IRC log of #maemo-ssu for Tuesday, 2013-08-27

*** nox- has joined #maemo-ssu00:00
*** arcean has quit IRC00:16
*** dafox has quit IRC00:17
*** mkaindl has left #maemo-ssu00:17
*** sunny_s has joined #maemo-ssu00:20
*** Pali has quit IRC00:22
*** Milhouse has quit IRC00:34
*** Vlad_on_the_road has quit IRC00:44
*** Milhouse has joined #maemo-ssu00:49
*** sunny_s has quit IRC01:00
*** mkaindl has joined #maemo-ssu01:19
*** povbot has joined #maemo-ssu02:12
*** ChanServ sets mode: +v povbot02:12
*** tom____ has quit IRC02:17
*** kolp has quit IRC02:27
*** DrCode has quit IRC02:47
*** DrCode has joined #maemo-ssu02:47
*** dos1 has quit IRC03:09
*** LauRoman has quit IRC03:14
*** dos1 has joined #maemo-ssu03:43
*** dos1 has quit IRC03:50
*** povbot_ has joined #maemo-ssu04:04
*** ChanServ sets mode: +v povbot_04:04
*** LaoLang_cool has joined #maemo-ssu04:28
*** LaoLang_cool has quit IRC04:40
*** M4rtinK has quit IRC04:53
*** amiconn_ has joined #maemo-ssu05:25
*** amiconn has quit IRC05:25
*** nox- has quit IRC05:39
*** FlameReaper has joined #maemo-ssu06:07
*** dafox has joined #maemo-ssu07:57
*** M13 has joined #maemo-ssu08:04
*** dafox has quit IRC08:37
*** FlameReaper has quit IRC08:44
*** LaoLang_cool has joined #maemo-ssu08:48
*** LaoLang_cool has quit IRC08:53
*** freemangordon has quit IRC09:11
*** dafox has joined #maemo-ssu09:22
*** LaoLang_cool has joined #maemo-ssu09:23
*** dhbiker has joined #maemo-ssu09:26
*** sunny_s has joined #maemo-ssu09:47
*** dafox has quit IRC09:48
*** FlameReaper has joined #maemo-ssu09:57
*** FlameReaper has quit IRC10:00
*** FlameReaper has joined #maemo-ssu10:13
*** kolp has joined #maemo-ssu10:56
*** LauRoman has joined #maemo-ssu11:06
*** Martix has quit IRC11:21
*** psycho_oreos has joined #maemo-ssu11:32
*** FlameReaper has quit IRC11:33
*** LaoLang_cool has quit IRC11:39
*** Milhouse has quit IRC12:02
*** Milhouse has joined #maemo-ssu12:14
*** joshgillies has joined #maemo-ssu12:19
*** Milhouse has quit IRC12:41
*** M4rtinK has joined #maemo-ssu12:42
*** LauRoman has quit IRC12:47
*** M4rtinK has quit IRC13:12
*** discopig is now known as bromide14:06
*** Milhouse has joined #maemo-ssu14:13
*** Pali has joined #maemo-ssu14:28
*** M13 has quit IRC14:32
*** dos1 has joined #maemo-ssu14:32
*** arcean has joined #maemo-ssu14:47
*** drathir has quit IRC15:40
*** sunny_s has quit IRC16:22
*** joshgillies has quit IRC16:37
*** joshgillies has joined #maemo-ssu16:40
*** sunny_s has joined #maemo-ssu16:55
*** xes has joined #maemo-ssu17:05
*** NIN101 has joined #maemo-ssu18:00
*** drathir has joined #maemo-ssu18:08
*** xes has quit IRC18:23
*** arcean has quit IRC18:37
*** joshgillies has quit IRC18:42
*** FatPhil has joined #maemo-ssu19:00
FatPhilDoes the CSSU pay any attention to the ARM errata for the OMAP?19:10
*** DrCode has quit IRC19:10
FatPhilFor example, enabling Thumb2 can cause heisenbugs19:11
*** DrCode has joined #maemo-ssu19:17
kerioFatPhil: cssu-thumb ships a kernel (which is a rebranding of kernel-power) with the errata workaround19:17
keriothere's some sort of plan to update the kernel for cssu proper19:17
FatPhilwhat's the giturl for the kernel, I'll grab a copy when I get home19:20
keriomh, poke freemangordon or Pali19:20
kerioPali: *poke*19:20
FatPhilAnd do you have a commit id for the fix (it might not be the same bug that I'm thinking of)19:20
kerioit's the same bug i think19:20
kerioerrata 430973 and errata 65741719:21
PaliFatPhil: gitorious.org/community-ssu19:21
FatPhilhmm, is there a problem with static.maemo.org ? wiki pages are taking for ever to load19:21
*** dafox has joined #maemo-ssu19:22
FatPhilOK, found git@gitorious.org:community-ssu/kernel-cssu.git ...19:25
FatPhiltime to go home, methinks...19:26
DocScrutinizer05?o/19:50
DocScrutinizer05\o/ even19:50
DocScrutinizer05FatPhil: freemangordon did an awesome work (rather fight) to fix the thumb heissenbug19:51
*** BCMM has joined #maemo-ssu19:51
DocScrutinizer05and he finally succeeded I'd say19:51
DocScrutinizer05:-D19:51
DocScrutinizer05static.m.o maybe had a temporary hickup. I can't detect any problems right now19:59
DocScrutinizer05FatPhil: if any such problems like slow static.maemo.org arise, please do: *) check http://monitor.maemo.org/ganglia/  *) ping me or warfare or freemangordon or merlin1991 on IRC  *) send a mail to techstaff@maemo.org20:01
FatPhilThe thunmb one (something to do with a cache boundary, but I forget what) was never hit in practice, but TI and ARM didn't have a work around, so we just disabled thumb, not wanting to risk it in the released product.20:01
DocScrutinizer05mhm, interesting new facts20:02
FatPhilSomething that happens only once a year for an engineer happens several times a day to 100000 customers.20:02
DocScrutinizer05yep, sure20:02
FatPhilIs anyone (still) in Nokia?20:02
DocScrutinizer05not afaik20:02
DocScrutinizer05well, we have one contact, or we had several months ago at least20:03
FatPhilThe channel's logged, though, so I shouldn't say anything too revealing.20:03
DocScrutinizer05hehe20:03
*** dafox has quit IRC20:04
DocScrutinizer05SiErr in TI and what needed to be done to cpe with them are not really highly secret (anymore)20:05
DocScrutinizer05cope even20:06
DocScrutinizer05particularly when they are published in TI's own SiErr list - which not all of them are, see the nasty one in OMAP4 I mentioned in #maemo20:07
*** Vlad_on_the_road has joined #maemo-ssu20:07
DocScrutinizer05We're currently wondering about some SiErr in twl4030, do you happen to know why e.g Nokia didn't use the audio of GAIA?20:09
DocScrutinizer05I gather the whole USB disaster been due to a SiErr in twl4030 PHY which failed to detect fastchargers20:09
DocScrutinizer05but why is nobody (Nokia, Pandora) using the audio of twl4030?20:10
FatPhilI see from linux-arm-kernel that freemangordon is in dialog with both Tony and Nishanth, so if they've seen his patches they're probably good.20:13
FatPhilOK, that git repo is diffs relative to some baseline - but what baseline?20:15
DocScrutinizer05probably stock20:24
DocScrutinizer05that's a big nasty in all maemo kernel I've been told by those who actually touched that stuff. Originally there been no git for kernel, so aiui everything got applied by a huge diff patchset20:26
FatPhilpullable from where? I have no idea how you outsiders get your kernels! ;-)20:26
DocScrutinizer05that's the crux20:26
DocScrutinizer05there's probably no "root" for that20:26
DocScrutinizer05(git noob here)20:26
*** M4rtinK has joined #maemo-ssu20:33
FatPhilDamn, I've forgotten everything20:43
*** M4rtinK has quit IRC20:43
FatPhilanyone know what version 21.2011.38-1 corresponds to?20:45
FatPhilThis new n900 has been quite pimped, I really have no idea what's on iyt20:45
*** MohammadAG_ has joined #maemo-ssu20:46
*** MohammadAG has quit IRC20:46
*** MohammadAG_ is now known as MohammadAG20:46
*** discopig has joined #maemo-ssu20:51
*** discopig has quit IRC20:51
ShadowJKMy guess is Probably 1.320:52
ShadowJKunless I'm confusing with Diablo20:52
*** bromide has quit IRC20:53
FatPhilI don't have a FIASCO that corresponds to that, alas20:53
FatPhilIt would be nice to know I could just restore it to the state before I start experimenting20:54
ShadowJK1.3.1 then? Minor update that never received an image20:54
keriothere's even pr13120:54
kerioShadowJK: there's an image20:55
ShadowJKok so 38-1 is pr 1.3.120:55
FatPhilfrom t.m.o, "21.2011.38-1 PR is the version that is installed in my N900 which is pr1.3.1"20:55
*** dafox has joined #maemo-ssu20:59
DocScrutinizer05~pr13121:01
infobotmethinks combined is the rootfs fiasco image of maemo. For N900 latest (PR1.3.1) see http://maemo.cloud-7.de/maemo5/fiasco+co/RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin  or  http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2011/9/--FID--A0A22MVWFVFAM/--LID--FiRe1317015685654/RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin21:01
DocScrutinizer05last "official" one been PR1.3 aka 2010.36-2 iirc21:01
FatPhilI have 36-2 PR and RD images, but nothing newer :-|21:02
DocScrutinizer0538.1 is just with some SSL cert patches21:02
DocScrutinizer05see21:03
DocScrutinizer05~cssu21:03
infobot[cssu] http://wiki.maemo.org/Community_SSU21:03
DocScrutinizer05also see21:03
DocScrutinizer05~bm21:03
infobotsomebody said backupmenu was http://talk.maemo.org/showthread.php?t=6397521:03
DocScrutinizer05(for "restore what been there")21:04
FatPhilI only just got the device, there's none of my data on it.21:04
DocScrutinizer05caveat: it does NOT backup7restore kernels21:04
FatPhilRestore for me just means a fresh OS21:04
*** psycho_oreos has quit IRC21:04
DocScrutinizer05BM does a full rootfs image21:05
DocScrutinizer05it however does NOT backup/restore kernel NAND partition21:05
FatPhilI was thinking it would be nice to take a device, and create a FIASCO that would regenerate that device.21:05
DocScrutinizer05~bm21:06
infobothmm... backupmenu is http://talk.maemo.org/showthread.php?t=6397521:06
DocScrutinizer05>>Can I restore my BackupMenu images through the Nokia Flasher?<<21:06
DocScrutinizer05>>It -is- possible to convert a BackupMenu image to a Nokia Flasher rootfs image, but it requires a Linux PC with mtd-utils.<<21:06
FatPhilwhat's a COMBINED fiasco?21:10
FatPhilWhat MMC image goes with the 131 rootfs?21:11
DocScrutinizer05there are two image types, COMBINED aka rootfs, and VANILLA aka eMMC21:12
DocScrutinizer05there are just 2 eMMC images, the newer one is unchanges since... errr PR1.1?21:12
DocScrutinizer05I *guess* /home/jr/n900/RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.bin is newest one21:14
DocScrutinizer05~flashing21:16
infobothmm... maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware21:16
FatPhilcool, I already have that fiasco21:16
FatPhilSo is CSSU just a series of package upgrades relative to 131? There's no "can't downgrade" NOLO security bullshit implications?21:23
FatPhildamn, root password isn't rootme...21:24
ShadowJKinstall rootsh21:28
ShadowJK"rootme" needs RD mode enabled via flasher, iirc, and comes with side-effects21:29
FatPhilit has sudo gainroot enabled, and I stuck it in rd mode21:30
FatPhilso my console's now active, yay21:30
FatPhilare there any tell-tale packages I can look for to see if I've got cssu already installed (and in which case, what version)?21:31
ShadowJKSettings - "About Community SSU",21:32
ShadowJKdpkg -l | grep cssu21:33
ShadowJKI guess21:33
FatPhilthat's what I tried first - nothing, and settings shows nothing21:33
FatPhilso previous owner just stuck 100s of packages on it (including gui tweaks that confuse me!)21:34
*** nox- has joined #maemo-ssu21:44
*** discopig has joined #maemo-ssu21:47
*** sunny_s has quit IRC21:48
FatPhilyeah, I think ARM errata 430973 was the Thumb issue I remember (or should I say, clearing the BTB could solve the issue that I remember)21:51
DocScrutinizer05yep21:53
FatPhiloooh, uname says "Linux Nokia-N900 2.6.28.10-power50 #1 PREEMPT Sun Mar 18 20:10:56 EET 2012 armv7l unknown21:53
DocScrutinizer05and it supposedly "can't be done" on N900 due to NOLO locking the HS or whatever. Fremangordon found some call to ROM-monitor to toggle the relevant bit nevertheless21:54
DocScrutinizer05KernelPower5021:54
DocScrutinizer05you should probably upgrade to 5221:54
* DocScrutinizer05 wishes Pali and/or FMG were aroundd21:55
*** sunny_s has joined #maemo-ssu21:55
DocScrutinizer05~botsnack21:56
infobotaw, gee, DocScrutinizer0521:56
kerioFatPhil: just reflash it21:57
FatPhilpackage manager is offering me updates, I see a 52 in apt-cache search21:58
keriowhat's this n900? bought used?21:59
kerioor just forgotten in a drawer?21:59
*** M4rtinK has joined #maemo-ssu22:01
FatPhilbought used22:02
FatPhilBought 2 used in the last month.22:03
kerioso... why aren't you reflashing it?22:03
FatPhilI'll probably reflash the other one.22:04
keriothey're used22:04
kerioyou can't know which rootkit the previous owner has installed on it22:04
FatPhilI worked with the previous owner at nokia, and work with him in my current job too. No rootkits22:05
FatPhilHe's put loads of toys on it, it might be nice to play with a few to see if I want to keep them22:05
FatPhilbugger, package conflicts22:05
FatPhilHas anyone got aptitude to work? (the command line works, but not the GUI, IIRC)22:06
*** amiconn_ is now known as amiconn22:09
DocScrutinizer05go for BM! really!22:12
DocScrutinizer05BM is your only chance to restore all those awesome toys22:12
DocScrutinizer05install backupmenu, make a backup of rootfs and /opt (actually /home) to MyDocs (or even better a uSD), backup the MyDocs content via mass storage mode. Then do a uname -a and label the backup with the kernel version to not mess up when you restore22:14
DocScrutinizer05restoring rootfs (incl modules) to a non-matching kernel causes bootloops22:15
*** Vlad_on_the_road has quit IRC22:16
DocScrutinizer05thus: label your backup with the kernel version and copy the backup and MyDocs to your PC. You're done for any disaster recovery22:16
DocScrutinizer05power users and devels say, resoring a BM backup is easier than reflashing22:17
FatPhilyeah, gonna install BM. BUt gonna use aptitude to install it, as I can't grok the bloody gui22:17
DocScrutinizer05as long as you don't do apr-get upgrade or apt-get dist-upgrade or apt-get autoremove, you're fine22:18
DocScrutinizer05mind you, maemo is NOT debian ;-)22:19
DocScrutinizer05though it almost looks alike22:20
FatPhilhmmm, root's PATH was bogus, that's odd22:22
DocScrutinizer05my KP install procedure: cd ~MyDocs; wget kernelpower kernelpower-flasher kernelpower-modules; while ! dpkg -i kernel*.dpk; do :; done22:23
DocScrutinizer05after the max 3rd dpkg invocation, all three pkgs should be installed properly22:24
FatPhilI'm confused by the optified nature of this device22:24
DocScrutinizer05hehehe22:24
DocScrutinizer05~optification22:24
infobotoptification is a inventive duct tape workaround to reclaim space in fs root, done due to the fact the systeminit *and* partitioning is FUBAR,  http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging,_Deploying_and_Distributing/Installing_under_opt_and_MyDocs, or ""OMG - I wish they looked into FHS and moved /usr to eMMC"", http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 bullet1,2 and fhs-2.3.html#PURPOSE16 dot3"22:24
FatPhillooks like my daily device is 1.3 and is not optified22:25
DocScrutinizer05WUT?22:25
DocScrutinizer05df -h /opt22:25
DocScrutinizer05hardly feasible22:26
DocScrutinizer05there's not enough space on NAND /22:26
DocScrutinizer05that's why Nokia invented optification22:26
*** Vlad_on_the_road has joined #maemo-ssu22:26
DocScrutinizer05since PR1.2 a device won't run without proper 7opt22:26
DocScrutinizer05/opt even22:27
FatPhil/dev/mmcblk0p2   2.0G   448.4M    1.4G   23%  /home22:27
DocScrutinizer05excellent22:27
FatPhilI kinda optified it myself about a year before nokia invented optification22:27
FatPhilI think22:27
FatPhilI have no idea wat device this is22:27
FatPhilI've had about 10 in my time, and did different things to each of them22:28
DocScrutinizer05don't tell me, I know that with another device ;-)22:28
* DocScrutinizer05 tries to count the Neo Freerunner he has or had22:28
FatPhilall I know is that it says 20.2010.36-2, and it doesn't have all the bloody bind mounts22:28
DocScrutinizer05o.O22:29
DocScrutinizer05http://paste.opensuse.org/2438843922:29
DocScrutinizer05for your reference22:29
FatPhilyeah, I have no /opt/py*22:30
* DocScrutinizer05 frowns at paste.opensuse22:30
DocScrutinizer05missing python upgrades22:31
DocScrutinizer05not fremantle core22:31
FatPhilI did my optification just using symlinks, assuming this is the one I did it to22:31
DocScrutinizer05symlinks is the way optification works22:32
DocScrutinizer05python is a special critter22:32
DocScrutinizer05pymaemo-optify - PyMaemo optification package22:33
DocScrutinizer05came with PR1.2 iirc22:33
DocScrutinizer05or shortly after22:33
FatPhilyeah, my old device is clear of that pymaemo-optify crap22:34
FatPhilI've never liked bind mounts anyway, but this is total abuse of them.22:35
DocScrutinizer05won't hurt until you want to runa any python based app22:35
*** Vlad_on_the_road has quit IRC22:35
* DocScrutinizer05 agrees in "abuse" notion22:36
DocScrutinizer05nfc why they didn't use symlinks22:36
DocScrutinizer05one possible explanation would be: a reduced set of python support crap is needed in early boot, so there actually is sth in /opt/pymaemo/*22:38
DocScrutinizer05later on they bindmount the true stuff over that22:38
FatPhil'cos I wasn't in the packaging team, that's why!22:38
DocScrutinizer05hehe22:38
FatPhilThen again, I'd have just got the partition sizes (and distribution across devices) right in the first place22:39
DocScrutinizer05indeed22:39
DocScrutinizer05againn (sorry)22:39
FatPhilSwap was bouncing between OneNAND and eMMC several times, they couldn't make their mind up. And when they finally decided on a layout, they got it wrong!22:39
DocScrutinizer05~optification22:39
infobotoptification is a inventive duct tape workaround to reclaim space in fs root, done due to the fact the systeminit *and* partitioning is FUBAR,  http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging,_Deploying_and_Distributing/Installing_under_opt_and_MyDocs, or ""OMG - I wish they looked into FHS and moved /usr to eMMC"", http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 bullet1,2 and fhs-2.3.html#PURPOSE16 dot3"22:39
*** Vlad_on_the_road has joined #maemo-ssu22:40
DocScrutinizer05~factinfo optification22:40
infobotoptification -- created by DocScrutinizer <~jr@openmoko/engineers/joerg> at Mon Mar 15 10:29:39 2010 (1261 days); last modified at Wed Nov 30 04:17:02 2011 by DocScrutinizer!~halley@openmoko/engineers/joerg; it has been requested 121 times, last by DocScrutinizer05, 22s ago; it has been locked by DocScrutinizer.22:40
FatPhilhmmm, I do have a single opt bind mount, just not the pymaemo ones. No idea how I managed to avoid those22:42
FatPhilSo clearly this wasn't the device I self-optified22:43
DocScrutinizer05never installed pymaemo-optify package?22:43
FatPhilif it's not essential generally I don't install anything22:43
FatPhilmy n900 is used for calls, smses, browsing, and ssh-ing to my home machine22:44
DocScrutinizer05it's not essential until you want to run apps that are based on python22:44
DocScrutinizer05all that works without python pretty much ;-)22:44
DocScrutinizer05I do the other way around usually, my home machine ssh into N90022:45
FatPhil14 python packages installed22:45
ShadowJKRe root's path, "sudo gainroot" is equivalent to "su", which retains user environment. "root" when you've got rootsh installed is equivalent to "su -", which gives root env and path22:45
DocScrutinizer05prolly none of them wirll work22:45
DocScrutinizer05ShadowJK: \o/22:45
DocScrutinizer05ack22:45
FatPhilshadowjk: this was root login from the serial console22:46
* DocScrutinizer05 has to leave for dinner22:46
ShadowJKI, for one, like having rootfs on nand, as ubifs is much more competent than ext3 on emmc22:46
DocScrutinizer05check your shell init scripts22:46
DocScrutinizer05oooh, dunno if you find anything useful in it (just occurs to me when I say "check shell init"):22:47
DocScrutinizer05~jrtools22:47
infobotfrom memory, jrtools is http://wiki.maemo.org/User:Joerg_rw/tools22:47
*** Vlad_on_the_road has quit IRC22:48
FatPhilmy device has a sensible ~root/.profile, the new device has none. I guess previous owner may have blatted some dirs22:49
* DocScrutinizer05 grins at >>All these hacks are tested on PR1.2(!!), but the cmdlines here are just off top of my head, so maybe typos or sth can be found. Lots of it will work on PR1.3 as well, but I haven't tested anything for PR1.3 yet<< proving he refused for quite some time to upgrade to PR1.322:51
ShadowJKIf I had tons of spare time, and someone made a modern kernel work, I'd probably convert my device to use f2fs for /home and MyDocs :)22:51
DocScrutinizer05f2fs? o.O22:52
ShadowJKflash friendly filesystem22:52
DocScrutinizer05obsoletes ubi?22:52
DocScrutinizer05sounds intriguing22:53
ShadowJKNo, it's for emmc/sd22:53
DocScrutinizer05aah22:53
ShadowJKnot for raw nand22:53
DocScrutinizer05I bet Pali will love to implement it into KP22:54
FatPhilshadowjk 3.10: boots on n90022:59
ShadowJKiirc there were modem issues?23:00
FatPhilwhether it works with hildon userspace is another matter. but you can boot to a minimal shell.23:00
FatPhilthere's a guy who tests every kernel release on his n810, his n900, his n950 and his n923:01
FatPhilbut being a kernel dev, he doesn't care about userspace - a shell's all he needs23:02
FatPhilHe might well be interested in contributing to Neo90023:03
DocScrutinizer05coool23:07
DocScrutinizer05our problem with fremantle port is: it depends on a certain look&feel of kernel API23:07
DocScrutinizer05actually our problem with all kernels23:08
DocScrutinizer05KP as well as Neo90023:08
DocScrutinizer05basically KP is all about keeping 2.6.28 API/ABI while backporting all the new good stuff23:08
DocScrutinizer05if we only could implement a compatibility layer between recent API (read: sysfs etc) and fremantle userland...23:09
DocScrutinizer05and the very few direct kernel function calls that may have changed23:10
FatPhilYou're not saying that kernel devs have broken userspace are you??!?!?!23:10
DocScrutinizer05LOL23:10
FatPhilBut teh Linus says that never happens!23:10
DocScrutinizer05no, it been evidently Nokia that created broken crappy closed userland apps23:11
DocScrutinizer05that are too silly to learn new sysfs paths etc23:11
DocScrutinizer05even worse: libs23:12
ShadowJKBesides sysfs, there was a bug in the kernel, I forget the details, which some of the closed source packages depend on23:12
DocScrutinizer05o.O23:12
DocScrutinizer05pthread?23:12
ShadowJKiirc23:12
DocScrutinizer05or what it's called?23:12
ShadowJKNo I donmt remember what area23:13
DocScrutinizer05aah no, poll23:13
DocScrutinizer05hardly the right one23:13
DocScrutinizer05since nothing can depend on it23:13
DocScrutinizer05it's simply broken23:13
FatPhilno syscalls have been removed, we'd never do that23:14
DocScrutinizer05yep23:14
DocScrutinizer05nobody claimed that23:14
DocScrutinizer05but aiui some libs (powervr?) are linked against a certain kernel ABI23:14
DocScrutinizer05and as long as you can't recompile them... dooom23:15
FatPhilThere was a change in the POLL_IN behaviour in poll on virtual files23:15
DocScrutinizer05some *poll* been emuated in libc, but not for ARM23:16
DocScrutinizer05emulated23:16
FatPhilAfter 2.6.29 you had to use POLL_PRI, otherwise you'd spin23:17
DocScrutinizer05ask the CSSU dudes, we're getting a new libc for that23:17
FatPhilsh: =: unknown operand23:17
FatPhilsh: 0: unknown operand23:18
*** DrCode has quit IRC23:18
DocScrutinizer05https://bugs.launchpad.net/ubuntu/+source/linux/+bug/31972923:18
DocScrutinizer05here you are23:18
FatPhilIf you want to stay sane - never look at the console...23:19
DocScrutinizer05dang, that took long23:19
DocScrutinizer05 pselect() and ppoll()23:19
*** DrCode has joined #maemo-ssu23:20
DocScrutinizer05FMG fixed it23:21
DocScrutinizer05afaik23:21
DocScrutinizer05now I hope we hooked you as kernel maintainer ;-)23:22
DocScrutinizer05since, if that's not your fav stuff, I dunno what else is23:22
ShadowJKThat must be some nightmare scenario, to get thrown back to a world of 2.6.2823:23
DocScrutinizer05ROTFL23:23
DocScrutinizer05well, come up with a nice compatibility layer (some symlinks?) and we happily migrate to kernel5.756, as long as our lib blobs still work23:24
ShadowJKI wonder if that can be fudged in userspace with LD_PRELOAD tricks23:25
FatPhil2.6.28's a bit outdated, yeah. Doesn't have some must-have features23:25
* ShadowJK vaguely recalls trying to convince microb in Diablo to not focus-steal, via LD_PRELOAD tricks23:26
DocScrutinizer05tbh we lack true kernel hackers23:26
DocScrutinizer05even pali and FMG are not the natural born kernel hackers23:27
DocScrutinizer05they do an awesome job, sure23:27
ShadowJKTurns out there are 97 different ways for an app to bring itself to front and steal focus, and mozilla/gecko/microb called everyone of them 7 times just to be sure.. so I gave up :)23:27
DocScrutinizer05lol23:28
FatPhilI wouldn't claim to be one myself, it just so happens I ended up having to work on it23:28
FatPhilShadowJK: I remember bugs being filed about that.23:28
FatPhilprocess 1356: arguments to dbus_connection_unregister_object_path() were incorrect, assertion "path != NULL" failed in file dbus-connection.c line 5567.23:29
FatPhilThis is normally a bug in some application using the D-Bus library.23:29
DocScrutinizer05anyway, i have to take care about my dinner now. For real. Please give FatPhil a cosy feeling and a heartly welcome23:30
DocScrutinizer05bbl23:30
FatPhilI need to get some food myself too23:31
ShadowJKYou'll saturate serial console with dbus and gtk warnings...23:31
FatPhilshadowjk: I was behind the great console spam cleanup operation. Must have filed hundreds of bugs against userspace.23:31
FatPhilYOu wouldn't believe some of the crap excuses they gave. Actually, you probably would, as you've probably seen their source code.23:33
*** arcean has joined #maemo-ssu23:34
ShadowJKI once tried to submit a bugreport+patch to gypsy. Though it was said in different words, the reply was basically "yes it's broken for real life, but it passes all our tests and the spec", bug closed :-)23:35
*** Vlad_on_the_road has joined #maemo-ssu23:36
*** Pali has quit IRC23:38
FatPhilSeen many a response like that23:39
FatPhilClearly sometimes tests don't include things like "must not emit any GTK assertion failures"23:39
*** LauRoman has joined #maemo-ssu23:39
ShadowJKso you redirect stderr to /dev/null on start?23:40
ShadowJK;D23:40
FatPhilI tried to persuade them to turn assertion failures into fatals, and their response was "but nothing would run!"23:40
FatPhilshadowjk: the bodgers in HArmattan time did exactly that - except they used "console none" in their upstart script23:41
ShadowJKlol23:42
*** dafox has quit IRC23:45
*** _rd has joined #maemo-ssu23:45
DocScrutinizer05just for some idea you can chew on while I'm off for diner: how about a FMRX replaced by a SDR, like http://sdr.osmocom.org/trac/wiki/rtl-sdr, but hooked up directly to some shared RAM that can get processed by DSP?23:48
*** NIN101 has quit IRC23:49
FatPhilShadowJK: do you know what kernel tree/tag the kernel-cssu is based on?23:49
ShadowJKMy guess would be that it's based on nokia's kernel as it is in the source deb?23:51
DocScrutinizer05prolly a terrible power hog (the E4000 is rated ~500mW iirc), but definitely a cool set op option you get23:51
*** arcean has quit IRC23:51
DocScrutinizer05RX from 20 to 2000MHz, SDR23:51
DocScrutinizer05means you even could do DVB-T23:52
DocScrutinizer05and maybe GPS as well23:52
ShadowJKBut could you do police radar warner?23:52
DocScrutinizer05;-D23:52
DocScrutinizer05sure23:52
ShadowJKWith infinite amounts of code..23:52
DocScrutinizer05for those <2Ghz23:52
*** dafox has joined #maemo-ssu23:53
DocScrutinizer05well the rtl2832 would do DVB-T decoding for you23:53
DocScrutinizer05but is has USB interface :-/23:54
FatPhilwhere's the bleedin nokia deb-src url?23:54
ShadowJKAn SDR module would be cool idea implemented in Jolla's "other half" style though23:54
DocScrutinizer05ok, on GTA04 we at least have full featured hostmode that *maybe* can do highspeed23:54
DocScrutinizer05forget that jolla isdea, you won't see the highspeed interface to OH23:55
DocScrutinizer05I still wonder if their interface is BT or NFC23:55
ShadowJKyeah they're probably doing NFC23:55
DocScrutinizer05screw dat23:56
DocScrutinizer05boring23:56
FatPhilIceweasel can't find the server at maemo.research.nokia.com23:56
DocScrutinizer05prolly scrapped23:56
DocScrutinizer05n.c tore down a lot of URLs23:57
DocScrutinizer05aka servers23:57
FatPhilindeed23:57
DocScrutinizer05even worse: some of them gone down without any prior notice23:58
DocScrutinizer05~skeiron23:58
infobotrumour has it, skeiron is the semi-official backup and emergency standin for all internet borne maemo resources: http://skeiron.org/tablets-dev/   http://talk.maemo.org/showthread.php?p=1315143#post1315143, or see: ~tabletsdev23:58
ShadowJKdeb-src http://repository.maemo.org/ fremantle/sdk free23:59
DocScrutinizer05seems I'm already too late for dinner. Might become a beer then23:59

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