**** BEGIN LOGGING AT Mon Jun 11 02:59:59 2012 Jun 11 04:35:52 anyone still looking into a compass for the N900? Jun 11 05:56:03 ds3: honestly I'm afraid there are too many magnets in N900 Jun 11 05:56:51 DocScrutinizer05: hmmmm... I was looking at a little plugin module on the outside Jun 11 05:57:28 place a paperclip near the N900 and see for yourself Jun 11 05:57:56 I can get a magnetometer around it to measure it Jun 11 05:58:35 but I suspect if the chip has a large enough dynamic range, that can be calibrated out Jun 11 05:58:48 http://www.engadget.com/2012/06/10/us-navy-picks-linux-for-unmanned-vtol-drones/ Jun 11 05:59:00 for some of the magnets you can calibrate out the offset if the magnetometer chip has sufficient dynamic range Jun 11 05:59:24 yep Jun 11 05:59:34 k273: someone should hit them up for a repo Jun 11 05:59:42 there is a project out there about using the audio jack to interface sensors Jun 11 05:59:44 for some you can't, like kbd slide, probably vibrator motor, stand magnet Jun 11 05:59:54 so a tiny little nub sticking out of the audio jack might be one way Jun 11 06:00:06 hi doc|home, this is DocScrutinizer05 from home, right? Jun 11 06:00:09 or no? Jun 11 06:00:14 k273: nope, different one Jun 11 06:00:17 ok Jun 11 06:00:17 no Jun 11 06:00:20 ok Jun 11 06:00:45 DocScrutinizer05: let's say that can be made to work, is there anyone else interested in having a compass? Jun 11 06:01:01 I'd bet there are Jun 11 06:01:11 I am interested, but doubt will use it that often, Jun 11 06:01:13 ... Jun 11 06:01:33 a compass would let you have a fixed orientation Jun 11 06:01:47 sensor at AV plug has a hard time getting powered Jun 11 06:01:51 much rather do things Linux style instead of the Android hack Jun 11 06:02:05 there is a design out there that can get about 5mW Jun 11 06:02:12 5mW can run some mag sensors Jun 11 06:02:22 hmm, possible Jun 11 06:02:43 was reading a paper where someone had a full setup Jun 11 06:02:58 Left channel for power, Right channel for data in, Mic for data out (relative to the sensor) Jun 11 06:03:13 for a mag, pull power from left and right together and double the power Jun 11 06:04:56 I'd rather use USB Jun 11 06:05:12 WAAAAY cleaner design Jun 11 06:05:25 but USB is much more complex Jun 11 06:05:32 and you need the host mode hacks Jun 11 06:05:38 and you loose the ability to charge Jun 11 06:05:54 hmm Jun 11 06:06:15 to be really useful, the mag needs its ownaccelerometer Jun 11 06:06:17 1) no. 2) correct. 3) well, depends Jun 11 06:06:37 you got a 1 chip solution for SPI or I2C to USB? Jun 11 06:06:47 most 1 chip sol'n are for UART Jun 11 06:06:51 several, iirc Jun 11 06:06:58 Oh? who? Jun 11 06:07:26 who makes one I mean Jun 11 06:07:50 you can bitbang I2C via FTDI chip, and that's a solution I see while not yet awake Jun 11 06:08:20 Hmmmm Jun 11 06:08:47 still remember battery life issues with host mode though...has that been solved? Jun 11 06:09:32 battery life issues quite likely are system immanent: 200mA@5V Jun 11 06:09:48 if your plug doesn't use those, they won't drain battery Jun 11 06:10:17 right now with my original battery, I am barely getting 4-5hours on 3G :( Jun 11 06:10:27 USB controller on SoC though is also a bit of a energy hog Jun 11 06:10:50 maybe it is what the code did, I remember having rapid battery drain after just trying host mode. a reboot seems to fix that Jun 11 06:11:02 similar to what I saw on the N800 Jun 11 06:11:10 :shrug: 5h on 3G sounds about right Jun 11 06:11:17 really? Jun 11 06:11:30 I used to see almost 8h when it was brand new Jun 11 06:11:54 depend on data volume Jun 11 06:12:05 pattern Jun 11 06:12:16 dist to BTS Jun 11 06:12:18 this is idle just checking email, didn't think that changed much Jun 11 06:12:30 does 3G vs 3.5G make a difference? Jun 11 06:12:40 you never know for sure about inbound data Jun 11 06:12:44 no Jun 11 06:12:59 I see Jun 11 06:14:12 the other argument for the audio jack is, might be able to get some interest from non N900 folks Jun 11 06:14:25 sure thing Jun 11 06:14:37 think host mode is still rare Jun 11 06:14:44 yup Jun 11 06:15:03 and I donno of anyone selling solder on USB micro connectors Jun 11 06:15:17 solder on male ones that is Jun 11 06:15:18 most nifty design however was the magmeter-uSD Jun 11 06:15:51 well, I got two builds here Jun 11 06:16:02 builds of? Jun 11 06:16:11 one was sold to me though with comment "last one" Jun 11 06:16:17 USB plugs Jun 11 06:16:34 and that one solders onto a PCB? Jun 11 06:16:41 nope Jun 11 06:16:57 oh Jun 11 06:17:09 you'd need to come up with a non-standard design Jun 11 06:17:14 seems easy Jun 11 06:17:19 it'd involve gluing Jun 11 06:17:26 solder it Jun 11 06:17:29 still Jun 11 06:17:41 the ones I have seen are not amendable to soldering Jun 11 06:17:46 there is a sheild blocking it Jun 11 06:18:12 solder the shield, connect by wire Jun 11 06:18:20 Ohhhh like that Jun 11 06:18:59 interesting Jun 11 06:19:09 but that'd require a pass through to handle power/charging, right? Jun 11 06:19:47 hmm, unless you build in a barrel connector or 2nd USB female for charger... Jun 11 06:20:24 hmmm that could work Jun 11 06:20:37 the charging current would be a nasty H field to calibrate out though Jun 11 06:20:39 I see another argument for AV: it's more sturdy, at least on N900 Jun 11 06:21:20 the AV jack route looks idiot proof for prototyping... build it, check to see if there is power...then the rest is a MSP430 talking to a SPI or I2C bus Jun 11 06:21:22 (field) not if you run VBUS and GND in parallel and close to each other Jun 11 06:21:53 I donno if htat is enough, saw 'issues' on another design relating to it Jun 11 06:22:28 I'm rather sure the N900 itself will give you more (field) trouble Jun 11 06:23:01 more reason to figure a way of quicky rigging something up to test that Jun 11 06:23:23 has anyone attempted the uSD mag design yet? Jun 11 06:23:31 mini toy compass ? ;-D Jun 11 06:23:45 no, with real mag chips Jun 11 06:23:54 hmm, I think SpeedEvil looked into it Jun 11 06:24:06 got access to a few PhDs that can help with the calibration issues Jun 11 06:24:14 so that part I can get help on Jun 11 06:24:22 gtggys Jun 11 06:24:24 cya Jun 11 06:24:57 as long as the fields are static and don't saturate the sensor, it's no real issue Jun 11 06:25:17 even opening a keyboard can be handled Jun 11 06:25:20 I have concerns against both preconditions Jun 11 06:25:33 it is 'static' in a way Jun 11 06:25:35 :nod: Jun 11 06:25:51 I doubt the magnets on there are more then 1-2G if even that Jun 11 06:26:53 there are common mags out there that can handle 8-16G's of dynamic range Jun 11 06:26:55 vib motor might be a bitch Jun 11 06:27:15 true...where is it physically? Jun 11 06:27:21 luckily vib is next to power button Jun 11 06:27:41 so we got stuff to the 3rd power working for us there Jun 11 06:28:58 what kind of update rate on the mag is reasonable? Jun 11 06:29:05 10Hz? 50Hz? 100Hz? Jun 11 06:29:08 :shrug: Jun 11 06:29:25 depends I'd say Jun 11 06:29:43 usually 10Hz should be more than enough, no? Jun 11 06:29:58 I have seen devices with updates as slow as 3Hz Jun 11 06:30:04 so I just donno Jun 11 06:30:41 if you wanna use it for Wii-like controller, you want 50..100 Jun 11 06:31:15 hmmm Jun 11 06:31:27 the problem is the faster it goes, the more power is needed Jun 11 06:31:38 obviously Jun 11 06:31:54 each measurement is about 12bytes Jun 11 06:32:05 the paper for the AV interface suggests a 300bps interface Jun 11 06:32:10 I know those from N9 Jun 11 06:32:30 LOL WUT? Jun 11 06:33:01 it is doing a modem with a MSP430 Jun 11 06:33:07 and running on like 5mW Jun 11 06:33:09 OUCH! Jun 11 06:33:50 I'd run RLL or MFM encoding straight away, via those audio lines Jun 11 06:34:05 should at least be worth 9.6k Jun 11 06:34:12 probably 19k Jun 11 06:34:21 maybe more Jun 11 06:34:24 hmmmm Jun 11 06:34:40 donno if enough CPUpower is available Jun 11 06:34:51 definitely Jun 11 06:35:09 unless you run that msp430 at 32k clock Jun 11 06:35:33 how are you thinking of generating the MFM signals? Jun 11 06:35:45 algorithm Jun 11 06:35:46 through the ADC block or? Jun 11 06:35:59 on N900 side? sure Jun 11 06:36:07 no, on the MSP430 side Jun 11 06:36:14 the N900 is pretty easy...lots of CPU there Jun 11 06:36:16 nah, plain GPIO Jun 11 06:36:38 got a library? signal processing is not my strong suit Jun 11 06:36:47 I can get the HW going and do ports but..... Jun 11 06:37:04 I am still looking for a simple way of doing FSK decoding :D Jun 11 06:37:26 FSK decoding way more complex than straight RLL decoding Jun 11 06:38:09 doesn't direct RLL require DC coupling? Jun 11 06:38:25 nope Jun 11 06:38:31 that's the point of RLL Jun 11 06:38:34 wiki Run Length Limited Jun 11 06:38:43 ~wiki Run Length Limited Jun 11 06:38:47 At http://en.wikipedia.org/wiki/Run_Length_Limited (URL), Wikipedia explains: "'Run length limited' or 'RLL' coding is a line coding technique that is used to send arbitrary data over a communications channel with bandwidth limits. This is used in both telecommunication and storage systems which move a medium past a fixed recording head. Specifically, RLL bounds the length of stretches (runs) of repeated bits during which the signal does not change. If ... Jun 11 06:39:09 the last sentence says it all Jun 11 06:39:10 I looked at that Jun 11 06:39:34 and I don't see a lower bound on what frequency the channel must pass Jun 11 06:39:50 the usage of the term 'DC' is a mistake on my part Jun 11 06:39:53 the max run length Jun 11 06:40:16 thinking of something like audio systems might high pass some at some frequency Jun 11 06:40:52 Oh you mean use the max lenght of the 0's or 1's as a limiter on the min. freq? Jun 11 06:40:54 I think usually max RL is ~5 bits Jun 11 06:41:05 yep Jun 11 06:41:14 hmm seems complicated Jun 11 06:41:21 ??? Jun 11 06:41:27 wouldn't know how to pick that out of a stream sampled by ALSA Jun 11 06:41:36 you don't use any limiter Jun 11 06:42:06 you get a square wave shaped signal and you detect edges Jun 11 06:42:20 simple as that Jun 11 06:42:37 yes but I don't see a square wave. I see a sampled version of that through ALSA. one that might be filtered and shaped by the audio circuit Jun 11 06:42:53 *sigh* Jun 11 06:43:03 this is driving the mic signal Jun 11 06:43:11 or do you have another way for the N900 to see that? Jun 11 06:43:38 tmo down with Database error? Jun 11 06:43:44 no, you'll have to go the hard way and inspect your .wav for FF and 00 values Jun 11 06:43:59 or rather >a0 and <30 Jun 11 06:44:07 OHhhhhhh Jun 11 06:44:29 Hmmm Jun 11 06:45:04 or, for s16le, for positive and negative Jun 11 06:45:19 well... if I can get it work w/the stock 300bps modem, I'll pester you on how to get RLL going :D Jun 11 06:45:27 it is all sw at that point Jun 11 06:45:39 modem is WAAAY harder Jun 11 06:45:52 yes, but the folks who did it wrote the code Jun 11 06:46:01 :shrug: Jun 11 06:46:08 I know how to port, I am not an expert at signal processing Jun 11 06:46:23 see this app called, err.... magreader Jun 11 06:46:33 all you need already in place ;-P Jun 11 06:46:38 incl AV Jun 11 06:46:43 oh okay Jun 11 06:47:44 you will be capable of creating a signal at least as stable and exact in timing and shape as that produced by a magnet head when you pull a credit card - will you? Jun 11 06:48:00 yes Jun 11 06:48:08 is a CC strip considered to be MFM or RLL? Jun 11 06:48:27 technicall I guess it's lilke MFM, roughly Jun 11 06:48:35 I've done mag strip stuff so I understand that :D Jun 11 06:49:10 MFM and RLL are to me hard drive encodings Jun 11 06:49:26 sure, they were mainly used for that Jun 11 06:49:48 but a HDD is also a mag medium pulled under a magsensor head Jun 11 06:49:50 ;-) Jun 11 06:50:09 good point... lost sight of those ancient drives Jun 11 06:50:13 and that's why they used MFM or RLL Jun 11 06:50:32 as a coil in magnet head needs flux change Jun 11 06:50:40 to create signal Jun 11 06:50:49 imanently DC averse Jun 11 06:50:55 yes. Jun 11 06:51:09 i remember that from mag strips... write HC11 code to decode them before Jun 11 06:56:28 2% of net income! Plus acknowledgements ;-P Jun 11 07:04:55 :) Jun 11 09:08:19 How can I read out a wpa key my n900 device has saved? Jun 11 09:13:15 stored in gconf Jun 11 09:13:38 I posted a mini-howto to tmo long ago Jun 11 09:13:58 tmo? Jun 11 09:14:06 If you mean talk.maemo.org, that's apparently down. Jun 11 09:14:15 hmmm Jun 11 09:15:04 But I guess I can do it with 'gconftool' ? Jun 11 09:15:28 o/ Jun 11 09:17:26 gconftool -R /system/osso/connectivity/IAP Jun 11 09:19:15 gconftool -R /|grep EAP_wpa_preshared_passphrase Jun 11 09:20:13 phryk: ^^ Jun 11 09:20:21 thanks Jun 11 09:20:37 o/ chem|st Jun 11 09:20:56 moo doc Jun 11 09:21:15 freemangordon: you settled your CSS issue? Jun 11 09:21:20 tmo down? Ha! Jun 11 09:21:40 database issues afaik Jun 11 09:22:04 Ah gimme a sec Jun 11 09:26:42 Ah that works, thanks a bunch. Jun 11 09:36:59 DocScrutinizer51: yeah server is back up but the db is still borked Jun 11 09:44:33 TMO is down? Jun 11 09:45:45 borked db Jun 11 09:45:46 yay Jun 11 10:17:47 Does reggie know? Jun 11 10:20:29 maybe it had this faulty mysql and someone stole our password hashes! Jun 11 10:42:36 ds3: calibration is near-trivial Jun 11 10:42:43 ds3: problem is thickness Jun 11 10:43:36 You're basically looking at having to wirebond bare dies, or perhaps get away with sanding down existing magnetometer chips Jun 11 10:43:47 Tmo back!! Jun 11 10:49:50 TMO back down! Jun 11 10:53:09 petteri: please elaborate. What's up with mysql? Jun 11 10:53:25 https://community.rapid7.com/community/metasploit/blog/2012/06/11/cve-2012-2122-a-tragically-comedic-security-flaw-in-mysql Jun 11 10:56:29 o.O Jun 11 10:57:14 DocScrutinizer51: http://www.heise.de/newsticker/meldung/MySQL-Datenbank-Zugang-auch-ohne-Passwort-1614987.html Jun 11 11:03:48 chem|st: petteri: many thanks Jun 11 11:19:45 it's not mysql problem Jun 11 11:19:53 it's libc and inconsistent documentation problem Jun 11 11:20:04 some man pages say it returns a char Jun 11 11:20:08 some say it returns int Jun 11 11:20:19 and in reality it's absolutely random Jun 11 11:20:31 and for example on my system it returns -1,0,1 and only those 3 Jun 11 11:20:39 it's not returning difference between those string Jun 11 11:21:05 OS community did really good job writing confusing documentation Jun 11 11:30:24 jacekowski: man memcpy, why would anybody care for the actual value, instead of just checking ==0 ? Jun 11 11:31:13 DocScrutinizer05: they do Jun 11 11:31:28 DocScrutinizer05: except they check for (char)memcpy(...)==0 Jun 11 11:31:31 so they do a cast to int8? Jun 11 11:31:42 dafaq Jun 11 11:31:46 but according to man it's supposed to return char Jun 11 11:32:31 on ubuntu it says int Jun 11 11:32:45 int memcmp(const void *s1, const void *s2, size_t n); Jun 11 11:33:10 yeah, but read the description Jun 11 11:33:10 s/memcpy/memcmp/. Jun 11 11:33:16 it says it's comparing bytes Jun 11 11:33:22 and it's returning byte compare value Jun 11 11:33:27 it *compares* bytes Jun 11 11:33:36 which is friggin irrelevant Jun 11 11:33:50 except for the way n is counting Jun 11 11:34:08 and for example Jun 11 11:34:09 http://www.manpagez.com/man/3/memcmp/ Jun 11 11:34:40 max os x? Jun 11 11:34:43 mac :) Jun 11 11:34:52 BSD Jun 11 11:35:08 and you have inconsistent behaviour across different linux distros Jun 11 11:35:18 jacekowski: still returns int Jun 11 11:35:27 Slackware manpage The memcmp() function returns an integer less than, equal to, or Jun 11 11:35:27 greater than zero if the first n bytes of s1 is found, respectively, to Jun 11 11:35:27 be less than, to match, or be greater than the first n bytes of s2 Jun 11 11:35:32 read the description Jun 11 11:35:33 and says nothing about the actual range Jun 11 11:36:21 fact is the function still is defined as int memcmp(const void *s1, const void *s2, size_t n); Jun 11 11:36:30 It's clearly possible to assume that it's the result of subtracting each char in each input string, and returning the result if diffefrent Jun 11 11:36:44 but that isn't specified Jun 11 11:36:47 doesn't matter Jun 11 11:36:47 jacekowski: maybe you are just one bitter mysql dev :P Jun 11 11:37:09 nah, you just hit some random behaviour that makes stuff go wobbly Jun 11 11:37:18 i've tried to repeat it Jun 11 11:37:34 and in 9 cases out of 10 all i was getting was -1 or 0 or 1 Jun 11 11:37:48 and then like every 10th try i was getting something different Jun 11 11:37:55 doesn't matter Jun 11 11:38:23 http://bazaar.launchpad.net/~mysql/mysql-server/5.1/revision/3560.10.17 Jun 11 11:38:50 ( DocScrutinizer05: except they check for (char)memcpy(...)==0) only a fool would cast a function defined as int to a char Jun 11 11:39:02 they're not really using the value to compute something it seems, its the return value of the function Jun 11 11:40:12 SHITE! again this 40s delay on clicking a URL Jun 11 11:41:45 yeah, they return an int in a function defined as char Jun 11 11:41:52 I guess Jun 11 11:42:13 yes, seems my_bool is really a char Jun 11 11:43:11 well, returning random shit to a bool (or char) is clearly something even lint would've noticed Jun 11 11:43:30 it's my_bool Jun 11 11:43:38 not a standard type Jun 11 11:43:42 if it were a real bool it would've worked Jun 11 11:43:55 but I don't really know what happens here Jun 11 11:44:19 you should Jun 11 11:44:40 you should take each opportunity to learn Jun 11 11:45:26 all hail to strict typing Jun 11 11:45:41 I was already looking but I'm getting a lot of result for people trying to convert number-to-string Jun 11 11:46:52 the nasty part is: all vulnerable MySQL servers may define all new passwords now Jun 11 11:47:29 http://stackoverflow.com/questions/5881895/truncating-an-int-to-char-is-it-defined Jun 11 11:47:36 seems to explain it Jun 11 11:48:20 friggin c-coders never doing proper strict typeing Jun 11 11:48:51 you simply don't do implicit typecasts - problem soved Jun 11 11:50:28 If everyone did proper strict typing, we would not have crashing machines anymore. Jun 11 11:50:47 We would have machines which always worked and never failed. Jun 11 11:51:25 Unfortunately, it seems that although everyone wants working systems, nobody builds them. Jun 11 11:53:55 I think the problem here is treating my_bool as the builtin bool type though (different rules apply) Jun 11 11:55:09 there is no bool in C Jun 11 11:55:19 In new C there is. Jun 11 11:55:29 nothing in C99 Jun 11 11:56:23 stdbool.h is in C99, AFAIK. Jun 11 11:56:54 Yes, and _Bool even without the header. Jun 11 11:57:32 (C99 6.3.1.2 "Boolean type".) Jun 11 11:59:57 There is no bool in C? Jun 11 12:00:00 ORLY? Jun 11 12:00:56 There is no spoon in C. Jun 11 12:02:10 But perhaps in C34 you can #include and get spoon, knife and fork types. (With spork as a GCC extension.) Jun 11 12:03:36 Beware! http://xkcd.com/419/ Jun 11 12:06:56 SpeedEvil: aaah, one of my very favourite XKCDs. Jun 11 12:09:42 C isn't even a formal language. Jun 11 12:10:11 I.e. you can say some things in C, but you have no idea whether the machine will respond to it. Jun 11 12:10:51 And the much touted 'but then you can get more performance' is just an argument which nobody has ever backed up. Jun 11 12:11:27 The very notion of accepting input without semantics is an extremely bad idea. Jun 11 12:11:38 in other languages it's "you have no idea whether the implementation will conform to spec" ;-) Jun 11 12:12:25 ShadowJK: well, not the languages I am interested in and/or using. Jun 11 12:13:06 Well, I also use sh/perl/python at the edge of certain systems. Jun 11 12:13:26 Still, the supposed fact that other also do it wrong, is no reason to also do it wrong. Jun 11 12:13:52 C was just the simplest thing that could possibly work. Jun 11 12:14:11 Where simple is defined as 'given the bias of the designers'. Jun 11 12:14:59 One fun thing about java is the java memory model is like only implemented on one obscure machine, and basically acts like C everywhere else.. But the java model would give you much performance boost in big SMP systems... I doubt any already written code would actually work though Jun 11 12:23:53 The architecture we use in computers also has nothing to do with the universe. Jun 11 12:24:32 99.99% of programming languages instead assume this architecture. Jun 11 12:25:02 Programming language cores could better be designed by physicists. Jun 11 12:25:27 ShadowJK: And quite a lot of smartcards. Jun 11 12:26:19 Way too many people don't understand underlying hardware. Jun 11 12:26:26 Well the language defines the architecture, but the real life architecture is nicer, so java unintentionally ends up behaving in a for the programmer nicer way Jun 11 12:26:38 - that random access speeds have barely changed for RAM since 1985. Jun 11 12:27:24 (They've dropped ~10fold, on the back of a ~millionfold expansion in size) Jun 11 12:28:12 I am talking about something like cellular automata as a base language. Jun 11 12:28:34 Synchronisation is a hard problem. Jun 11 12:28:42 cockless design isn't fun. Jun 11 12:28:44 clock Jun 11 12:28:51 Who cares about fun? Jun 11 12:29:09 It's our task to create optimal devices. Jun 11 12:29:11 Fun = tractable. Jun 11 12:29:22 Chuck Moore created some. Jun 11 12:29:38 Still, that's not radical enough. Jun 11 15:34:13 does N900 have any kind of un-resettable lock? Jun 11 15:34:21 if so, how do I verify I have the password for it? Jun 11 15:49:55 luke-jr: while unlocked you can access the storage of the lock-code which can be cracked in seconds Jun 11 17:17:20 <_freemangordon> In the meanwhile: http://talk.maemo.org/showthread.php?t=84829&goto=newpost Jun 11 17:26:24 freemangordon: cool stuff! is this something that will eventually make its way into stable cssu? Jun 11 17:27:00 dafox: it depends on the results from the testing stage Jun 11 17:28:16 obviously :p But this is really awesome. But it would make the cssu quite big, no? Jun 11 17:28:44 if it turns out to be stable, if there are no conflicts, etc, etc, then we could start thinking. It is too early now. After all it might be that I am wring and stable thumb2 is impossible on n900 Jun 11 17:28:54 dafox: no, it will make it smaller :P Jun 11 17:29:18 I mean, I assume ideally every [stock] application that can be recompiled would be 'updated' to this, so many more applications to install? Jun 11 17:29:48 dafox: it is not for the applications, but for the system stuff, ie bins in /rootfs Jun 11 17:31:02 It is up to application developers to decide whether to thumb-compile their SW ( reducing the user base to CSSU-devel users only) or to put the stuff in maemo-extras. On the other hand there is COBS comming, so who knows ;) Jun 11 17:31:03 rootfs is the core system right? but wouldn't you want to recompile everything that runs? Jun 11 17:31:12 yep Jun 11 17:31:20 but it takes time ;) Jun 11 17:31:28 ofcourse Jun 11 17:31:35 I hope it works, you have my vote :) Jun 11 17:31:49 then try it if you feel brave :P Jun 11 17:32:56 unfortunately I'm not that brave :p I use my n900 as my 'real' phone Jun 11 17:35:05 dafox: so am i :) Jun 11 17:35:16 but it is up to you of course Jun 11 18:02:50 dafox: CSSU never will force a new kernel on users, since we can't force commercial providers of kernel modules to support that kernel Jun 11 18:04:41 the idea of CSSU is *not* to be the better new OS version, CSSU is all about keeping *original* maemo alive Jun 11 18:25:41 why are 'commercial providers of kernel modules' needed? can you please give me an example? Jun 11 19:09:11 hi is there a way to run whatsapp or wazapp on N900? Jun 11 19:23:01 freemangordon: HOLY SHIT. Jun 11 19:23:48 Is there any other dogma you two want to smash? Jun 11 19:31:12 hi there, is there anything wrong with maemo.org? I am absolutely unable to register an account Jun 11 19:32:38 mira_, you should file a bug. Jun 11 19:32:40 * GeneralAntilles cackles. Jun 11 19:33:12 :P Jun 11 19:37:45 mira_: do you mean the forum? Jun 11 19:37:57 maemo.org Jun 11 19:38:03 mira_: do you mean the forum? Jun 11 19:38:06 yes Jun 11 19:38:25 talk.maemo.org is a different account from maemo.org Jun 11 19:38:26 I was told to wait an hour Jun 11 19:39:03 did you register for talk.maemo.org or maemo.org? Jun 11 19:39:11 maemo.org Jun 11 19:39:21 and you want to use talk.maemo.org? Jun 11 19:39:28 (the forum)? Jun 11 19:40:13