IRC log of #harmattan for Wednesday, 2012-01-11

ieatlintraspberry pi went into full production00:01
*** hhartz has quit IRC00:01
*** ajalkane has joined #harmattan00:02
admiral0C# ?00:02
admiral0oh lord why?00:02
djszapibecause it is a real OOP language unlike C++00:04
admiral0it's too verbose00:05
* ieatlint senses a language war approaching00:05
* ieatlint runs00:05
*** djszapi has left #harmattan00:06
*** Venemo_N950 has joined #harmattan00:06
*** Venemo has quit IRC00:07
*** admiral0_n9 has quit IRC00:08
*** Venemo has joined #harmattan00:09
*** NIN101 has quit IRC00:10
*** admiral0_n9 has joined #harmattan00:12
auenffinally, PR1.1 OTA00:21
ieatlintmake sure you backup *everthing* :P00:22
*** admiral0_n9 has quit IRC00:22
ieatlintthe pr1.1 ota update seems notorious for failing and forcing a full reflash00:22
mgedmingood thing I didn't know THAT when I upgraded00:25
mgedminmy upgrade went smoothly00:25
admiral0it failed for me00:25
mgedminI don't know what I'd do if I had to reflash00:25
admiral0had no data on my n9 though :D00:26
*** diggy has quit IRC00:26
mgedminand I've no firmware image, and I'm on linux, so no navifirm00:29
*** zarlino has quit IRC00:31
admiral0mgedmin: there is a navifirm clone in C++ that you can run under wine00:33
admiral0tested and works for me00:33
mgedminI sincerely hope I'll never need to experience it00:34
admiral0flashing is quick and painless if you have backed up00:34
*** diggy has joined #harmattan00:35
*** xTs has quit IRC00:43
ieatlinthah, intel demoed a reference platform smartphone based on the atom chip00:45
leinir*nods* That's the one which /would/ have run meego...00:46
ieatlinti suspect this was to be a meego phone a few months ago :P00:46
leinir(hardly a secret, now, but let's say i'm probably a bit more annoyed about that than most people here)00:46
ieatlintlooks like android in the photos00:46
ajalkaneDoesn't look anything special from photos :(. Was hoping more.00:47
mgedminis atom really power-efficient enough these days to be used in smartphones?00:47
SpeedEvilNot atom00:47
ieatlintajalkane: just a reference platform, it's not supposed to be a retail device really00:47
ieatlinterr, right, SpeedEvil is correct, i said atom and am wrong00:47
ajalkaneieatlint: yeah, true. MeeGo proper didn't look anything too special either from photos.00:47
ieatlintajalkane: that was also just a reference UI, and wasn't supposed to be for retail devices00:48
ajalkaneBut the APIs that were published seems lackluster also.00:48
ajalkaneBut I'm still hopeful. Lack of Qt is a major bummer though.00:49
ieatlintyou mean tizen?00:49
ajalkaneare talking about something else?00:49
mgedminhuh? the engadget link claims " 1.6GHz Intel Atom Z2460"00:50
ajalkanes/are/are we/00:50
infobotajalkane meant: are we talking about something else?00:50
ajalkaneoh right, that engadget link seems to claim it's Android.00:51
ieatlintmgedmin: i don't even now know00:51
ieatlintmy intel lingo is outdated... is medfield a line of atom cpus?00:52
kimjuthe whole platform, including cpu and peripherals.00:55
*** Iktwo has quit IRC01:01
*** wirwe has quit IRC01:06
*** zk8 has quit IRC01:09
ajalkaneah fuckity doo. I've been trying to figure out for 3 days what's wrong why my code does not work. Seems like it was qt-components bug with TextField, compiling from snapshot solved the problem. So beware Nokia FN-PPA qt-components package, it's out-dated.01:19
*** rm_work has quit IRC01:36
ieatlintyeah... i hate it when that happens :(01:46
*** ajalkane has quit IRC01:49
*** Venemo has quit IRC01:50
*** Venemo_N950 has quit IRC01:51
*** tbf_ has quit IRC02:01
*** diggy has quit IRC02:03
*** Natunen has quit IRC02:16
*** diggy has joined #harmattan02:21
*** norayr has joined #harmattan02:25
*** diggy has quit IRC02:46
*** M4rtinK has quit IRC02:52
*** Sazpaimon has quit IRC02:59
*** Sazpaimon has joined #harmattan02:59
*** norayr has quit IRC03:07
*** JackaLX has quit IRC03:16
*** JackaLX has joined #harmattan03:19
*** Smtih has joined #harmattan03:21
*** adlan has joined #harmattan03:24
*** adlan has quit IRC03:28
*** storm_ has joined #harmattan03:28
*** adlan has joined #harmattan03:28
*** storm_ has quit IRC03:29
*** storm900 has joined #harmattan03:29
*** foobar_ has joined #harmattan03:30
*** foobar_ is now known as Guest7171803:31
storm900hello all, just installed nemo on my n900 and loving it !!!03:32
*** Guest71718 has quit IRC03:33
*** adlan has quit IRC03:39
*** adlan has joined #harmattan03:39
*** storm900 has quit IRC03:44
*** Smtih has quit IRC04:04
*** adlan has quit IRC04:09
*** adlan has joined #harmattan04:09
*** JackaLX has quit IRC04:31
*** Sazpaimon_ has joined #harmattan04:34
*** Sazpaimon_ has quit IRC04:34
*** Sazpaimon_ has joined #harmattan04:34
*** Sazpaimon_ has quit IRC04:34
*** Sazpaimon_ has joined #harmattan04:34
*** Richlv has quit IRC04:35
*** Richlv has joined #harmattan04:35
*** Sazpaimon has quit IRC04:37
*** npm_ has joined #harmattan04:50
*** npm has quit IRC04:51
*** JackaLX has joined #harmattan04:55
*** JackaLX has quit IRC05:06
*** JackaLX has joined #harmattan05:35
*** JackaLX has joined #harmattan05:35
*** Natunen has joined #harmattan05:49
*** JackaLX has quit IRC05:50
*** JackaLX has joined #harmattan05:53
*** risca has joined #harmattan05:53
*** djszapi has joined #harmattan06:52
djszapiieatlint: "hah, intel demoed a reference platform smartphone based on the atom chip" -> do you have a link by hand ?06:54
*** JackaLX has quit IRC07:00
*** JackaLX has joined #harmattan07:03
*** risca has quit IRC07:27
djszapianybody from the US ?07:29
*** xarcass has joined #harmattan07:30
Mekyes, why?07:31
*** hardaker has quit IRC07:37
*** djszapi has left #harmattan07:49
*** admiral0 has quit IRC07:50
*** admiral0 has joined #harmattan07:50
*** luke-jr is now known as luke-jr_07:54
*** luke-jr_ is now known as luke-jr07:54
*** vladest has quit IRC07:57
*** admiral0 has quit IRC08:06
*** mece has joined #harmattan08:07
*** trx has quit IRC08:10
*** mece has quit IRC08:26
*** DocScrutinizer has quit IRC08:26
*** DocScrutinizer has joined #harmattan08:26
*** mece has joined #harmattan08:30
*** JackaLX has quit IRC08:36
*** beford has quit IRC08:44
itsnotabigtruckSpeedEvil: What matters here is what's inside: this 4-inch handset packs a single-core 1.6GHz Intel Atom Z2460 chip, XXM 6260 modem and Intel GMA graphics < yes atom08:58
*** ZogG_N9 has joined #harmattan09:06
ZogG_N9no talks here lately09:07
*** ZogG_N9 has quit IRC09:07
*** zarlino has joined #harmattan09:13
*** zarlino_ has joined #harmattan09:17
*** zarlino has quit IRC09:18
*** zarlino_ is now known as zarlino09:18
*** Tuju has quit IRC09:27
*** sroedal has quit IRC09:27
*** sroedal has joined #harmattan09:27
*** fuz_ has quit IRC09:36
*** fuz_ has joined #harmattan09:41
*** gabriel9 has joined #harmattan09:57
*** norayr has joined #harmattan09:58
*** norayr has left #harmattan09:59
*** zarlino has quit IRC10:07
jonniZogG_laptop: did you manage to get right orientation for your CameraCapture.qml yet?10:12
jonniZogG_laptop: (I have a working patch for you if you havent...) :)10:13
*** JackaLX has joined #harmattan10:15
*** mece has quit IRC10:18
*** lamikr has joined #harmattan10:27
*** DocScrutinizer has quit IRC10:42
*** DocScrutinizer has joined #harmattan10:42
*** Stecchino_ is now known as Stecchino10:44
*** tbf has joined #harmattan10:44
*** djszapi has joined #harmattan10:46
djszapiAvengence: if you can guess key that was used to encrypt luks filesystem that is in backup, then it's possible. used key is random number that is stored on calibration area, and others using sim card10:46
djszapiin pr1.1 database is encrypted with aegis, and only signond can open it. another possibility is to have r&d device where you can install modified signond or create client with enough security tokens to be able to query those passwords. also replacing authentication plugins with ones that prints out given passwords is possible10:46
*** jreznik has joined #harmattan10:47
*** vladest has joined #harmattan10:58
*** trx has joined #harmattan11:19
*** jpnurmi has joined #harmattan11:35
djszapiieatlint: someone is buying the Lumia 800 for 450 EUR on a Finnish forum :)11:42
*** correne has joined #harmattan11:55
*** vladest has quit IRC11:56
*** diggy has joined #harmattan11:56
*** Avengence has quit IRC11:57
*** correne is now known as Avengence11:58
Avengencedjszapi: calibration area is in the device only or patr of the backup?11:58
*** mece has joined #harmattan11:59
*** rcg1 has joined #harmattan12:02
djszapiit is not tied to the backup. :)12:05
Avengencelooking through dmesg last night, i see onenand spitting out correctable ECC error 96 times in about 12 hours. the addr1 and addr8 are always the same. is that really one block that keeps having the read area, and if so, why is the data not being moved to a new location and that location added to the badblocks list?12:07
Avengencedjszapi: so, device backup is then not portable to a new device.12:07
Avengencedjszapi: when you asked about someone from US, you mean living there now or had come from there at some point?12:08
djszapiI have had a license issue for 1-2 months about a Qt Project, and Nokia proper does not seem to give it too high priority. I tried to write an email to the company, called Creative Labs, but they did not answer it either. I was about to call them, but wanted to first know whether the call prices are the same inside the US altogether so that I could ask a familiar for this call.12:10
djszapi(Creative Labs seems to reside in the US for making it explicitely clear)12:10
Avengenceah, calls from abbroad to US are generally all the same price. if you have ability to call from a phone in US, then cost is either free or whatever the rate is for long distance according to the service on that phone.12:12
Avengenceis it an 800/888/866 number?12:13
*** mega2000 has joined #harmattan12:14
djszapieither 800 or 405:
*** mega2000 has quit IRC12:14
*** lucido has quit IRC12:15
*** mega2000 has joined #harmattan12:15
djszapi -> Their mailing lists are down, forum has been closed. They do not answer for contact emails. Difficult case. :)12:16
Avengenceok, 800 is totally free to call from anyt phone in the us, but you can't call it internationally. 405 you can call international, you can call from the US and there may be a charge if the number you are calling from is outside 405 area code. depends if you pay for long disttance by use, typical for non-business landline, or pay for all calls by time, typocal for cell phones12:17
*** mega2000 has quit IRC12:19
*** mece has quit IRC12:19
djszapiok, thanks. I can ask one friend to call them since it is probably at least 10-20 minutes initially for them to realize where to redirect the person in such a big company. =)12:19
djszapiAvengence: back to your question. It is portable if you use the same sim.12:21
*** mece has joined #harmattan12:23
Avengenceah, so the sim holds the key. not sure how hard it is to read that out12:26
*** lizardo has joined #harmattan12:27
Avengenceif you have a gmail account that google hasnt yet flagged as outside the US, you can call with google voice from the gmail page12:29
djszapisome constant data is encrypted with sim to get key, so you need to use sim to get key, so it cannot be read out12:29
Avengencethat's what i was thinking, have to ask sim to decrypt some wrapped key but can't actually get anything out of the sim12:30
djszapiI have a gmail address ;-)12:30
Avengenceso try calling that way, it'll either be free or if they think you are outside US then they will say you have to pay 1cent/minute12:31
*** adlan has quit IRC12:32
Avengencei guess the restore backup to different device with same sim works for if the phone dies, but still leaves you stuck if the phone gets stolen or the sim is otherwise unusable12:32
djszapiyou can use up to 7 sims, in stolen device case most important is that passwords cannot be get out of device than user losing backuped passwords12:37
Avengenceideal design: ALL user data is on encrypted filesystem. filesystem keys are wrapped using login password. login password unwraps keys to mount filesystem, but password isn't stored anywhere. the backup is the fs image plus wrapped keys. backup is fully portable and user data is very difficult to retrieve from stolen device, especially if too many failed attempts blow away the wrapped keys12:41
*** Venemo has joined #harmattan12:41
*** Venemo has joined #harmattan12:41
djszapiyes, that was on original design, backup had passphrase to open it, but to make ui simple, that dialog was dropped. And leaving tie to sim as only portable way to open backup12:46
SpeedEvilSIM or IMSI?12:48
Avengencewhen you said that loosing stored passwords from backup in stolen case is ok, my thought is what about the rest of the backup? is the rest of the data not encrypted, or is it also inaccessible from backup in that case?12:48
SpeedEvilAvengence: Read errors are expected.12:49
Avengencethere is a LOT more than just stored passwords that I want to have inaccessible to anyone else that might take physical posession of my phone. and by inaccesible, I mean encrypted and potentially overwritten in case of attack12:49
SpeedEvilAvengence: The onenand is specified something like 'xMb of blocks without errors exceeding the threshold after 100000 writes'12:50
AvengenceSpeedEvil: read errors from flash is expected, but in that case the block should be removed from use. so I was wondering, does linux report address accurately, meaning that is same block failing again and again, and if so, whu isn't it reallocated?12:50
SpeedEvilSubthreshold errors do not trigger marking as a bad block, as it's not.12:50
SpeedEvilIt's only once they get close to being uncorrectable that they do12:51
djszapimetadata is plain text, but passwords are in encrypted, that protection is done by lock code12:51
Avengenceif it requires ECC correction to read, then its going bad and better to get the data out now rather than wait until enough bits are bad that it can't be read without errors12:51
SpeedEvilAvengence: Doesn't work that way12:51
SpeedEvilAvengence: There are specified error rates for 'good' blocks.12:52
djszapiand it does what you ask for, but without lockcode most of data can be get out of phone.12:52
SpeedEvilAccording to the manufacturer of the NAND.12:52
* Avengence wonders if adherence to the specified error rates have anything to do with the miserably short life of SSDs and SD cards before unrecoverable read errors make the devices useless12:53
SpeedEvilThere are specified in the datasheet thresholds, and what you should do on a correctable block.12:53
Avengencedjszapi: metadata is in clear, but what about stuff like contact list, mailbox, etc? are those encrypted or clear?12:53
SpeedEvilWith SD/SSD, you're more vulnerable than with onenand.12:53
SpeedEvilWith onenand, blocks go explicitly bad.12:53
djszapiif you insert lock code incorrect multiple times, clear device is used to clear all stored data from it12:54
djszapithis is called "wiping".12:54
SpeedEvilWith SSD/SD - blocks are always fictionally ideal, and work fine, up until the spare space to replace failed blocks is exhausted at which point you get irrecoverable block errors with no warning.12:54
djszapiand there is a syslog entry nowadays about it as well, why the swiping happened.12:55
SpeedEvilWith unpredictable other side-effects.12:55
Avengencedjszapi: lock code is good an all, but other concern is when the flash is read directly. is that user data stored encrypted or clear on the flash?12:56
djszapiI cannot remember which part of data is protected with aegisfs, but some of those are12:57
djszapi think in backup those are all readable, so phone without lock code is open to read most of things12:57
AvengenceSpeedEvil: when I worked with raw nand some time ago, one error was enough for us to flag it bad and reallocate. it was not this onenand, but I don't really trust manufacturer claims about flash cells. I see errors far too easily and the results are always bad12:58
SpeedEvilAvengence: Well - one error is specified as OK. You're supposed to use the ECC.12:58
djszapiin terminal: "df" -> and looks like MyDocs is aegisfs12:59
AvengenceI've had SD cards start corrupting data within a month of use and formatting sometimes helps but only for a while. I've had SSDs fail in weeks, one within 2 days, and one was basically DOA. All those failed SSDs returned bad data or no data on reads with no warning and apparently no attempt to reallocate at all13:00
djszapi(so most of things are encrypted on flash)13:00
SpeedEvilAvengence: Problem is you don't know that.13:00
SpeedEvilAvengence: You have no insight as to what the underlying controller is doing to the raw NAND.13:00
jonnibasicly passwords are the only data that is quite safe, all the meta data is quite open (since you can bypass devicelock screen with few tricks)13:00
Avengencedjszapi: would be MUCH better is /home/user was encrypted. MyDocs has little of importance, compared to say /home/user/.qmf which is not protected as far as I can tell13:01
jonniand would be much slower :)13:01
SpeedEviljonni: Not really.13:01
SpeedEvilMaybe use more power13:02
AvengenceSpeedEvil: I have no insight on SSD/SD, but since I know how raw nand works and have written code to deal with it many years ago, I know that these devices are clearly NOT dealing with it correctly, otherwise they would not have the massive failure rate I've seen13:02
SpeedEvilBut the chip has a crypto core13:02
*** snowpong has joined #harmattan13:02
*** vladest has joined #harmattan13:02
SpeedEvilAvengence: It depends.13:02
Avengenceflash is already SLOW. the time to decrypt/encrypt on cpu is less than the read/write time of flash13:02
SpeedEvilAvengence: Are you seeing errors that are in fact due to controller error, or hard unspecified errors of the flash chip.13:03
*** lucido2 has joined #harmattan13:03
SpeedEvilcontroller error as in handling the raw NAND incorrectly13:03
SpeedEvilIt would be so nice if SD/SSD had a 'get out of my way and show me the raw NAND'13:03
AvengenceSpeedEvil: you mean on my N9 or on the SSd/SD devices I've seen crap out FAR faster than any spinning disk?13:04
djszapibest would be that everything is encrypted, but that will also cost in performance13:04
SpeedEvilAvengence: both13:04
Avengencei can't tell in either case, not enough data is exposed13:04
SpeedEvilAvengence: Unless you have both source, and logs from the SD/SSD controller, you can't speculate meaningfully.13:04
lucido2hello, can I modify the /etc/ssh/sshd_conf to allow my subnem or will it trigger some aegis lock?13:04
SpeedEvilI note that at least some failures on n900 are simply bond failures, and reflowing the eMMC makes it work again13:05
SpeedEvil- one user reported this13:05
AvengenceI know in one case the controller on the SSD failed, leaving the only course to physically destroy the device and forgo warranty since there was no way to erase the data before returnig it13:05
djszapithere is accelerator for crypto, but it consumes battery too.13:05
Avengencei can't accurately speculate, but it meaningful to say they are doing something just plain wrong13:05
jonnilucido2: that file is not protected by refhashlist, so you can edit it. you can check with grep filename /var/lib/aegis/refhashlist13:06
SpeedEvilAvengence: It might be, but you can't really say that, over mechanical/electrical defects.13:07
SpeedEvilAvengence: Also - encryption13:07
Avengencefor myself personally, I refuse to use SSD for anything until I get one that exposes the raw nand. I don't trust the controllers, the firmware on all of them is loaded with bugs and adds complication. the performance and flash life would both improve if they were directly used by a filesystem that was written for flash. using a filesystem designed for spinning disk atop flash with a remapper is completely stupid.13:07
SpeedEvilvery much so.13:08
*** lucido2 has quit IRC13:10
Avengenceugh, why are there no manpages on N9?13:13
jonnitakes too much room13:13
SpeedEvilUsers are all expected to be experts, so not need manpages.13:13
djszapiit was a high demand 2 years ago to reduce those.13:14
astraljavaWho users manpages locally anymore, when you got google & all the forums in the world? Besides, aren't we supposed to have upgraded to TexInfo already? :D13:15
AvengenceI have 64GB of space, its mostly empty. manpages are important when using stuff like busybox where its up to the whim of the person compiling it what optiopns actually are present13:15
SpeedEvilThe manpages are the least of my worries about the platform.13:15
Avengencegnu info can fuck off, along with most other gnu trash. when I encounter software that has a manpage saying simply "see info", then I delete that garbage and find a better alternative13:16
jonniAvengence: most of them are massstorege fat, only couple gigs of ext4, so man pages eat too much space13:16
Avengencegoogle is barely useful thanks to seo fags. forums rarely have any correct) solutions13:17
Avengencefat and ext4 on nand? wtf13:18
astraljavaWell, if we're being serious13:18
astraljavaIf you can google, manpages are online.13:18
astraljavaWell okay, not always, but most often.13:18
Avengencemanpages for the N9 specifically?13:18
astraljavaAhh... yeah. Sorry, I didn't realize which channel I was on. Sorry for the noise.13:19
Avengenceor manpages for linux whatever and then just guess at how applicable they are to this instance13:19
jonniAvengence: basicly you have 60gigs vfat which is your mass storege, 2gigs ext4 as /home/ and 2gigs ext4 as /13:19
Avengencewhy cfat instead of ext4? why either of those instead of something like jffs2?13:19
jonniAvengence: because of windows users :)13:20
Avengencei'm so very tempoted to just reroll the whole OS on this thing... homedir is too small, root is too small, most of the storage is trapped in a useless place using a crap filesystem13:21
Avengenceffs, it's a linux device, its going to write to a dos filesystem and not be able to store any proper metadata or have to package that up in some awful hack. that's just.. dumb13:22
jonniAvengence: device has sfdisk for repartioning, and there are few wiki pages out there for resising the N9 partitions13:22
jonniresizing even13:22
djszapiAvengence: I can do a community manpage package for you :)13:23
Avengenceresize is one thing, but I'm looking at changing the layout completely. basically, I was 2 filesystems, root that is in the clear and /home/user that is encrypted (plus whatever stuff like /dev and /tmp that are not real)13:24
SpeedEvilAvengence: There is no good solution.13:24
SpeedEvilAvengence: The only alternative for 'real' users would be to build a fake filesystem on the fly, on plug-in.13:24
Avengencedjszapi: if they are accurate manpages installable from a .deb, that would be great. don't forget to include man itself with the pages13:24
SpeedEvilWhich would imply that the fs would still have to be effectively unmounted.13:24
SpeedEvilI assume there are no manpages for internal nokia stuff alas.13:25
djszapiit is called SDK13:25
djszapibut an average user does not need a manpage after, maximum helps which are available.13:26
Avengencethose would be most useful, but at least some of the nokia stuff has reasonable usage output. i.e. qmftool prints enough to write a basic manpage for it. the stuff in busybox on the other hand doesn't13:26
djszapiafter all*13:26
Avengenceaverage user would have probably hit reset long ago when mail stopped working instead of flipping on dev mode and digging into the problem. after all, that's the only solution posted anywhere on the forums (or the variant, delete all accounts, delete .qmf dir, reboot, try again, then reset device if still borked)13:27
*** lucido2 has joined #harmattan13:29
djszapiwell, a developer can type "man ls" into google.13:29
jonniand you can just scp -r QtSDK/Maemo/4.6.2/sysroots/fremantle-arm-sysroot-20.2010.36-2-slim/usr/share/man  to your device and you have man pages.13:30
djszapi(or just on the host system, so why waste space unneccesarily?)13:30
Avengencebut I also think average user doesn't buy this phone for trhe reason I did. I got pissed at iphone rubbish as apple kept taking away work features (added by hacks) and adding trash i don't want, I bought N9 because I want a prong wayhone open enough that I can control the direction it goes when Nokia drops it or tries to go the w13:30
djszapiyou can control it; I do not understand the problem.13:31
djszapiI am pretty sure (as I said yesterday) there is going to be a post-Harmattan community team making releases.13:31
jonniNemo is then the way to go if you want full control :)13:31
Avengencewell, to be more explicit, the df in busybox on the N9 is missing a whole pile of options and it doesn't match df or man df on my host system13:31
djszapiNemo is a Ui, not platform13:31
djszapiI would use the term "Mer" ;-)13:31
jonniAvengence: you can always download N9 sources for busybox and see it from there :)13:32
Avengenceyes, that was part of the buying decision. I know some people with N800 and N900 who use community built OS instead of what Nokia gives. I intend to do the same, just didn't expect to need to do it so soon13:32
Avengencejonni: I will probably be downloading all the sources Nokia shares once I am back home13:33
djszapiwhen did you jump into the Harmattan world ? (It has been pretty known for a while what is going on) :)13:33
Avengencethis week, when the dead laptop forced me to figure out what is going on with this thing sooner rather than later13:33
*** arekinath has quit IRC13:36
Avengencequick count, df is missing 8 of the 14 flags documented in the 'host' manpage. If replacing the OS, I would most likely drop busybox and include all the real utilities rather than stripped stuff13:36
*** arekinath has joined #harmattan13:37
*** superdump has quit IRC13:37
djszapiAvengence: df is probably busybox df, not really your host linux df.13:39
*** JackaLX has quit IRC13:39
djszapiI would expect the same opportunities, otherwise why busybox after all if not truncated in senses ?13:40
djszapiwould not*13:40
djszapi -> df is properly documented here.13:40
*** superdump has joined #harmattan13:41
*** xmlich02 has joined #harmattan13:41
*** vladest has quit IRC13:43
*** tbf has quit IRC13:43
*** ab has quit IRC13:43
*** merlin1991 has quit IRC13:43
*** leio has quit IRC13:43
*** phunguy has quit IRC13:43
*** nibbler has quit IRC13:43
*** tomma has quit IRC13:43
*** zz_gri has quit IRC13:43
*** Milhouse has quit IRC13:43
*** vladest has joined #harmattan13:47
*** tbf has joined #harmattan13:47
*** ab has joined #harmattan13:47
*** merlin1991 has joined #harmattan13:47
*** leio has joined #harmattan13:47
*** phunguy has joined #harmattan13:47
*** nibbler has joined #harmattan13:47
*** tomma has joined #harmattan13:47
*** zz_gri has joined #harmattan13:47
*** Milhouse has joined #harmattan13:47
Avengencedjszapi: it is documented there, in a way so abbreviated that the original manpage is still needed for the details on the options. I guiess that's to be expected, minimal implementation with minimal documentation. can you tell I hate busybox?13:47
Corsacgrmbl, all sms from/to a contact are magically gone13:48
djszapiAvengence: I think you are upset, and you now find every minor missing bit as a huge issues. Might go away when you chill out. ;)13:49
AvengenceI like the device experience as a user for first month or so, but shortly after PR1.1 I had no working email and no useful solutions on bug tracker, forums, etc so I went at it myself and came here when I figured out what went wrong but hjadn't figured out why. Since then, I've been poking and the more I do so the more I realize that it's pretty on the surface but ugly underneath, meaning more rework than anticipated. Some of this I'm13:51
Avengencejust trying to understand why and some I'm just checking to ensure my interpretation is correct.13:51
djszapithe why for a man page is that what has been told above: no need for average user, so why waste space by default ? Developers can make a manpage package. I think it is reasonable in /my/ opinion.13:53
AvengenceIt might have helped to install the SDK and read it's docs. I wanted to do that before, but it's HUGE and if your only internet access is slow wwan with charge per MB used, downloading a >1GB SDK for the docs is a lottle excessive13:53
Avengencereasonable: make manpages one of the packages installable after switching on developer mode13:53
djszapiyou can read the SDK online, I guess.13:53
djszapiwhy ?13:54
djszapiI can install a manpage in "user" mode.13:54
*** lucido2 has quit IRC13:54
jonniand you can ask nokia to send N9 source packages to you with DVD, and they send it for free to your postal address.13:54
AvengenceIt's only a little less than a week ago when I hopped to next location I have use on dsl line (768k down) without small usage cap, so I downloaded that monster and installed it but then that was the laptop that crapped out 2 days later. So I have to fetch it again on this one, whihc I haven't bothered with yet.13:55
Avengencejonni: relies on having postal address that is mine. I might just do that when I get back home13:55
djszapiI have never downloaded the SDK, just read online, but as jonni said: feel free to request a dvd to your current reside (does not matter who is the apartment owner).13:56
Avengencedjszapi: why developer mode? because without it I don't have terminal or ssh. What I mean is I hit that dev mode switch, and I get options to install stuff like debugger, profiler, etc. Why no docs/manpages option on that list?13:56
*** lucido2 has joined #harmattan13:56
djszapiAvengence: because not enough interest in that from users ?13:57
djszapiAvengence: also, you can install store packages at least which should be doable after all.13:57
Avengencedjszapi: it took a month for first mail to arrive at last address, where I was only at for a month. Now stuff is hitting there and I'm elsewhere. what fun it is to be on ass end of nowhere13:58
djszapiif you change apartments repeatedly every single month, you do not just have a problem with Nokia DVD delivery ;-)13:58
Avengencewhere does nokia look for user interest? I'll go file a need for it (and reiterate the need for CardDAV)13:58
Corsacpff, having the sms is the tracker db is really a pain13:58
AvengenceI know, but at end of this month I get to go home (visas granted just a few days ago, have to arrange for earlier return than anticipated)13:59
jonniyou can file in there that you want developer-mode to include man-pages13:59
*** lamikr has quit IRC14:00
djszapiAvengence: and spam everybody to vote for your report then ;-)))14:00
Avengencethey actually read that? I hadn't seen much action there14:04
djszapiof course, they do.14:04
djszapiafaik, lot of issues were fixed from that.14:04
djszapiand also some of them said wontfix.14:05
Avengenceto statr, please vote for
_MeeGoBot_Bug 128 enh, ---, ---, ext-risto.lahti, ASSI, More synchronization options for agenda/contacts14:05
CorsacAvengence: I just commented, but syncml over bluetooth works for me on N914:09
Avengencewell, I really want CardDAV as I noted in my response, I could care less about syncML as I don't have anything that works on the other end of that. I figured better to add to existing than create another bug to just be labeled duplicate and ignored.14:10
*** JackaLX has joined #harmattan14:17
*** JackaLX has quit IRC14:17
*** JackaLX has joined #harmattan14:17
*** adlan has joined #harmattan14:32
*** snowpong has quit IRC14:39
*** snowpong_ has joined #harmattan14:40
*** M4rtinK has joined #harmattan14:42
*** lamikr has joined #harmattan14:47
*** iHile has joined #harmattan14:54
*** iHile has quit IRC14:55
*** tbf_ has joined #harmattan15:04
*** snowpong_ is now known as snowpong15:04
marsjeI'm trying to mount a directory on my server on my n9 using sshfs. It gives me: fusermount: mount failed: Operation not permitted15:05
marsjeanyone knows why?15:06
marsjeI tried as root (devel-su) and as user15:06
*** povbot has joined #harmattan15:08
*** ChanServ sets mode: +v povbot15:09
marsjethat is a bit disappointing to found15:09
djszapiwell, who forced you to pay for this phone ? ;)15:09
marsjeafter I bought it15:09
djszapicannot you send it back ?15:09
djszapiI guess there is some time after the buying.15:09
marsjeI thought I would be real root and have ultimate power15:09
*** MFaro-Tusino has joined #harmattan15:09
jonniif you boot to openmode you can get to root, which can mount15:09
*** mgedmin has joined #harmattan15:09
djszapimarsje: -> he sent back15:09
_MeeGoBot_Bug 128 enh, ---, ---, ext-risto.lahti, ASSI, More synchronization options for agenda/contacts15:09
djszapiit is always worth paying attention for the time after the buy in which period you can send back. You can test it out inasmuch.15:10
*** MFaro-Tusino has quit IRC15:11
marsjedjszapi: damn15:11
marsjewhy all these silly restrictions?15:11
marsjeit makes no sense to me...15:12
djszapiyou mean to have a secure platform ?15:12
Avengencewhat is dangerous about mounting things on the phone?15:12
marsjeas long as I am the one doing it15:12
djszapibecause that is the point of an average user. You can use open mode with loosing warranty, if you wish.15:12
marsjenot sure what openmode is15:12
SpeedEvilOpen mode cannot void warranty legally in many aspects.15:12
marsjebut it soundslike something I want15:12
Avengenceits easier to install packages from 3rd party sources. that is more 'dangerous' than mounting the filesystem of your choice on your device (which will usually be a fs you control)15:13
djszapihow to install third party apps on your device with Nokia softwares and OVI checked ones ?15:13
djszapithat should not be possible either.15:14
djszapijust on your wish.15:14
Corsaccan I copy/paste the number I got a sms from?15:15
djszapiSpeedEvil: as far as I can tell, I get a warranty void screen.15:16
Avengencedjszapi: err, i can flip one switch in settings (not even turn on dev mode), ok the warning, copy any .deb file i want over usb in mass storage mode, use any of the free filemanagers from ovo store to 'run' it and the package intalls15:16
SpeedEvilWarranty void screen does not actually mean much legally.15:16
marsjethis openmode sounds a lot like "rooting" on a iPhone or android phone15:16
SpeedEvilNokias legal department does not always properly understand national laws.15:16
marsjecan I have this openmode permanently?15:16
djszapiwhy not ? just flash it ?15:17
marsjeanyone here is using this?15:17
Avengenceopenmode sounds like what I want, to bad it isn't better documented anywhere. the first question is, how does upgrade to 1.2 happen? I have to flash to stock 1.2 and then open 1.2, or straight 1.1 open to 1.2 open?15:18
djszapiAvengence: mass storage is not explicit, you need to allow it. Not sure if it can be changed from code in the settings.15:18
Avengencemass storage is an option when i plug in usb. it asks sync or mass storage every time15:18
djszapi(my guess would be not)15:18
djszapiAvengence: such a file manager should not pass OVI for fun.15:18
djszapisince it is dangerous in the sense as you have just mentioned above.15:19
jonniyou can just use search instead of filemanagers anyways15:19
AvengenceFile Manager Beta, featrured app in ovi store, will install .deb files with single tap15:19
djszapiAvengence: -> it /is/ documented there.15:19
djszapikeep reading.15:20
*** Smtih has joined #harmattan15:20
djszapialso, copying third party deb should also be checked in OVI15:20
djszapiso I do not see theoritically how it could occur, if OVI works fine.15:20
jonnibut yes in 1.2, you need to go for stock 1.2 first, wait that someone ask PR1.2 kernel sources from Nokia, someone to compile new kernel, and then switch again to open mode15:20
djszapiso OVI should check many things over there already.15:21
djszapibut theoritically staright-forward defense mechanism.15:21
Avengencejonni: thanks, the page I read (linked by djszapi) only says I will haefv to reflash eventually, but not explicitly to what15:21
AvengenceOVI waht is checking? the store is checking what an apps does after its installed, or does OVI cover more than the store?15:22
jonniAvengence: changing -b flag to -f flag, will flash the kernel instead just loading it to ram15:22
djszapiFirst, OVI is the store itself15:22
AvengenceI mean the OVI term, which seem ambigous and I thought they were dropping anywhat, called Nokia Store now15:22
djszapiSecond, they have binary scanner tools, and all that jazz, thus it is a bit more than you think.15:22
djszapiAvengence: I do not see the problem with the instruction above, works fine for 1.115:23
djszapithey will probably update it when needed in PR1.215:23
Avengenceso anyway, the store includes at least one app that will install an .deb file if its on the device and I have set the switch to allow installing non-store apps. flipping that switch gives some silly warning, says nothing of warranty15:23
djszapianyway, it is a weak reasoning to say if there is a hole "A", we should make much more.15:24
AvengenceI can shortcut the process further if I turn on developer mode as I can take the link from the builtin browser that is crippled and paste to terminal where i use wget to fetch the deb and tehn install, no need for desktop and usb15:24
djszapi(even if I do not see the hole "A" here)15:24
AvengenceI can speed it up even more by installing firefox that way and then use it to download .deb files and go 'run' them to install15:25
djszapihow can you turn the developer mode on from code, and why do you think if there is such an option, OVI does not check ?15:25
djszapiwhat you do manually is your decision after all; let it open mode, developer mode or so forth (inside the available frame)15:26
Avengencethe point is, mounting a filesystem of the users choice is akin to inserting an sd card, and actually probably less dangerous than using mass storage mode (since in that case a windows box could delete everything, corrupt files, try to install virus, etc)15:26
djszapiI could mount /anything/ in.15:26
*** xarcass has quit IRC15:26
SpeedEvilMounting noexec - sure15:26
SpeedEvillike vfat15:26
Avengencethere is little to no danger in mounting a filesystem15:27
Avengencethere IS threoretical danger in installing arbitrary software that isn't signed15:27
SpeedEvilMounting, and paying attention to execution permission bits - well that's a clear risk.15:27
Avengencethe former is not allowed, but the later is by flipping one switch and OKing a warning15:27
djszapiAvengence: no there is no theoritical danger.15:27
djszapithere is (if any) only practical in the OVI QA process.15:28
AvengenceI'm talking about software NOT through OVI, gezz15:28
djszapiand how would that software go to your device ?15:28
Avengenceyou see firefox in ovi store? nope, but I installed it15:28
djszapiwell, it is up to you what you do with your device.15:29
Avengenceffs, read what I typed. i got it there htrough terminal and wget, but could do the same with mass storage mode15:29
djszapibut such a software will not come to OVI15:29
djszapi(a.k.a. to average user)15:29
Avengenceaverage user will find instructions for instaling with mass storage mode and file manager on many sites15:29
Avengencethe whole point is, what 'danger' could mounting a filesystem present? NONE15:29
djszapino, the average user, like my mom, will not do such things15:30
djszapithey will just well use ovi, and nothing anything fancy on a developer website.15:30
Avengencethe average user, when they can't freakin' copy text from a website on N9, will google for how, and they will find instructions for installing firefox and opera mini exactly as i described on multiple sites as both those will solve their problem15:30
djszapiI could mount anything, as I said above.15:30
djszapiso if an executable reads a file in a directory for its process, I mount something completely irrelevant15:31
djszapiwhich messes up the operation of the process into malicious usage.15:31
djszapiI am kinda out-of-the-luck.15:31
Avengencemount anything requires going open mode. that's a lot more of a barrier than it is to install anything I please15:31
djszapithat is what the open mode was designed for ...15:31
Avengenceoh, I can understand protecting mounting over /usr/bin for example, but not allowing to mount into the homedir is crazy15:32
djszapiwhy ?15:32
CissWitAvengence: how did you install firefox on your n9 ?15:32
djszapiI write an application getting the parameters or anything from that folder, and then you mount something.15:32
djszapithe process gets messed up.15:32
Avengencenameone app that will magically run a binary without user request just because it's in a folder under their homedir15:32
AvengenceCissWit: I fetched the .deb file with wget and used file managger to install the deb. i could have done same without dev mode by putting the file on it with usb mass storage mode and then same install from file manager15:33
djszapiI would be unhappy if you replace my vim configuration file by mounting something else in there.15:34
djszapiand my vim starts working completely differently.15:34
Avengenceyou would be the one replacing your own vim config. you are the user15:35
CissWitAvengence: where did you get the .deb from ? From ?15:35
Avengenceremeber, the user owns the device, not the mnufacturer or developers15:35
jonniAvengence: only if you enable 3rd party debs from the ui, unsinged debs otherwise are not installed even if you start it from filemanager.15:35
djszapiAvengence: except that, it would be a plus check in OVI15:35
djszapifor /every/ mount15:35
*** lamikr has quit IRC15:35
Avengencejonni: yes, i know, i've said that. why do we get that swith but we dont get a switch that says "Let me mount filesystems FFS"15:35
Avengencedjszapi: except OVI doesnt have jack shit to do with moutign filesystems15:36
djszapiand I bet, they do not have enough resource for that, and it could not be probably properly done without forbidden certain external fses where the contents is dynamic anyway.15:36
jonnibecause you can brick the device if you mount to wrong place, installing deb package ususally does not brick15:36
djszapiyes, you can brick the device15:36
djszapithat is also a use case along with all the malicious and other use cases.15:37
AvengenceCissWit: first google hit
Avengencejonni: explain how mounting under the homedir will brick the device15:37
jonniI said wrong place, not homedir15:38
*** mece has left #harmattan15:38
AvengenceI already said I can understand mounting over root not allowed, but mounting into homedir is always relatively safe so saying no mounting anywhere is just dumb15:38
djszapiagain, I would not like any apps mounting into my home.15:39
Avengencethen don't use them15:39
Avengenceyour choice as a user15:39
djszapino, it is not my choice15:39
djszapiif OVI does not make proper check15:39
Avengenceright now, nokia hasn't given us the choice to mount15:39
jonniwell in homedir usecase, you mount to homedir/foo, make deb package which installs itself to homedir/foo, you happen to lose tcpip connection -> device is brick since refhaslist check fails15:39
djszapiand OVI says it is safe.15:39
djszapialso, if someone makes a symlink to home from etc15:40
djszapiyou can also brick the stuff15:40
djszapiI think it was way easier when OVI decided about this credential than having this risk in OVI apps.15:40
Avengencesomeone please explain MALF since the accronym is not expanded where its used15:40
jonniusually happens if files sha1 checksum does not match reflashlist, or if you modify system critical binaries or configs15:42
Avengenceso when it says I'll get MALF when going back to stock from open mode, which is apparently required, what is the reprecussion of that?15:42
infobothmm... malf is
djszapithat is a malf screen :)15:42
jonnithat malf happens, if you install any deb package durin open mode session, because refhashlist is signed with open mode key, and going back to secure mode, will lead signature not matching.... well not exactly accurate, but about :)15:43
Avengenceso to recover from that, what is the solution?15:45
jonni(actually its files.tcb which is signed,but thats details) :)15:45
jonniAvengence: recovering from open mode malf, -> complete reflash to stock15:45
Avengenceor in other words, what is the practical solution to get from open 1.1 to open 1.2 without loosing everything15:45
jonnigood backups with tar :)15:46
djszapithere is no "open 1.2" in the first place.15:46
djszapibut obviously backups as jonni said.15:46
Avengencewell, when 1.2 comes out, there will be open 1.2 as there's open 1.1. it would be desirable to upgrade without passing through stock 1.2 regardless whether there is malf15:48
djszapino need to speculate what is gonna happen, we will see.15:48
jonniah finally my ovi qa passed trial version for my app15:48
djszapicook with the existing ingredients ;-)15:48
Avengenceit sounds a lot like the experience i had on iphone and the challenge to restore things to some similar state after eahc os update, which usually involved more than one reflash cycle15:49
jonniso can put next version in the queue15:49
djszapijonni: congrats =)15:50
AvengenceI have a feeling that what I will end up doing is something like rsync the whole mess out somewhere, play with open 1.1, and then confirm I can build that myself and proceed to build something a bit more sane but still able to run the regular apps (both nokia and 3rd party, store or not) and hope the process works for 1.2 when it happens15:50
jonnipersonnally I just use stock firmware which I have hacked open, so I dont have to worry about restoring things, but unfortunately thats not what normal users can do.15:50
*** stroughtonsmith has joined #harmattan15:51
jonniah, time flies, weekly floorball session... bbl15:52
Avengenceon the topic of filesystems, if its ext4 and vfat that are used, neither of which have a clue about flash, are they running raw or is there a black mapping underlying them, and if the later what is it and how good is it at remapping?15:57
Avengenceseparately, is there a good reference somewhere on the hardware in the N9? what I've found poking around raises questions. i.e. The TI wlan chip has bluetooth, but I've read in several places that the bluetooth used is a broadcom chip which also includes FM Rx (the model up of the TI chip has FM Rx/Tx and GPs, all in one RF)16:01
xmlich02Hi, I tried to create orientation lock application based on javispedro's glol orientation faker. It is not working well after the faker disconnecting, I am not sure how to fix it.
lizardoAvengence: N9's BT chip is from TI (WL12xx)16:06
*** hardaker has joined #harmattan16:07
Hq`Avengence: like SSDs, eMMC hides the raw flash interface and has wear-leveling on hardware (or controller firmware) level16:10
*** Sazpaimon_ has quit IRC16:21
djszapiit is a good question why we stopped using UBIFS.16:24
Avengencelizardo: so the WL1271 is the bluetooth as well as wlan? any idea then why turning on FM Rx also turns on bluetooth? what I read said it was because they were same chip and it was all on/all off (and also mentioned broadcom)16:25
AvengenceeMMC = onenand?16:26
*** Venemo has quit IRC16:27
Avengenceany way to bypass the mess and go direct, or at least to get real stats on what is going on in the mapper so the system can be aware of the health (and potential override it's decisions, i.e. avoid writing a weak area it doen't want to take out of service)16:27
*** Venemo has joined #harmattan16:27
*** Venemo has joined #harmattan16:27
djszapiAvengence: we have eMMC and onenand, too afaik16:27
Avengenceah, two flash chips. basically onenand is root and eMMC is user space?16:28
djszapimmm, we have not used ubifs, good question why not.16:29
lizardoAvengence: to enable FMrx , you had to issue a vendor specific HCI (bluetooth)  command, because they share the same chip. But I think there are other people here more familiar with FMrx on N9 (javispedro comes to mind, but he is not online ATM)16:32
djszapiAvengence: cat /proc/mtd16:36
Avengenceis onenand only for kernel/initrd images? assuming the eMMC is 64GB, adding up the space of /, /home and /home/user/MyDocs adds up to about 64(fake)GB16:36
djszapiand: cat /proc/mounts | grep mmc16:37
djszapiI think your best bet is to simply let the flash translation layer do all the wear levelling for you. If you repeatedly write to the same block, it will slowly migrate around the device naturally.16:37
Avengenceis /dev/root really /dev/mmcblk0p2?16:37
*** jpnurmi has left #harmattan16:38
SpeedEvilIt does not migrate around the whole device - at least according to public documents.16:38
*** adlan has quit IRC16:38
SpeedEvilThe device is typically segmented into (say) thousand block (130K blocks) regions.16:38
Avengenceok, so the blast of onenand errors is not so bad as that arae is infrequently rewritten. all the active FSes are on eMMC and if I trust its magic I might be ok (so can probably juts go pure ext4)16:39
SpeedEvilThe wear leveling happens over these block segments.16:39
SpeedEvilAvengence: you can monitor the bad block level - or at least you can on the n90016:39
*** adlan has joined #harmattan16:39
AvengenceSpeedEvil: i can monitor that for teh onenand or the eMMC?16:39
SpeedEvil /sys/class/ubi/ubi0/bad_peb_count16:39
SpeedEvilyou ahve no insight into the emmc16:40
djszapiI do not have "/sys/class/ubi/ubi0/bad_peb_count" at least here.16:40
Avengencei assume that all mtd partitions are onenand and all /dev/mmcblk* stuff is the eMMC16:40
Avengencethere won't be any ubi anything if mtd us used16:40
*** etrunko has joined #harmattan16:41
djszapiSpeedEvil: Ah, OK. Maybe I was misinterpretting something one of the FS guys said. They tried to induce failures by writing to a single block repeatedly, and said somehting like that after several million iterations, every eraseblock had been used about 10 times.16:41
Avengenceso, eMMC lies about space like most disk storage, and the datastorage page reflects that (to coincide with marketing)16:41
Avengenceis it documented anywhere waht pars of the UI use fakeMB/fakeGB and waht parts use realMB/realGB?16:42
SpeedEvileMMC probably has a couple of extra percent of space.16:42
SpeedEvilThis is split over some thousands of wear leveling sectors.16:42
SpeedEvilSo there are a few blocks spare per leveling sector16:42
djszapiIt might be that these segments are larger than our whole onenand, for example.16:43
Avengenceso basically the real flash is real 64GB, but to make room for wear level, they only expose about 60GB (64fakeGB) and call it 64 because then it matches to lying harddrive manufacturers16:43
SpeedEvilSo you might have - for example - 130M of wear leveling segment, with 24 spare blocks - 2M16:43
SpeedEvilIf you write to this wear leveling sector a single block of 130K, repeatedly, then you have about - if the endurance of the flash is 100kcycles - about 100m writes.16:44
Avengencethat would explain a bit why when I work directly with flash the sizes make sense, but all these devices (SD,MMC,SSD,etc) seem to be missing space16:45
SpeedEvilThe above is all pieced together from various sources.16:45
SpeedEvilIt's not actually documented by teh makers.16:45
Avengencebut anyway, how much of the N9 UI uses real measurement and how much uses fake?16:46
djszapithe eraseblocks are 256KiB according to the above, so it only takes 16Ki of them to cover the entire device16:46
SpeedEvilThe really fun bit would be if you could reprogram the little micro in the SD card.16:46
Avengencethe controller in sd cards is often not reflashable. all its login is in a mask rom for cheapest mass production. same for usb mem sticks16:47
*** rcg1 has quit IRC16:47
*** lucido2 has quit IRC16:47
SpeedEvilBut sometimes flash can be cheaper than mask.16:47
AvengenceSSD gets a bit more complicated and and so more necessary to make them reprogrammable to fix some of the most serious bugs. also, plenty of room to store the firmware with all that flash so controller only needs to be smart enough to read that out at power on16:48
djszapiI presume their operations are very hard coded. They're not really servicable parts.16:49
SpeedEvilThe above is largely pointless.16:49
SpeedEvilEven if they are reprogrammable - which is not impossible - they will be completely locked.16:49
Avengenceugh, File Manager doesn't use real sizes16:51
*** rcg1 has joined #harmattan16:51
AvengenceFilebox on the other hand does use real sizes16:52
marsjeregarding opensh: if I understand correctly, I can boot one time from a aegis-less kernel, install opensh, then boot from the normal kernel, and still have REAL root access using opensh?16:52
Avengenceso I guess I have to do some digging to figure out which apps I can trust and which I can't16:53
djszapiAvengence: yes, that is what I was talking about previously.16:53
djszapiyou might now get my point ;-)16:53
Avengencedjszapi: point about what earlier?16:54
Avengencesince there have been numerous topics, be specific about which point and to what earlier16:55
djszapimmh, if you do not ring a bell, it might be that you do not still understand. :)16:55
djszapiit does not ring a bell*16:55
Avengenceno, it's because you are being vague.16:55
djszapibut ideally, people would like to trust the OVI coming out as much as possible. Disabling the mounting is a very quick way of doing it.16:56
djszapiOVI application*16:56
Avengenceoh, you mean about trusting apps. I am talkig trust in a different sort. These are two OVI apps, but one liers aboutfile size, thus I can't trust it to accurately tell me the size16:57
djszapithat is the same trust indeed.16:57
Avengencewhich raises the question, are these both doing it in the app, or is one using an API and then which one does it which way and is that API the one that is accurate or inaccurate16:57
Avengenceit might be the case the platform is wrong (e.g. Mac OS x 10.6/10.7) and then only those which don't use the platform API are trustable16:58
*** rcg1 has quit IRC16:58
*** adlan has quit IRC16:59
*** lucido2 has joined #harmattan17:00
Avengenceis this onenand 512MB or 1GB? (mtd6 is half the space or is the whole device itself)17:00
Avengencedo I understand this right? the eMMC has wear leveling so using ext4 and vfat on it is considered safe. the onenand is raw flash with some buffer and somewhat smarter error handling, but has no remapper on it (just mtd) but a large chunk of that is used for swap.17:02
Avengenceit would seem that is just asking to burn up the swap area of the onenand unless the swapper was taught to not abuse the flash17:02
lucido2when debugging on n9 I get this: /home/chris/QtSDK/Madde/sysroots/harmattan_sysroot_10.2011.34-1_slim/usr/lib/qt4/plugins/inputmethods/" is not at the expected address (wrong library or version mismatch?)17:04
lucido2what does it mean?17:04
djszapiat one point, it was used for swap, at the moment it is not.17:05
djszapiI think rover has swap on onenand, so it was not considered that much of a problem.17:06
djszapiah, they moved it off there to the emmc.17:07
*** minimec has joined #harmattan17:09
djszapiAh, now I remember - there is an mtdswap module that provides the smarts.17:09
Avengencehow to get swapstats?17:10
Avengencemount doesnt show any swap but not sure if mount on linux ever shows that17:12
djszapi"free" ?17:13
*** rcg1 has joined #harmattan17:13
djszapior cat /proc/mem, I guess17:13
Avengencefree doesnt say where the swap is mounted17:13
Avengence /proc/mem doesnt exist it says17:14
SpeedEvilcat /proc/swaps17:14
djszapimeminfo, use tabs on Linux ;-)17:14
*** Sazpaimon has joined #harmattan17:14
djszapibut SpeedEvil is right17:14
Avengence /dev/ramzswap0 = what real dev?17:16
SpeedEvilLook to see if you can find configuration for the ramswap module,17:16
deramcompressed part of real ram?17:16
SpeedEvilramzswap is block-backed compressed swap17:17
*** leinir has quit IRC17:19
*** leinir has joined #harmattan17:19
*** leinir has joined #harmattan17:19
djszapiVenemo: mmm, there is an interesting bug in irc chatter. It did not connect to the channel that I typed into the configuration screen, but it connected to freenode.17:21
djszapiSpeedEvil: it doesn't have to be backed. ours is not17:21
*** djszapiN9_ has joined #harmattan17:22
djszapiVenemo: I have just tried to join explicitely by typing "/join gluon", but does not work. :/17:23
djszapialso, "/join #gluon".17:23
*** briglia has joined #harmattan17:23
djszapiVenemo: after closing the part (gluon) on the Ui, joining works fine. There is an error in the main tab, btw: [ERROR] An error occured! Error code is: 442 (red letters)17:24
*** djszapiN9_ has left #harmattan17:24
*** zarlino has joined #harmattan17:26
djszapiVenemo: Also, I have an internet connection (3.5G), but irc-chatter reboots the connection for some reason, at least according to what I see at the top bar. The relevant icon vanishes, and then it reveals again in few seconds.17:29
Venemohello djszapi17:29
Venemodjszapi, that icon vanishes randomly on my N950 when it feels like.17:30
djszapiand now stuck at "Connecting, please wait...", but google works just fine in grob.17:30
Venemoplease add the issues you found to the wiki.17:32
VenemoI can't really do anything about them for at least a week.17:32
*** vladest has quit IRC17:37
*** vladest has joined #harmattan17:37
*** rm_work has joined #harmattan17:39
*** rm_work has joined #harmattan17:39
kulveare there any (good) examples of the Settings applets using the declarative language?17:56
*** Iktwo has joined #harmattan18:01
*** vladest has quit IRC18:02
*** tbf_ has quit IRC18:02
Avengenceso, if its greater than 1milKB then there is 256MB of compressed swap in ram, otherwise the compressed swap lives atop the mtd swap partition (with encryption)18:03
Avengencethe commment in the script says 1GB, but the test is incorrect in two ways18:03
*** gabriel9 has quit IRC18:04
*** lucido2 has quit IRC18:08
*** zarlino has quit IRC18:20
*** zarlino has joined #harmattan18:26
*** lucido has joined #harmattan18:29
lucidoif I change the sshd_conf on my N9 will aegis lock the device?18:30
lucidoI want to be able to access it from other ips than 192.168.18:31
Avengencewould also be good to bump up the server key size to something less pitifully small18:32
Piruit won't18:32
Piruyou can modify the file18:33
Avengencei wonder, why does the starup script pass a parameter to disallow root login when the sshd_config already has PermitRootLogin no18:33
PiruAvengence: to lock root login to devs only?18:34
Avengenceis that some double-fix after they shipped it initially with PermitRootLogin yes?18:34
Piruapparently /usr/sbin/rdc_cert_verify checks for some cert.. if you don't have it, tough luck, you can't login as root18:35
*** stroughtonsmith has quit IRC18:36
*** stroughtonsmith has joined #harmattan18:37
Avengenceit's make more sense to just use a sane config file and change it when adding the rdc cert, but alas, anything sane is not the solution so make hackery (or kill aegis)18:37
Avengencei'm really wondering now what (if any) useful things aegis does18:38
*** zarlino has quit IRC18:50
*** lucido has quit IRC18:52
*** lucido has joined #harmattan18:53
*** heymaster has quit IRC18:56
*** heymaster has joined #harmattan18:59
*** Venemo has quit IRC19:06
*** NIN101 has joined #harmattan19:11
*** adlan has joined #harmattan19:11
*** stroughtonsmith has quit IRC19:12
*** oytunx has joined #harmattan19:12
*** heymaster has quit IRC19:17
*** heymaster has joined #harmattan19:18
*** zarlino has joined #harmattan19:21
*** jreznik has quit IRC19:26
kulveI've tried to show my own icons in several places without success. I just get a red square. Should I update some icon cache or something to get those visible?19:28
tommawhich places? some places doesn't support svg icons19:32
*** cfetzer has joined #harmattan19:34
kulveI'm using png19:34
kulveand currently in share-ui19:34
*** vladest has joined #harmattan19:35
*** xTs has joined #harmattan19:37
*** faenil has joined #harmattan19:37
*** ten0r_ has joined #harmattan19:41
*** ten0r has quit IRC19:44
*** lucido has quit IRC19:45
*** lucido has joined #harmattan19:46
*** rcg1 has quit IRC19:55
*** Pali has joined #harmattan19:56
*** djszapiN9 has joined #harmattan20:00
*** stroughtonsmith has joined #harmattan20:01
*** lucido has quit IRC20:13
*** lucido has joined #harmattan20:14
*** snowpong has quit IRC20:15
Avengencesome of the things installed after dev mode (energy profiler, maybe terminal) had red squares until next reboot. the qml examples have rounded green icons from install and across boots20:19
djszapiAvengence: the number beginning with 800 has been disconnected. The other number would incur long distance charges.:/20:21
*** hhartz has joined #harmattan20:21
*** jkt has joined #harmattan20:22
Avengencedjszapi: try goog voice20:22
jkthi there, I'm trying to debug a segfault in my pretty basic app which just uses the front camera to show a mirrored face20:22
jkt is the farthest I was able to go20:22
jktany idea where can I get the debug symbols from? And yes, I did use Settings -> Security -> Developer mode and installed the "Debugging" thing from there20:23
ZogG_laptopjkt: there is working example on gitorious20:25
ZogG_laptopjkt: mirror20:26
jktZogG_laptop: ?20:26
*** jreznik has joined #harmattan20:26
*** snowpong has joined #harmattan20:29
jktZogG_laptop: interesting, that application doesn't segfault :)20:30
ZogG_laptopjkt: it's in ovi even btw20:35
ZogG_laptopi checked it for code but it's irellevant to me20:36
jkthe's wrapping it inside the PageStack, but apparently that isn't the cause here20:37
jktlet's see if my QML qorks with his C++ implementation of the class20:37
*** lucido has quit IRC20:38
*** hhartz has quit IRC20:39
ZogG_laptopit qorks =)20:40
ZogG_laptopi qorked it myself =)20:40
*** Smtih has quit IRC20:40
*** hhartz has joined #harmattan20:43
*** faenil has quit IRC20:44
*** admiral0 has joined #harmattan20:46
*** mgedmin has quit IRC20:46
*** mgedmin has joined #harmattan20:46
ZogG_laptopmgedmin: hmm, nice vhost20:47
*** djszapi has left #harmattan20:48
*** chris_ has joined #harmattan20:48
jktas soon as I import his C++ class into my QML items, I get the segfault again20:48
jktso I suspect that one has to deal with the resulting QML item rather carefuly20:49
ZogG_laptopjkt: fail =)20:49
* jkt really wonders how come that MADDE generates object files with "illegal instructions"20:49
jkteither that, or a really nasty memory corruption in there20:49
admiral0dudes, how do you do a modal dialog in qml? after i tap on it , it closes -.-20:50
*** chris_ has quit IRC20:53
*** iHile has joined #harmattan21:01
*** ZogG_laptop has quit IRC21:02
*** snowpong has quit IRC21:02
*** Avengence has quit IRC21:05
*** ZogG_laptop has joined #harmattan21:08
*** chris_ has joined #harmattan21:09
jktokay, as soon as I switch from the QtCreator-provided QWidget-based viewer to a custom one which uses QGLWidget, everything works as expected21:11
jktthis smells of a low-level bug21:11
* RST38h moos sadly21:13
*** BluesLee has joined #harmattan21:13
*** BluesLee has quit IRC21:13
jktwhat's the usual way of reporting these bugs?21:14
*** vladest_ has joined #harmattan21:14
jktshall I file that as a Qt one, or send it somewhere else? I've seen traces of that proprietary OMAP's video driver in valgrind's output :(21:15
*** zk8 has joined #harmattan21:15
*** vladest has quit IRC21:18
*** vladest_ is now known as vladest21:18
*** jreznik has quit IRC21:19
*** benares_98 has joined #harmattan21:22
ZogG_laptopjkt: btw
ZogG_laptopbak amera21:24
*** iHile has quit IRC21:25
*** zarlino has quit IRC21:27
* DocScrutinizer51 moos at RST38h 21:31
ieatlintthe last work day before vacation always seems to go slower21:34
*** lucido has joined #harmattan21:34
*** vladest has quit IRC21:36
*** zarlino has joined #harmattan21:41
*** beford has joined #harmattan21:41
*** etrunko has quit IRC21:49
*** zarlino has quit IRC21:51
lucidoI'm getting a buch of wrong librar, version mismatched? errors when running in debugging mode like this one: "/home/chris/QtSDK/Madde/sysroots/harmattan_sysroot_10.2011.34-1_slim/usr/lib/qt4/plugins/inputmethods/" is not at the expected address (wrong library or version mismatch?)21:55
lucidowhat is that?21:56
*** SqRt7744 has joined #harmattan21:56
*** stroughtonsmith has quit IRC22:05
*** rZr is now known as RzR22:10
*** vladest has joined #harmattan22:10
Sazpaimonso what's this developer-mode update Im seeing, 1.35+0m622:19
deramwhy isn't there a way to see changelog22:20
deram(yeah I know, this is for users, not powerusers/developers)22:20
*** rcg1 has joined #harmattan22:22
*** wirwe has joined #harmattan22:34
tabaskodoes anybody know when we can expect mer/nemo on n9? :)22:41
Sazpaimonderam, dont changelogs exist in the package source?22:43
Sazpaimoni dont think they exist in the binary package22:43
*** vladest has quit IRC22:44
*** vladest has joined #harmattan22:47
deramprobably yes, but is there a way to see them before updating the packages in phone?22:49
*** zarlino has joined #harmattan22:50
*** admiral0_ has joined #harmattan22:54
*** admiral0 has quit IRC22:55
*** lizardo has quit IRC23:00
*** wirwe has quit IRC23:01
*** Natunen has quit IRC23:01
*** wirwe has joined #harmattan23:01
*** Natunen has joined #harmattan23:04
*** wirwe has quit IRC23:07
*** wirwe has joined #harmattan23:08
*** wirwe has quit IRC23:09
*** jpnurmi has joined #harmattan23:11
*** jpnurmi has left #harmattan23:11
*** admiral0_ has quit IRC23:13
jussican anyone poin out to me how to enable the always on clock on the n95023:16
*** vladest has quit IRC23:18
dm8tbrI don't think you can. the n950 has a LCD as opposed to the OLED of the N923:23
dm8tbrthat said if you are keeping the device powered, you could install one of the night clocks23:24
dm8tbrlike e.g. 'bedside'23:25
*** npm_ is now known as npm23:26
*** hhartz has quit IRC23:29
*** briglia has quit IRC23:38
*** oytunx has quit IRC23:39
benares_98would there be any reason to use community edition/nemo on n9? I had thought it was meant to allow n9 apps on n900.23:39
*** javispedro has joined #harmattan23:47
ZogG_laptopnight pals23:56

Generated by 2.15.1 by Marius Gedminas - find it at!