DocScrutinizerhttp://www.cyberciti.biz/tips/re-read-the-partition-table-without-rebooting-linux-system.html10:44
DocScrutinizerman partprobe10:47
DocScrutinizerit's highly unclear (yet, to me) who (which programs - like sfdisk - or which daemons) does inform kernel to reread partition tables, esp on fremantle10:51
DocScrutinizeralso it's unclear how much of the infra needed to support all this is actually implemented in fremantle kernel, not to mention whether all that got tested and is bugfree or not10:52
DocScrutinizeranyway keep in mind there's - afaik - nothing in kernel or userland that's supposed to keep sticky as in across-boot any info about partition tables of attached devices10:54
DocScrutinizermaybe except /dev fs itself10:54
DocScrutinizerdepending on what the system is utilizing to create device nodes (udev, devtmpfs, ke-recv, manual mkdev calls, whatever), you easily might find an obsolete devicenode in /dev, or miss a devicenode for a new device or partition. Esp when "hotplugged" or - worse - created "on the fly" as new partition on a storage dev10:58
DocScrutinizeralso I think that while /dev/sdx (example) gets created by e.g. a udev rule, yet /dev/sdxN etc most likely are created vy the kernel driver itself, on invocation11:11
DocScrutinizers/vy/by/.11:12
*** juzn900 has joined #maemo-ssu11:16
*** juzn900 has left #maemo-ssu11:21
*** LaoLang_cool has quit IRC12:03
*** Guest50125 has joined #maemo-ssu12:25
DocScrutinizeror put simply: "<merlin199> runing blkid as root shows you some info about all partitions in the system. but when I reformatted my sd card to only have one partition I still have ghost entries in the output from blkid. even after a reboot[!!!]" and would you comment/correct my thoughts regarding that, as of http://mg.pov.lt/maemo-ssu-irclog/latest.log.html#t2012-04-10T09:20:06 ff, please12:26
DocScrutinizeroops12:27
DocScrutinizerECHAN12:27
*** StyXman has joined #maemo-ssu12:38
*** arcean has joined #maemo-ssu12:40
*** M4rtinK has joined #maemo-ssu12:52
*** dafox has quit IRC12:57
*** ivgalvez has joined #maemo-ssu12:58
*** dafox has joined #maemo-ssu12:58
*** schen has joined #maemo-ssu13:25
*** lizardo has joined #maemo-ssu13:29
*** schen has quit IRC13:29
*** juzn900 has joined #maemo-ssu13:43
*** schen has joined #maemo-ssu13:47
*** wmarone_ has quit IRC13:49
*** wmarone_ has joined #maemo-ssu13:49
ivgalvezfreemangordon: ping13:51
*** Guest50125 has quit IRC14:00
freemangordonivgalvez, pong14:19
*** juzn900 has left #maemo-ssu14:29
*** arcean_ has joined #maemo-ssu14:30
*** arcean has quit IRC14:33
*** arcean_ is now known as arcean14:45
ivgalvezI have commented some questions about promotion of KPv50 in http://talk.maemo.org/showpost.php?p=1190361&postcount=18315:03
ivgalvezI'm afraid some people may complain about promoting it having issues with commercial applications15:04
*** schen has quit IRC15:08
freemangordonivgalvez, I saw that, but what are the issues , I am not sure that abc or bless n900 have problems because of KP15:09
freemangordonand joikuspot already should be ok15:09
*** merlin_1991 is now known as merlin199115:13
*** merlin1991 has joined #maemo-ssu15:13
*** M4rtinK has quit IRC15:24
ivgalvezI didn't know Joikuspot was OK with KP5015:30
ivgalvezbut it wasn't at least until v48 right?15:30
ivgalvezand with BlessN900 / ABC it seems that reinstalling fcam drivers from devel is needed.15:32
ivgalvezas it's not a User's package, most people doesn't have the update15:33
merlin1991ivgalvez: about blessn90015:34
freemangordonivgalvez, see my post on TMO15:34
merlin1991or rather fcam drivers15:34
merlin1991that's actually f00bar on the side of the fcam drivers package15:34
freemangordon:nod:15:34
ivgalvezok I've read your response15:35
ivgalvezthanks that's clarifying15:35
freemangordonwell, there will be issues with KP50 going to extras because of the changed OC frequencies, but I will ask Pali (once again) to make some changes to kernel-power-settings, so old frequencies to be automatically replaced with new15:37
ivgalvezare you going to attend the mer-meeting about OBS?15:37
freemangordonivgalvez, what am I supposed to do there? My question is serius15:38
freemangordons/serius/serious/15:38
infobotfreemangordon meant: ivgalvez, what am I supposed to do there? My question is serious15:38
ivgalvezyou are more capable than most of the tipycal chit chatters15:38
ivgalvezs/tipycal/typical/15:39
infobotivgalvez meant: you are more capable than most of the typical chit chatters15:39
freemangordonwell, i *might* be more capable as a coder, but still...15:39
ivgalvezand having a Community OBS for Maemo is something that in the short term could help in development15:39
ivgalvezI'm thinking about the possibilities of upgrading the toolchain15:40
ivgalvezAapo was trying to upgrade gcc to 4.6 with no luck in the autobuilder15:40
freemangordonagree, but AIUI SD69 will attend the meeting, and I expect more people from Maemo/Harmattan to join15:41
ivgalvezat least things are beginning to move again15:42
freemangordonivgalvez, lets look it from that angle - noone can expect a couple of people to do all the work, if there is no enough active community members, then let it be that way15:42
freemangordonyeah, lets see in what direction15:43
freemangordon:D15:43
freemangordonand re new toolchain - CSSU does not rely/use on autobuilder and if a new toolchain is needed that could be easily achieved15:44
ivgalvezgood to know15:47
freemangordonin fact I have working 4.6.2 in one of mine SB targets :)15:51
ivgalvezone more question, I have seen recent commits for tracker package15:52
ivgalvezare you going to upgrade the deprecated version we are using? I've seen a lot of supposed improvements in performance and battery consumption15:53
ivgalvezin latest tracker versions15:53
freemangordonivgalvez, better ask Pali, I am not aware what is the case with tracker :)15:56
ivgalvezok, thanks15:58
*** ivgalvez has left #maemo-ssu16:01
gulywhich twitter client do people here use?16:03
freemangordonmIRC16:03
freemangordonoops16:03
freemangordontwitter?16:03
freemangordonsorry, don't use any :)16:04
gulymirc with pizza script? ghgh16:04
merlin1991guly: I use tweego16:16
merlin1991btw anybody got the obs meeting time and date at hand?16:17
DocScrutinizerbtw any comments on my comments?16:19
merlin1991yeah16:23
merlin1991it makes no difference if it's our ke-recv or the pr1.3 ke-rcv16:23
merlin1991blkid outputs nonsense regardless16:23
Lava_Croftfor the fans: http://forum.xda-developers.com/showthread.php?t=154740816:47
DocScrutinizermerlin1991: so the problems are genuine maemo problems16:50
DocScrutinizerquite expected16:50
DocScrutinizersince renaming and changing partitions doesn't even work under a lot of "desktop" linux'16:51
DocScrutinizernevertheless, as mentioned this morning, your "sticky" problem is a class on its own, and needs some more investigaton16:52
DocScrutinizerfirst of all I'd like to see it reproduced/confirmed, and the simple swap-SD-to-other-rootfs/device test done, to rule out it's just your partitioning tool that is messing up part-tbk16:54
DocScrutinizertbl*16:54
Lava_Crofthttp://news.bbc.co.uk/2/hi/programmes/click_online/9711150.stm16:55
Lava_Crofter16:55
merlin1991DocScrutinizer: you can reproduce it without even using an sd card16:59
merlin1991if you create a swap device with ramzswapd16:59
merlin1991remove it16:59
merlin1991reboot16:59
merlin1991blkid still thinks you have it16:59
DocScrutinizerI'm not particularly interested in uSD-free reproduction of the bug, since uSD is the most convenient and easy case to triage17:00
merlin1991well anyway it got reproduced with and without uSD on 3 devices17:00
DocScrutinizerand I wouldn't bet a fly shit on what ramzswap does to system17:01
DocScrutinizerafter all that's SWAP17:01
DocScrutinizermight easily get an own early init job to create that partition during boot, or whatever17:02
merlin1991well to make you happy I've swapped the sd card to my pr1.3 n90017:02
merlin1991and blkid only reports the single partition that should be there17:02
Lava_Croftwas it DocScrutinizer asking for new N900's to buy, not too long ago?17:02
merlin1991Lava_Croft: yes17:03
Lava_CroftAccording to my wife, MediaMarkt here still sells them17:03
Lava_Croftfor 379euros orso17:03
DocScrutinizerand when you insert another SD to the afected device now, will it still show ghosts?17:03
Lava_Croftat least the store nearby here, in Alkmaar apparently has a few left17:03
DocScrutinizerLava_Croft: yep17:03
merlin1991DocScrutinizer: it even shows the ghosts if I don't insert it and reboot the affected device17:03
DocScrutinizerLava_Croft: o.O17:03
Lava_Croftexactly my reaction17:03
DocScrutinizermany thanks17:04
Lava_Crofti have not seen it myself17:04
DocScrutinizeralas ATM I'm broke, since my employer thinks I'm cheaper than a bank17:04
Lava_Croftand while my wife certainly knows what my n900 looks like, i cannot be sure 100% until i saw it myself17:04
Lava_Croftand she can read too17:04
Lava_Croftfine wife of mine17:04
merlin1991:D17:04
DocScrutinizerthank her from me :-D17:05
Lava_Croftmediamarkt website here in NL does not show any n900's17:05
Lava_Crofti will17:05
merlin1991DocScrutinizer: I jsut removed the sd card from the pr1.3 n900 and rebooted17:05
merlin1991blkid still reports the sdcard partition17:05
DocScrutinizerI wonder if MM here in germany can order goods from MM NL17:05
Lava_Croftno idea17:06
Lava_Crofti dont go to mediamarkt17:06
DocScrutinizermerlin1991: AHAAAA!17:06
Lava_Croftmerlin1991: magic17:06
DocScrutinizernot magic, probably friggin static /dev/*17:06
merlin1991no17:06
DocScrutinizerNO??17:07
merlin1991 /dev does not have the partition anymore17:07
Lava_Croftheh17:07
merlin1991ONLY blkid still shows it17:07
DocScrutinizerW?T!F?!17:07
merlin1991see my pastebin from earlier today http://pastebin.com/qR5iMx0117:07
DocScrutinizermerlin1991: that looks like... thoroughly fsckedup :-/17:09
DocScrutinizermerlin1991: please rm  /etc/blkid.tab17:11
DocScrutinizeror rather17:12
DocScrutinizerls  /etc/blkid.tab17:12
DocScrutinizerls -l17:12
merlin1991oh that file contains all the evil entries17:12
merlin1991removing it yields in a perfect list17:12
DocScrutinizerand17:13
DocScrutinizer^W17:13
DocScrutinizer:-D17:13
DocScrutinizerbug squashed? ;-D17:14
merlin1991nah17:14
merlin1991I have to investigate if it is the desired behavious of blkid to store the info in such a way17:15
merlin1991but I smell bug17:15
merlin1991because there also is a /etc/blkid.tab.old file17:15
DocScrutinizerwhat's file perms?17:15
DocScrutinizerDUH?17:15
merlin1991which automatically gets created when you run blkid and already have a .tab file17:15
DocScrutinizershiiiiiit!17:15
merlin1991so my bet is that blkid should write the proper new info but doesn't17:16
DocScrutinizer -w writecachefile17:16
DocScrutinizer              Write the device cache to writecachefile instead of writing it to the default  cache  file  /etc/blkid.tab.17:16
DocScrutinizer              If  you  don’t  want to save the cache to the default file, specify /dev/null.  If not specified it will be17:16
DocScrutinizer              the same file as that given by the -c option.17:16
*** Pali has joined #maemo-ssu17:17
merlin1991hm running blkid -g17:18
merlin1991followed by blkid17:19
merlin1991gives actually proper results17:19
merlin1991now is that weird cache behaviour intentional or not is the question17:19
DocScrutinizerI smell a bug introduced by a "fix" supposedly introduced during maemo adaption of blkid. This adaption probably be done due to >> If  none is given, all devices which appear in /proc/partitions are shown, if they are recognized.<< while we all know (and see in your pastebin) that /proc/partitions on maemo is "wrong"17:20
DocScrutinizerooh, double invokation of blkid resores the correct result? kinda anticipated behaviour, when yu mentioned .old17:21
merlin1991nah first invoking blkid with "-g"17:22
merlin1991which is epxlained as "garbage collect the cache file"17:22
DocScrutinizerBWAHAHA I missed that17:22
DocScrutinizer-g     Perform a garbage collection pass on the blkid cache to remove devices which no longer exist.17:22
* merlin1991 idly wonders WHO WANT A FRIGGIN REPORT FOR DEVICES THAT DON'T EXIST ANYMORE17:22
DocScrutinizerthat's ot intentional, that's to fix the known issue we ran into17:23
DocScrutinizernot*17:23
DocScrutinizersince probing all devices takes eternities, blkid uses a cache17:24
merlin1991hm I notice no difference between invoking blkid without a cache or with the cache17:24
DocScrutinizerusing this cache per definitionem forbids automatic updates of the cache when devices changed17:24
merlin1991in otherwords invoking it after removing the cache file, or when it still has one17:25
merlin1991DocScrutinizer: wrong17:25
merlin1991it actually overwrites in the cache17:25
merlin1991it only doesn't delete17:25
DocScrutinizermerlin1991: blkid is a general tool, no specialized one for maemo17:25
merlin1991I've repartitioned my sd from 2 times ext + 1 vfat (in that order) to 1vfat17:25
merlin1991and then blkid reported 1 vfat 1 ext 1 vfat17:26
DocScrutinizerhmm, it has intelligent cache ;-P17:26
merlin1991it has a foobar cache :D17:26
DocScrutinizerwhatever, you correctly stated it doesn't matter for maemo at all, how fast this shitty thing performs. So we want to wap it up with a wrapper script that changes "blkid $@" to "blkid -c /dev/null $@"17:29
DocScrutinizerwrap*17:29
DocScrutinizerwhich is an exemplary good CSSU fix of a maemo bug17:29
DocScrutinizermv /sbin/blkid /sbin/blkid.real; echo -e '#!/bin/sh\n/sbin/blkid.real -c /dev/null "$@"\n" >/sbin/blkid;chmod 755 /sbin/blkid17:32
DocScrutinizer>>17:34
DocScrutinizer-c cachefile17:34
DocScrutinizer              Read from cachefile instead of reading from the default cache file /etc/blkid.tab.  If you  want  to  start17:34
DocScrutinizer              with  a  clean  cache  (i.e.  don’t report devices previously scanned but not necessarily available at this17:34
DocScrutinizer              time), specify /dev/null.17:34
DocScrutinizerof course it needs thorough testing nevertheless, maybe some idiotic script relies on the caching17:36
DocScrutinizeractually we should review rather than test17:37
DocScrutinizerall locations where blkid gets invoked17:37
DocScrutinizerand no, I'd not suggest to fix the intiscripts that invoke blkid, the wrapper solution is way more robust17:38
Palifreemangordon, if you finding upstream git repo of osso-gnome-vfs2, it is here: http://maemo.gitorious.org/maemo-af/osso-gnome-vfs2/commits/fremantle17:43
DocScrutinizeractually make that  echo -e '#!/bin/sh\n/sbin/blkid.real -c /dev/null\nif [ 4 -eq $? -o x$1 == x-h ]; then echo "via wrapper script $0 to pass -c /dev/null"; fi; "$@"\n" >/sbin/blkid17:48
DocScrutinizerugh17:50
DocScrutinizeractually make that  echo -e '#!/bin/sh\n/sbin/blkid.real -c /dev/null "$@"\nret=$?\nif [ 4 -eq $ret -o x$1 == x-h ]; then echo "via wrapper script $0 to pass -c /dev/null"; return $ret\nfi; \n" >/sbin/blkid17:52
merlin1991I think first we have to investigate what in maemo actually uses blkid17:55
merlin1991:D17:55
DocScrutinizereasy, grep is your friend17:55
merlin1991sure :)17:55
merlin1991grep -r is running soon17:55
DocScrutinizerI gues you can limit it to /etc /var17:56
DocScrutinizerthough... ohwell t900 has nothing to do anyway ;-D17:57
freemangordonPali, wrong, it is here git@gitorious.org:community-ssu/osso-gnome-vfs2.git :P18:04
Palifreemangordon, what is wrong?18:05
DocScrutinizermerlin1991: meh, seems blkid is used nowhere18:18
DocScrutinizerlibblkid is used 'a lot'18:18
*** dafox has quit IRC18:39
freemangordonPali, I was joking, we already have those sources in CSSU (cloned from the address you pointed) plus a fix for wrong partition labels18:40
Palifreemangordon, then OK18:40
freemangordonbtw you can use cssu-devel repo to test the fix18:41
freemangordonmerlin1991, did you send HOWTO use -devel repo to others?18:41
amiconnWhat does maemo use blkid for, btw_18:42
amiconns/_/?/18:42
infobotamiconn meant: What does maemo use blkid for, btw?18:42
amiconnDamn keymap18:42
merlin1991freemangordon: I've sent it to Pali, still have to explain it to arcean18:44
freemangordonok18:44
merlin1991btw I think I've made a config mistake18:44
freemangordonwhy?18:45
merlin1991"distribution" is community-devel instead of fremantle18:46
freemangordonhmm, why is that?18:47
merlin1991because I configured it that way (by mistake)18:47
merlin1991technially it's no problem18:47
merlin1991it only looks "wrong" in the cataloges list18:47
freemangordonno, i meant why it should not be fremantle?18:47
merlin1991it should be fremantle18:47
freemangordonaaah, ok18:48
freemangordonmy old eyes18:48
freemangordonwell, I don't think anyone has this repo set up, so you still can change it18:48
freemangordonand maybe make an announcement thread on TMO after that18:49
freemangordonis it hard to change distribution to correct one?18:50
merlin1991nah should be easy18:50
merlin1991jsut need to find the right options18:50
freemangordondo it then :)18:50
merlin1991the config files are not exaclty straight forward :D18:50
*** Pali has quit IRC18:55
DocScrutinizermerlin1991: I'm inclined to close your virtual ticket as INVALID19:08
DocScrutinizermerlin1991: seems nothing in maemo fremantle is using blkid. So if it doesn't behave as *you* expect, that's no valid ticket19:08
DocScrutinizermerlin1991: I thought you're describing an actual error showing up in maemo $whatever app19:09
DocScrutinizerlike e.g. your ghost partitions would show up in HFM19:11
freemangordonmerlin1991, BTW did you try osso-gnome-vfs2 from cssu-devel according to yesterday's plan?19:11
DocScrutinizerwhich, btw, made me enter panic mode19:12
DocScrutinizer;-P19:12
merlin1991yeah it seems to be fsckd by design19:12
freemangordonmerlin1991, this is a FEATURE, not a BUG :D19:13
freemangordonyou being a developer should know that19:13
freemangordon:P19:13
DocScrutinizeralso you should know how to set a proper SEVERITY tag to your error reports ;-)19:14
merlin1991I never said it is baaaaaaaaaaaaad :D19:17
freemangordonHmm, was it Raimu to report that renaming a partition usign filemanager screws volume labels?19:19
freemangordonRaimu, ping19:20
DocScrutinizermerlin1991: you actually said nothing regarding that19:25
DocScrutinizerI read that in context of our current ke-recv mess, and filled in the missing SEVERITY tag - obviously incorrect, due to missing input. Your name was the guidance19:26
merlin1991tbh I hade no idea about severity, I came across it when debugging ke-recv stuff19:27
DocScrutinizer:nod:19:27
DocScrutinizercommunication failed. Always a problem of both ends ;-)19:28
DocScrutinizeranyway a great opportunity to learn some stuff about the topic at large19:28
freemangordonmerlin1991, what exactly ke-recv stuff are you debugging?19:29
freemangordonthe rename bug?19:29
merlin1991nope had someone where ke-recv wouldn't lead to moutns at all19:29
merlin1991even with proper init.d script19:30
freemangordonaah, ok19:30
freemangordonPali, ping19:30
merlin1991freemangordon: please reupload your package19:42
merlin1991I've changed the distribution and everything but lost your package on the way19:42
freemangordonok, give me a minute19:42
freemangordonready19:44
freemangordonBTW I think all of us should have that repo enabled19:44
merlin1991dang19:45
merlin1991it went back into community-devel19:45
merlin1991seems there still is some gotcha19:45
freemangordonmerlin1991, seems it went in BOTH repos19:47
merlin1991ls19:48
merlin1991oops :D19:48
freemangordon.19:48
freemangordon..19:48
freemangordon0 bytes in 0 files19:48
freemangordon:D19:48
merlin1991hm it is in the fremantle dist19:49
freemangordonnp then19:49
merlin1991is it only 1 package or a few?19:50
freemangordon5 .debs19:50
merlin1991okay please upload once more :$19:52
freemangordonok19:52
merlin1991if it doesn't go to community-devel this time it's fine :)19:52
merlin1991okay seems the setup is proper now19:55
merlin1991btw I've created a .install file aswell http://maemo.merlin1991.at/cssu/community-devel/community-devel-fremantle.install19:56
freemangordonyep19:56
freemangordonyeah,i were just browsing from my n900 to use ot :D19:56
* merlin1991 watches ham19:57
freemangordonme too :)19:58
merlin1991it's currently at preparing installation, I guess that is when it adds the repos and refreshes19:58
freemangordonyeah, the same19:58
DocScrutinizersomebody over at #maemo mentioned there's allegedly some kind of dependencies conflict with new KP50 and KP50-multiboot|uboot (can't remember)19:59
merlin1991oh yes, apt-worker going insane19:59
merlin1991ham rally needs a busy indicator in the main window for when apt-worker is running in circles20:00
DocScrutinizerYEP!20:00
DocScrutinizeradd a statusbar icon to apt-worker20:00
merlin1991I always open a xterm and run top to see when I finally can use the ui again20:01
DocScrutinizerwill help as well on closing HAM20:01
DocScrutinizerwell, I got loadapplet for that ;-D20:01
freemangordonloadapplet?  and how is the battery life with it?20:01
DocScrutinizeran applet that should've been in maemo default metapr20:02
freemangordonisn't that python?20:02
DocScrutinizernah, don't think so20:02
freemangordonhmm,ok, if you say so20:02
merlin1991damn I want the connect now widget to work for bluetooth too :D20:03
DocScrutinizerbut actually you're right it's probably eating a few cpu cycles and causing wakeups20:03
merlin1991the only applet out there for that IS python20:03
*** BCMM has joined #maemo-ssu20:03
merlin1991aka evil20:03
DocScrutinizerI don't mind if it's python, the problem is refresh cycle and suspend or not, and when20:04
merlin1991okay freemangordonwhich package do I have to upade? all of them or only libosso-gnomevfs2-0?20:04
freemangordonfew CPU cycles? possible. And...err... what about memory?20:04
DocScrutinizerI got "plenty of memory" here20:04
freemangordonn950? :D20:04
DocScrutinizerway more than average joe20:04
DocScrutinizeror maybe not... I'll investigate. Thanks for pointing me at it20:06
merlin1991DocScrutinizer: gimme a free -m on your device when it's idle20:06
merlin1991I wonder how much it differs from a mostly unmodified device20:06
DocScrutinizerIroN900:~# free20:06
DocScrutinizer              total         used         free       shared      buffers20:06
DocScrutinizer  Mem:       245540       242880         2660            0         517620:06
DocScrutinizer Swap:       786424       129540       65688420:06
DocScrutinizerTotal:      1031964       372420       65954420:06
DocScrutinizernot what I expected20:07
DocScrutinizerthus "or maybe not..."20:07
freemangordonNokia-N900:~# free20:08
freemangordon              total         used         free       shared      buffers20:08
freemangordon  Mem:       235260       197044        38216            0          49220:08
freemangordon Swap:      1107688       218728       88896020:08
freemangordonTotal:      1342948       415772       92717620:08
merlin1991             total         used         free       shared      buffers20:08
merlin1991Mem:        235260       219768        15492            0         376820:08
merlin1991-/+ buffers:             216000        1926020:08
merlin1991Swap:       786424        76216       71020820:08
freemangordonbut i have PhoneME with Opera loaded and microb and 4 accounts (skype,facebook,gtalk and irc) connected20:08
merlin1991my pr1.3 device has a better ratio though20:09
BCMMDocScrutinizer: my gf's phone has the same ke-rcev bug. i followed the steps that fixed mine, and it says mp-fremantle-community pr is alread the newest version20:09
BCMMdo i continue with dsmetool anyway?20:09
merlin1991BCMM: then she updated already20:09
merlin1991this way you need20:09
merlin1991dpkg --purge --force-all20:10
merlin1991apt-get install ke-recv20:10
merlin1991dsmetool -b20:10
freemangordondpkg --purge --force-all ke-recv20:10
merlin1991yeah20:10
merlin1991what freemangordon said :D20:10
merlin1991silly me20:10
BCMMthanks, doing that20:11
freemangordonDocScrutinizer, you are on omap kernel?20:11
DocScrutinizerfreemangordon: stock kernel20:12
BCMMah, now i forgot to clear the mount point...20:12
merlin1991did it reboot already?20:12
BCMMoh silly me i can just expose it again with mass storage mode20:12
merlin1991well even if it did, you can just unmount clean remount20:12
BCMMnah, playing boot video now20:12
freemangordonyea, that is why your memory is 245540 vs 235260 on mine and merlin's20:12
merlin1991hm I just realized I have pk on my -stable, but omap on my -testing n900 xD20:13
BCMMthanks, mydocs mounts now20:14
DocScrutinizerhow would I spot the process of loadapplet? or is it just a "subroutine" of hildon-foo?20:14
BCMMfwiw, this phone had the same issue with rotation20:14
freemangordonshould make no difference(beside those additional 10MB for DSP and framebuffer)20:14
freemangordonDocScrutinizer, it is hildon-home20:14
DocScrutinizer:nod:20:14
DocScrutinizerthought as much20:14
freemangordonwait, hildon-status-manu20:14
DocScrutinizerwhaever20:15
merlin1991BCMM: I guess the issue is gone now?20:15
DocScrutinizerooh wait, we could compare length :-P err values of hildon-status-menu20:15
freemangordonmerlin1991, new gnome-vfs-whaecer installed?20:16
merlin1991yea20:16
freemangordonhow the things look like now?20:16
merlin1991now trying to find that usim20:16
freemangordonusim?20:16
merlin1991err usd20:16
freemangordonwhat for?20:16
merlin1991whatever :D20:16
freemangordonaah :D20:16
DocScrutinizergrrr20:16
DocScrutinizer 1556 user       712  20   0  3932   532   528 S  0.0  0.2  0:00.00 /usr/bin/hildon-status-menu20:17
DocScrutinizer  PID USER      PPID PRI  NI  VIRT   RES   SHR S CPU% MEM%   TIME+  Command20:18
DocScrutinizer0.0 CPU% 0.2 MEM% I'm not at all concerned20:19
*** javispedro has joined #maemo-ssu20:23
BCMMmerlin1991: yeah works now20:31
*** CepiPerez has joined #maemo-ssu20:33
Raimufreemangordon, pong20:35
freemangordonraimu, was that you who had problems with volume labels incorrectly shown in filemanager after volume rename?20:36
freemangordonRaimu ^^^20:37
merlin1991hm I missed the #mer-meeting20:39
* DocScrutinizer MEHs20:39
DocScrutinizermer vanished from my radar completely20:40
DocScrutinizerI just can't see it leading anywhere20:40
Raimufreemangordon: Yes.20:40
merlin1991DocScrutinizer: I read the minutes, and it looks like nokia is looking into funding parts of their obs structure20:41
freemangordonRaimu, ok, there is supposed to be a fix for that, dare to enable community-devel repo to test it?20:41
DocScrutinizerthat's ewlwvant how?20:41
Raimufreemangordon: Hokay.20:41
DocScrutinizerdang20:41
merlin1991and it also seems that nokia/maemo.org people seem to be interested into replacing autobuilder with obs aswell20:41
DocScrutinizerrelevant*20:41
merlin1991s/into/in/20:42
infobotmerlin1991 meant: and it also seems that nokia/maemo.org people seem to be interested in replacing autobuilder with obs aswell20:42
freemangordonRaimu, here http://maemo.merlin1991.at/cssu/community-devel/20:42
freemangordonthere is .install file to be opened directly from your n90020:42
Raimufreemangordon: Ahah, I got it from here - now what's the file I'm to update?20:42
freemangordonosso-gnome-vfs, but you need 2 .debs, better use HAM to upgrade20:43
merlin1991why did  I decide to get a dns entry with my nick20:43
merlin1991plenty of useless highlights on irc xD20:43
freemangordonyou want to be famous :D20:43
DocScrutinizerhaha20:44
freemangordonmerlin1991, you know, beware what you wish for20:44
freemangordon:D:D:D20:44
merlin1991:D20:44
merlin1991apart from the youtube results, everyhting for merlin1991 (I get) is actually me :)20:45
freemangordonis afk20:45
merlin1991wtf is that "is afk" you keep writing lateley :D20:46
merlin1991*lately*20:46
DocScrutinizer~afk20:50
infobotfrom memory, afk is Away From Keyboard20:50
*** trbs has joined #maemo-ssu20:53
*** CepiPerez has quit IRC20:53
*** CepiPerez has joined #maemo-ssu20:53
*** Pali has joined #maemo-ssu20:56
Raimufreemangordon: Are you quite sure it can be found via HAM?21:00
*** javispedro has quit IRC21:01
merlin1991DocScrutinizer: but he should /me that and not privmsg it21:04
DocScrutinizer??21:05
merlin1991hm which sys / dev node had the fm_tansmitter power level?21:05
DocScrutinizermhm, yeah21:05
merlin1991DocScrutinizer: I meant in that case one writes21:05
* merlin1991 is afk21:05
freemangordonRaimu, just add the repo and update21:07
DocScrutinizererrr, really??21:08
merlin1991Raimu: apt-get install libosso-gnomevfs2-021:08
DocScrutinizerthat's more like it21:08
* freemangordon is not afk21:08
freemangordonmerlin, happy naow?21:08
merlin1991:D21:09
*** mirandir has joined #maemo-ssu21:09
DocScrutinizerinsane, instagram: 30E6 app users, selling for 1E9 USD21:11
DocScrutinizerthat's 33USD/customer21:12
DocScrutinizersuckerberg21:12
RaimuUpgrading.21:12
*** BCMM has quit IRC21:12
*** mirandir has left #maemo-ssu21:13
* DocScrutinizer wishes somebody would pay 33USD per h-e-n user21:13
merlin1991hrhr21:14
*** mirandir has joined #maemo-ssu21:14
DocScrutinizerBWAHAHA another friggin portrait bug21:17
DocScrutinizeropening call history: LS. clicking an entry->PRT. closing this requester by clicking background-> call history in PRT. taskswitcher, select call history: LS21:19
DocScrutinizercongrats portrait mode, that's really fsckdup21:19
Raimufreemangordon: Okay, done. What to test now?21:28
DocScrutinizerhehe, the requester that appears after clicking an entry in call history also rotates back to LS, when you lock and unlock21:30
DocScrutinizerX-P21:30
DocScrutinizerdid I mention I *adore* portrait mode?21:31
Raimufreemangordon: Well, after the update renaming partitions randomly doesn't suddenly give all other partitions names too.21:33
Raimufreemangordon: Looks like fileman doesn't display the new names though they're renamed properly.21:34
RaimuI can live with that21:34
*** freemangordon_ has joined #maemo-ssu21:36
DocScrutinizerwell, I bet there's no method in filemanager to rescan names of directories and partition names, in case they get renamed. It should, however, be aware of such renames when they are done inside filemanager itself21:39
*** M4rtinK has joined #maemo-ssu21:40
freemangordon_actually there is, the problem is that even hal is not aware of the rename21:40
*** nox- has joined #maemo-ssu21:55
*** dafox has joined #maemo-ssu21:55
DocScrutinizerhow could it, it's not designed to monitor for changes ;-D21:56
DocScrutinizerat least not for changes of partition names21:56
DocScrutinizerrather some ahl-addon-foobar I guess21:58
DocScrutinizerhal*21:58
*** freemangordon_ has quit IRC21:59
*** freemangordon_ has joined #maemo-ssu22:16
*** gri has quit IRC22:29
*** gri has joined #maemo-ssu22:29
*** M4rtinK has quit IRC22:44
*** Kaptenen has joined #maemo-ssu22:49
*** Guest50125 has joined #maemo-ssu22:58
*** M4rtinK has joined #maemo-ssu22:59
*** CepiPerez has quit IRC23:01
*** Guest50125 has quit IRC23:06
*** Guest50125 has joined #maemo-ssu23:06
*** javispedro has joined #maemo-ssu23:16
Palifreemangordon, what is state of RE cellular control panel plugin?23:17
freemangordon_bad23:17
freemangordon_it will take lots of time23:18
Paliok23:18
*** mirandir has quit IRC23:30
*** M4rtinK has quit IRC23:43
*** M4rtinK has joined #maemo-ssu23:44

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