IRC log of #harmattan for Sunday, 2011-11-06

*** niqt has joined #harmattan00:03
*** tbf has joined #harmattan00:10
* DocScrutinizer points Venemo to and waves with a large stick00:10
*** tbf has quit IRC00:10
*** rcg has quit IRC00:41
*** freemangordon__ has quit IRC00:49
VenemoCorsac, yes, 39-500:54
VenemoDocScrutinizer, I used to follow that thread00:56
VenemoDocScrutinizer, what's new?00:56
DocScrutinizernothing yet, your contribution missing ;-P00:57
DocScrutinizerthough starting arbitrary binaries as *user* seems isn't all new00:58
Venemoit was only news for me01:00
Venemoworks out of the box if you flash N950 with 39-5 and enable developer mode01:01
Venemoif you updated with ota, you will need to reinstall a bunch of packages and then it'll work too.01:01
Venemowell, I'm only interested in this from app developer point of view, but you can find my info about this here:
*** Robot101 has quit IRC01:05
DocScrutinizerVenemo: thanks for that link01:13
SpeedEvilVenemo: 39.5 is the most recent?01:14
*** Robot101 has joined #harmattan01:17
VenemoSpeedEvil, 39-5 is the most recent afaik.01:20
VenemoSpeedEvil, see /topic01:20
*** piggz has quit IRC01:27
VenemoDocScrutinizer, you are welcome :)01:32
VenemoI must go to sleep now. good night :)01:33
*** Scifig has quit IRC01:33
*** Venemo has quit IRC01:33
SpeedEvilthanks ven01:35
*** npm_ has joined #harmattan01:39
*** npm has quit IRC01:39
*** niqt has quit IRC01:56
*** messerting has quit IRC02:07
*** toninikkanen has quit IRC02:08
*** toninikkanen has joined #harmattan02:08
*** NIN101 has quit IRC02:21
*** Natunen has quit IRC02:34
DocScrutinizer>> <KRF> hey guys, i installed the package 'meegotouch-qt-style' this morning, restarted my device and now it's display a "device is malfunctioning, security problem (devicelockd): .../" message on startup.  it wants me to re-flash the device. so, how i am supposed to fix this?<< *cough*02:38
*** mtnbkr has quit IRC02:41
*** arcean has quit IRC02:42
*** pawky has quit IRC02:44
*** mtd_ has quit IRC02:47
*** mtnbkr has joined #harmattan02:53
*** wook has joined #harmattan02:54
wookI was wondering, how to configure .desktop so it can run terminal app with command line parameter02:55
*** pawky has joined #harmattan02:55
*** Soder has joined #harmattan03:01
* rzr is finally flashing Linux_OCF_39-5_RM680-RM680-OEM1.bin03:04
*** RST38h has quit IRC03:05
*** vladest has quit IRC03:09
* wook want to flash liinux_ocf too :-O03:10
wookWhere to get it?03:11
*** sandst1 has quit IRC03:19
*** sandst1 has joined #harmattan03:20
rzrit's still alive03:23
wookaaaaah, stupid wook....i thought its linux os :D03:23
wookYes, i flashed yesterday :S03:24
rzrit is03:24
rzrgnu is the os03:25
rzrlinux is just the kernel :)03:25
*** sandst1 has quit IRC03:25
*** sandst1 has joined #harmattan03:26
wookHehehehe, i thought pure one :)03:26
wookBtw, 3.5mm jack don't work after flashing with 34_2 (as i remember)03:27
wookIs this only my case or?03:27
rzrwas this reported ?03:27
wookDidn't because its on n903:28
wookAnd didn't knew since yesterday either03:28
wookSeems to me that there's no any voltage on its wires03:30
rzrdo u want me to test the headphones ?03:31
wookSure, if it is not problem :)03:32
wookBut, i don't use them much, most of my music is on e7 ;)03:32
rzri have a e7 too03:32
rzrbut i dont use it03:32
wookSend it to me :D03:33
wookI am using it at the moment :)03:34
wookI tweaked it sooo hardly that is faster than any pc or phone i ever used before :)03:34
rzri launched nfs03:36
rzri can hear the music in headphones03:36
wookHmm, then something went wrong with flashing in my case03:37
wookI will reflash it tomorrow :D03:37
*** mtnbkr has quit IRC03:37
wookBtw, how to config .desktop file so it can run new terminal window with command line parameter? I want to .desktop python :D03:39
*** M4rtinK has quit IRC03:40
wookRzt thank you for your help, sleep time here, good night/day, enjoy it ;)03:48
*** wook has quit IRC03:48
*** mtd has joined #harmattan03:49
*** mtnbkr has joined #harmattan03:49
*** mtd has quit IRC04:12
*** vladest has joined #harmattan04:16
*** strannik1 has joined #harmattan04:18
*** Guest53415 has joined #harmattan04:22
*** mtnbkr has quit IRC04:37
*** Scifig has joined #harmattan04:38
*** mtnbkr has joined #harmattan04:49
*** Scifig has quit IRC04:52
*** leinir has quit IRC04:59
*** Guest53415 has quit IRC05:10
*** strannik1 has quit IRC05:29
*** mtd_ has joined #harmattan05:30
*** mtd_ has quit IRC05:35
*** mtnbkr has quit IRC05:35
*** Soder has quit IRC05:41
*** mtd_ has joined #harmattan05:43
*** pcfe has quit IRC05:46
*** mtnbkr has joined #harmattan05:47
*** mtd_ has quit IRC05:48
*** mtd_ has joined #harmattan05:52
*** mtnbkr has quit IRC05:54
*** pcfe has joined #harmattan05:56
*** mtd_ has quit IRC06:05
*** mtnbkr has joined #harmattan06:06
*** Natunen has joined #harmattan06:17
*** Guest47737 has joined #harmattan06:33
*** Guest47737 has quit IRC07:04
*** dymaxion has quit IRC07:10
*** DocScrutinizer has quit IRC07:13
*** DocScrutinizer has joined #harmattan07:13
npm_so can one put the n950 into init 3 to do a proper backup of /home/user ?07:19
*** npm_ is now known as npm07:19
npmoh now i'm me07:19
psycho_oreosI could if I had one ;p07:27
*** Strannik has joined #harmattan07:28
StrannikNow i am07:28
DocScrutinizernpm: hmm, you mean UID:user will log out, so root could do a backup of /home/user?07:30
DocScrutinizerI can't see root doing a restore of such backup, as root got no permission to write to /home/user AIUI07:31
DocScrutinizeranyway init 5 proprerly fsckd up the device here07:33
*** Natunen has quit IRC07:38
*** roop has joined #harmattan07:39
npmDocScrutinizer: yes07:39
npmi'm too wimpy to try it, so i'm just doing a live backup as 'user' of /home/user07:39
npmthen as 'developer' of /home/developer07:40
npmalthough i probably should stop it and rerun the 'tar' part on the handset...07:41
npm^C at 339M07:41
*** Strannik has quit IRC07:43
*** Natunen has joined #harmattan07:44
*** rzrmob has joined #harmattan07:46
*** mtd has joined #harmattan07:49
npmtar cf - user | ssh npm@usb0 "gzip -c > /media/gnulem-npm/home/npm/n950/home-user.tgz"07:50
npmtar: can't open 'user/.signon/signonfs-mnt/signon.db': Permission denied07:52
*** mtd has quit IRC07:53
npmnow dat's security!07:53
*** crevetor has quit IRC08:06
*** hardaker has quit IRC08:06
*** mtd_ has joined #harmattan08:10
*** mtd_ has quit IRC08:19
*** rzrmob has quit IRC08:22
*** roop_ has joined #harmattan08:56
*** roop has quit IRC08:59
*** Chaz6 has joined #harmattan09:00
*** dang_ has quit IRC09:04
*** rzr has left #harmattan09:08
*** clbr has joined #harmattan09:28
*** mtd has joined #harmattan09:32
*** mtd has quit IRC09:40
*** vladest has quit IRC09:59
*** vladest_ has joined #harmattan09:59
*** vladest_ is now known as vladest09:59
*** mtd has joined #harmattan10:00
*** mtd has quit IRC10:05
KRFDocScrutinizer: what's so "*cough*"  about that?10:07
*** aheinecke has joined #harmattan10:11
*** rcg has joined #harmattan10:14
*** Jonno has joined #harmattan10:21
*** Guest79350 has joined #harmattan10:27
*** roop_ has quit IRC10:27
*** Guest79350 has quit IRC10:31
*** seif has quit IRC10:58
*** seif has joined #harmattan10:59
*** qronic has joined #harmattan11:00
qronicdoes somebody know where I can tune the number of alarm sound repeats?11:00
qroniccurrently alarm plays the sound only twice11:01
Chaz6Dont know, but reminds me of a bug i found11:04
Chaz6I had an sms tone i used fine in my n8 that didnt work on n911:04
Chaz6It was quite short, so to get it to work i had to add some silence to the end of it11:04
Chaz6Without the silence I could not select it11:04
qronicI'm using standard from the factory alarm sound11:04
qronicso I guess it's not quite it11:05
qronicI've 'find / -name alarm*' but nothing good11:05
*** mtd_ has joined #harmattan11:08
*** Rizzer has joined #harmattan11:09
*** niqt has joined #harmattan11:23
*** GNUton-BNC has quit IRC11:24
*** GNUton-BNC has joined #harmattan11:24
*** leinir has joined #harmattan11:32
*** wook has joined #harmattan11:37
wookHow to set command line param so that python will be runned under meego-terminal?11:38
*** wook has quit IRC11:48
*** Jonno_ has joined #harmattan11:48
*** zz_gri is now known as gri11:48
*** Jonno has quit IRC11:50
sandst1wook: /usr/bin/meego-terminal -e /usr/bin/python -i /home/user/MyDocs/myscript.py12:00
*** mtd_ has quit IRC12:09
*** vandenoever has joined #harmattan12:20
*** mtd has joined #harmattan12:24
*** mtd has quit IRC12:29
*** Guest64572 has joined #harmattan12:32
*** vladest has quit IRC12:35
*** Venemo has joined #harmattan12:48
*** Guest64572 has quit IRC13:00
*** NIN101 has joined #harmattan13:03
*** pirut has joined #harmattan13:33
*** Jonno_ has quit IRC13:35
*** mtd_ has joined #harmattan13:45
*** vladest has joined #harmattan13:45
*** mtd_ has quit IRC13:49
*** djszapi has joined #harmattan13:51
*** djszapi has left #harmattan13:51
*** arcean has joined #harmattan13:56
*** niqt has quit IRC14:03
*** mtd_ has joined #harmattan14:07
*** M4rtinK has joined #harmattan14:11
*** mtd_ has quit IRC14:12
*** aheinecke has quit IRC14:23
*** aheinecke has joined #harmattan14:25
*** Rizzer has quit IRC14:26
*** tomhut has left #harmattan14:41
*** strannik1 has joined #harmattan14:56
*** mikelima has quit IRC15:06
*** razvanpetru has joined #harmattan15:08
razvanpetruhi, is there an easy way to recreate the native colored header?15:08
*** mtd_ has joined #harmattan15:23
*** roop_ has joined #harmattan15:25
*** Natunen has quit IRC15:30
*** natunen has joined #harmattan15:34
*** strannik1 has quit IRC15:37
*** pirut has quit IRC15:39
*** natunen has quit IRC15:41
*** NIN101 has quit IRC15:42
*** berndhs has joined #harmattan15:44
*** razvanpetru has left #harmattan15:49
*** piggz has joined #harmattan15:54
piggzis it possible from QML to display a message...the type that some apps display in a box at the top of the screen?15:56
vandenoeverpiggz: probably easiest to add it yourself, make such a text box and add states that show and hide it when the msg text is changed15:59
piggzvandenoever: yeah, shouldnt be difficult, just wondered if it was a qt-components feature15:59
piggzvandenoever: found hthis
vandenoever is in MeeGo.Components 0.1,16:00
*** Stecchino_ has joined #harmattan16:00
vandenoeverModalMessageBox is in MeeGo.Components 0.1,
trxpiggz InfoBanner16:00
*** niqt has joined #harmattan16:00
qronichow to check if some file is protected by aegis?16:01
qronicI remember there's some file with hashes of "important" files16:01
trxpiggz dont forget to import 1.0, its in there16:02
piggztrx: thx16:02
piggztrx: is there docsadding to my bookmarks16:03
*** Stecchino has quit IRC16:03
*** Venemo_ has joined #harmattan16:07
*** Venemo_ has joined #harmattan16:07
DocScrutinizerKRF: the mere fact that Nokia seems lazy on that16:07
*** Venemo has quit IRC16:10
*** NIN101 has joined #harmattan16:12
fralsqronic: /var/lib/aegis/refhashlist or something like that16:22
vladesthow to best implement settings page in QML?16:25
vladestwith expanding sections16:25
*** Natunen has joined #harmattan16:36
*** hardaker has joined #harmattan16:38
KRFDocScrutinizer: i see ;)16:48
infobotmay-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not"16:49
DocScrutinizerqronic: ^^^16:49
DocScrutinizer~literal may-i-edit16:50
infobot"may-i-edit" is "<reply>may-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not" || echo "edit if you feel venturous"; }"16:50
DocScrutinizerinfobot: no, may-I-edit is <reply>may-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not" \|\| echo "edit if you feel venturous"; }16:51
infobotokay, DocScrutinizer16:51
DocScrutinizer~literal may-i-edit16:51
infobot"may-i-edit" is "<reply>may-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not" |\| echo "edit if you feel venturous"; }"16:51
infobotmay-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not" |\| echo "edit if you feel venturous"; }16:51
DocScrutinizerinfobot: no, may-I-edit is <reply>may-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not" \|| echo "edit if you feel venturous"; }16:52
infobotDocScrutinizer: okay16:52
infobotmay-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not"16:52
* infobot starts crying and hides from docscrutinizer in the darkest corner of the room. :(16:53
DocScrutinizerinfobot: no, may-I-edit is <reply>may-I-edit () { grep `basename $1` /var/lib/aegis/refhashlist && echo "probably not" && exit; echo "edit if you feel venturous"; }16:53
infobotokay, DocScrutinizer16:53
qronicDocScrutinizer: thanks!16:57
*** dang_ has joined #harmattan17:13
*** dannybe has quit IRC17:20
*** dannybe has joined #harmattan17:20
qronicoh, one more question, what's dui stands for?17:20
DocScrutinizer~wtf dui17:22
infobotGee...  I don't know what dui means...17:22
qroniccat /etc/meegotouch/themedaemonpriorities.conf17:22
qronicsysuid = 9017:22
qronicduihome = 8017:22
arceandui = Direct UI17:23
qronicthanks, do you know what's that for?17:24
arceanthat's all I know :)17:24
qronicok :)17:25
Venemo_DUI (Direct UI) is the old designation of the MeeGo Touch Framework17:25
Stskeepsor, developing under the influence17:25
Venemo_but why "developing under influence"? :D17:27
*** niqt has quit IRC17:27
Stskeepsmore of a historical thing17:29
DocScrutinizerdialer, calculator etc are kinda freaky UI - tap and hold a digit, sile off zhe digit and back on to it - watch what happens17:36
DocScrutinizerin calculator deleting results by the digit via "backspace" is nonsense17:38
DocScrutinizercool: they were clever enough to disable "standby screen" (the clock display) when proximity sensor suggests device is in pocket17:40
DocScrutinizernot cool: the display update of standby clock screen - on new minute - is freaking jerky17:42
DocScrutinizerlooks like there are 3 refreshes done, with short blackouts in between17:43
DocScrutinizeraah, and that standby screen still seems to contribute to >30% battery consumption/day, without even WLAN enabled17:45
frals30% battery comsumption when nothing else is consuming battery...17:45
DocScrutinizersmaller fontsize will *drastically* reduce power consumption17:46
*** roop_ has quit IRC17:46
DocScrutinizerdue to way less bright pixels17:46
fralsbecause i assume you didnt mean it lost 30% of a full charge in one day with everything else off17:46
fralsand you somehow pinpointed those 30% to standby screen :)17:47
DocScrutinizerthat's exactly what I meant, frals17:47
fralshow did you attribute this to the standby screen?17:47
DocScrutinizerI never did17:47
fralssince i have a device on with inet connection disabled and it has never lost 30% in one day when its not being poked17:48
DocScrutinizer>>  seems to contribute to >30% battery <<17:48
DocScrutinizerfrals: stock system, no SIM, no installed apps, no installed WLAN connectivity, charged to 100% some <20h ago, since then it sits there idle with time display - now down 67%17:50
DocScrutinizerlemme redo this check - I'm not sure I never touched it and played with camera whatever since I unplugged it17:53
DocScrutinizerso - scratch the whole comment about 30% - for now17:53
DocScrutinizerI may be wrong17:54
*** stroughtonsmith has quit IRC17:55
SpeedEvili'd be interested on the 'one pixel on' consumption.17:56
DocScrutinizerplugged it to charger now - which btw is a really funny design17:56
DocScrutinizerwow, I'm still online?17:57
DocScrutinizerjust had a lag of 150s17:57
*** gri is now known as zz_gri17:58
*** Scifig has joined #harmattan18:00
DocScrutinizeruhuh, chanlog tells me I *was* offline18:01
DocScrutinizer[2011-11-06 16:55:17] <DocScrutinizer> plugged it to charger now - which btw is a really funny design18:01
Venemo_DocScrutinizer, how's N9(50) hostmode going?18:02
*** RST38h has joined #harmattan18:02
DocScrutinizermaking friend with N9, but not to a degree where I could do anything regarding messing with hw yet18:02
RST38hMoo, Doc18:04
RST38hHow is N9?18:04
DocScrutinizernot even dared to go online with that device for any prolonged time, to avoid possible forced updates to any even newer fw version18:04
DocScrutinizerRST38h: shiny18:04
DocScrutinizerfriggin shiny18:04
Venemo_DocScrutinizer, "to avoid possible forced updates to any even newer fw version" -> what's the point in that?18:05
RST38hyeah - handled it in a store yesterday18:05
RST38hDoc: Nokia Russia seems to be treating it as a flagship btw18:05
RST38hbased on the amount of advertising...18:05
DocScrutinizerVenemo_: what's the point of my N950 still on 22-6? Paegis :-/18:05
Venemo_DocScrutinizer, well, I'll want hostmode on the newest firmware.18:05
DocScrutinizerVenemo_: you won't get that, any time18:06
SpeedEvilDeveloping hostmode is a seperate thing from implementing it on user devices.18:06
DocScrutinizerhostmode is impossible on *any* stock firmware18:06
Venemo_DocScrutinizer, why is that18:07
DocScrutinizerwe'll inevitably need a hostmode enabled kernel for that - at least that's the scientific recent knowledge about the issue18:08
Venemo_so what.18:08
Venemo_you make an aegis manifest file for your package18:08
Venemo_and it'll work18:08
SpeedEvilErr - no.18:09
SpeedEvilNot to install kernel modules.18:09
DocScrutinizerso I'm not going to update devel devices to any more restrictive stock FW for the shits'n'giggles18:09
SpeedEvilin addition, you can't get all tokens.18:09
SpeedEvilEven through ovi store, you can't get kernel module install token.18:09
DocScrutinizerand for now I'm still missing any official stock Nokia HARM image to reflash N918:10
DocScrutinizerso messing with kernels etc on that N9 was rather insane a thing to do right now, in my book18:11
DocScrutinizerI *might* have to go the meegoCE path even, to do *anything* for HEN918:12
DocScrutinizerI got a "cracked" 22-6 N950 where I can mess with hw the way I like, to check what are the chips and how they work - not so on N918:13
*** vladest has quit IRC18:14
DocScrutinizerthe odd thing is it's quite likely N9 has significantly different hw for USB than N918:15
*** Jonno has joined #harmattan18:17
DocScrutinizerso I couldn't even test if my N950 success on enabling vboost 5V-VBUS would work at all on N9, due to newer FW which I don't exactly feel like trying to free it up / de-Paegis it right now18:18
*** crevetor has joined #harmattan18:19
*** vladest has joined #harmattan18:19
DocScrutinizerand our master dragonfighter aegiskiller javisp isn't around to share his secret wisdom to me18:20
RST38hHe used enemas.18:20
DocScrutinizerThe Illinois enema bandit?18:21
RST38hDunno, ask!18:24
* DocScrutinizer prefers to listen to
*** aheinecke has quit IRC18:29
VelmontAny idea howtoshare the net fromthe Nokia N9 to the computer using USB cable? Wifi sharing drains too much battery, my old android does that automatically.18:30
SpeedEviluse it asd modem18:31
VelmontThe N9 kinda does it as well, it seems, I get an IP adress from the N9, but it doesn't come with correct gateway and it doesn't really NAT me through.18:31
VelmontSpeedEvil: Aahhh. OK. So modem, not NAT, then.18:31
Venemo_Velmont, use it as a USB modem...18:31
ScifigHas anyone ported duplicity to harmattan or working on a port?18:35
Venemo_wut's duplicity?18:38
ScifigVenemo_, rsync like backup with encryption.
ScifigIt was available in N900. Don't remember who ported it.18:40
ScifigVery useful to take scheduled remote backups of  device18:41
*** NIN101 has left #harmattan18:41
*** NIN101 has joined #harmattan18:42
VelmontI'm looking into using unison myself. It's very handy, more like dropbox, in that it does two-way syncing.18:45
fluxscifig, actually would you call it rsync-like?18:46
fluxscifig, I'm under the understanding that you cannot really do a 'full backup' each time, that only transfers modifierd parts, like you can with rsync18:47
pinheiromy n950 now decided to keep asking me for my passwords all the time18:47
pinheirowhats up with that?18:47
fluxso you typically transfer only increments, but because you cannot keep doing that either (eventually the increments are larger than multiple fulls), you need to fall back to doing a full and that again takes time a lot :)18:47
Scifigflux, I should correct myself. It is based on rsync. And yes, it only transfers modified parts which is ideal to save on network traffic if you are taking remotebackups. You can configure it take full backups after x backups or y months AFAIK18:50
fluxscifig, so when doing you want the full image to arrive to the backup site, it can optimize the transfers so that it doesn't copy everything, if it already has a previous image?18:51
fluxhmm, then I should reinvestigate duplicity18:52
DocScrutinizerrsync reduces data volume on channel by only transferring diffs, but the mirror is no incremental backup, it's always a full representation of the source - plus files that got deleted on source but stay on mirror18:52
*** Jonno has quit IRC18:53
fluxI thought duplicity simply used rsync's diff algorithm for storing incrementals, so that it doesn't maintain 'a full copy' image on the server18:53
*** arcean_ has joined #harmattan18:54
*** arcean has quit IRC18:54
*** arcean_ is now known as arcean18:54
DocScrutinizernfc what duplicity does18:54
DocScrutinizerrsync however creates an exact copy of source on mirror, and simply keeps it in sync. No versioning18:55
ScifigMy configuration was  --full-if-older-than 2W18:55
Scifigfor duplicity18:55
fluxdocscrutinizer, yes, but rsync doesn't do encryption, which is great to have if you don't 100% trust the storage18:55
DocScrutinizerif I don't trust the storage I'll use cryptfs18:56
DocScrutinizerand mount that remotely18:56
fluxbut does rsync work efficiently in that case?18:56
fluxbecause I would imagine rsync expects fast access to 'local files'18:56
DocScrutinizernothing can work "effectively" on an untrusted storage18:57
ScifigDocScrutinizer, Does cryptfs encrypt a folder/partition or the complete filesystem?18:58
fluxI cannot see how that be an axiom.18:58
fluxit may not be easy, and it may not be as efficient as rsync currently is (ie: file-level), but a useful solution could exist, IMHO18:59
DocScrutinizerflux: just think a bit about how you'd attack such a setup, then you'll find it's (once more) the problem of "trusted computing"19:00
fluxdocscrutinizer, well, think of a setup where you locally first encrypt a file with a symmetric key, and rsync that to the remote system19:01
fluxdocscrutinizer, if you do modify the file (a block-based cipher), you can efficiently rsync that again to the peer19:01
fluxactually, you could do that with ecryptfs already, hm :). you would just rsync the encrypted files with the remote, which doesn't know anything about encryption.19:02
DocScrutinizerif your encrypted file is easily diff-able than your encryption is next to useless19:02
flux(actually encfs)19:03
fluxwhen you modify a file on an encrypted filesystem, do you expect that the whole file gets modified within your block storage?19:03
Scifigflux, or rsync an un-encrypted file to a remote filesystem/volume(?) encrypted using cryptfs?19:04
fluxscifig, something like that. but actually now that I think of it, it would sort of need to be the reverse of encfs, unless you want to encrypt the files locally19:04
fluxscifig, so you could mount an 'encrypted view' of a plain text filesystem19:05
fluxsomething that mounted with (say) encfs gets decrypted properly19:05
DocScrutinizerif you don't encrypt locally, you got a huge open attack vector, just strace the remote process19:05
*** Jonno has joined #harmattan19:06
fluxI never intended in my scneario that plain text would leave the source host19:06
ScifigDocScrutinizer, Didn't think of that. Good to know.19:06
fluxbasically: view-as-encrypted-mount /tmp/plaintext /tmp/encrypted; rsync -a /tmp/encrypted remotehost:19:07
fluxbut, it would need to store some additional data in /tmp/encrypted, such as file-specific keys19:07
fluxand it would need to create such keys if they don't already exist19:07
*** NIN101 has quit IRC19:07
fluxnow I understand there some chance that some cryptoanalysis could be performed if you encrypt the same block twice with the same key..19:08
fluxbut perhaps there are some ways to solve that19:08
flux(same block as in the same offset of the file but with perhaps different content)19:09
DocScrutinizeruse an incremental backup scheme locally to create a local "diff", encrypt it, save it to any storage by any means, no need and use for rsync via open channel in that case19:11
*** Venemo_ is now known as Venemo19:12
ScifigDocScrutinizer, Doesn't rsync need access to the un-encrypted backup image to calculate the diffs?19:12
fluxthe problem (solvable?) is that you need to have the previous image for that19:13
fluxmaybe you can keep a snapshot around with lvm19:13
DocScrutinizerScifig: where's rsync used in my last posted proposal?19:13
DocScrutinizerflux: you usually keep a list of mtimes/md5sums19:14
DocScrutinizerand "backup" whole files for the changed parts19:15
DocScrutinizerif you want to do true diff then obviously you need a full version of the old state locally19:15
fluxdocscrutinizer, would the other end then finally have lists of incrementals?19:15
fluxfor example, my local backup system has dozens of 'full' images of my system with diffs stored with the rsync algorithm, and I never need to actually transfer everything again19:17
fluxI wonder how easily that would be implemented with an encrypting system19:17
DocScrutinizeryou do a find / -newer lastbackuptagfile -exec add_to_tarball, then you encrypt the tarball19:18
fluxeventually the incrementals will end up being larger than the original base image, though19:19
DocScrutinizerflux: sorry you completely lost me on "diffs stored with the rsync algo"19:19
DocScrutinizerflux: BS19:20
DocScrutinizerthis is impossible to happen19:20
fluxdocscrutinizer, how do you choose which tar incrementals to remove from the remote server?19:21
DocScrutinizerif you simply create a tarball of only the changed files, this tarball never can get larger than the original set of data19:21
fluxbut the base image + n incrementals can get a lot larger than 2 times the base image19:21
fluxif you never 'reset' and transfer the base image again19:21
DocScrutinizerso what? that's how backup works19:22
*** Scifig has quit IRC19:22
DocScrutinizerit has more info than the current original, it has history19:22
fluxwhat you are missing is that eventually the oldest backups are purged19:22
fluxmaking more space into the system19:22
fluxwhich incrementals can you remove?19:22
DocScrutinizerthat's up to you, the admin19:22
*** Scifig has joined #harmattan19:22
DocScrutinizerthere's NO rule for that19:23
fluxwell I don't unfortunately have infinite storage available, so I need to purge old backups at soem point19:23
DocScrutinizeryou purge when you think there's no way you ever will need any of that stuff again19:23
berndhsask a lawyer, see what you can get away with19:24
fluxwhat can you purge? the oldest incrementals? the base image?19:24
fluxassuming you don't want to lose any files that are still in the file system, and that you don't want to transfer them again either19:24
DocScrutinizerplease read a good manpage about backup, there are schemes suggested19:24
*** Scifig_ has joined #harmattan19:26
fluxafaik BackupPC works by storing the latest image and previous versions are always diffs to the next version, this way it can just remove the oldest incremental19:26
fluxoops, that wasn't coherent19:26
fluxdocscrutinizer, what those backup tools always require is a periodic full transfer19:27
fluxand I would rather not do it, if transferring a full image takes a week, or a month.19:27
*** Scifig has quit IRC19:28
fluxdump is written in the age of tape drives. you don't really do tape backups without making a full in a periodic fashion. unless you have inifinite supply of takes.19:30
fluxtapes, even.19:30
DocScrutinizerflux: basics of backup management are completely understood and haven't changed since age of Hollerith punchcards19:32
DocScrutinizerand dump isn't doing useless stuff for the mere "because we *can* (and are too lazy to think about better ways)"19:33
DocScrutinizerif you bother to read and *understand* that manpage, you'll see dump provides a few elementary methods and suggests a scheme to use them. Both the methods and the scheme are valid no matter what's your backup storage media, or your channel, or whatever19:34
*** NIN101 has joined #harmattan19:38
*** Venemo has quit IRC19:40
*** berndhs has left #harmattan19:41
fluxdocscrutinizer, let's say I have a file system to backup with 365 pieces of 1 GB files, and I've managed to transfer it once through some means to another site. each day I wish to modify one of these files, in sequence, and backup the file system to the site, transferring at most one gigabyte, and keep 14 latest versions of my current file system in a restorable fashion in the site. how can I do it so that I never store more than 365 + 14 gigabytes on ...19:42
flux... the remote site? which files does the site have in its storage at the end of the year?19:42
fluxperhaps I really do not internalize this subject then, but I've really given it a thought during the years.19:42
*** ab has quit IRC19:44
*** ab has joined #harmattan19:45
*** ab has joined #harmattan19:45
DocScrutinizerflux: as you can not consolidate on your mirror server (it's "untrusted", mind?), there's no other way than either store all 356 incremental backups (at that point you could actually purge the original full image as it has no more relevant data in it, which is known only by your source system as mirror server doesn't know about filenames and relations between them and full image) - or you do another full image every 14 days19:47
*** kgutteridge has joined #harmattan19:48
DocScrutinizeror you remote mount (fuse) a cryptfs file from your mirror server to local, and rsync to that19:48 that local cryptfs volume19:49
DocScrutinizeryou may want to keep local incrementals (like unionfs etc) to keep versions, each mounted to one of 14 cryptfs files on your mirror19:50
Scifig_DocScrutinizer, FWIW duplicity works similar to how u suggested earlier. That is, generate diffs,  encrypt the diffs (it uses gpg) and transfer them to the remote location. It has option to create full backup every 'x' days/weeks/months19:52
fluxdocscrutinizer, in the first solution purging of the original data would need to occur in steps, or at day 364 there would be 1.99 times the original data. but something like the cryptfs would work, but it requires keeping the data encrypted locally as well.19:53
DocScrutinizeron a trusted mirror things are of course way simpler, as you could consolidate there, means create a true open image of the full data that always represents state of 14 days ago, and then you "restore" the oldest 14th incremental backup to the mirror full image and then purge that incremental backup file19:53
fluxbut I'm hoping someone, some day, will write a simple tool that lets me mindlessly run commands and it just works ;)19:53
DocScrutinizerrestore *locally* - no data transfer involved as the mirror "updates" local full image from now-15d to now-14d20:01
DocScrutinizerfor obvious reasons all this can't be done if you store crypted data on your backup mirror20:02
*** mtd_ has quit IRC20:04
DocScrutinizeri.e >> purging of the original data would need to occur in steps << is impossible on encrypted data20:04
*** mtd has joined #harmattan20:05
DocScrutinizerunless you crypt every file separately and thus at very least reveal info about the size and last-modified date of your files, even when you keep a pseudo filename for all of them and crypt the true nametable20:06
*** khertan__ has joined #harmattan20:07
*** Jonno has quit IRC20:07
fluxdocscrutinizer, at least the size could be inexact and the times could be encrypted20:08
flux(but the peer could just find out which files you modify, so it would affect only attacks that occur later than updating the backup data)20:09
*** khertan_ has quit IRC20:11
*** mtd has quit IRC20:25
*** Jonno has joined #harmattan20:30
*** mtd has joined #harmattan20:32
*** stroughtonsmith has joined #harmattan20:43
*** dymaxion has joined #harmattan20:46
dymaxionhi, there's a bit of contradiction on the meaning of the Qt deploy targets.   Meego 1.2 Harmattan API, vs Harmattan Platform API,   (once says that the former is Meego1.2+Nokia,  other pages say that the former is just Meego 1.2 and only the "harmattan API" is Meego+Nokia )  confused!  thanks20:47
*** kgutteridge has quit IRC20:52
DocScrutinizerharmattan is maemo for this, anyway. And there's no harmattan without Nokia21:01
*** dominikb has joined #harmattan21:04
DocScrutinizerI.E. meego.*harmattan is NOT meego21:11
*** mardy has joined #harmattan21:11
DocScrutinizerthat's why this irc channel exists ;-)21:12
piggztrx: either the documentation is horribly wrong, or im doing something horrible wrong....InfoBanner has a 'timeout' property of type int .... ive assigned to it, yet i get an error saying no such property21:12
Scifig_dymaxion, IMO in context of packages available with Qt SDK,  "Meego 1.2 Harmattan API" is bare minimum harmattan API. For example, it does not install gconf header files in the harmattan target. "Harmattan Platform API"  has gconf.21:19
Scifig_I recommend installing both21:19
piggztrx: ah, looking at the source, it appears the documentation i was reading is very wrong
Kypelipiggz: I completely omitted the (anyway lacking) documentation and looked only at QmlComponentsGallery source code21:48
*** harbaum has joined #harmattan21:53
*** Jonno has quit IRC21:56
*** Scifig_ has quit IRC21:59
*** rzr has joined #harmattan22:18
*** fuz_ has quit IRC22:22
*** zz_gri is now known as gri22:22
*** fuz_ has joined #harmattan22:27
*** stroughtonsmith has quit IRC22:29
*** Stecchino_ is now known as Stecchino22:30
*** harbaum has quit IRC22:30
*** artemma has joined #harmattan22:41
artemmaHi All22:41
artemmaHas anybody tried setuid on harmattan?22:41
artemmaI can see the setuid bit, but can't see any effect of it. However, I am not very familiar with anything *nix, so maybe I am missing something very basic22:42
*** NIN102 has joined #harmattan22:43
*** NIN101 has quit IRC22:43
*** mtd has quit IRC22:48
*** mtd has joined #harmattan22:51
merlin1991artemma: you're missing aegis22:54
artemmamerlin1991: could you expand?22:54
artemmaI can see the setuid bit set properly (as it looks to me), but whoever launches the binary, binary is launched under this "whoever" indeed22:55
artemmanot under the binary owner as setuid wants (I think)22:56
merlin1991I'm not sure if aegis (the security framework) denys this but I think it does, ask someone else though22:57
artemmaproblem is that my knowledge is limited regarding both aegis and general unix rights management :)22:57
artemmaso I am hoping for somebody knowledgeable to help indeed22:58
artemmaAnother question BTW: Does anybody know a [relatively] simple public project with a UI app AND a server that is started on a bootup?22:59
artemmaI'd like to clone something like that. That would be an alternative to setuid :)22:59
griartemma: ?23:02
gri"On Harmattan devices, the config files of 3rd party daemons must be placed into the /etc/init/apps/ directory when they are installed on the device. At the end of the device boot sequence, the Upstart framework executes applications within that directory."23:03
grireads like what you want?23:03
artemmagri, indeed, this is what I want. Oh well, I want a compilable example of it :)23:04
artemmaDocs are also very correct that SyncFW is actually way more appropriate for me, but I failed to make SyncFW work23:05
artemmaso for start daemon would be good enough (code is ready and working, I just want to daemonize it). Later I'll research SyncFW again23:05
griwhat type of process do you daemonize?23:06
artemmaflickr wallpaper downloaded23:07
artemmaright now it is started by alarmd everyday (or every hour, etc)23:07
grifor example if you create a dbus-controlled service, this gets started automatically when someone tries to call functions on your service and it's not running23:07
grihmm ok23:07
artemmaso alarmd started service can write only to /var/tmp that is fine.. until a reboot when tmp is cleaned23:07
griwell, you can work around that with dbus23:08
gribut it may not be the perfect solution23:08
artemmaso I thought about creating a server that would listen to message on gconf (or dbus, but I know only about how to make  gconf work)23:08
artemmathe real proper solution is SyncFW, just too complex for me right now23:08
artemmadbus-controlled service is something I didn't know about23:08
artemmasounds proper enough if it can start services indeed23:09
grithe other way: start only a bash script that calls dbus-call23:09
artemmado you know under which user such dbus-service is started?23:09
grino, but I assume user23:09
artemmaas my service would need to save downloaded files to some permanent place23:09
artemmahopefully not under whoever called dbus (that would be nobody)23:10
* artemma went to googling about autostarted dbus services and harmattan-specific code examples23:10
griovi store client was not started before23:11
griand my ssh user is developer23:11
griand the process is started by "user"23:12
griso that looks correct23:12
artemmaI don't know what qdbus is, I assume it is some kind of command-line interface to dbus23:12
griyes, it lets you call functions on dbus by commandline23:12
artemmaI can't see on which line you called that function23:12
artemmalooks like you only listed the methods available23:13
gribut to list the methods it already needs to start the application23:13
*** dominikb has quit IRC23:13
grithe only thing dbus knows is which application it has to start when someone asks for ""23:13
* artemma is trying same via ssh23:14
griyou can also do this with calc etc23:14
artemmaoh. So just a simple message can do that!23:14
artemmaSo I guess all you need to do is to somehow register to dbus and have some special entry point in your server (or maybe even in the same app binary)?23:15
gritake a look at the files in /usr/share/dbus-1/services/23:15
grino special entry points23:16
artemmalooks very simple, indeed23:16
artemmaif I've got it correctly, you just specify which a command line to run and it can include arguments (way for my code to check if it's called via service or not)23:17
artemmait would still be cool to find a working example project (to cut the research time), but that is already something I can start researching23:18
artemmathanks a lot, gri!23:18
artemmathen my alarmd script could just send a message to dbus. Maybe the very same qdbus would be good enough for it23:19
* artemma hopes that user nobody can send messages to dbus23:19
artemmahey, I can verify it right now, by running qdbus under nobody! :)_23:19
artemmaIt works!23:20
grione thing left: I don't know if qdbus comes with developer mode23:21
artemmahmm, I have developer mode activated, but I didn't install anything from it23:21
artemmaexactly to stay as close to final N9 as possible23:21
artemmaAnybody with a retail N9 here?23:21
grime but also developer mode active23:22
griI hope it's included without developer mode23:22
artemmaasked on twitter for somebody with N9 to try that command :)23:23
griif not, there's also "dbus-send" which is mentioned in the harmattan docs of SyncFw23:23
griwell problem is: if you don't turn on developer mode, you have no terminal to try23:23
griso someone would need to create a test package and run this on a phone without developer mode (browser downloading)23:23
grior a nokian replys to your tweet :)23:24
artemmacan you install app out of ovi store if your developer mode is off?23:25
griI think so23:25
artemmaoh well, I can test it on my own N950 by switching dev mode off23:25
artemmamaybe things are a little different if user NEVER switched dev mode on23:26
griI could reflash the N9 tomorrow to test if no one replied23:26
grihaven't even used it right now since the n950 has the newer firmware23:27
artemmaoh well, the change needed for my current alarmd-nobody based implementation is so small that I may be able to prepare a test package very soon23:28
artemmawill probably make no final sense, but can be used for testing the approach23:28
*** Sazpaimon__ has quit IRC23:31
*** Sazpaimon__ has joined #harmattan23:31
*** niqt has joined #harmattan23:42
*** hardaker has quit IRC23:45
*** Venemo has joined #harmattan23:46
*** javispedro has joined #harmattan23:47
javispedroI had happily prepared my N950 for a small weekend trip, loaded songs, maps, etc. etc.23:50
javispedroand like 1 hour after exiting home, MALF.23:51
*** stroughtonsmith has joined #harmattan23:51
DocScrutinizeryou beat me on suggesting AMLF :-P23:51
DocScrutinizerwhat's been the "offending MALF component" according to msg?23:52
javispedrothe MALF's message been changed23:52
javispedroshorter now23:52
javispedrofun thing, this was PR1.1, where I had no crack available for because they closed my beatiful one23:52
*** artemma has quit IRC23:53
javispedroeither way, at that moment, only the thought of the $500 fine prevented me from smashing the N95023:54
* javispedro is not happy.23:54
DocScrutinizerI think smashing isn't punishable23:54
DocScrutinizeras long as you keep the pieces23:54
DocScrutinizerdon't shoot the hw when the sw is the culprit, though23:55
javispedrogood point.23:55
DocScrutinizer(though I feel with you, had similar ambitions several times)23:56
griIs there anyone doing the community-app downloader?23:57
psycho_oreoswhich one would you prefer smashing? a rare N950 device which actually has a hardware keyboard and is probably more loyal to you? or a common N9 device that doesn't even have hardware keyboard?23:57
mgedminit has something even better: BEAUTY! ;)23:58
psycho_oreoswhich has? N950? no doubt. N9? no there's no beauty in N9 when you have lumia 800 looking very similar to it23:59
*** artemma has joined #harmattan23:59

Generated by 2.15.1 by Marius Gedminas - find it at!