**** BEGIN LOGGING AT Sun Dec 27 02:59:59 2015 Dec 27 06:34:28 hi Dec 27 07:50:44 plz save me Dec 27 07:50:53 Sorry, it seems that you are using an IP address or a proxy that is listed in the forum anti spam blacklist. Dec 27 07:50:56 Feel free to contact our staff on irc freenode #maemo channel. Dec 27 07:51:05 188.165.30.151 Dec 27 07:51:12 ~unbanip Dec 27 07:51:12 i guess unbanip is please contact techstaff maemo org with your request, or see ~techstaff Dec 27 07:53:16 okay emailed, thx! Dec 27 08:24:45 pali, why do you want to rely on .n.c/m$ anyway? Dec 27 08:25:15 imo there should be cssu release as a flashable firmware Dec 27 08:25:26 call it 1.4-cssu1 or something Dec 27 08:25:45 or just 1.4 Dec 27 08:25:46 KotCzarny: we cannot redistribute nokia bins via r.m.o Dec 27 08:25:56 only unofficial Dec 27 08:26:03 can't you do it unofficially? Dec 27 08:26:11 how? Dec 27 08:26:25 I want to have one "official" location Dec 27 08:26:29 for all maemo stuff Dec 27 08:26:35 and it is r.m.o domain Dec 27 08:26:45 make the image, put on the internet, google will direct people easily Dec 27 08:26:46 it is "trusted" by all n900 users Dec 27 08:26:54 as most just want 'latest' Dec 27 08:26:55 hi all :) Dec 27 08:27:59 KotCzarny: ask merlin if he can host such firmware on site where is cssu-devel Dec 27 08:28:27 and once we will have location where to host, I could prepare "firmware" file Dec 27 08:28:50 hmm, cssu-thumb might be nice candidate because 'its incompatible with regular stuff' Dec 27 08:29:09 in cssu-devel is prepared new kernel Dec 27 08:29:14 and 'its untested/unstable' Dec 27 08:29:18 with thumb support Dec 27 08:29:25 pali, i know, but if anyone asks why ;) Dec 27 08:29:30 and after testing I would suggest to mark it as stable Dec 27 08:29:52 vahe: hi you Dec 27 08:30:05 because kernel contains also fixes Dec 27 09:15:28 * jonwil wonders what the minimum set of files required to get a N900 to boot is and how many of those remain closed source' Dec 27 09:15:54 probably more than I thought (e.g. ICD is necessary) Dec 27 09:16:05 boot is one thing, boot maemo is another Dec 27 09:16:07 to boot what? Dec 27 09:16:08 yea Dec 27 09:16:14 for minimal set see rescueos Dec 27 09:16:26 NOLO is closed i think Dec 27 09:16:38 yeah it is closed unfortunately :( Dec 27 09:17:43 so Dec 27 09:17:56 that's it really Dec 27 09:18:11 and we can't run uboot directly because nolo does some hardware initialization that we're not aware of Dec 27 09:21:24 well if we can clone MCE and camera-ui and the various other packages, I am sure we could figure out the important bits of what nolo is doing. If I knew how to get nolo into IDA etc I would take a look myself. Dec 27 09:22:37 nolo is arm binary without any particular format, executed directly by cpu i think Dec 27 09:23:07 well Dec 27 09:23:19 ROMBL probably does stuff Dec 27 09:23:19 yeah but you need to know where in memory it sits and what's code and what's data and etc which isn't easy Dec 27 09:23:48 if you can get uboot into ida, then nolo is probably few first kbytes from the nand Dec 27 09:24:23 and i guess you will need to know asm well enough Dec 27 09:25:05 HexRays will help a bit with some of the work if you can get everything lined up properly Dec 27 09:27:02 i wonder if qemu would be able to boot from exact nand dump Dec 27 09:27:46 probably not since you dont have rombl Dec 27 09:27:49 and whatever rombl does Dec 27 09:28:08 http://wiki.maemo.org/N900_Hardware_Hacking/serial_dump Dec 27 09:29:06 Über-cool backlight fade-in took 9 ms Dec 27 09:29:08 hehe Dec 27 09:30:19 KotCzarny: qemu implement only GP device (general purpose), not HS (high secure) Dec 27 09:30:47 so qemu ROM cannot parse special header in x-loader where are certificates and RSA keys Dec 27 09:30:49 pali, did someone actually tried installing generic boot loader? Dec 27 09:31:13 KotCzarny: nokia got us x-loader & nolo for GP devices Dec 27 09:31:18 and this is used in qemu! Dec 27 09:31:40 Pali: pls implement nolo in uboot :3 Dec 27 09:31:54 kerio: do it yourself :-) Dec 27 09:32:07 nu :c Dec 27 09:36:57 jonwil: also here: http://wiki.maemo.org/Firmware_hacking Dec 27 09:39:13 DocScrutinizer05: if you actually do it, please use a CNAME to maemo.muarf.org or whatever *.maemo.muarf.org subdomain, not just "muarf.org" :) Dec 27 09:40:54 the muarf.org home page lists a SHA1 sshfp for a DSS key Dec 27 09:41:10 do two wrongs make a right? Dec 27 09:41:14 or is it just a double wrong Dec 27 09:41:20 hmmm? Dec 27 09:41:39 ah you mean sha1 is broken and dss should no longer be used? :D Dec 27 09:41:41 it's also got a CAcert https certificate :\ Dec 27 09:41:42 muarf.org uses an invalid security certificate. Dec 27 09:41:45 The certificate is not trusted because it was signed using a signature algorithm that was disabled because that algorithm is not secure. Dec 27 09:41:51 I haven't updated it for a long time Dec 27 09:41:54 oh god, a sha1 cert Dec 27 09:42:23 bencoh: letsencrypt? :3 Dec 27 09:42:31 kerio: no fscking way Dec 27 09:42:35 3: Dec 27 09:42:40 not for now at least Dec 27 09:42:44 y not Dec 27 09:43:04 their "automagic" stuff is .... awful Dec 27 09:43:12 so don't use that client Dec 27 09:43:16 there's plenty of other ACME clients Dec 27 09:43:35 yeah sure, like I want to let a cron update a file on my server so that they can fetch it - wtf is wrong with them? Dec 27 09:44:04 1. I dont like the idea 2. it's http-centric Dec 27 09:44:12 3. I dont like automagic stuff Dec 27 09:44:25 4. I dont need my cert to change every 3 months Dec 27 09:47:07 wut Dec 27 09:47:48 bencoh, still, you should update that cert to something more secure, otherwise there is no point in https/ssh Dec 27 09:48:24 * jonwil is bored Dec 27 09:48:25 i mean, dns verification is coming soon, afaik Dec 27 09:48:28 * jonwil has run out of N900 stuff to do Dec 27 09:51:44 Signature Algorithm: sha256WithRSAEncryption Dec 27 09:51:49 according to openssl (?) Dec 27 09:52:11 using non-own CA is unsecure Dec 27 09:52:12 and short-lived certs are a workaround for the unreliable revocation of certs in the TLS PKI Dec 27 09:52:49 just own certificates could be secure by definition Dec 27 09:52:58 secure according to what Dec 27 09:52:59 Pali: :) Dec 27 09:53:07 kerio: according to "who can I trust" Dec 27 09:53:13 secure according from client to server connection Dec 27 09:53:19 trust no 1 Dec 27 09:53:43 no authentication Dec 27 09:53:53 if I'm connecting from my mobile to server XYZ, then of course verisign or other CA has nothing to do with connection Dec 27 09:54:22 it is peer-to-peer connection Dec 27 09:55:01 yes, but there's no reasonable way to get your server to be trusted by a majority of clients without using a PKI Dec 27 10:00:53 Lets see what www.ssllabs.com has to say about maemo.muarf.org :P Dec 27 10:04:38 it points to a bunch of dangerously weak cyper suites and gives it a fail Dec 27 10:05:57 jonwil: and now you have to wonder what maemo actually supports, with an aging openssl Dec 27 10:06:16 *aging NSS Dec 27 10:06:18 but yea Dec 27 10:06:38 depends whether you are talking SSL for the browser (which is NSS) or SSL for other things like apt-transport-https (which probably uses openssl) Dec 27 10:07:04 or wget,curl,ssh :) Dec 27 10:07:14 (the latter is not so relevant for http) Dec 27 10:07:33 anyway, openssl 0.9.8 supports a vaguely decent cipher in DHE-RSA-AES256 Dec 27 10:07:51 (or aes128 i guess) Dec 27 10:08:12 SSH is in extras so its age is less of a factor (its not installed by default and someone could update it to something modern without breaking things) Dec 27 10:08:24 i updated mine to something modern :3 Dec 27 10:08:48 i like me some ed25519 Dec 27 10:09:10 yeah apt-transport-https does use openssl Dec 27 10:09:35 we really ought to start using openssl 1.0.2 for extras Dec 27 10:10:02 we should ship latest openssl in community-ssu and recompile everything that uses it that is FOSS (or has been cloned) against that version too Dec 27 10:10:29 let's move to libressl instead :> Dec 27 10:11:46 even better :) Dec 27 10:12:02 i have to run openssl for my tls terminator :< Dec 27 10:12:28 libressl doesn't have SSL_CTX_add_server_custom_ext Dec 27 10:12:49 file a bug report Dec 27 10:12:56 it's already filed Dec 27 10:13:19 make a pull request Dec 27 10:13:34 kerio: would be good to push out open/libressl and then a recent ssh Dec 27 10:13:59 jonwil: +1000 Dec 27 10:16:27 12:10 < kerio> let's move to libressl instead :> Dec 27 10:16:36 and rewrite everything? yeah, sure, good luck :p Dec 27 10:16:55 it's very api-compatible Dec 27 10:17:05 I am going to go through my on-disk copy of the stock N900 root filesystem, identify everything that links to openssl and make a post about it Dec 27 10:17:05 how very is very? Dec 27 10:17:11 documenting all the binaries that link to openssl Dec 27 10:17:19 whether they are open or closed Dec 27 10:17:21 what package they are in Dec 27 10:17:23 KotCzarny: you could be missing some include or something like that Dec 27 10:17:35 and if they are closed, what openssl functions they call Dec 27 10:17:46 kerio: is there any distro that uses libressl by default? Dec 27 10:17:50 openbsd Dec 27 10:18:01 and for linux world? Dec 27 10:18:07 linux sucks Dec 27 10:18:10 yeah but no sane people use openbsd ;) Dec 27 10:18:39 having linux distro patches would ease eventual fixing Dec 27 10:18:39 os x's openssh is statically linked against libressl Dec 27 10:18:49 KotCzarny: oh there's a ton of those actually Dec 27 10:18:53 bencoh: libressl works quite well Dec 27 10:19:02 openbsd ports and the like Dec 27 10:19:03 bencoh: gentoo has it for most major packages too Dec 27 10:19:20 freebsd ports, too! Dec 27 10:19:35 KotCzarny: perhaps alpinelinux uses it, otherwise perhaps voidlinux Dec 27 10:19:43 KotCzarny: and for gentoo you can just set USE="libressl" Dec 27 10:22:21 We dont need to replace every single binary that calls openssl in order to improve security (for example its unlikely to matter so much if something like the ActiveSync/Exchange support is still using weak old OpenSSL 0.9.8 Dec 27 10:22:59 it does for people who actually use that Dec 27 10:23:09 that and you can have both openssl 0.9.8 and libressl installed Dec 27 10:23:27 (or openssl 0.9.8 and openssl 1.0.2) Dec 27 10:26:11 It also depends on exactly which calls in openssl each closed-source binary is making Dec 27 10:26:40 If its only calling functions to do, say, SHA1 hashing and isn't using any of the actual SSL logic, it probably isn't a security problem. Dec 27 10:35:32 It would be great if it could all be switched over to a new libre/openssl though Dec 27 10:35:50 If then mainline kernel is managed somehow, you'd have one even more epic system :) Dec 27 10:37:59 then all we would need is thinning/removing apps Dec 27 10:38:16 having 128M of ram free would make it usable Dec 27 10:38:36 jonwil: what were those 2 functions that were waiting to be REed in devicelock? Dec 27 10:38:57 and could I have your IDA DB, to not reinvent the wheel Dec 27 10:39:33 its libcodelockui that needs work done Dec 27 10:39:54 ok, whatever it is, just hand me the DB :) Dec 27 10:40:43 I believe I gave you my latest IDB before Dec 27 10:40:53 but yeah let me get it anyway Dec 27 10:41:22 the missing ones should be marked with todo in the code IIRC Dec 27 10:41:49 https://drive.google.com/file/d/0B9idqO9KygGsdXltRVBEYWtUQmM/view?usp=sharing Dec 27 10:41:51 latest db Dec 27 10:42:14 ok, thanks Dec 27 10:43:44 coderus seems gone ..next time time i hope he would provide real ip while asking for a check of the ban.. Dec 27 10:43:49 jonwil: wht is the "todo mark", 'z' at the beginning of the name? Dec 27 10:44:13 cant remember off the top of my head Dec 27 10:44:24 ok, will check against the source code Dec 27 10:46:53 xes, most of those people are from shared ips with reported spammers anyway Dec 27 10:47:27 xes, maybe you should add info about it too, and probably link to google in form: http://google.com/?q=$ip Dec 27 10:53:37 KotCzarny: yes, i know. I want give support explaining and checking. ..But that ip isn't banned and never asked for a TMO page Dec 27 10:54:06 xes, i think you can do /whowas coderus and use that one Dec 27 10:54:14 or just reply to his mail Dec 27 10:57:33 in fact, coderus is logged in TMO now... Dec 27 11:03:03 so, either a spammer or clueless user Dec 27 11:03:11 jonwil: maybe a priority regarding ssl is QT itself Dec 27 11:03:18 - coderus is ~coderus@static.88-198-208-108.clients.your-server.de (backdoor) Dec 27 11:03:35 QT is open source so its easy to fix Dec 27 11:03:36 nice 'realname' Dec 27 11:03:46 i know for a fact that by default it tries to use unsecure algos Dec 27 11:04:12 I am concerned about the closed source bits so we know what might be worthwhile targets for cloning Dec 27 11:04:24 on yappari i had to forbid it from using ssl3 beause the servers were rejecting that kind of connection Dec 27 11:05:10 i guess that gtk/qt will make the majority of ssl users Dec 27 11:05:42 but certain closed sources could use openssl by themselves, that's true Dec 27 11:07:42 wow, maemo.sexy domain is free to register Dec 27 11:07:58 is.my.maemo.sexy ;) Dec 27 11:07:58 lol Dec 27 11:47:45 http://talk.maemo.org/showthread.php?p=1492735#post1492735 Dec 27 11:50:04 So there are a total of 9 binaries that are closed source and call openssl functions Dec 27 11:50:43 how did you make that list? ldd? Dec 27 11:51:04 no Dec 27 11:51:22 I have a local copy of a stock untouched (e.g. no optification) root filesystem on my disk Dec 27 11:51:32 I searched that for everything that linked to libssl and libcrypto Dec 27 11:51:55 but how did you actually test if package links to libssl Dec 27 11:53:24 Its easy enough to tell by searching for the right strings (i.e. libssl.so and libcrypto.so) Dec 27 11:54:00 So then I matched each of those binaries (the ones referencing libsso and libcrypto in their strings) to a package Dec 27 11:54:08 and then for the closed source ones I dumped a list of symbols Dec 27 11:54:13 list of imported symbols Dec 27 11:54:24 and compared it to the list of symbols exported from libssl/libcrypto Dec 27 11:54:25 uhum Dec 27 11:54:50 trust me when I say that the info in the forum thread is accurate and complete Dec 27 11:54:51 Use ldd? Dec 27 11:55:03 Ah Dec 27 11:55:06 Sorry. Didn't read up. Dec 27 11:55:29 jonwil: is that only for the default rootfs, though? Dec 27 11:55:36 yes only for the stock rootfs Dec 27 11:55:48 I dont see anywhere else where there are closed source binaries that need examination Dec 27 11:56:34 idk, the nokia repo Dec 27 12:01:42 jonwil: can you look at http://talk.maemo.org/showpost.php?p=1491455&postcount=476 ? Dec 27 12:02:43 I tried to find where nokia debugging messages come from but was unable to properly reverse engineer the relavent bits (my clone does the same job as stock but it does it in a different way) Dec 27 12:03:08 just add them into your clone Dec 27 12:03:30 run binary and copy strings (if you do not know from where comes from) Dec 27 12:03:36 this should be easy Dec 27 12:09:22 Pali you still need this available-notifications file? Dec 27 12:16:37 ok, so I checked the list of packages that http://mirrors.muarf.org/maemo/apt-mirror/mirror/downloads.maemo.nokia.com/fremantle/ssu/mr0/ (nokia repo) says it provides against the list of packages installed on the root filesystem image I have Dec 27 12:17:01 Then for the ones on the repo but not the root FS I checked if they were in the SDK repo as source Dec 27 12:17:21 and for the few that aren't FOSS, they aren't using OpenSSL (I checked) Dec 27 12:17:25 Halftux: of course :-) Dec 27 12:18:04 so http://talk.maemo.org/showthread.php?p=1492735#post1492735 contains the complete list of all the closed-source things talking to openssl unless there is a repo somewhere I dont know about Dec 27 12:19:48 and we now have enough information (hopefully) to evaluate each of those closed binaries and figure out what the security risks might be going forward if those binaries continue to talk to openssl 0.9.8 Dec 27 12:22:22 hmmm, I never knew modest used microb-engine... Dec 27 12:22:52 almost everything in maemo uses microb engine Dec 27 12:23:01 at least if there is html involved in app Dec 27 12:23:07 yeah :) Dec 27 12:25:48 Pali: I had a look at my phone which had ota updates from Nokia. But the file was not there. Then the yellow update notification pops up and a file available-notifications.tmp was created. I think when the blinking stops the file disappears. After ignoring the updates a seen-updates file was created. Don't know if the temp file is of any help sadly I was to slow to copy it. Dec 27 12:28:14 The good news is that everything talking to NSS seems to be FOSS Dec 27 12:28:38 microb-engine and bits of modest being the main ones Dec 27 12:29:03 so it should be possible to replace NSS with something newer (and make any necessary changes in microb-engine etc) easily enough Dec 27 12:29:54 I guess it's very hard to port microb engine to use newer firefox engines Dec 27 12:30:01 yes it is Dec 27 12:30:27 for a bunch of reasons including Flash Dec 27 12:31:01 but updating NSS and the security sensitive stuff should be possible without breaking things Dec 27 12:31:29 modest depends on NSS? Dec 27 12:31:36 Flash? Who is so mad to use it? Dec 27 12:31:39 yes it does Dec 27 12:31:43 oh and, seriously, flash... Dec 27 12:32:32 I'm sure most of us would be happy to trade a flash-enabled browser with a real "fast" html5-compatible browser Dec 27 12:32:42 Halftux: .tmp file is created and HAM try to download new version to it from nokia server (which is down) Dec 27 12:33:38 jonwil: I think gnash does better flash than the maemo flash :) Dec 27 12:35:50 Fennec probably has just as many security issues as microb but since its not an official part of the system (not even sure where it comes from) its not within the scope of CSSU Dec 27 12:35:58 Pali: ok I see Dec 27 12:36:01 and therefore CSSU people dont need to care about fixing it Dec 27 12:37:43 main thing is, we have target list for closed source binaries that may present security risk if not moved to newer openssl Dec 27 12:39:41 :) Dec 27 12:44:02 hmm, looks like curl on maemo doesnt make any use of tlsv1 (?) Dec 27 12:44:45 error SSL routines:SSL23_GET_SERVER_HELLO ... Dec 27 13:11:36 would be good to see security improvements for Maemo Dec 27 13:12:10 but its a matter of finding people who understand the libraries and code involved (openssl, libcurl, microb, nss etc) Dec 27 13:12:36 in case of libressl it could be as simple as recompiling Dec 27 13:12:50 for the things that are FOSS yes Dec 27 13:12:53 (and source available) Dec 27 13:13:10 but for the closed source things we need to figure out which ones are using openssl in a way that matters for security Dec 27 13:13:49 also we need to deal with how certain things use openssl (e.g. lubcurl and libqt4-network) to make sure they use it in a way that is secure Dec 27 13:14:08 i.e. selecting the right algorithms and stuff and not claiming to support the old insecure stuff Dec 27 13:17:47 is libcurl part of the stock system? Dec 27 13:18:00 yes it is Dec 27 13:18:43 ah. Dec 27 13:19:50 its used by (among other things) apt-transport-https Dec 27 13:19:57 and nokia maps Dec 27 13:20:08 and flash Dec 27 13:20:29 and activysync daemon Dec 27 13:20:35 activesync Dec 27 13:21:09 hmm Dec 27 13:21:40 hmm what? Dec 27 13:22:06 then we need to either patch it to use the tlsv1 function or try to pull a recent version from upstream Dec 27 13:22:21 first option is probably the easiest one Dec 27 13:23:39 yeah once we have newer openssl/libressl we need to fix up libcurl so that it makes the right calls to openssl to do all the correct security stuff (and none of the stuff we dont want like sslv3) Dec 27 13:27:16 we dont even need to upgrade for that Dec 27 13:27:29 (ssl*_ Dec 27 13:51:13 iirc curl in cssu is "newer" and should behave better Dec 27 13:51:36 we do, we dont just want TLS1.0 we want whatever the latest TLS standard is Dec 27 13:51:39 TLSv2 or whatever it is Dec 27 13:51:57 merlin1991: in -testing maybe Dec 27 13:52:03 (-testing/-devel) Dec 27 13:52:04 bencoh: yep Dec 27 13:52:08 (testing) Dec 27 13:52:14 iirc luf did that Dec 27 13:52:28 luf? Dec 27 13:54:16 hm stable also has curl 7.26 Dec 27 13:58:41 yeah, that's what I have here as well Dec 27 14:50:47 qt4 doesn't support tlsv1.2 i think Dec 27 14:50:56 or at least it needs openssl 1.0.1 Dec 27 14:51:01 so... Dec 27 14:51:12 sometime in the future we may be not even able to access https sites? Dec 27 14:53:34 depends what browser you are using Dec 27 14:53:38 microb needs nss updated Dec 27 14:54:01 i was trying to get at why getmewheels wasn't working Dec 27 14:54:05 fennec is likely using its own copy of NSS (although by all accounts there are builds of Fennec for the N900 that are a fair bit newer than microb) Dec 27 14:54:08 and it's because car2go site uses tlsv1.2 Dec 27 14:54:19 getmewheels an interface to an http based api Dec 27 14:54:29 so... this application is now useless Dec 27 14:55:20 its not useless, it just needs someone to bring newer OpenSSL to Maemo then someone to fix QT to use the new OpenSSL and pick the best/most secure/etc settings Dec 27 14:55:45 i don't think it's going to be very easy, isn't it? Dec 27 14:56:14 it probably wont be as hard as it looks... Dec 27 14:57:18 sorry to go back to 'old' news ... i don't understand the significance of the file Pali was working on for HAM. we've been getting updates without it. what is it important for? Dec 27 14:58:19 Sicelo: update for stock n900 devices Dec 27 14:58:48 not apt update, just pop-up note with URL link Dec 27 15:02:24 ok, zzz time Dec 27 15:02:27 cya later :) Dec 27 15:02:28 sicelo: it was to notify users that dont know about t.m.o or w.m.o or cssu that there is something they can install Dec 27 15:03:45 pali, but if the sources for ham are available, maybe its just as simple as pushing updated ham into extras? Dec 27 15:04:26 KotCzarny: no, HAM does not allow updates for NokiaSSU packages which are not from NokiaSSU repo Dec 27 15:04:34 it has special pinning in HAM for it Dec 27 15:04:42 uhum Dec 27 15:04:45 reason why we need to patch everything in CSSU Dec 27 15:04:58 CSSU installer adds new repo with higher priority Dec 27 15:05:14 and so this repo can update packages also from NokiaSSU repo Dec 27 15:05:36 it is security framework, something like in harmattan, but fully open and configurable Dec 27 15:05:42 btw. if m$ ever decides to revive nssu repos and push rogue things there to install bugs, will they succeed? Dec 27 15:18:51 why would they do that? Dec 27 15:19:14 thanks for explanation by the way :) Dec 27 15:19:24 sicelo, because evil people often do evil things Dec 27 15:20:24 there also is http://maemo-repos.com/apt-mirror/ Dec 27 15:20:27 http://talk.maemo.org/showthread.php?t=95870 Dec 27 15:20:45 wtf was that? Dec 27 15:21:17 Singapore calling... Dec 27 16:06:08 DocScrutinizer05 and infobot: Thanks! Will do that when my N900 arrives. Dec 27 16:07:42 infobot is a bot Dec 27 16:07:52 infobot: botsnack Dec 27 16:07:53 aw, gee, KotCzarny Dec 27 16:07:59 aw, gee, KotCzarny Dec 27 16:08:04 Doesn't mean I can't say thanks :-) Dec 27 16:08:08 wizzup: rollover Dec 27 16:08:18 \o/ Dec 27 16:08:19 bob: sure, why not Dec 27 16:08:38 :-P Dec 27 16:16:23 Luke-jr: It sounds like you've already made up your mind about Ethereum, but for everybody else's benefit, let me reply. Dec 27 16:16:23 In summary, Bitcoin introduced two new innovations at the same time: Dec 27 16:16:24 Block-chain. Decentralized open ledger, consensus mechanism. Dec 27 16:16:25 Crypto-currency built on top of that block-chain. Dec 27 16:16:25 Great stuff. However, the "VM" for the Bitcoin block-chain was designed only for that single purpose. You can transfer crypto-currency excellently, but using it as a consensus mechanism for other purposes is hard. You see a bunch of other "Bitcoin 2.0" projects coming into existence which are piggy-backing on the Bitcoin block-chain for things like land registry, digital assets, insurance, etc. The problem they Dec 27 16:16:26 all have is that the Bitcoin block-chain really wasn't designed for that. Dec 27 16:16:50 Ethereum block-chain is Turing-complete, so you can build anything on it. Dec 27 16:17:34 Now of course, it's not at all intended as a replacement for local CPUs. That would be insane. It's probably trillions of times slower. Obviously. Dec 27 16:18:06 It's just intended for consensus Dec 27 16:18:15 Accounts. Who owns A, B, C. Dec 27 16:19:24 And then the "new stuff" over Bitcoin is that you can build (simple) behaviours on top. AKA - "smart contracts", which are self-executing. Like our friend "infobot" Dec 27 16:19:41 ie. If Condition A then pay money to account B. Dec 27 16:19:42 Escrow Dec 27 16:19:44 Insurance Dec 27 16:19:47 Shares Dec 27 16:19:58 Voting mechanisms Dec 27 16:20:00 blah blah blah Dec 27 16:20:48 But all just on the Ethereum block-chain using the basic functionality of the system, rather than wedged into the Bitcoin block-chain with a crow-bar and external systems. Dec 27 16:20:57 That's about it. Dec 27 16:20:58 http://devcon.ethereum.org Dec 27 16:21:12 Not a shitcoin. It' Dec 27 16:21:19 It's a technology Dec 27 16:21:41 You have a real absence of the speculators in Ethereum. It's all technologists. Dec 27 17:16:36 M-bobsummerwill: lol Dec 27 17:17:14 M-bobsummerwill: maybe if any of the people working on Ethereum ever did Bitcoin development, they would have realised Bitcoin already does basically everything useful that Ethereum aims to do Dec 27 18:11:11 Luke-jr: I don't know quite why you are being so aggressive and religious here. Vitalik Buterin, the creator of Ethereum, co-founded Bitcoin magazine in 2011, and is no crypto-dummy. It must be delightful for you to have such utterly certainly in your own opinions that you can discard the efforts of thousands of people to advance the state of technology. Also, delightful that there is "one true way". Dec 27 18:18:25 -> private messages? Dec 27 18:19:44 Or "let's agree to disagree", I think. Dec 27 18:20:25 Anyway ... I'll get going with my N900 development as-and-when. Best wishes, everyone! Dec 27 18:39:23 what is this? Advertising channel? Dec 27 18:40:17 What does it look like? Dec 27 18:41:12 looks like proud software dev overdefending his work Dec 27 18:41:28 an advertising monologue? Dec 27 18:41:36 + backlock reading Dec 27 18:41:40 *backlog Dec 27 19:24:56 hey everyone, i had power kernel and decided to return back to stock, so i launched Nokia kernel from menu and restored stock kernel (got successful message) after that removed power kernel settings package. but after reboot device went to reboot loop: nokia logo screen with usb icon appears on pale screen. after that black screen lights up and after few seconds reboots. any suggestions? Dec 27 19:29:28 useretail: it might be that kernel could not find its modules, try to flash kernel-power again, by using the flasher Dec 27 19:33:23 freemangordon: ok, so what do i need to flash exactly? looks like i had that problem some time ago. i have uImage-2.6.28.10-power52 and zImage-2.6.28.10-power52 images Dec 27 19:37:17 useretail: maybe first try to use zImage, and tell the flasher to only boot that kernel Dec 27 19:37:25 then you can flash wuth uImage Dec 27 19:37:31 *with Dec 27 19:46:21 or just try to use flasher to BOOT without flashing Dec 27 19:46:31 right. Dec 27 19:46:51 KotCzarny: how? i can't get it stop rebooting Dec 27 19:47:16 KotCzarny: please help him, as I don;t have time now Dec 27 19:51:39 useretail, grab the kernel power image package on pc, unpack it Dec 27 19:52:12 KotCzarny: yeah, i have done that already Dec 27 19:52:46 flasher-3.5 -k kernelimgfile -l Dec 27 19:54:23 there is also -b param to set default cmdline for kernel Dec 27 19:55:54 By using the "-l" option, we do not flash the kernel or initrd image. Dec 27 19:56:06 It only loads the kernel into RAM. No modification on the NAND or bootloader happens. Dec 27 19:56:30 0xFFFF also can load kernel Dec 27 19:56:42 0xFFFF -m kernelfile -l Dec 27 19:57:19 i assume you know how and when to connect n900 Dec 27 19:57:45 if not, run flasher/0xffff command first, then connect (powered off) n900 via usb cable to computer Dec 27 20:05:14 ok, flashed the kernel. but it still doesn't boot and more importantly doesnt charge from wall charger Dec 27 20:08:22 did you use -l ? Dec 27 20:08:47 -l mean to 'load/boot only' not actual flashing Dec 27 20:09:09 well i actually flashed: flasher-3.5 -k zImage-2.6.28.10-power52 --flash-only=kernel -f -R Dec 27 20:09:16 eh Dec 27 20:09:32 anyway, you can still try to boot Dec 27 20:09:49