IRC log of #maemo-ssu for Tuesday, 2011-05-31

*** lizardo has quit IRC00:04
*** freemangordon has quit IRC00:12
*** TurboVomit has quit IRC00:26
*** jamie_t has joined #maemo-ssu00:30
*** mr_jrt has quit IRC00:31
*** fatphil has joined #maemo-ssu00:36
*** M4rtinK has quit IRC00:38
fatphil^5 kirma00:40
*** m0use has joined #maemo-ssu00:44
*** jamie_t has quit IRC00:49
*** wmarone has quit IRC00:49
*** mr_jrt has joined #maemo-ssu00:50
*** wmarone has joined #maemo-ssu00:54
*** mr_jrt has quit IRC00:56
*** the_lord has quit IRC00:59
*** mr_jrt has joined #maemo-ssu01:00
*** mr_jrt has quit IRC01:04
*** mr_jrt has joined #maemo-ssu01:08
*** M4rtinK has joined #maemo-ssu01:11
*** wmarone has quit IRC01:14
*** wmarone has joined #maemo-ssu01:16
*** jamie_t has joined #maemo-ssu01:38
*** trbs has quit IRC01:38
*** mr_jrt has quit IRC01:40
*** jamie_t has quit IRC01:42
*** mr_jrt has joined #maemo-ssu01:47
DocScrutinizera silly question: is (will be) mediaplayer part of CSSU, according to CSSU def to include no things that could go to extras?01:47
thpDocScrutinizer: afaik mediaplayer might be 'protected' because it's coming from nokia repos, so we might need to use the cssu repo if we want to replace that01:49
thpi.e. an extras repo can't replace a nokia-provided package for security reasons01:49
thps/extras repo/package in extras/01:50
* DocScrutinizer scratches head, as there are plenty of mediaplayers. Just wondering what makes this one special (except of course MohammadAG's love)01:50
DocScrutinizerthat's not a proper rationale: "we need to include it into cssu, because we want to replace the stock one"01:51
DocScrutinizerthe question is: why do we need to replace stock mediaplayer?01:53
MohammadAGportrait mode is enough reason imo01:54
DocScrutinizerno way01:54
MohammadAGof course, when it offers the same functionality01:54
*** mr_jrt has quit IRC01:54
MohammadAGyes way, if it's as stable01:54
MohammadAGwhich right now, it isn't01:54
MohammadAGno UPnP support is a blocker for example01:55
DocScrutinizerso are you going to replace xterm, when you manage to get another one that can do pink-on-skyblue?01:55
DocScrutinizerthe foundation rule been: nothing goes to cssu that could as well go to extras(-*)01:56
MohammadAGno, but maybe skyblue on pink01:56
*** mr_jrt has joined #maemo-ssu01:57
* DocScrutinizer starts feeling totally angry and pissed, and considers better starting his planned walk to the cosy night out there01:57
fatphilOK, I'm new to all these discussions, but why is the nokia closed status important - can't you reverse engineer interfaces?01:58
DocScrutinizerjus to finish on my post 2 above: and definitely nothing that could go extras will go cssu to kick stock stuff out for no reason but the mere "we did it" thrill01:58
*** mr_jrt has quit IRC02:03
*** mr_jrt has joined #maemo-ssu02:07
*** mr_jrt has quit IRC02:12
*** mr_jrt has joined #maemo-ssu02:15
*** mr_jrt has quit IRC02:22
*** mr_jrt has joined #maemo-ssu02:25
*** mr_jrt has quit IRC02:30
*** mr_jrt has joined #maemo-ssu02:34
*** mr_jrt has quit IRC02:38
*** mr_jrt has joined #maemo-ssu02:42
*** mr_jrt has quit IRC02:46
*** mr_jrt has joined #maemo-ssu02:50
*** mr_jrt has quit IRC02:54
*** mr_jrt has joined #maemo-ssu02:58
*** mr_jrt has quit IRC03:03
*** mr_jrt has joined #maemo-ssu03:06
*** peetah has quit IRC03:11
*** mr_jrt has quit IRC03:11
*** peetah has joined #maemo-ssu03:11
*** mr_jrt has joined #maemo-ssu03:15
*** peetah has quit IRC03:16
*** mr_jrt has quit IRC03:19
*** M4rtinK has quit IRC03:22
*** mr_jrt has joined #maemo-ssu03:23
*** mr_jrt has quit IRC03:27
*** mr_jrt has joined #maemo-ssu03:32
*** peetah has joined #maemo-ssu03:33
*** mr_jrt has quit IRC03:39
*** mr_jrt has joined #maemo-ssu03:42
*** mr_jrt has quit IRC03:47
*** mr_jrt has joined #maemo-ssu03:51
*** mr_jrt has quit IRC03:58
*** mr_jrt has joined #maemo-ssu04:01
*** mr_jrt has quit IRC04:06
*** wmarone has quit IRC04:07
*** wmarone has joined #maemo-ssu04:07
*** mr_jrt has joined #maemo-ssu04:09
*** mr_jrt has quit IRC04:14
*** mr_jrt has joined #maemo-ssu04:18
*** mr_jrt has quit IRC04:22
*** mr_jrt has joined #maemo-ssu04:26
*** mr_jrt has quit IRC04:31
*** mr_jrt has joined #maemo-ssu04:34
*** mr_jrt has quit IRC04:39
*** mr_jrt has joined #maemo-ssu04:43
*** mr_jrt has quit IRC04:48
*** mr_jrt has joined #maemo-ssu04:52
*** mr_jrt has quit IRC04:56
*** mr_jrt has joined #maemo-ssu05:00
*** GonzoTheGreat_ has joined #maemo-ssu05:03
*** GonzoTheGreat_ has quit IRC05:04
*** mr_jrt has quit IRC05:04
*** GonzoTheGreat_ has joined #maemo-ssu05:04
*** GonzoTheGreat_ has quit IRC05:06
*** GonzoTheGreat_ has joined #maemo-ssu05:07
*** BCMM has quit IRC05:07
*** mr_jrt has joined #maemo-ssu05:08
*** mr_jrt has quit IRC05:12
*** GonzoTheGreat_ has quit IRC05:15
*** GonzoTheGreat_ has joined #maemo-ssu05:15
*** mr_jrt has joined #maemo-ssu05:16
*** mr_jrt has quit IRC05:23
*** mr_jrt has joined #maemo-ssu05:26
*** mr_jrt has quit IRC05:31
*** mr_jrt has joined #maemo-ssu05:34
*** mr_jrt has quit IRC05:41
*** mr_jrt has joined #maemo-ssu05:45
*** mr_jrt has quit IRC05:49
*** mr_jrt has joined #maemo-ssu05:54
*** mr_jrt has quit IRC06:03
*** mr_jrt has joined #maemo-ssu06:04
*** mr_jrt has quit IRC06:08
*** mr_jrt has joined #maemo-ssu06:12
*** mr_jrt has quit IRC06:17
*** DocScrutinizer has quit IRC07:34
*** DocScrutinizer has joined #maemo-ssu07:34
*** Khertan has quit IRC08:18
*** crashanddie has quit IRC09:03
*** paiburio has quit IRC09:03
*** frostbyt1 has joined #maemo-ssu09:19
*** myuu has joined #maemo-ssu09:35
*** GonzoTheGreat_ has quit IRC09:45
*** M4rtinK has joined #maemo-ssu10:21
*** m0use has quit IRC10:41
*** rantom has joined #maemo-ssu10:46
rantomI've just got a quick question : is there, by any chance, a chance to have "pages" from MeeGo's Applications list in Maemo5?10:48
*** crashanddie has joined #maemo-ssu11:07
*** crashanddie has joined #maemo-ssu11:07
thprantom: you mean have the application launcher icons "snap to page"?11:28
*** m0use has joined #maemo-ssu11:29
thprantom: there's but it just arranges the icons paginated, it does not yet "snap to page" on scroll11:30
*** M4rtinK has quit IRC11:33
rantomthp: Not quite11:40
rantomHang on, I'll post a screenshot11:40
rantomSimply put : X amount of items for one page, then move to another one (e.g. 15, as it is now)11:43
thprantom: well, that's basically it (just vertical instead of horizontal) and missing the snap-to-page as i said11:44
rantomAh, so that's what snap-to-page meant11:45
rantomMy bad :P11:45
rantomBut it might/is/will/could be possible to add it to Maemo 5?11:45
rantom(That might though require rewriting to the Applications list editor too, since it doesn't work in portrait without forcing the rotation and it doesn't support folders nor pages)11:47
myuuwow somehow updating to CSSU11:51
myuufixed a problem i was having with feedback on certain headphones11:51
*** VanpoofFluffy-ku has joined #maemo-ssu11:51
fatphilIs CSSU known for killing the battery?11:52
fatphilI'm down to 25h idle time now.11:53
*** paiburio has joined #maemo-ssu11:58
*** ThreeM_ is now known as ThreeM12:08
psycho_oreosI can't see how it would kill battery.12:11
fatphilI never really used that device before putting CSSU on it, so I don't have anything to compare too.12:12
psycho_oreosthat could be your reason why. I've hardly seen mine using more battery than prior to installing CSSU, and I've been using my device literally everyday12:13
fatphilI'll do a powertop run of several hours as soon as it's charged again, that should highlight any bad boys in userspace.12:14
Damnshockfatphil: sometimes is not that simple ;)12:14
fatphilthen I'll reflash the stock FIASCO again and compare.12:14
psycho_oreosusing power kernel may also affect the performance12:15
psycho_oreosand any other tweaks/customisations you have setup12:15
fatphilNever used the power kernel (if I want a custom kernel, I'll build my own), and I have no customisations.12:16
psycho_oreosnot even adjusting for instance the brightness display level?12:16
fatphilThis was device idle. Screen off the whole time12:16
fatphilDamnshock - it's very hard for userspace to hide from powertop. I have an i2c sniffer too which can help identify who is behind the various I2C interrupts too.12:18
psycho_oreosmy other guesses might be with constant checking of updates. CSSU has their own repository which may use up a bit more battery if anything12:18
fatphilno SIM, no WLAN.12:18
Damnshockfatphil: on my experience userspace is not usually what drains the battery12:18
Damnshockrather than some problem with a connection or whatever12:18
Damnshockon, no sim and no wlan12:19
Damnshockthen you're right ;)12:19
psycho_oreosiirc the device on default setup would search for wireless networks and if there's any open ones it might randomly associate. Otherwise it would wait for the user to input the various password so that the device could login to the network12:20
psycho_oreosUnless of course you have blacklisted wl12xx :)12:21
psycho_oreosthat's one ugly method, the other might be to just go through settings and disable it from there12:21
fatphilYou can set teh settings to not automatically scan and connect to wlan12:21
psycho_oreosright which was what I said however should you have any open wireless networks nearby, the assumption of no wireless LAN would have been wrong. Otherwise if you set the settings to not automatically scan and connect to wireless LAN, you would have contradicted your saying of not customising your device :)12:23
merlin1991oh lol12:24
merlin1991going to some crappy exam in 2 hours12:24
merlin1991there's supposed to be an example of it online12:24
merlin1991only the link 404s12:24
fatphilSetting the "always ask" setting is not customisation, IMHO12:25
psycho_oreoseither way I'll be interested on the results of your outcome if you: 1) don't actually have any open wireless networks nearby and that you do not take your device anywhere else, 2) maybe disable the automatic scanning/connecting of wireless LAN on both occasions of having CSSU installed and not having CSSU installed.12:25
psycho_oreosto me it is :)12:25
psycho_oreosany form of modification from the standard form, irregardless if its deemed as tailoring, personalising or even customising is still more or less the same because it is not completely original from manufacturer's form apart from the usage of CSSU and its repository12:27
fatphilIt's a device I don't use, so I can happily reflash, install CSSU, etc. any number of times12:27
Damnshockwhat's the exam about merlin1991?12:28
fatphilpsycho_oreos: so adding a single contact is customising?12:28
Damnshockwow, I wish I had a spare device12:28
merlin1991Damnshock: sql :P12:28
Damnshockgood luck with that merlin199112:28
psycho_oreosfatphil, yup12:28
psycho_oreosin other words I'll be looking forward to this proof of CSSU chewing up more battery than usual with scrutiny12:29
fatphilI've not made such a claim. It was worded as a question.12:31
merlin1991fatphil: it would still be interesting to see the comparison12:32
fatphilI have no idea whether the base FIASCO I flashed onto it was what other people use anyway.12:32
Damnshockmerlin1991: indeed12:32
Damnshockwhat I really would like to see though12:32
merlin1991cssu changes quite a lot on the system level, so it could eat more battery12:32
fatphilOK, I'll try to be as scientific about is as possible.12:32
Damnshockis the ability to pipe the received email through an external program12:32
psycho_oreosit wasn't phrased in one line as a claim :) it was more of the fact that you seem to indicate CSSU as being the only culprit as to why your device's usefulness being decreased12:32
Damnshockpsycho_oreos: i did not read that12:33
psycho_oreosDamnshock, or you weren't following the coversation12:34
psycho_oreoshe said apart from installing CSSU, he has not changed anything and so he could only get 25hr idle time12:34
Damnshockso what?12:34
Damnshockdude, what can come from anywhere12:34
Damnshockhe said he is going to test if it comes from the CSSU12:35
Damnshocknothing else, nothing more12:35
Damnshockyou are reading more than it is on his statements12:35
Damnshockas he already told you12:35
fatphilmore than not changed anything - I didn't even use it before putting cssu on it. It's not a comparison before/after12:35
psycho_oreosDamnshock, more like reading between the lines12:36
Damnshockanywahy, there's probably something wrong because I can get far more than 25h of idle time12:36
*** trbs has joined #maemo-ssu12:36
Damnshockeven with a 2 years old battery12:36
Damnshock<psycho_oreos> Damnshock, more like reading between the lines <--- then again: he already told you that was not his intention12:40
psycho_oreosDamnshock, intention of?12:41
Damnshockintention of indicating CSSU as being the only culprit of... blablablabla12:42
Damnshockit is a possibilty, ain't it?12:42
Damnshockmight be not likely, but it still is a possibility12:42
Damnshockthere's nothing wrong in checking12:42
fatphilThe fact that nobody answered immediately "nope" implies that nobody's been paying that close attention to such things12:42
fatphilthe fact that nobody answered "yup" implies that there's no evidence of a widespread problem likewise12:44
merlin1991fatphil: agreed12:45
psycho_oreosDamnshock, I never said there was anything wrong with checking I'm just merely speculating on the possibilities of one noting the cause12:45
fatphilGIven that this device does almost nothing apart from sit on my desk, I'm much more likely to notice idle-time issues12:46
fatphildo you guys have pmtrackerd on your devices?12:48
*** rantom has left #maemo-ssu13:02
*** BCMM has joined #maemo-ssu13:39
*** lardman|gone is now known as lardman13:52
*** chi_Wubbie_Baby_ has joined #maemo-ssu14:00
lardmanIs there a CSSU category on the bugtracker?14:09
lardmanI think not, would it be useful? I see there is a thread on TMO which contains people moaning about CSSU bugs (or in many cases probably things that are nothing to do with CSSU)14:09
merlin1991it's under extras14:09
lardmanmerlin1991: oh really, I must have missed it14:09
merlin1991Maemo community14:09
lardmanah ok14:09
*** the_lord has joined #maemo-ssu15:13
*** lcuk2 is now known as lcuk15:14
*** lcuk has quit IRC15:14
*** lcuk has joined #maemo-ssu15:14
merlin1991thp: I'm trying to watch the that rabbit game youtube vid, but buffering doesn't really work @ U2 karlsplatz :P15:19
thpmerlin1991: 3g or q/spot?15:26
thpmerlin1991: although i think q/spot is resselpark only15:26
merlin19913G :P15:29
merlin1991It actually did work15:29
merlin1991After I did reload the page :P15:29
merlin1991Damnshock: sql exam turned out to be a real joke :P, one of the most easy ones so far15:35
*** RST38h has joined #maemo-ssu16:25
*** BCMM has quit IRC16:26
DocScrutinizerI have standby times like >25h here - WITH wlan, WITH xchat monitoring #maemo* and some other chan, and WITH GSM registered so I can get called every time16:37
DocScrutinizerobviously on cssu -
fatphilI've removed pmtrackerd, which I suspect was part of the problem, and am now full charged. I'll wait for 'pew's...16:42
DocScrutinizerhehe, pmtrackerd - nice16:42
DocScrutinizerIroN900:~# powertop16:43
DocScrutinizerPowertop 1.13.316:43
DocScrutinizerstatus: Unknown job: pmtrackerdaemon16:43
DocScrutinizerthat's considered "normal"16:43
DocScrutinizerwhat's that i2c-tracker you mentioned, fatphil ?16:44
fatphilkernel patch16:44
DocScrutinizeris it configurable?16:44
DocScrutinizeras then we'd probably like it to go to PK16:45
fatphilI'm not sure if I can release it, as I did it during billable hours. I'll ask my boss16:46
DocScrutinizerok :-)16:46
DocScrutinizeryour boss will get free advertisment in the (C) notice ;-)16:47
fatphilI think Nokia are already mentioned, aren't they?16:48
DocScrutinizersponsoring almost always pereived rather positive in community, no mater how 'big'16:48
DocScrutinizerthey are, for the good as for the bad ;-)16:49
fatphil^ selotape over the mouth16:49
DocScrutinizerouch, that hurts ;-D16:50
DocScrutinizeryou're probably free to do a 'similar' project in your leisure time16:51
DocScrutinizerit's not really rocket science to patch the kernel16:51
DocScrutinizeryou just had to do it "in a cleanroom" - I can see that'S not very tempting a leisure activity ;)16:53
fatphilIs CSSU stuck at 2.6.28?16:53
DocScrutinizernuttin we can do about it16:53
DocScrutinizerexcept binnin whole maemo and rewriting it from scratch16:54
DocScrutinizerwe are missing ~30% sources so we could recompile on recent kernels - a HUGE PITA16:54
DocScrutinizerwell, we somewhat "learned to live with it", backporting kernel patches to powerkernel, and taking 2.6.26maemo as an invariant16:57
DocScrutinizerreally annoying: meego is on 2.6.34++ now, so no more direct compatibility16:58
DocScrutinizere.g maemo doesn't profit from the maemo mce sources getting disclosed for meego, as Nokia decided to "update" to new kernel *prior* to disclosing it17:00
DocScrutinizerthat's as I understand it17:00
fatphilIf I remember the innards of the MCE architecture, the kernel dependency is absolutely minimal17:01
fatphilstrings and strace should tell you everything you want to know17:01
DocScrutinizerI'm not saying it's major problem for mce particularly, more I want to show a basic issue17:01
DocScrutinizerconstant source of mild annoyance, related to maemo being tied to 2.6.2817:02
DocScrutinizerwe can cope with all that, just it's so odd some way17:03
fatphilHmmmm, if I'm bored I might try flashing a more modern kernel and seeing what goes wrong17:03
DocScrutinizerwords of a truly great man17:04
fatphilI've not achieved anything yet.17:05
DocScrutinizeryou're rising mood to the better here, wiith a few lines of posting17:05
DocScrutinizerwe can use that, a lot of it17:06
fatphilug. comparing defconfigs won't be fun...17:07
DocScrutinizer[2011-05-31 11:48:19] <mrmoku> Configuring kernel-module-musb-hdrc.17:09
DocScrutinizer[2011-05-31 11:48:19] <mrmoku> FATAL: Module musb_hdrc not found.17:09
DocScrutinizer[2011-05-31 11:48:32] <mrmoku> that probably is your usbnet problem17:09
DocScrutinizeris what you're facing when trying to get recent kernels to N90017:09
*** lizardo has joined #maemo-ssu17:09
DocScrutinizerNB not for maemo17:10
DocScrutinizerthat's : and those pals are rather good at getting sth running on N90017:11
DocScrutinizeruuhm a lil bit OT here17:11
DocScrutinizerrealy nice guys, if you want to have a short chat about things like defconfig - /join #openmoko-cdevel, just ask17:13
*** crashanddie has quit IRC17:19
fatphilyou don't need to be dour here, it's those inside nokia who should be dour:
fatphilWhat's the identifier of the FIASCO everyone uses?17:20
fatphilas in a week & build identifier?17:20
DocScrutinizermoment please17:21
DocScrutinizerfatphil:  RX-51_2009SE_20.2010.36-2_PR_COMBINED_MR0_ARM.bin17:24
fatphilOK, I have a 36-2 RD to hand17:25
DocScrutinizershould be recent17:25
DocScrutinizerfatphil: (dour) we try to not be - yeah NOK :-S17:28
DocScrutinizermost baffling thing: there's been nothing worse than a simple >>Nokia to Support Symbian System During Shift to Windows, Chief Elop Says [Bloomberg - May 27, 2011]<<17:30
DocScrutinizerso maybe -15% is completely unrelated, I'm not much into stock exchange obscurities17:31
DocScrutinizerwell, Q2 outlook published/adjusted17:35
RST38hHmmm, with the new CSSU, the status bar has got some garbage17:37
RST38hEvery time I run an application, the title bar is covered with a bit of garbage17:38
RST38hmodest-home-applet is gone again17:40
*** paiburio has quit IRC17:40
fatphilarse, the only spare hardware I have is too old, and unsupported by 36-2. I'll need to visit the office...17:41
DocScrutinizerfatphil: you're making my day17:45
DocScrutinizerreally encouraging to see somebody with a clue and access to unspecified but undoubtedly promising resources is caring about maemo-cssu17:47
RST38hDoc: Let the guy phantasize a bit :)17:48
fatphilI've got a rx51_defconfig from Jun 2010 here ...17:50
fatphilI can try a kernel build with that17:51
DocScrutinizerwell, here should be a couple of users who know a lot more about defconfig (esp for rx51) than I do - MohammadAG lcuk Jaffa come to mind17:53
fatphilThat defconfig will probably not work, as it's got GPIO configuration changes, and there's no way to change NOLO17:54
fatphilI wish Nokia had just released a completely open device :-(17:55
DocScrutinizergot GPIO changes?? o.O17:55
DocScrutinizerfatphil: we're all on same page wrt that :-D17:56
fatphilSee, IIRC, Jani Nikula vs. Dmitry Torokhov on the linux-input mailing list, regarding GPIO changes.17:57
fatphilThe single best thing about Nokia is that they have the "upstream first" policy regarding kernel development.17:57
fatphilSo if you want to know what nokia are working on, just read lkml, linux-omap, linux-input, linux-usb, linux-mtd etc...17:58
DocScrutinizerindeed - the more a pity a relatively sane driver (c) Nokia didn't make it upstream as there's already some botch occupying the field there - see lis302dl.ko (vs lis3lv02.ko)17:59
DocScrutinizerthat's however a general linux/FOSS problem it seems to me18:01
DocScrutinizernot Nokia's fault18:01
fatphilI disagree with upstream regarding the architecture for those hwmon drivers18:03
fatphilThey work on the policy that if the module's loaded it should be polling constantly.18:03
fatphilwhich doesn't work on an embedded system18:04
fatphilFelipe Balbi pretty much completely rewrote that module one weekend as it was a mess.18:04
fatphilHis patches are lost for ever :-(18:04
DocScrutinizerfatphil: my exact words on this topic :-D18:06
DocScrutinizerpolling on an embedded device that supports IRQ correctly is insane and really poor a concept18:07
DocScrutinizerwell, polling is *generally* a poor concept18:09
DocScrutinizerno matter which platform and which level of the API18:10
DocScrutinizeron embedded it's fatal though, while on servers you may get away with it, unnoticed18:10
DocScrutinizerand obviously there's a difference between a battery-state applet polling /sys/foo/bar once a minute, and a kernel module polling a chip @ 100/s18:11
DocScrutinizernevertheless the concept of polling is always last resort, if you're out of ideas how to implement things in a more smart way18:13
DocScrutinizeron a general rule if the hardware can do without polling, you better find a way for your software using that hw is doing without polling as well18:14
DocScrutinizerI always thought that'S plain obvious and best common sense18:15
DocScrutinizerseems I'm mistaken18:16
fatphilMost things were finally done right though. The kernel's power management team have done an astonishingly good job.18:21
*** fatphil2 has joined #maemo-ssu18:23
fatphil2Indeed, rx51_defconfig doesn't even build. Hmmm, revert or bisect?18:24
DocScrutinizerrx51_defconfig with upstream kernel? I'm a noob about those things, but I think that'S doomed to fail18:26
DocScrutinizerfatphil2: not that *I* would understand exactly what's it all about ;-) but maybe SCM has some hints about defconfig18:30
DocScrutinizere.g;a=commit;h=a78d411c6518eb7dd038eef6668f78a7e6c14ec4  find defconfig (bear with me, I admit I have no clue at all)18:38
fatphil2OK, that's something I can check out too.18:46
fatphil2note that one shouldn't do versioning of .config, as that's a target file. the defconfig is what you should be modifying and diffing.18:48
fatphil2At least my 1st hiccup is passed with a very simple patch. I'm sure there will be many more18:49
fatphil2I've not flashed a kernel onto the n900 for ages, I'm not sure I can remember how to do it!18:49
DocScrutinizerfatphil2: PaulFertser knows his way around for all I can tell, and I'm confident he did the right thing. If you find anything fishy, I'm happy to ask him about it18:50
fatphil2eww, that's definitely fishy.18:50
fatphil2does he hang around here?18:50
DocScrutinizerfatphil2: that's what he said as well, many times. This is based on powerkernel and that'S known to do some odd things regarding maintenance/git18:51
DocScrutinizerpaul was swearing quite a bit when trying to get stuff up and running the way he likes things to be18:52
fatphil2who's the actual kernel maintainer? does he have a public git tree?18:52
DocScrutinizerbut honestly, I have not even a faint clue about all that18:52
DocScrutinizerfatphil2: of which kernel?18:52
fatphil2good question18:53
DocScrutinizerthe h-e-n kernel got discontinued and patches are integrated to powerkernel18:53
fatphil2so powerkernel is what you're running with then?18:53
DocScrutinizerpowerkernel been orphaned for several months, I'm unsure who's maintaining it now - if anybody18:53
*** lardman is now known as lardman|gone18:54
fatphil2Does CSSU ship with with the powerkernel, or are the upgrades independent of each other?18:54
DocScrutinizerafaik powerkernel is maintained in a non-git way18:54
DocScrutinizerask MohammadAG18:55
DocScrutinizerI'm ignorant of those bits18:55
DocScrutinizerafaikt cssu is based on stock kernel18:55
fatphil2OK, good to know. I haven't done a dpkg -l to see what it contains.18:56
ThreeMruns also on powerkernel18:56
DocScrutinizeryeah, it shouldn't really care about kernel18:56
DocScrutinizercssu are mainly fixes to some hildon stuff etc18:57
DocScrutinizerplus of course apps18:57
DocScrutinizercore apps18:57
fatphil2Good. Too many dependencies are a bad thing18:57
DocScrutinizerthings like audio etc18:57
fatphil2Hahah - my next build failure is because of a patch that has my Acked-by!18:58
DocScrutinizeryou're ok with "how to flash kernel" now?18:58
DocScrutinizerthere are two ways: easy one via flasher-3.5 with some parameters, and a SSU way that's so tricky I never wrapped my head around it19:00
DocScrutinizeragain MohammadAG should know details19:01
fatphil2I used some internal tools. I no longer have them, but I can get them again.19:02
DocScrutinizerno need for19:02
DocScrutinizerflasher-3.5 works just great19:02
DocScrutinizerbut take care about modules ;-)19:02
fatphil2The tools were a wrapper around flasher19:03
DocScrutinizerI see19:03
fatphil2I'll not flash this, I'll boot from RAM with an initrd19:03
DocScrutinizerfair enough19:03
DocScrutinizerflasher can do this19:03
DocScrutinizernot exactly initrd but aiui mere loading of kernel to ram, rather than flashing it19:04
DocScrutinizer  might have some small hints about it19:05
DocScrutinizer>>  NOTE: Here we do not flash the kernel [or initrd to the device, but only load them to the device memory during the boot. Original images are restored when the device is restarted. ]19:07
DocScrutinizer-k is the key it seems19:08
* DocScrutinizer idly wonders if multiboot abomination could do this, instead of flashing kernel on each boot. Not that multiboot is needed at all anymore in times of properly working uBoot19:11
DocScrutinizermultiboot - such a brilliant nasty hack, abused by nitdroid in a manner that earned it flames for doing sth it probably never been meant to do19:13
fatphil2multiboot looks pretty cool,19:22
*** lardman has joined #maemo-ssu19:24
*** lardman has quit IRC19:24
*** lardman has joined #maemo-ssu19:24
DocScrutinizermultiboot is a cool hack19:24
DocScrutinizerbut it tends to mess up system for ignorant users19:24
DocScrutinizerobviously a concept of flashing kernel during each boot doesn't comply exactly with usual ways to update kernel19:25
DocScrutinizerthe worst thing about it: if it does, you're back to a rootfs+kernel reflash19:26
DocScrutinizerusual diagnose: kernel doesn't match modules -> bootloop -> you have to reflash19:27
DocScrutinizerrecover other than reflash is difficult at best, usually mere impossible19:28
DocScrutinizerstandard user quote: "I installed nitdroid, then I installed powerkernel, now device does bootloop. HLP PLZZZ!"19:29
DocScrutinizers/installed powerkernel/updated to new powerkernel/19:30
DocScrutinizerthere are an arbitrary number of variations to the theme ;-)19:31
DocScrutinizerone of the more popular ones: "I unistalled nitdroid, now device bootlops"19:33
*** fatphil2 has quit IRC19:33
DocScrutinizera funny one: "I uninstalled powerkernel, now doesn't boot anymore"19:34
DocScrutinizer(not multiboot related :-D )19:34
lardmanwhat does uninstalling powerkernel actually do though?19:35
lardmanpresumably it should put back the original kernel19:35
*** fatphil2 has joined #maemo-ssu19:37
DocScrutinizerit doesn't19:41
DocScrutinizerI'm not sure what it does, but I think uninstalling a kernel can't that easily trigger installation of another one19:41
DocScrutinizerprobably an apt thing, about the way uninstalling works19:42
*** psycho_oreos has quit IRC19:42
lardmanjust thinking that as installing it overwrites the existing one, it would be good to be able to uninstall it and get the original back19:43
lardmanor simply not be able to uninstall it19:43
lardmanI found the whole power kernel installation rather opaque - i was worried about which components did what, and I do pretty much know what I'm doing19:44
DocScrutinizerindeed, forbidding uninstall seems a sane thing to do19:44
lardmanor provide an original kernel package that can be installed and uninstall the power kernel stuff perhaps19:45
*** m0use has quit IRC19:45
DocScrutinizerthat's the recommended way19:45
lardmanoh I didn't realise there was one19:45
DocScrutinizerseems PK howto is telling exactly that - even the warning you have to click OK when installing PK says it iirc19:46
*** fatphil2 has quit IRC19:46
DocScrutinizerlardman: ponder that: pk pkg doesn't know what's been your "previous kernel", so there is no sane way to revert to the previous state by "uninstalling"19:48
*** fatphil2 has joined #maemo-ssu19:52
fatphil2looks like meego have a 2.6.35 kernel for n900. I wonder if that's a good starting point?19:52
fatphil2though 2.6.35 isn't long term support, is it?19:52
DocScrutinizerno idea. What I seem to know is meego-kernel is missing quite some maemo-specific bits, which are making for the smoother over-all experience on maemo, and the main issue are incompatibilities between the kernel APIs19:55
DocScrutinizeragain look at lis3lv02 in meego. I bet there are a couple more of that class19:56
MohammadAG2.6.37 in MeeGo iirc19:57
DocScrutinizerhey MohammadAG \o/19:58
MohammadAGheya DocScrutinizer19:58
DocScrutinizerMohammadAG: can you help out answering a few more of those questions to fatphil2 ?19:59
wmaroneSome of the IVI ARM things are using 2.6.35, so it is a valid kernel (just FYI)19:59
MohammadAGSure :)20:00
DocScrutinizerfine, as I'm partially utterly ignorant on that20:00
fatphil2dang, nothing I used to use works any more! I will need to visit the office and get access to my old PC...20:07
*** WellMux has quit IRC20:11
MohammadAGdefconfigs, right20:12
fatphil2A better way to look at the issue is to ask "is the current kernel broken", and not change it unless it is.20:12
MohammadAGMeeGo went through 2.6.32 to 38 afaik20:12
MohammadAGall of them booted on the N90020:12
MohammadAGso their defconfigs should work fine, for basic stuff20:12
fatphil2Note - that '.config' issue I highlighted earlier is a real one.20:13
MohammadAG(sorry, I was away, studies)20:13
fatphil2.config has been put under version control, which is a bad thing20:13
MohammadAGkernel-power got a git tree as of v47 afaik20:14
MohammadAGthe new maintainer does stuff The Right Way (TM), at least with sources20:14
fatphil2OK. That's good.20:16
DocScrutinizerthat's nice to hear20:16
DocScrutinizerMohammadAG: who's this new maintainer?20:16
MohammadAGoh, so .config wasn't cleaned from sources, damn20:16
MohammadAGDocScrutinizer, pali, he's had some CSSU patches20:16
MohammadAGI could link you, but I'm on the N900 in a 2G area, opening tmo won't be fun20:17
DocScrutinizerno need for that right now20:17
DocScrutinizerMohammadAG: all fine with your exams?20:18
fatphil2do you have a giturl for kernel-power?20:18
MohammadAGyep, I'm pretty much done for a week or two20:18
DocScrutinizerfatphil2: not me20:18
* MohammadAG attempts to flip to 3G20:18
DocScrutinizerMohammadAG: stay with us!!! :-P20:19
MohammadAG3/4 bars are good enough I guess20:20
*** VanpoofFluffy-ku has quit IRC20:20
DocScrutinizerlcuk: ping20:21
* lcuk looks suspiciously at DocScrutinizer 20:21
DocScrutinizerlcuk: your expertise would be appreciated, if you got no other pressing things to do20:22
MohammadAGgit clone
lcukexpertise in what?20:22
DocScrutinizerjust see if you feel like commenting :-)20:22
* lcuk is busy working20:22
DocScrutinizeraah, ok20:22
lcukhi fatphil2, spotted you earlier pottering about, did you get an image sorted for your machine?20:23
fatphil2lcuk: The only spare hardware I have lying around currently can't be flashed as it's not supported any more by flasher. I'll have to wait until I've finished some pm tests on another device, and use that.20:35
fatphil2so I have an image, I just can't flash it yet20:35
MohammadAGflash it on device20:36
MohammadAGunless you want to load it20:36
fatphil2It doesn't have a suitable bootloader20:36
fatphil2ewww. the patches in kernel-power aren't patches to the kernel, they're patches to debian's package.20:38
MohammadAGhe picked up the old source20:38
MohammadAGand that had patches in debian/patches (quilt)20:38
fatphil2wouldn't it have made sense to just apply the patches one by one and commit at each step?20:40
fatphil2could probably be scripted.20:40
* MohammadAG agrees, but that's how he picked it up I guess20:40
*** TurboVomit has joined #maemo-ssu20:41
*** wmarone has quit IRC20:42
*** wmarone has joined #maemo-ssu20:42
fatphil2Ug. Do people build in scratchbox?20:42
fatphil2my scratchbox is totally fucked20:43
DocScrutinizerisn't it that way from day one? ;-D20:44
fatphil2It did work on and off for several months20:44
MohammadAGwhat's the problem?20:44
fatphil2I'm configured for Harmattan20:44
fatphil2and at that, I'm misconfigured20:45
DocScrutinizereeew err20:45
DocScrutinizerrestore a backup ;-)20:45
fatphil2My g/f has a Fremantle scratchbox though.20:45
fatphil2Is there a way to just apply the patches, and not do the build?20:46
MohammadAGcheck debian/rules20:47
MohammadAGQUILT_PATCHES=debian/rules quilt push I think20:47
MohammadAGjust check debian/rules though, not that fluent in quilt20:47
fatphil2never used it. never needed to use it. bleh20:49
MohammadAGit's the debian way of shipping patches without changing upstream source20:50
DocScrutinizerit feels somewhat like "duh, I don't need that any second time"20:56
fatphil2If you're using a git repository, you're not changing upstream source20:56
*** paiburio has joined #maemo-ssu20:57
MohammadAGfatphil2, the old maintain3r never used git properly20:58
MohammadAGmaintainer*, damn hildon-input-method20:59
*** TurboVomit has quit IRC20:59
fatphil2Ah, it's not a kernel git repository anyway.21:01
fatphil2It seems to presume you have a kernel source tree somewhere else, apparently21:01
*** TurboVomit has joined #maemo-ssu21:02
*** TurboVomit has left #maemo-ssu21:02
DocScrutinizerfatphil2: that's what I got about it as well, yes21:06
DocScrutinizerfatphil2: and what PaulFertser tried to deal with and "fix" in h-e-n SCM21:07
DocScrutinizerfinaly fixing this abomination of powerkernel maintenance for good surely would be appreciated by a few people21:09
fatphil2The problem is that kenel devs (like me) just want to use git, and not worry about packaging. But distributions only care about packages. Either someone budges, or you stick a packaging layer in between the two.21:12
*** BCMM has joined #maemo-ssu21:27
*** Kaptenen has quit IRC21:35
*** m0use has joined #maemo-ssu21:38
lcukfatphil2, that is historically the way indeed21:42
lcukmake the changes you need, put them into git repository, advertise whereabouts and let distributions handle the rest would be ideal21:42
lcukis your proto a pretty one or ugly shell version?21:43
fatphillcuk - do you know what tag the quilt patches are based on?21:44
fatphilFor the released products I have the real thing.21:45
lcukfatphil, was just wondering with you mentioning builds and flasher versions etc21:45
lcukand no, I don't even know which kernel you are hacking on21:45
lardmanDocScrutinizer: sorry for the slow reply: re previous kernel - yeah, but it should at least the device in a workable state either by agreement for where to store old kernel or simply by assuming the vanilla version21:45
fatphilsame one that everyone posting to lkml with an address is working on21:46
DocScrutinizerlardman: only sane way I see is to disable uninstalling of kernels21:47
MohammadAGor to flash the stock kernel in prerm21:52
DocScrutinizerMohammadAG: you ever tested that?21:52
lardmanprevent uninstalling, but allow new ones to be installed21:52
lardmanthough then you end up with left over kernel modules21:53
lardmanreally something should depend on the package, something that can't be removed21:53
MohammadAGDocScrutinizer, no, but flashing is done in postinst so21:53
MohammadAGno reason for it not to work in prerm21:53
lardmananyway, deb packaging magic is a pita ;)21:53
* lardman heads back to the TV, bbl21:54
DocScrutinizerMohammadAG: I see problems with module dependency21:54
MohammadAGhow so?21:55
MohammadAGafaik modules load from /lib/modules/$(uname -r)21:55
DocScrutinizerthat'd be nice, if only we could rely on it, and also on proper matching /lib/modules/$(uname -r)/* being installed (just think "not enough storage")22:00
MohammadAGif exit code is 1, don't remove the kernel22:00
DocScrutinizerwhatever you do, you can't handle another package thus important and elementary as kernel/modules from a random prerm in any failsafe manner22:02
DocScrutinizerthat's my 2 cents22:03
*** wmarone has quit IRC22:04
*** wmarone_ has joined #maemo-ssu22:04
DocScrutinizerlrwxrwxrwx 1 root root 12 2009-10-30 12:32 /lib/modules/current -> 2.6.28-omap122:06
DocScrutinizeryou're sure about that $(uname -r) thing?22:07
*** wmarone has joined #maemo-ssu22:08
*** wmarone_ has quit IRC22:08
MohammadAGyes, at least on my system22:08
DocScrutinizera system might even have dependencies that forbid indiscriminately installing another kernel22:11
MohammadAGcan't be done, you'd flash directly with flasher22:12
DocScrutinizerit's not a sane way to install a new kernel when user decides to do sth he's not supposed to do - uninstalling a kernel22:12
DocScrutinizermere exiting with a "WTF?!" message when trying to uninstall a kernel should be all that's needed to deal with this22:13
MohammadAGI wonder what debian does22:15
*** chi_Wubbie_Baby_ has quit IRC22:16
*** M4rtinK has joined #maemo-ssu22:19
DocScrutinizerMohammadAG: deinstalled your server kernel?22:39
DocScrutinizerwait, doesn't apt have some --simulate-it flag?22:40
fatphildoesn't andre__ work with these things?22:52
andre__maybe andre instead of andre__ :)22:53
fatphilHaven't you dealt with packaging?22:54
andre__fatphil, no?22:55
andre__I like bugs though, for breakfast!22:55
fatphilI'm sure you've filed bugs that were to do with packaging22:55
andre__I want proof! ;-)22:55
andre__well, might be... when I'm drunk I do funny things sometimes, my friends say.22:56
fatphilYou've filed ones about dependencies, at least.22:57
fatphilwhich we closed as invalid!22:58
fatphilHowever, "forwarding from b.m.o" implies that you were just the messenger22:59
andre__fatphil, got one bug ID?23:00
fatphil206478 and 167483 are to do with dependencies23:03
DocScrutinizer2011-05-31 21:50:23] <keks-n> I've got a reboot loop   [2011-05-31 21:50:28] <keks-n> with kernel power [2011-05-31 21:50:48] <keks-n> It booted perfectly with multiboot before [2011-05-31 21:51:09] <keks-n> But I've removed multiboot and installed kernel-power and kernel-power-flasher23:04
* DocScrutinizer headdesks23:04
* DocScrutinizer headdesks a lil bit longer23:05
*** frostbyt1 has left #maemo-ssu23:07
andre__fatphil, well, that's internal Nokia stuff. I rather expected something maemo related.23:09
andre__I don't really care about what Nokia screws up internally.23:09
fatphilI've only ever seen your name in the context of nokia internal stuff.23:09
andre__fatphil, you can see my name a bit more often in and I hope.23:10
fatphilI've not had any reason to look there.23:11
*** paiburio has quit IRC23:11
andre__I see.23:13
ArkenoiWeird, now i cannot update anything via app manager. It says "Not enough space in application install memory". wtf is "application install memory"? I have plenty of space in all filesystems.23:14
fatphilSomeone needs to get aptitude working (interactive part, that is)23:16
ArkenoiIs there any known quick fix?23:16
fatphilcan you use apt-get instead?23:17
fatphil(or aptitude's command line)23:17
DocScrutinizerandre__: isn't this flattering you do a job so good that nobody needs to leave nb tracker ;-)23:18
andre__DocScrutinizer, errm... No. ;-)23:18
andre__DocScrutinizer, it's a pity that the video of my talk at MeeGo conf Dublin is not online.23:19
DocScrutinizerindeed, I'd be interested in it for sure23:19
* Arkenoi is going to try apt-get upgrade, hope nothing will break23:19
andre__but slide 7 describes the mindset of some internal Nokia people quite well23:20
fatphilInternal nokia people range from black to white, there's no way you can classify us!23:20
andre__fatphil, "some" I wrote.23:21
andre__but I certainly got a list of people I'd love to see fired immediately because of how much time they waste discussing crap with me.23:21
andre__instead of getting work done.23:21
DocScrutinizerfatphil: that's obvious and a usual mistake made on each side23:24
DocScrutinizerwhen I', swearing about Nokia I'm for sure not shooting @ all, but at a particular aspect as perceived "on the outside"23:25
DocScrutinizersame way when andre__ says "some" he clearly doesn't include *@nokia23:26
andre__I also have good experience with some Nokians. Mostly the subcontractors that understand open source though, instead of the code monkeys that just want to get paid for writing some hackish code that kind of works and try to close every bug report as quickly as possible to reduce their workload.23:27
fatphilI have much to say about nokia obviously, but ...  :-//23:28
DocScrutinizerlet's just stop this, as otherwise I'm prone to "act erratically" once more :-)23:28
lcukDocScrutinizer, that is my department23:29
lcukI even have a note on the wall about it.23:29
fatphilandre__ - on the idea of a 'fired immediately' list, I'd run out of paper ...23:29
fatphilThis IRC channel's archived, isn't it?23:29
fatphilDo I shouldn't say things like "there are enough guns and gun nuts in Finland, I can't believe no-one's taken a pot shot at Elop yet"23:30
fatphilThe last asehullu whackjob in Finland was a Nokia employee23:30
DocScrutinizerindeed it's probably no good idea to say that ;-)23:31
andre__it's archived, I think. however I say anyway what I want to say.23:31
fatphilI'm reminded of
*** wmarone_ has joined #maemo-ssu23:36
*** wmarone has quit IRC23:36
* DocScrutinizer has strange fantasies of cops wearing a t-shirt: if you say <120PT>PLEASE SOMEBODY KILL THE U.S. PRESIDENT</120PT> then I'll have to arrest you23:38
*** Kaptenen has joined #maemo-ssu23:46
*** paiburio has joined #maemo-ssu23:47

Generated by 2.13.0 by Marius Gedminas - find it at!