**** BEGIN LOGGING AT Fri May 24 02:59:58 2013 May 24 03:50:57 Well, glad to see the room is hoppin tonight May 24 05:32:45 Hey, random n900 camera app question. May 24 05:34:38 Is there a way to enable the camera flash and video recording so that the flash provides a small floodlight for the recording? May 24 05:35:08 I noticed my spouse's Android camera has that feat and thought it was neat. May 24 05:55:35 Raimu: with nicocam, press t May 24 05:55:43 for "torch", i guess May 24 05:56:09 Hehey, thanks! Somehow I'd totally missed that. May 24 07:58:39 A wonderful good Morning on this splendid Prickle-Prickle. May 24 09:06:04 anyone remember that article, a few years back, about an IT guy who worked for some mafia? May 24 09:12:49 stupid but possibly fun Maemo question... May 24 09:12:59 is it possible to leave USB mass storage mode without unplugging the USB cable? May 24 09:13:02 (to go back to charging only) May 24 09:14:17 and if it is possible (I would guess it is, if necessry, by unloading this USB storage provider module then moutning the filesystems again by this *-mmc-*.sh script) May 24 09:14:24 why isn't it offered from the status menu2~? May 24 09:16:04 divVerent: can you not just eject (unmount) it on the computer? May 24 09:16:22 sixwheeledbeast: I did May 24 09:16:26 but then the N900 stll doesn't mount it May 24 09:16:32 and it is right there - I might want to mount it again May 24 09:16:40 divVerent: rmmod g_file_storage May 24 09:16:54 freemangordon_: that was my idea, but it probably doesn't notify the status applet right May 24 09:17:06 divVerent: it does May 24 09:17:14 and it then automatically mounts everythign again? May 24 09:17:19 no May 24 09:17:27 you're asked May 24 09:17:37 but ok May 24 09:17:37 well, at least you should be asked May 24 09:17:39 so all of this is possible May 24 09:17:50 if I read you question right, yes May 24 09:17:52 why doesn't the status menu offer it? May 24 09:18:01 NFC May 24 09:18:23 divVerent: write a replacement, we'll include it in CSSU :p May 24 09:18:32 oh, this part is still closed source May 24 09:19:45 ok, this rmmod command was not wise to do :) May 24 09:19:54 it from then on also didn'rt detect plugging in USB any more May 24 09:20:08 but that was to be expected, that something like this would happen when the system services don't know about it properly May 24 09:20:36 so possibly a wiser way to do this would be to somehow fake the USB unplug notification (I bet this works by dbus, like everything else on Maemo) May 24 09:28:33 divVerent: just do insmod after that May 24 09:29:04 divVerent: see the parameters in osso-usb-enable.sh (or whatever was it called) May 24 09:35:21 ah, that was the script name May 24 09:35:45 one of the osso scripts :) May 24 09:35:49 divVerent: iirc it is in /usr/sbin May 24 09:37:04 /usr/sbin/osso-usb-mass-storage-enable.sh May 24 09:37:24 ah, yes May 24 09:37:32 it tells the module to remove the USB devices May 24 09:44:28 sixwheeledbeast: thedead1440: both of you are councilors, ain't? May 24 09:44:50 freemangordon: yes i'm for MCC and 6wheeled is for HFC May 24 09:45:18 ^^^ May 24 09:45:22 ok. I need some "official" to take it from here http://talk.maemo.org/showpost.php?p=1345746&postcount=1214 May 24 09:46:11 at least we need that discussed first, beforo contacting Ittiam again May 24 09:47:26 If the project "Porting Ittiam HD codecs to N900" kicks, we'll have to have agreement inside the community that we want to do it May 24 09:47:39 and some fundraising should be done May 24 09:48:00 Bring it up in #maemo-meeting, I may not be available again. Due to RL plus internet intermittent issues. Only just catching up on logs of last week. May 24 09:48:04 freemangordon: you have the best technical knowledge of those codecs so i think during today's council meeting you can bring this up for a council response May 24 09:48:34 unfortunately I'll be travelling by the time of the meeting May 24 09:49:01 that is why I am contacting you now :) May 24 09:49:13 ok i'll bring it up during council meeting then May 24 09:49:17 to include that as a topic in today's meeting May 24 09:49:19 yep May 24 09:49:33 I might be able to be around May 24 09:50:04 i guess we'll need some poll on TMO, as if it turns out only 5 people are interested... ;) May 24 09:50:47 freemangordon: polls don't get that many votes; let council come up with a response and when it comes to dollars and cents interested people will raise their voices via their wallets May 24 09:51:36 ok, I am fine either ways May 24 10:05:01 thedead1440: BTW what has happened with Bod elections? May 24 10:12:30 freemangordon: please register and id with nickserv! May 24 10:12:48 DocScrutinizer05: how to do that? May 24 10:12:53 google? May 24 10:13:05 /msg nickserv help May 24 10:13:23 DocScrutinizer51: ok. though I don't have time now May 24 10:13:40 it's a 60s job May 24 10:14:06 /msg nickserv help register May 24 10:44:29 measured by irc users, jolla is now more alive than #maemo ;p May 24 10:56:31 more people in room != "more active". May 24 10:59:25 hehe May 24 10:59:44 I wonder how soon the jolla devices will be rooted and all that May 24 11:25:41 I hope it's not needed May 24 11:27:00 by that I mean I hope it's like Maemo: download root May 24 11:27:06 from appstore May 24 11:50:52 13:27:00 ShadowJK | by that I mean I hope it's like Maemo: download root May 24 11:50:54 13:27:06 ShadowJK | from appstore May 24 11:51:07 it will still have aegis, a quite effective sandbox/binary signing system May 24 11:51:08 I'd hope it was better than maemo May 24 11:51:17 however, the big question is - is root enough for what I would do EVEN with aegis active May 24 11:51:20 it may be just all I need May 24 11:51:34 I mean, on Maemo, I wouldn't really need root for what I do May 24 11:51:42 apart from the command line apt-get interface I don't do much as root May 24 11:51:57 and the hack to disable IPv6 due to a bug in cutetube to be fixed soon :) May 24 11:52:02 for accessing certain /dev/ files etci t is nice May 24 11:52:21 nice but not necessary for productive use May 24 11:52:55 to compare, on iOS a jailbreak is necessary for productive use (to be able to share data between apps), too bad it turns off the sandboxing entirely May 24 11:53:26 on Android, one can live pretty well without rooting it, as one STILL has a command shell, can run one's own code (via third party app installation) May 24 11:53:39 and apps can share data via e.g. the SD card area May 24 11:54:17 so whether that Jolla thing will be actually useful without rooting is not known yet :) May 24 11:54:35 divVerent: well, for some programs /dev/ access is required... May 24 11:54:39 what I also need is SOME good file synchronization method over WLAN usable by Linux May 24 11:54:45 divVerent: sshfs? May 24 11:54:46 like, ssh May 24 11:54:51 it'll run ssh obviously May 24 11:54:52 yes, but for that you need a sshd May 24 11:54:59 Why would it not? May 24 11:55:00 and be able to log in as a user May 24 11:55:06 that's not supported on iPhone May 24 11:55:17 iPhone doesn't use Mer... May 24 11:55:19 sure May 24 11:55:21 thing is May 24 11:55:25 how hardened IS that jolla thing May 24 11:55:31 that is still to be found out, or even, to be decided May 24 11:55:54 they, as the (now) developer, can choose to include thsi functionality or not May 24 11:55:57 also, I need rsync too :) May 24 11:56:05 ideally, that is May 24 11:56:09 on Maemo it all is available May 24 11:56:28 if there is no way to make a rsync binary work (on ANdroid there is, if needed, by self built apk) May 24 11:56:29 divVerent: I'm not sure what you're rambling on about. I doubt this will be much of a problem. May 24 11:56:35 one can still use sshfs + a local rsync run May 24 11:56:41 Wizzup: it just isn't known yet May 24 11:56:46 Jolla keeps pretty quiet about details May 24 11:56:52 They use Mer. May 24 11:57:00 but they can screw it up any way they want May 24 11:57:04 they also use Aegis to lock it down May 24 11:57:11 so their "lockdown" concept may just not allow SSH May 24 11:57:11 yes, and the world may explode any second May 24 11:57:25 thing is, Apple makes such a locked down device and is successful May 24 11:57:30 the world exploding is preferable to suffering with aegis May 24 11:57:39 thing is, WOULD I suffer with aegis :) May 24 11:57:50 it all depends on how well the rest works for my purposes May 24 11:57:57 if the device is locked down, I can't go beyond what the vendor intended May 24 11:58:45 unrooted Android is e..g open enough for my purposes May 24 11:58:48 I imagine there would certainly be a way around it. And that's all you can wonder. May 24 11:59:42 I would prefer it without needing a way "around it" May 24 11:59:49 as these "ways around it" tend to come with disadvantages May 24 11:59:58 e.g. if you jailbreak iOS, you also lose app sandboxing, which is quite good to have May 24 12:00:08 I would actually WANT app sandboxing, BUT I'd like to have control over it May 24 12:00:11 like on Blackberry :) May 24 12:00:42 Android has a pretty weak sandbox concept (due to which any app can read your photos, like it or not) May 24 12:01:03 because SD card data is not protected in any way May 24 12:01:13 (fixable if you have root, obviously) May 24 12:01:37 actually, MAYBE reformatting the SD card with ext3 and setting proper file system ACLs on the root node would suffice :) May 24 12:02:13 to implement a rule like "only the creator of a folder can use it unless administrator sets the ACLs otherwise" May 24 12:02:47 I will soon have an Android device and then try out these things to lock it down a bit further May 24 12:03:05 care to link to this aegis project? May 24 12:03:10 google only turns up games. May 24 12:06:21 Wizzup: either scroll up here, it gets mentioned often May 24 12:06:29 or look for aegis on github, there is a derived project there May 24 12:06:42 basically, it is yet another executable sandbox May 24 12:06:48 and the claim is that Nokia did it right, as opposed to Apple May 24 12:08:22 one thing they do that Apple does not is on-boot integrity checking of startup related files May 24 12:08:39 so even if you can break it, it'd be quite challenging to make the "jailbreak" permanent May 24 12:08:53 divVerent: fyi, the entire boot chain on iOS devices is signed May 24 12:09:06 rikanee: apparently, not completely May 24 12:09:09 but ok, then Apple tried the same May 24 12:09:24 apparently, because how else do jailbreaks get stored persistently? :) May 24 12:09:34 kernel is signed, boot loader is read only May 24 12:09:41 and kernel does binary verification on all executables May 24 12:10:13 divVerent: AMFI is patched to ignore those checks either through misplaced binaries, or misplaced executables May 24 12:10:23 *configuration May 24 12:10:55 I just don't know why you're doubting the people behind Jolla, I mean, sure, I have my doubt about the company as a whole as well. But I don't doubt that they will make an open device. May 24 12:11:30 Wizzup: there has been some interview with the CEO that basically said that the device will be locked down, BUT Jolla will also release a "developer version" May 24 12:11:53 and I would expect this developer version to be quite a lot more expensive than the normal one May 24 12:12:41 Why do you expect it to be? May 24 12:12:47 divVerent: that'd harbor negative attention from Maemo and MeeGo users though. May 24 12:14:13 divVerent: also, regarding sandboxing, I agree that Android did it right, but it didn't go far enough. May 24 12:14:44 iOS makes apps write all their data to their own private folder, Android should enforce that too May 24 12:15:21 data shared with other apps e.g. photos, music, etc. should be exposed via API to standard locations in /sdcard. May 24 12:17:29 rikanee: just read about AMFI... May 24 12:17:38 so the main issue is that launchd.conf is not integrity checked :) May 24 12:17:54 together with a symlink traversal issue in unpacking backups May 24 12:18:01 damn, the symlink issue I almost found before... May 24 12:18:09 used it to browse the iPhone FS via a FTP server app May 24 12:18:11 divVerent: technically, only binaries are checked May 24 12:18:20 by putting a ../../../../../ symlink into an app's Documents folder May 24 12:18:34 I should have tried making a zip file that ALSO has a symlinkname/foo file :) May 24 12:18:55 rikanee: which is the mistake :) May 24 12:19:05 one thing Aegis does extra is some on-boot verification of most /etc stuff May 24 12:19:21 divVerent: it'd be too CPU-intensive to sum and sign every single file, so Apple depends on the fact that binaries are normally not touched and touchable by anything May 24 12:20:15 given that all text files aren't signed, the hilarious part of the evasi0n jailbreak was that it used a shell script to remount / as rw May 24 12:20:34 haha, next iOS bug was that signing only signs text... wouldn't have expected Apple to be THAT clueless May 24 12:20:52 rikanee: I considered that the least interesting part May 24 12:20:54 :) May 24 12:21:02 as it shows no OS flaw May 24 12:21:29 if I had coded binary signing, I probably would have made it verify the executable header AND all sections except for the signature section May 24 12:21:54 just signing .text... I guess I now know how to crack ads out of iOS apps :) May 24 12:22:19 (ok, my previous way was to block the ad servers on my DNS) May 24 12:24:26 Wizzup: May 24 12:24:29 ~aegis May 24 12:24:30 http://www.developer.nokia.com/Community/Wiki/Harmattan:Developer_Library/Developing_for_Harmattan/Harmattan_security/Security_guide , or "The purpose of this framework is: ... to make sure that the platform meets the requirements set by third party software that requires a safe execution environment.", or http://en.wikipedia.org/wiki/Trusted_Computing#Criticism, or http://en.qi-hardware.com/w/images/1/10/ME_382_LockedUpTechnology2.gif May 24 12:25:04 divVerent: last info is "the recent SDK for sailfish has NO aegis-alike functionality" May 24 12:25:19 well, the SDK doesn't necessarily need to have it May 24 12:25:27 if signing is performed when uploading an app to the "store" May 24 12:25:39 but, we'll see in the end May 24 12:25:46 it may even be open enough without any rooting/jailbreaking May 24 12:25:46 yes it does, if you want to write apps that actually work on the final platform May 24 12:26:02 only if these apps use aegisfs May 24 12:26:10 i.e. store/use secret data May 24 12:26:11 no May 24 12:26:18 otherwise, they'd be signed when you upload them May 24 12:26:21 or not? May 24 12:26:31 a security platform on the kernel can work with the OS to redirect calls transparently. May 24 12:26:39 that's the whole point of this shit that you need permissions/tokens for about everything you do May 24 12:26:49 right, THAT part can not be tested then May 24 12:26:55 that is, if you asked for all necessary permissions May 24 12:27:04 DocScrutinizer05: thanks, I'm always interested in the low-level side of sandboxes May 24 12:27:25 but in the end we'll see May 24 12:27:31 it MAY even be open enough without even root May 24 12:27:37 divVerent: actually, another approach would be to discard asking for permissions altogether, and simply disallow certain calls/redirect filesystem access on the fly May 24 12:27:50 rikanee: which is what iOS does May 24 12:28:12 that you see apps asking for permission before reading your photos is only a requirement of the store, but not technically enforced May 24 12:28:17 (enforced by review only, that is) May 24 12:29:16 divVerent: it'd be like SELinux, if implemented May 24 12:29:16 they of course use some "red flags" in the review process, i.e. the permission dialog is shown by a known API, and the images can only be accessed via another API May 24 12:29:31 making it quite straightforward to see if an app needs to be tested for doing this right or not May 24 12:29:42 if it doesn't use the image retrieval API, nobody will test if a dialog pops up May 24 12:30:04 rikanee: with the one exception of who's under control :) May 24 16:34:20 I giggled at the [X] I am a spammer May 24 16:38:13 DocScrutinizer05: okay, made the garage account, have initial tgz and other files made with py2deb, now I'll wait for my account to be allowed to upload to extras-devel :-) May 24 18:05:18 ~botsnack May 24 18:05:18 :), DocScrutinizer51 May 24 21:45:30 hi guys May 24 21:46:20 i was planning to migrate from multiboot to u-boot... then i read this http://talk.maemo.org/showpost.php?p=1344957&postcount=2 May 24 21:46:37 It's totaly useless when booted from u-boot (no wifi, battery meter, charging...) May 24 21:47:53 is it true? any one with nitroid switched to uboot here? May 24 22:28:34 darodi: May 24 22:28:38 >> May 24 22:28:40 To instal kernel you have to: May 24 22:28:41 mount /home /and May 24 22:28:43 first May 24 22:28:45 << May 24 22:29:41 no more to say. All that stuff looks completely gibberish to me. I don't think I could install anything useful following those instructions, and I usually know my way around unix/linux May 24 22:32:09 alas I can't help about your original question, since I never touched nitdroid - sorry. Installing uBoot on N900 is not that hard, it even tries to convert/migrate existing multiboot/bootmenu entries May 24 22:38:58 God gracious, this stuff makes my eyes bleed! May 24 22:39:12 bzip2 -d N12_UMay.tar.bz2 cd / May 24 22:39:24 mount /home /and May 24 22:40:04 cd /and May 24 22:40:36 despite the whole thing is tagged "PHP code:" I think they really mean it May 24 22:41:18 the bzip and mount parameters make no sense to me May 24 22:41:42 except somebody did a damn good job to obfuscate stuff May 24 22:42:36 I wouldn't hope bzip2 takes a COMMAND "cd" as parameter May 24 22:43:18 and I never heard of a device /and you could mount to /home, for whatever purpose May 24 22:43:41 if that's a bindmount that it should need more parameters May 24 22:43:52 at least --bind May 24 22:45:30 even worse, mount syntax is >>mount [-fnrsvw] [-t vfstype] [-o options] device dir<< so the device was /home which for sure is no device on any linux I ever seen May 24 22:46:03 DocScrutinizer05: perhaps that's why people have been filling up their rootfses? May 24 22:46:29 lol, I can't even figure what might happen when trying to follow that instruction May 24 22:47:31 honestly, have a look at this: http://talk.maemo.org/showthread.php?p=1346033#post1346033 for a laugh or a cry, depending on your mood May 24 22:51:03 of course "cd" *might* be a filename for bzip2, but honestly who names his file "cd" ? May 24 22:51:28 sigh, that entire thread is just sad. May 24 22:51:31 ~rtfm May 24 22:51:31 from memory, rtfm is Read The F*cking Manual (TM). It is a suggestion to do your homework before posting a question. Sometimes used as RTFM $SPECIFIC_MANUAL to refer to a specific source of information. See also http://en.wikipedia.org/wiki/RTFM May 24 22:51:35 or it is a option parameter to bzip2 with a typo: -cd May 24 22:52:26 but then I'm afraid user wouldn't have any joy in watching the decompressed output spamming his stdout May 24 22:53:15 only explanation is that a newline is missing, but then the bzip2 cmd makes no sense without parameters May 24 22:53:48 oooor the whole shit is some extremely nifty PHP code indeed, that I simply fail to grok May 24 22:54:03 ;-P May 24 22:54:37 in te end I'd fail to enter the green "/" anyway X-D May 24 22:54:38 DocScrutinizer05: they simply forgot to include exec(); May 24 22:54:41 ;) May 24 22:57:50 ohmy android. Too complicated for me ;-P May 24 22:58:50 mount /home /and is just a bind as you say ...never seen that syntax before May 24 23:01:46 i think i'll just remove nitdroid and go for uboot. i'm using maemo 99% of the time and just reboot to nitdroid when my wife desesperately needs to play a stupid game :P May 24 23:03:41 aaah, maybe get her a cheap droidfon May 24 23:04:42 darodi: make sure you boot to maemo, so multiboot cleans up its shit May 24 23:05:12 well... she also has a n900 with a gamegripper but pollutes mine with nitdroid :)) May 24 23:05:13 (proper kernel modules and kernel flashed - yes, multiboot *flashes* kernel on each boot) May 24 23:05:42 then you should have no problems installing most recent uBoot May 24 23:06:18 and if you're lucky, uBoot will be able to boot into android on uSD out of the box May 24 23:07:03 if not, I think the tweaks needed to uBoot config files and maybe android fs on uSD are minimal May 24 23:08:07 you might need to get the kernel uimage of android kernel and put it somewhere from where uBoot can load it May 24 23:14:22 are there pre-made uimage already or do i need to convert the vmlinuz kernel to uimage? May 24 23:14:39 i'll have a look in the nitdroid forums May 24 23:23:14 there are pre-built uimages for about everything May 24 23:24:03 ask Pali, he should know a lot more about all that stuff - after all he's maintainer of uBoot May 24 23:24:21 hmm, lemme try sth... May 24 23:24:23 ~uboot May 24 23:24:24 N900 uBoot is a siamese twin binary [uBoot+stockMaemoKernel] that resides in kernel NAND partition /dev/mtd3 aka "kernel". You can't uninstall it, rather you'll nuke it when you flash/install another kernel like stock maemo kernel or powerkernel. To start other than stock maemo kernel via uBoot, you have to provide the according kernel image files May 24 23:25:02 hmmpf, thought there might be an URL to a proper wiki/tmo page about uBoot and how-to-install May 24 23:25:05 if you put linux (z)image to /boot or /opt/boot it will be automatically converted to uimage format by u-boot-update-bootmenu May 24 23:25:16 wow May 24 23:25:25 and stored in MyDocs May 24 23:25:44 :-D May 24 23:25:54 scary :-> May 24 23:25:58 Pali: thanks for the info May 24 23:29:00 Pali: seems Nokia never will fix that signing key issue May 24 23:29:20 did you got any new email? May 24 23:29:41 Pali: should we get rip-offs of the Nokia repos and publish a tool to patch the catalogs? May 24 23:29:45 no new mail May 24 23:30:19 maybe even a fiasco image with patched catalogs list May 24 23:30:42 the run our own core repos May 24 23:30:45 then* May 24 23:30:46 Pali: i have cssuthumb and kp 51 in nand right now. should i reboot to omap1 before uninstalling multiboot? May 24 23:31:02 ohohoh May 24 23:31:04 15.5 I got email from Mikko that he going to fix it May 24 23:31:14 :-o May 24 23:31:47 darodi: a thumbified system will crash on a non-thumb-enabled kernel May 24 23:32:10 that's what i thought May 24 23:32:29 alas uBoot comes with stock kernel May 24 23:33:10 uhmf May 24 23:33:12 so installing uBoot on a thombified system is quite tricky - you easily end in a reboot loop May 24 23:34:17 you basically have to make sure that your uBoot can boot from a powerkernel uimage before you install uBoot May 24 23:34:55 otherwise you end up with a system with uBoot and stock kernel that can't boot up to fix it any easy way May 24 23:35:06 I will write to Mikko if he is really going to fix it or not May 24 23:35:25 make sure you have the right "To:" ;-) May 24 23:36:01 and maybe add council to CC May 24 23:36:25 for sure council wants to help on that issue May 25 00:26:58 DocScrutinizer05: so... to summarize May 25 00:26:58 1. uninstall multiboot May 25 00:26:58 2. install uboot May 25 00:26:58 3. upgrade kp51 to kp52 May 25 00:26:58 4. create bootmenu items May 25 00:26:58 5. run update boot menu May 25 00:26:59 6. reboot May 25 00:26:59 i am right? May 25 00:27:41 0. backup everything :P May 25 00:29:22 hmm, to me that lloks good, but that doesn't mean a thing. I'm no expert May 25 00:30:06 though... upgrade KP nukes uBoot May 25 00:30:28 you need to install KP-uimage May 25 00:32:47 that's what i meant with the point 3 May 25 00:32:48 install uimages for kp and create valid bootmenu items May 25 00:32:52 the idea is that uBoot loads different uImages depending on kernel you want to boot. But the one "built-in" kernel that ships with uBoot is a stock maemo kernel that doesn't work with thumb. So you need the kp52-uimage in place before you install uBoot May 25 00:33:10 yep May 25 00:34:55 s/install/flash/ May 25 00:35:18 I guess I found something to do tomorrow May 25 00:48:46 thanks again DocScrutinizer05 May 25 00:48:46 good night every one **** ENDING LOGGING AT Sat May 25 02:59:57 2013