**** BEGIN LOGGING AT Sat Jan 19 02:59:59 2013 Jan 19 03:28:25 ~mirror Jan 19 03:28:26 mirror is, like, http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ Jan 19 03:30:24 ~literal mirror Jan 19 03:30:25 "#maemo mirror" is "http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/" Jan 19 03:31:28 * grummund thinks light text on dark bacground is so much easier to read Jan 19 03:31:32 ~ no, #maemo mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 03:31:33 DocScrutinizer05: what are you talking about? Jan 19 03:31:49 infobot: no, #maemo mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 03:31:49 DocScrutinizer05: okay Jan 19 04:24:21 Anyone have experience trying to compile sudo and having it make and make install just fine segfault whenever executed with no error messages? Jan 19 04:25:27 s/fine segfault/fine but then segfault/ Jan 19 04:25:28 MentalistTraceur meant: Anyone have experience trying to compile sudo and having it make and make install just fine but then segfault whenever executed with no error messages? Jan 19 04:27:56 All the google results are either unrelated programs segfaulting (since like every other command is preceded with sudo nowadays, so google matches those pages), or segfaults from badly set up /etc/sudoers files, but I haven't modified sudoers at all from the one make install creates. Jan 19 04:35:15 Are there any key combos or gestures for Home, PgUp, PgDn? Jan 19 04:35:36 grummund, for the N900 you mean? Jan 19 04:35:57 Not that I know of, though it's fairly easy to remap hardware keyboard to add them. Jan 19 04:36:32 n900 yes. Jan 19 04:36:49 in particular to send the maemo web browser to top of page. Jan 19 04:38:15 Oh. Yeah, standard Home key (which I havmapped to Fn+Left Arrow) works to do that. Jan 19 04:38:37 Want me to find the wiki page for the remapping keyboard how-to? Jan 19 04:38:50 grummund: shift+space works as pageup in the browser Jan 19 04:39:39 khm: Doesn't seem to do so for me. Jan 19 04:39:54 weird Jan 19 04:41:05 http://tinyurl.com/n900keys Jan 19 04:41:12 it's listed in the official user guide Jan 19 04:41:15 and it works here. Jan 19 04:41:42 grummund: shift+up and shift+down are supposed to act like home and end in the browser Jan 19 04:41:52 but something may have broken these shortcuts apparently Jan 19 04:42:44 Now Shift+Up and Shift+Down work for me. Jan 19 04:43:38 excellent, thank you :) Jan 19 04:43:39 Granted, only in the browser as far as going to the top and bottom of screen goes (not in text boxes) Jan 19 04:44:03 I presume you no longer need this, then, but to anyone interested: http://wiki.maemo.org/Remapping_keyboard Jan 19 04:44:22 In case someone idling in here needed it and didn't already know about it or how to find it easily. Jan 19 04:44:46 khm: works for me too (in maemo browser) Jan 19 04:45:11 but only as simultaneous key press Jan 19 04:46:16 correct Jan 19 04:48:34 I honestly have no idea why it wasn't self-evident to Nokia to map PgUp, PgDn, Home, and End, to some keys. I get that for the layouts with only two arrow keys they've have had a harder time, but everything else can easily accomodate it on Fn+arrows. Jan 19 04:50:37 Even if you suppose users wouldn't type enough on-device to need them (a flawed supposition at best), they could've then used those standard keys for their expected behavior, instead of adding their own app-specific key combinations. Jan 19 04:52:43 I can't really say I use home or end ever Jan 19 04:52:53 99% of the time I'm in a terminal and ^a ^e work Jan 19 04:55:20 Meh. Point is, it wouldn't have harmed anyone, because I doubt you ever found yourself needing to hold Fn while pressing arrow keys, where-as my use of the N900 is vastly improved by having those keys, and since I've seen other people use the same mappings, I know I'm not the only one. Jan 19 04:56:45 As for myself specifically, my typing on the N900 is about evenly split between commandline and not (I like coding in the primitive busybox vi, actually) and other text input. Jan 19 04:57:46 For instance, I have never been on FreeNode's IRC using anything other than an N900 that I can remember. And as you can see, I'm quite verbose, so it greatly helps having those keys. Jan 19 04:58:29 Though ideally, I would prefer that almost everything was left to user choice... so my ideal setup would have provided users with a clear GUI to remap the keyboard for themselves to begin with. Jan 19 05:03:44 if I were a project manager and a software engineer told me with a straight face he wanted to include an on-device xkb map editor in the product, I would fire him before he could finish the sentence Jan 19 05:07:56 *Shrug*. That most likely (but would really depend on circumstance) if I were creating a phone operating system, I probably wouldn't want you as a project manager. Jan 19 05:08:43 Granted, I see your point, it's effort on a feature that most people wouldn't use, and which a decent amount would manage to misuse into screwing up their devices. Jan 19 05:12:27 So I wouldn't give it great priority, nor would I think it should be particularly easy to get to. And I would put a warning pop-up on it so those who don't know what they're doing are encouraged to just back out and leave it alone. And on further thought, I wouldn't mind it being optware, rather than prebundled... In fact that's probably better. But provided to the users in a easy-ish way to get to, I think it should b Jan 19 05:12:28 e. Jan 19 05:12:51 it wouldn't sell phones. Jan 19 05:17:27 I think little touches like that - the ability to make the system behave how you want it, helps make the difference between a typical phone that you're fine with, but will upgrade from to any other manufacturer without a second thought, and a phone that you remember as one that did what you wanted, how you wanted it. So in the short term? Of course it wouldn't sell anything. But in the long term, assuming the rest of t Jan 19 05:17:27 he OS was actually good and not just a hacker/nerd phone OS, it would provide for long-term loyalty. Jan 19 05:18:13 nope Jan 19 05:18:44 not for any measurable percentage of any market in the world Jan 19 05:22:01 Disagree. It's a long term thing, but if you managed to survive long enough as a company, it would pay off. Modern consumer culture and typical business-think might get away with forgetting that users are more than just mindless buyers of the latest shiny, but it doesn't mean there's nothing there. Jan 19 05:22:52 users are not more than just mindless buyers of the latest shiny. Jan 19 05:23:00 And whether or not it's measurable is unlikely to be conclusively demonstrable at this time, because I don't believe in modern times any company has ever tried it. Jan 19 05:35:21 They are, though (more than mindless buyers of the latest shiny). Or more accurately, they can be, if treated right. There are many flaws with humans, en-masse and as typical individuals, but I am not so cynical as to forget that all of those people are indeed people, with various feelings and ideas. I cannot count the number of times I have seen frustration from people over the fact that their device couldn't do what Jan 19 05:35:22 they wanted, and a decent amount of the time, it is not necessary that it be so limited, it was only corporate selfish bottom-line pursuing that caused the maker of the thing in question to not bother, either to concieve of the possibility that someone might prefer the option, or to invest the time into adding it. I'll certainly agree that on average, especially in current cultures, users do for the most part mindlessl Jan 19 05:35:22 y buy the latest shinies - but it's not like they're realistically ever given the opportunity to become anything else, because most devices focus on being just that - shinies. And sure, for every little choice you spend the time on enabling, only a handful of users will bo swayed. But if the rest of the system is good, and on top of that user choice is respected and enabled systematically in as many places as possible, Jan 19 05:35:27 it's inevitable that over time people would notice, that something they wanted to do, and couldn't on their other shinies, was not only possible, but relatively easy, on that one shiny. And sure, I could be wrong, and if I ever see sufficient evidence of that, I will change my views, but thus far, that has not been the case. Jan 19 05:35:45 Alright, I imagine you are bored of this discussion, and I am getting very wordy. Jan 19 05:35:50 it's moot Jan 19 05:35:55 since the market spake Jan 19 05:36:02 and lo, nobody paid Jan 19 05:36:16 The market never spake, because it has never been done properly yet. Jan 19 05:36:24 people say that about tons of things Jan 19 05:36:28 it doesn't matter Jan 19 05:36:48 socialism, kindness, charity, religion Jan 19 05:36:53 No device has thus far enabled user choice without making it require above-average tech-savvy-ness. Jan 19 05:37:07 'never did it just right' where 'right' has approximated 'my way' Jan 19 05:37:28 (No religion, actually, I am very much a-religious, and don't consider it to be connected to the other three) Jan 19 05:37:33 the flaw there is that desire for choice has tech-savvy as a build-time dependency Jan 19 05:37:51 Except it doesn't, and therein is the flaw in your reasoning. Jan 19 05:37:58 you never want a choice if you don't think you know better than the product engineer Jan 19 05:38:10 but I'm not here to convince you Jan 19 05:38:19 because again, it's moot. Jan 19 05:38:58 when you find a company who wants to ship an xkb map editing gui on a telephone, let me know. I'll send you a hundred dollars and make fun of them on twitter. Jan 19 05:44:38 *Shrug* K. If it ever happens, will do. Won't get my hopes up, but that's hardly sound proof in itself. One last note on the tech-savvy bit though: There's plenty of people who want thing X to do thing Y without being of the opinion that they can design the entirety of X better. E.g Windows users not liking the Start Menu going away, without having even minimal tech-savvy. In a sense, yes, they might think they are kno Jan 19 05:44:39 w better than the product engineer about whether there should be a start menu, but that's a conceptual difference, and does not depend on the capacity to program anything, or delve into system internals on the command-line. It just requires the concept of wanting the start menu. Jan 19 05:45:28 ...which you learned about from previous tech experience. Jan 19 05:45:32 got it. Jan 19 05:45:43 totally consistent, nothing circular there. Jan 19 05:46:12 There isn't, because again, that's not tech-experience of the not-needing-GUI form. Jan 19 05:46:34 you're conflating tech experience with command-line usage? Jan 19 05:46:38 eesh, I'm out. Jan 19 05:47:13 Nay, I am using command-line usage as the obvious alternative to a provided GUI. Jan 19 05:47:41 Editing system files and the like are others, but whatever. Jan 19 05:50:05 Anyway, if this ever interests you as a discussion topic again, by all means, bring it up. Though I suspect it will not, which fine by me as well. Jan 19 05:51:35 And on that note, I will be heading out. Have a good day/evening/whatever-time-it-is. Jan 19 05:51:40 * MentalistTraceur waves. Jan 19 05:53:09 no repos, no builder :( Jan 19 05:53:39 no TMO mail updates Jan 19 05:53:47 relax Jan 19 05:53:55 dns changes can take a long time Jan 19 05:54:04 which means shaking out problems can take even longer Jan 19 05:54:20 it'll get there. Jan 19 06:56:42 ARe the repositories down for anyone? Jan 19 06:56:47 *Are Jan 19 06:57:05 sgosiaco: see /topic Jan 19 06:57:22 ahh Jan 19 06:57:38 ~mirror Jan 19 06:57:39 methinks mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 06:57:51 thanks Jan 19 06:57:51 sgosiaco: ^^^ Jan 19 11:29:12 any idea when the dns server will be up ??? Jan 19 11:29:34 what DNS server? Jan 19 11:33:17 The DNS server for the repos Jan 19 11:33:35 or is it just my n900 Jan 19 11:50:32 repo mirrors anywhere, link to ? Jan 19 11:51:22 ~mirrors Jan 19 11:51:22 from memory, mirrors is Official Debian archive mirrors get an address of the form ftp://ftp..debian.org. These are the best advertised and most used sites; the face of Debian for most people. A list is available at http://www.debian.org/mirror/official/ Jan 19 11:51:25 hrmpf Jan 19 11:51:26 ~mirror Jan 19 11:51:27 extra, extra, read all about it, mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 11:51:41 infobot: #maemo mirrors is see mirror Jan 19 11:51:41 okay, kerio Jan 19 11:51:44 ~mirrors Jan 19 11:51:45 mirror is probably http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 11:52:20 maemo mirrors Jan 19 13:28:38 any news ??? Jan 19 13:29:17 Tek: lance armstrong admitted to doping Jan 19 13:39:33 Can someone help me fix all my repos Jan 19 13:42:38 Dont even know where to begin its all one big mess Jan 19 13:43:49 zomg!1!1 Jan 19 13:44:05 repository.maemo.org is down Jan 19 13:46:36 >:-| Jan 19 13:47:05 is repository.maemo.org down? Jan 19 13:47:12 (joke :P) Jan 19 13:49:07 NIN101: i'd say fuck off, but i can't ever be mad at you <3 Jan 19 13:50:09 someone needs to update that post Jan 19 13:50:27 last update says everything is sorted Jan 19 13:50:38 kerio :-) Jan 19 13:51:21 XDS2010_: the TMO migration is sorted Jan 19 13:51:35 but is the migration complete ? Jan 19 13:51:59 and what will become of the .org domain ? Jan 19 13:52:02 XDS2010_: is every service working at the moment? Jan 19 13:52:10 will we need to change repo addresses ? Jan 19 13:52:39 i picked a hell of a time to come back to maemo development Jan 19 13:52:58 will we need to change repo addresses ? !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Jan 19 13:53:03 HELLO Jan 19 13:53:24 calm down, geez Jan 19 13:53:24 IS ANYONE HERE !? Jan 19 13:53:48 No more caffeine for you. Jan 19 13:53:50 it's good that you're shouting. Otherwise no one would notice that you're writing messages to a channel with 268 people Jan 19 13:54:00 HELLO !? Jan 19 13:54:13 *cough267cough* Jan 19 13:54:18 6 in fact Jan 19 13:54:26 is this thing on!? Jan 19 13:54:32 :P Jan 19 13:54:39 XDS2010_; repos down waiting for dns changes by nokia. will return on old normal address Jan 19 13:54:52 ShadowJK: ty Jan 19 13:55:01 this is last I heard Jan 19 13:56:02 in other news my sleep cycle is all fubar Jan 19 13:56:33 and water bugs have invaded my kitchen Jan 19 13:59:54 that might be the hallucinogens Jan 19 14:00:08 drugs are bad, mmkay Jan 19 14:22:41 hello, what's with the holdup on the repositories? Jan 19 14:24:30 waiting for dns changes from nokia dnsmaster Jan 19 14:31:57 XDS2010_: http://talk.maemo.org/showthread.php?t=88707 <-- Read it and implement it on your system Jan 19 14:47:46 o/ Jan 19 14:56:10 \o Jan 19 14:57:33 \o/ Jan 19 14:57:42 sup? Jan 19 15:38:03 ~mirrors Jan 19 15:38:03 i heard mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 16:01:04 wtf is Pageant failed to answer challenge? Jan 19 16:01:17 trying to ssh into N900 and getting this... Jan 19 16:02:59 could be a bad command line or else a misconfigured server Jan 19 16:03:03 try again with ssh -v Jan 19 16:04:00 hmm, from terminal I can get in fine Jan 19 16:04:10 it's filezilla that refuses to connect Jan 19 16:05:48 khm, thanks, it's probably key mismatch Jan 19 16:06:01 and terminal falls back to password Jan 19 16:14:11 Hi. Anyone uses Kasvopus on n900 here? Jan 19 16:30:59 qwazix, ping Jan 19 16:35:06 ShadowJK, ping Jan 19 16:35:20 pong-ish Jan 19 16:35:33 could I use your expertise about filesystems and flash for a while? Jan 19 16:36:16 I'm trying to tune filesystem to avoid those damn fsck-will-nuke-our-flash-filesystem errors Jan 19 16:36:53 ...at the same time, tuning for max performance in case of our flash storage. EasyDebian partition (not file, obviously)* is my guinea pig Jan 19 16:36:56 Estel_: tune2fs -i 0 -c 0 Jan 19 16:36:57 or something Jan 19 16:37:01 now, to the question: Jan 19 16:37:27 if you're worried about the actual NAND, then don't worry, ubifs doesn't even have a fsck Jan 19 16:37:39 in case of our flash storage, does data=ordered have any sence over data=writeback Jan 19 16:37:40 ? Jan 19 16:38:02 or data=journal Jan 19 16:38:27 AIUI, we don't want fsck to touch filesystem with a long stick, and rely on big journal Jan 19 16:39:47 (my ed partition got just messed up by fsck again, so I though that it's good moment to re-create it, prior to restoring, with correct options, at last, and modify mounting) Jan 19 16:39:59 hi there, do you guys know what is the best way to install "less" on my n900 ? Jan 19 16:40:21 the armel deb from debian should work ? Jan 19 16:40:25 massoud: sudo apt-get install less Jan 19 16:40:47 kerio: my sources.list doesnt find less Jan 19 16:40:48 or busybox-power, but that's a bit weirder Jan 19 16:41:19 massoud: it's in extras(-something) Jan 19 16:41:21 same question goes for barriers, as we're definitely battery powered - there is chance for crash, but rather not for complete, rapid power loss - we should be safe to do barrier=0 yep? Jan 19 16:41:22 but the repos are down lately Jan 19 16:41:49 kerio: yes thats what i figured out :( Jan 19 16:41:56 massoud: use a mirror, i suppose Jan 19 16:42:24 massoud, fear not, repos are down, but you can use tmo, what else you need? ;) Jan 19 16:42:27 massoud: http://maemo.merlin1991.at/apt-mirror/extras-devel/pool/fremantle/free/l/less/less_429-2maemo2_armel.deb Jan 19 16:43:15 Estel_; as for journal options, I'm not sure Jan 19 16:43:33 anybody knows when the official software repositories will be available again? Jan 19 16:44:13 Estel_; barriers will severely slow things down I feel. As for performance, I'd investigate the raid stripe options, and tune for 64k/4M Jan 19 16:44:24 nevermind, saw the link above... Jan 19 16:44:29 Estel_; ext4 will perform much better I'm told Jan 19 16:44:44 i tried data=journal recently and it didn't impress Jan 19 16:45:07 also, I think for file integrity, flash is more or less screwed whatever you do Jan 19 16:45:14 ShadowJK: how do you raid on the n900? :o Jan 19 16:45:49 i didnt say that Jan 19 16:46:47 ext3 has options at mkfs time to give it hints about the underlying raid topology. We dont have raid, but in some ways the flash behaves somewhat similarily Jan 19 16:47:08 kerio : excellent thanks, i wanted to use git-core on my n900 and less was missing ... now i can git log projects such as openbsc on my n900 Jan 19 16:47:10 i see Jan 19 16:47:28 massoud: fwiw, extras-devel-light is a much better option Jan 19 16:47:29 ShadowJK, so better to keep data=writeback for performance, or data=ordered may force things to be less screwed, a little (by forcing writing to filesystem before journal commit)? Jan 19 16:47:29 i wish we could retrieve layer2,3 messages of gsm stacks on the n900 Jan 19 16:48:04 md2bus gives some dbus messages but not actual layer2,3 messages Jan 19 16:48:55 ShadowJK, excuse my ignorance, how to setup raid stripes properly on N900? wihout creating real raid? we're talking about ext4 here, of course Jan 19 16:49:08 well nokia used writeback, so dunno :) Jan 19 16:49:45 i'd say nobarrier data=writeback commit=600, because we're on battery Jan 19 16:52:15 kerio, isn't ext4 accepting commit as seconds, not miliseconds? Jan 19 16:52:25 yep Jan 19 16:52:26 10 minutes Jan 19 16:52:28 commit=600 = 10 minutes seems a little out of place Jan 19 16:53:01 in case of crash, being on battery won't save us from losing data from last 10 minutes, no fun Jan 19 16:53:13 crash of what? Jan 19 16:53:20 a kernel panic? Jan 19 16:53:38 for example Jan 19 16:53:50 what else? Jan 19 16:54:03 hm, do watchdog reboots go through the kernel? Jan 19 16:54:32 I dont thing changing it from 30 seconds to 10 minutes makes any performance difference, although, I may be wrong Jan 19 16:55:24 from 5 seconds to 30 yes, but more is rather numerelogy, isn't it? Jan 19 16:55:32 as for watchdog, no idea Jan 19 16:56:41 ShadowJK, have you tried those raid stripes on N900 yourself? Jan 19 16:57:17 nope Jan 19 16:57:26 mkfs has options for it Jan 19 16:57:54 * Estel_ nods Jan 19 16:58:15 searching for approriate options re ext4, now Jan 19 16:58:19 hm, why are mkfs.foo in /usr/sbin? Jan 19 16:58:31 any comment on our commit seconds discussion? Jan 19 16:58:34 nokia picked commit=1, and also tweaked vm to begin writing back early, so that we dont get massive stalls when huge amount of data would wait in ram and be written out in a burst Jan 19 16:59:36 ShadowJK: that's more or less irrelevant if you have the swap on usd, right? Jan 19 16:59:59 no Jan 19 17:00:08 funny enough, to make swap more usable and stable (proven, real-life tested) we need to make it as big burst as possible :P Jan 19 17:00:18 but filesystem is not swap, after all Jan 19 17:00:49 ShadowJK: won't home and swap being on different devices avoid those stalls? Jan 19 17:01:01 you know what I mean on the above, despite it's being mental shortcut :P Jan 19 17:01:21 well.. swap on fremantle, before fragmentation sets in, is sequential.. with good cards able to keep 5 AU's open simultaneously, we could in theory write to swap, and *a* file simultaneously without speed penalty Jan 19 17:01:57 with the best known cards, and no swap, we could do like 3 files without penalty! Jan 19 17:02:06 ;) Jan 19 17:02:59 So I gather the aim was to trickle files to storage instead of trying to update 600 files all at once every 10 minutes Jan 19 17:05:10 64k/4m stripe apply to ext4 as much as ext3? Jan 19 17:06:16 at openmoko we had a performance optimization team which did weeks and months of tests and tweaking. I guess Nokia most likely did same, alas afaik all that never got public, neither the design goal specs nor any results of tests nor the consequences taken from those. So we're doomed to go thru all that again. So much for FOSS development in huge companies Jan 19 17:07:32 Estel_; yeah it's just a handwave guess of the flash properties Jan 19 17:08:13 TomTom publishes results of their "repeatedly forcible remove SD card while writing to card", iirc Jan 19 17:08:34 basically only vfat gave vaguely safe results Jan 19 17:09:02 :o Jan 19 17:09:10 KISS approach Jan 19 17:09:21 surely ext4 with data=journal is safe by design Jan 19 17:09:32 bwahaha Jan 19 17:09:37 rofl Jan 19 17:09:38 hm, maybe not, nobody knows what SDs do Jan 19 17:09:46 surely it isn't ;) Jan 19 17:10:45 some of us have a certain idea of a few details of what an SD is doing internally Jan 19 17:11:12 At best, ext* assumes 4k blocks like recent HDs have.. on flash the blocksize is essentially unknown, but modern day SDs have empirically been determined to have 4M, 6M, 8M, 12M blocksizes.. if block is halfwritten on powerloss, worst case entire block is garbage, best case you get an old version of it, or then a mix :) Jan 19 17:11:23 and if kswapd also had those ideas, it would benefit swap performance a friggin lot Jan 19 17:11:52 ShadowJK: can't you specify a certain block size? Jan 19 17:11:54 DocScrutinizer05; the fremantle swap is quite close, it's only missing gc Jan 19 17:12:16 kerio; no it's a physical/electrical property of the flash Jan 19 17:12:33 no, i mean, to extfs Jan 19 17:12:50 rtfm? Jan 19 17:12:54 well the choices are 2k or 4k, iirc. Jan 19 17:13:02 aka man mkfs Jan 19 17:13:38 this is why I suggested using the raid options, as raids have some similar properties such as too small write triggering read-modify-write Jan 19 17:13:43 no way, even the blast vfat have more choices of block sizes Jan 19 17:14:24 Well also using a 12M filesystem blocksize would mean storing a 3 byte file would use up 12M of space Jan 19 17:14:57 -b block-size Jan 19 17:14:57 Specify the size of blocks in bytes. Valid block-size values are 1024, 2048 and 4096 bytes per block. Jan 19 17:15:03 FFS Jan 19 17:15:06 RTFM Jan 19 17:15:24 took me like 13.5s Jan 19 17:15:31 ,and what? it doesn't help us in any way, + what ShadowJK said. Jan 19 17:16:34 [2013-01-19 18:13:43] no way, even the blast vfat have more choices of block sizes -- ANd WHAT?! Jan 19 17:16:42 ShadowJK, I'll test that raid stripes empirically Jan 19 17:16:51 DocScrutinizer05, out of context ;) Jan 19 17:17:03 more = bigger than 4096 bytes Jan 19 17:17:21 well the choices are 2k or 4k, iirc. Jan 19 17:17:29 DocScrutinizer05> Specify the size of blocks in bytes. Valid block-size values are 1024, 2048 and 4096 bytes per block. Jan 19 17:17:47 = same, except for 1k addition in the latter Jan 19 17:17:51 anyway Jan 19 17:18:02 exactly, while your rant been mere bogus Jan 19 17:18:52 I'm not the one ranting about reading fuckin manual ;) back on track, please? Jan 19 17:19:18 you're the one posting nonsense Jan 19 17:21:17 ShadowJK, any ideas about modyfing buffer heads options to acompany data=writeback, and it's feasibility? Jan 19 17:21:51 bh is default option, while nobbh sounds more sensible for data=writeback Jan 19 17:21:57 IIUIC Jan 19 17:23:04 quoting, ""nobh" option tries to avoid associating buffer heads, (supported only for "writeback" mode." Jan 19 17:25:10 last but not least, ShadowJK, there is also mount-time option "stripe=", which seems to be doing same thing. Any comment on differencies between using this and mkfs-time one? that was last question, I promise ;) Jan 19 17:26:33 nobh is deprecated, afaik Jan 19 17:27:37 hm, for what reasons? it's still in kernel documentation for mount ext4 Jan 19 17:31:05 which kernel? Jan 19 17:31:25 Estel_; buffer heads options? what are those Jan 19 17:32:10 less /etc/mke2fs.conf Jan 19 17:34:47 ext4 associate (or associated, if it's really depreciated) buffer heads to either cache disk block mapping information, or link pages into transaction (if -bh) or Jan 19 17:34:51 whatever ikt means Jan 19 17:35:21 -nobh means that it tries to avoid associating buffer heads Jan 19 17:36:19 no, i think the issue was that you still had buffer heads, even with -nobh Jan 19 17:38:57 ShadowJK, mayday, but I'm ignorant about proper options for our "raid" in mkfs time, should I use: Jan 19 17:39:52 stride= Jan 19 17:39:58 or stripe-width Jan 19 17:39:58 ? Jan 19 17:40:25 both Jan 19 17:41:42 set -b 4096 to be sure, then stride 16, stripe-width 4M/4k, whatever that works out to Jan 19 17:41:45 ough, how should it look for your example? stride=64 and stripe-width=4096 Jan 19 17:42:06 ough Jan 19 17:42:09 * DocScrutinizer05 idly wonders if email notifications of edits on wiki are broken too. could some nice soul test it please? Jan 19 17:42:19 in units of fs blocksize, so 1 = 4k, 16 = 16*4k = 64k Jan 19 17:42:41 thanks a lot Jan 19 17:42:43 also who the heck are our wiki admins? Jan 19 17:42:52 now it's understood Jan 19 17:42:56 DocScrutinizer05: were admins nominated? Jan 19 17:44:15 kerio: I guess somebody instantiated some admins/mods for wiki Jan 19 17:44:29 like andre__ is for bugs Jan 19 17:45:36 I'd really like to see all this documented somewhere Jan 19 17:47:31 if it's only a matter of me looking at the right location to see it, please toss a pointer to that page. Otherwise all info welcome, so I can aggregate it Jan 19 17:48:54 andre__: could you check bugs.maemo.org if everything back to normal from your POV, please. Jan 19 17:51:49 Hello everyone, trying to reflash my N900 (eMMC & Combined, yes I've already read the wiki page), but I keep getting this error: "Error claiming USB Interface: Device or resource busy" Jan 19 17:53:17 I'm using openSUSE 12.2 64 bit and I already blacklisted cdc_phonet and phonet Jan 19 17:53:31 Also already tried all my USB ports (3) Jan 19 17:53:59 tried starting flasher-3.5 with sudo and with plain su Jan 19 17:54:10 (as root) Jan 19 17:57:30 Superpelican: hmm, apart from you picked a really bad moment to do that reflash, due to migration killed repos temporarily, I can't see anything you missed to try Jan 19 17:58:12 Superpelican, do you have ia-32libs? Jan 19 17:58:21 oooh 64 bit Jan 19 18:00:34 Does the 900 display usb-connected icon.. Jan 19 18:02:21 sorry, have to go. Jan 19 18:02:49 ShadowJK, picked up 100MB journal for 3GB filesystem, it should be enough to tell fsck to fckd out, do ya think? :P Jan 19 18:03:37 at least as much as it is possible Jan 19 18:04:08 I'd pick largest possible journal Jan 19 18:04:25 it's 400 MB Jan 19 18:05:00 doesnt sound like a little waste out of 3GB partition? Jan 19 18:05:17 maybe :) Jan 19 18:05:19 or I'm missing something? Jan 19 18:05:38 if you're experimenting and benchmarking perf, also try nilfs2 :) Jan 19 18:05:51 you might have to compile the utilities for it though :/ Jan 19 18:06:04 yea Jan 19 18:13:19 qwazix:looks like there is no ia32-libs for opensuse :( Jan 19 18:13:39 qwazix:but I did install 32 bit libusb Jan 19 18:14:11 Hmmm... Never used opensuse but on 64bit ubuntu installing ia32 is needed Jan 19 18:14:27 though I'm not sure that is the exact message when missing Jan 19 18:16:41 well I can try flashing with my Xubuntu netbook Jan 19 18:16:54 and of course install ia32 on the netbook Jan 19 18:18:15 Superpelican, you have a amd64 distro on a netbook? Jan 19 18:18:42 (I've read that newer atoms support x64 but my lenovo refuses to install x64 distros) Jan 19 18:26:55 qwazix: yes, 64 bit, Asus Eee PC 1005PE (although I'm not sure if it was that exact type, but at least a 1005 series)with Atom N450, 1 GB Jan 19 18:27:36 in fact I'm installing ia32-libs and all of it's 244 mb dependencies on the netbook Jan 19 18:28:22 Superpelican, same processor of my S10-3T, wonder why the different reaction to x64 Jan 19 18:28:42 qwazix:read that it has something to do with the motherboard Jan 19 18:29:02 qwazix:Manufacturers block x64 via the motherboard on purpose Jan 19 18:29:23 qwazix:Not all, but some Jan 19 18:29:23 hate crippleware Jan 19 18:29:32 The first generation atom also had x64-enabled CPUs :) Jan 19 18:29:58 qwazix, ShadowJK:I believe it does not have anything to do with the CPUs Jan 19 18:30:08 It's the motherboards Jan 19 18:30:11 Yes, some genuinely don't have x64 Jan 19 18:30:40 the original netbook-atom didn't, but the desktop version of same cpu had it Jan 19 18:31:20 don't understand why installing ia32-libs automatically pulls in stuff like 32 bit gtk2, vorbis encoder etc. Jan 19 18:31:36 Is it installing a complete 32 bit copy of my system or something? Jan 19 18:31:39 lol Jan 19 18:32:00 Hi, can someone please tell me the working n900 repositories? Jan 19 18:32:04 tbh I dont thik flasher would even execute if you didnt have libs already Jan 19 18:32:21 true Jan 19 18:32:41 You do have flasher-3.5, right Jan 19 18:33:07 iirc I had to run it as root under Arch to make it flash properly Jan 19 18:33:23 I'm using cssu stable but I'm not sure the repository is in there.. and extras, extras testing... are there any others? Jan 19 18:33:41 extras* isn't working right now Jan 19 18:34:00 you mean temporarily? Jan 19 18:34:01 ~mirror Jan 19 18:34:01 rumour has it, mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 18:34:12 Gear;yes temporarily Jan 19 18:34:23 Yay, it's flashing my N900 :D Jan 19 18:34:26 as in the next few hours or longer? Jan 19 18:34:39 days Jan 19 18:34:57 Conclusion:It seems that indeed the lack of ia32-libs etc was causing the flash problem Jan 19 18:35:01 ok, but what are the repositories that work from now on Jan 19 18:35:15 since there is no official nokia support anymore Jan 19 18:35:43 repository.maemo.org WILL come back, just not during the weekend Jan 19 18:36:01 so which ones should I have in my app manager? Jan 19 18:36:02 Hopefully nokia dnsmaster wakes up on monday and performs the requested changes Jan 19 18:36:17 Superpelican, yay! Jan 19 18:36:43 Dunno, random people have personal mirrors, but I don't renmember the details for anyone.. maybe search tmo Jan 19 18:37:15 I can wait, which ones should I have in there from now on? Jan 19 18:37:37 now flashing rootfs/combined :) Jan 19 18:37:40 uhm, same as before? Jan 19 18:38:17 I have like fremantle ssu apps, ssu mr0, 1.2 ovi, extras, and extras testing Jan 19 18:38:28 sounds fine Jan 19 18:38:34 those all work? Jan 19 18:38:54 Gear: look at topic, first link in it:) Jan 19 18:39:05 thanks Jan 19 18:39:11 anytime:) Jan 19 18:39:57 wait, I'm not really interested in the mirrors Jan 19 18:40:02 N900 rebooting, now I only have to install CSSU and for the first time KernelPower ;) Jan 19 18:40:03 I don't get why the poster assumes every device tries to update once a day Jan 19 18:40:03 I just mean the ones that permanently work now Jan 19 18:40:33 if you really need to update now, you need a mirror, if you don't, then i guess you can wait until maemo.org repos are back up Jan 19 18:40:37 I can wait a few days until they work Jan 19 18:40:46 Same here:) Jan 19 18:40:49 can you quickly tell me which repositories work? Jan 19 18:41:02 The Nokia ones Jan 19 18:41:06 the maemo.org ones don't Jan 19 18:41:09 Superpelican: I am not sure you'll be able to install CSSU right now, repo, you know :) Jan 19 18:41:19 I mean... like 2013 Jan 19 18:41:49 do I delete fremantle ssu mr0 and 1.2 ovi ? Jan 19 18:41:55 no Jan 19 18:42:03 oh? Jan 19 18:42:07 there is basically no need to delete anything, if im correct Jan 19 18:42:10 you can't delete them, only disable Jan 19 18:42:12 just wait until the repos are back up Jan 19 18:42:35 um I think I deleted them from the hildon app manager Jan 19 18:42:48 think again Jan 19 18:42:51 :) Jan 19 18:42:53 I need to put the cssu one in fapman Jan 19 18:43:07 ~fapman Jan 19 18:43:07 it has been said that fapman is Faster Application Manager, a bad package manager that causes problems, don't use it, ever Jan 19 18:43:19 Gear: And DON'T USE FAPMAN FOR SYSTEM UPDATES Jan 19 18:43:26 yeah true Jan 19 18:43:32 that advice is better than the infobot's Jan 19 18:43:33 that's not even easy to do, is it Jan 19 18:43:42 stick to fapman for 'simple' installing of apps etc Jan 19 18:43:46 fapman is great for single apps, but definitely might cause issues for system updates Jan 19 18:43:47 not for important stuff Jan 19 18:43:51 hah Jan 19 18:44:21 freemangordon:When will the repo's be completely up again? Jan 19 18:44:25 (estimated) Jan 19 18:44:43 NFC, I guess in the upcoming week Jan 19 18:45:03 I deleted some from app manager I think Jan 19 18:45:17 you can't delete Nokia repos Jan 19 18:45:24 I mean disabled Jan 19 18:45:33 delete != disable ;) Jan 19 18:45:38 freemangordon: you can, with 1337 haxx ;) Jan 19 18:45:42 I disabled some Jan 19 18:45:59 mozilla multi catalogue works? Jan 19 18:46:07 merlin1991: well, I guess I can rm -rf / too :D Jan 19 18:47:24 nokia applications, nokia system software updates, ovi, mozilla multi catalog..... which ones no longer work? Jan 19 18:47:37 nokia* Jan 19 18:47:52 heh Jan 19 18:47:59 freemangordon, do you, by any chances, read maemo-developers mailing list? Jan 19 18:48:08 or used to do, when mailing list worked? Jan 19 18:48:10 and ovi works? Jan 19 18:48:15 can you quickly tell me which repositories work? Jan 19 18:48:18 The Nokia ones Jan 19 18:48:20 the maemo.org ones don't Jan 19 18:48:42 sorry I'm just trying to sort out hildon app man Jan 19 18:48:50 as it's the important one Jan 19 18:48:50 Estel_: yeah, I read it when it works :D Jan 19 18:49:11 Estel_: I vaguely remember a mail from you Jan 19 18:49:15 freemangordon, any use for those qf9700 modules I've submitted there? Jan 19 18:49:17 lemme check Jan 19 18:49:19 yea ;) Jan 19 18:49:24 aah, i remember now Jan 19 18:49:35 are there any other repositories which I should put in there? Jan 19 18:49:42 sure thing, just wanted to be sure it haven't ended in /dev/null by mistake Jan 19 18:49:51 Gear: what are you trying to do? Jan 19 18:50:15 Estel_: no idea, if someone buys such a device, maybe :) Jan 19 18:50:18 just disable the repositories that don't work, and put any useful repositories that do in there Jan 19 18:50:28 Gear: what for? Jan 19 18:50:30 freemangordon, I meant chances to include modules in KP Jan 19 18:50:44 for software Jan 19 18:50:45 I *know* that module work Jan 19 18:50:50 and I'm using cssu Jan 19 18:50:54 Estel_: didn't checked, but I see no reason why not Jan 19 18:51:01 thanks :) Jan 19 18:51:22 though I guess it will take a while. We need that autobuider/repo mess sorted out Jan 19 18:51:31 afaik it just need compiling against kp headers and you know, regular inclusion Jan 19 18:51:33 sure Jan 19 18:51:35 so ovi doesn't work and nokia* doesn't either Jan 19 18:51:47 Uh oh, the main reason that I reflashed my N900 was that the maemo.org repo's didn't refresh anymore Jan 19 18:51:55 But now I suddenly understand why Jan 19 18:51:56 Superpelican: hrhr Jan 19 18:52:00 It was the phone... Jan 19 18:52:01 Superpelican: smart ;) Jan 19 18:52:04 I'll be happy if it end up in kernel as of 2013 Jan 19 18:52:10 Superpelican: use one of the mirrors ;) Jan 19 18:52:27 ~mirror Jan 19 18:52:27 extra, extra, read all about it, mirror is http://maemo-archive.wedrop.it/ http://maemo.merlin1991.at/apt-mirror/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 Jan 19 18:52:29 merlin1991: did anyone fix the hash mismatch error? Jan 19 18:52:43 afaik my mirror has no hash mismatch Jan 19 18:52:46 I don't need to put a mirror in Jan 19 18:52:52 Oh well I was also planning to install KernelPower some day but I had that special bootloader by Stskeeps installed for booting Nemo Mobile and Cordia Jan 19 18:52:55 if it has somebody please slap me and I'll fix it Jan 19 18:53:13 considering that autobuilder/cobs/repo mess will probably take half a year, *if* there still be anyone interested after, I'll be glad and thankful to see this module in kp as of 2013 :D Jan 19 18:53:13 ok Jan 19 18:53:17 Superpelican: you can use the latest u-boot from Pali for that Jan 19 18:53:37 are there some repositories that I might not know of? Jan 19 18:53:38 Yeah, but read about lot's problems with U-Boot from Pali Jan 19 18:53:42 Estel_: naah, once we have the control, I guess it will be fixed in 1-2 weeks Jan 19 18:53:44 atm it's fine Jan 19 18:53:47 it is in extras-devel Jan 19 18:53:48 there used to be problems Jan 19 18:53:50 Gear; afaik ovi repo works Jan 19 18:54:15 other old versions of u-boot has problems Jan 19 18:54:16 freemangordon, hope so ;) Jan 19 18:54:20 I know about extras, testing, and devel (I disable devel) Jan 19 18:54:27 Estel_: come on :) Jan 19 18:54:34 are there any others that I don't know of? Jan 19 18:54:48 Gear: you may want to disable -testing too Jan 19 18:54:48 that have some good stuff Jan 19 18:54:52 why? Jan 19 18:55:02 I use some software from that Jan 19 18:55:12 Sure, but don;t keep it enabled Jan 19 18:55:23 enable it when you really need to Jan 19 18:55:23 why not? Jan 19 18:55:53 freemangordon, Pali, btw I've just come across things from 2.6.37 kernel, that could benefit our filesystem performance a much... Jan 19 18:55:54 Gear: read the wiki article what extras-testing and extras-devel are Jan 19 18:56:02 and I'm not sure if it was backported Jan 19 18:56:10 take a look at this quotation, please: Jan 19 18:56:12 yeah I know what they are, a lot of things are still in testing Jan 19 18:56:14 Estel_: Pali is on 3.8 already :D:D:D Jan 19 18:56:22 Estel_, we are using 3.8-rc3 Jan 19 18:56:34 .37 is old Jan 19 18:56:37 does it really slow the phone down if I have testing enabled? Jan 19 18:56:45 you can bet Jan 19 18:56:57 i guess estel means backport those things from .37 to pk? Jan 19 18:57:00 merlin1991:Your mirror also includes maemo-extra? Jan 19 18:57:16 ShadowJK: ofc, I am kidding him :) Jan 19 18:57:45 ;) Jan 19 18:57:46 http://sebsauvage.net/paste/?5fce2900ad3e7057#swwkW7ONu40N0TgpMeE4pW9OPJ8ZJQ/ugnCVNeiRvNc= Jan 19 18:57:52 why does extras testing need to be disabled? Jan 19 18:57:57 it's quote from this interesting and long document: Jan 19 18:58:01 Superpelican: look at the page ;) it's cssu cssu-testing extras and extras-devel Jan 19 18:58:09 http://monolight.cc/2011/06/barriers-caches-filesystems/ Jan 19 18:58:12 the apps I use most are actually from testing Jan 19 18:58:13 merlin1991:get 404 not found when refresing your repo Jan 19 18:58:34 Superpelican: use one of the subfolders ;) Jan 19 18:58:37 Estel_, I understood that this is in upstream kernel Jan 19 18:58:43 so is included in 3.8 already Jan 19 18:58:56 considering how our flash media like barriers, and that disabling barriers doesnt work anymore, replacing 4thlvl barrier with emopty one as per fixes from 2.6.37 might be very beneficial Jan 19 18:59:02 Pali, yes Jan 19 18:59:07 Estel_: sounds very dangerous to me Jan 19 18:59:22 but as ShadowJK said, I mean backporting it to KP Jan 19 18:59:22 "In fact, the responsibility for safe request ordering has been completely moved to filesystems. " Jan 19 18:59:33 if I disable extras testing will it actually make a lot of difference? Jan 19 18:59:50 this barriers thing is mostly irrelevant for sd Jan 19 18:59:58 sd/emmca/flash Jan 19 19:00:07 merlin1991:What's the "community" subfolder? Jan 19 19:00:18 cssu stable Jan 19 19:00:19 Gear: try it, I have NFC what "a lot of difference" means to you Jan 19 19:00:26 irrelevant as in "disable it as it only kills performance for no benefit", or as in "doesn't matter at all"? Jan 19 19:00:36 ShadowJK, ^ Jan 19 19:00:39 as in "does not matter" Jan 19 19:00:47 kills how much performance? Jan 19 19:01:07 and my most useful stuff is from testing still... like modrana and pierogi and some others Jan 19 19:01:12 so explicit flushes and FUA requests instead of barriers are not enough as in dangerous? Jan 19 19:01:17 merlin1991:I shouldn't add cssu repo manually right? Installing the CSSU installer package will automatically do that, right? Jan 19 19:01:33 yep, but it adds the other repo Jan 19 19:01:38 which errm does not exist :D Jan 19 19:01:43 Estel_: I don;t know if filesystems in 2.6.28 will work correctly if we apply this patch Jan 19 19:02:01 thus "dangerous" Jan 19 19:02:06 Estel_; kills perf for no benefit I would say Jan 19 19:02:43 I was hoping someone had made an archery app Jan 19 19:02:57 it is related to ext4 only, because ext3 doesnt have barriers Jan 19 19:03:04 by default Jan 19 19:03:19 merlin1991:So I should install the CSSU installer, then delete the maemo.org cssu repo and add the mirror "community" repo manually? Jan 19 19:03:28 freemangordon, so, I don't see how it could affect other filesystems, i.e. default maemo ones Jan 19 19:03:53 and as ShadowJK said, barriers are PITA for us, hurting performance, without ensuring file integrity, anyway Jan 19 19:04:13 new barriers are probably just as pointless too Jan 19 19:04:32 Estel_: without the patch in front of me I can't tell more (and maybe even with the patch in front of me :) ) Jan 19 19:04:37 ;) Jan 19 19:04:49 thats why I've mentioned it as worth looking into Jan 19 19:05:05 Estel_: not to say that first thing we should do is a kernel in cssu Jan 19 19:05:07 merlin1991:adding /extras results in a "hash sum mismatch" error Jan 19 19:05:07 ;) Jan 19 19:05:10 Also, I'm not sure if barriers are supported on mmc :) Jan 19 19:05:11 it won't improve integrity of files, but wont hurt them too, gaining performance (in optimistic scenario) Jan 19 19:05:33 s/first/the first/ Jan 19 19:05:44 Estel_; currently barriers aren't used, so there'd be no performance gain? Jan 19 19:06:33 so I should disabe extras testing until I need it, are there any repositories I dont' know about that have cool things in them? Jan 19 19:07:04 qwazix: ping Jan 19 19:07:07 ShadowJK, in ext4, they're used Jan 19 19:07:14 by default Jan 19 19:07:28 and ext4 gains quite a lot re our flash storage and performance Jan 19 19:07:35 without barriers, it could add more Jan 19 19:07:48 makes more sense to mount with nobarrier then :) Jan 19 19:07:53 is ext4 windows compatible? Jan 19 19:07:54 :D:D:D Jan 19 19:07:56 doesnt work Jan 19 19:08:03 Gear_: no Jan 19 19:08:12 Gear_: not even close Jan 19 19:08:17 poop, I have to transfer pictures to windows computers Jan 19 19:08:17 nobarriers doesnt work without that path afaik Jan 19 19:08:25 with it it doesnt work too :P Jan 19 19:08:28 at best you get a fairly decent read-only mode with external drivers Jan 19 19:08:32 Gear_: relax, you don;t have ext4 on your device Jan 19 19:08:36 but uses other, less performance hurting method Jan 19 19:08:47 I don't? I use power kernel and cssu stable Jan 19 19:08:53 so? Jan 19 19:08:58 so I dunno. Jan 19 19:09:11 so you don't Jan 19 19:09:16 okely dokely Jan 19 19:09:30 kernel-power supports ext4, but ou need to create it manually, if you want to have it Jan 19 19:09:44 and then I'll lose windows plug and play Jan 19 19:09:51 nope ;) Jan 19 19:09:55 it's different thing Jan 19 19:10:06 Gear_: we are talking rootfs here (i guess) not your MyDocs partition Jan 19 19:10:10 you're talking about mydocs which is plug and pl.ay container Jan 19 19:10:15 no rootfs too Jan 19 19:10:18 home ;) Jan 19 19:10:23 /dev/mmcblk0p2 Jan 19 19:10:26 Estel_: aah, yes Jan 19 19:10:34 rootfs is ubifs and i wont change it with a long stick Jan 19 19:10:34 * freemangordon is stupid today Jan 19 19:10:37 is converting difficult? Jan 19 19:10:49 Gear_: yes Jan 19 19:11:05 I can probably do without it then Jan 19 19:11:17 AFAIK only a couple of people did it Jan 19 19:11:34 how do I clear all the unused crap from my device Jan 19 19:11:49 ext4 has barrier option just like ext3, are you saying passing barrier=0 to ext4 has no effect? Jan 19 19:12:01 to be absolutely honest (and very *not* politically right) it's easy as a piece of cake, but still, only a dozens of people or so use it Jan 19 19:12:29 * freemangordon is tempted to give Gear_ vi's medicine :D Jan 19 19:12:33 ShadowJK, now I'm not sure anymore. I'm sure that -nobh option doesn't work at mkfs time Jan 19 19:12:58 but nobarrier at mount time Jan 19 19:13:04 would I see a performance increase if I did use ext4 wherever you're talking about? Jan 19 19:13:07 i was thinking of disabling the ubifs compression Jan 19 19:13:08 barrier=0 Jan 19 19:13:12 should work, me think, which would nullify need for "my" patch Jan 19 19:13:25 kerio: do we have so much free space already? Jan 19 19:13:34 maybe you don't, but i do Jan 19 19:13:47 are you taunting freemangordon ? :P Jan 19 19:13:50 aah, yes, you've removed "nokia hands" video :P Jan 19 19:13:56 I don't have a hell of a lot of free rootfs or optfs space left Jan 19 19:14:17 and optified /var/lib/dpkg, /var/lib/apt, locale-archives, /usr/lib/gcc Jan 19 19:14:20 Estel_: no, he is just trolling me :) Jan 19 19:14:44 I disabled extras testing but it didn't say "blah blah cleared" Jan 19 19:14:50 kerio: good for you that you've optified /var/lib/apt Jan 19 19:14:53 Gear_, ext4 benefits performance quite a lot, but unless you correct swap settings, and do other many things like reading a wiki, you won't see any gain due to "crap on yourf device" ;) Jan 19 19:15:34 there is only stuff I use, but I'm not sure how to check for stuff I don't use or things that are left over from... things I'm unaware of Jan 19 19:15:34 freemangordon, I rather meant kerio teasing you by "maybe yoiu don't, but I do" Jan 19 19:15:37 nvm Jan 19 19:15:38 kerio: it seems you did that on my device too :P Jan 19 19:15:45 :o Jan 19 19:15:54 really, i remember doing that Jan 19 19:15:58 as my /var/lib/apt is optified too Jan 19 19:16:11 oh well Jan 19 19:16:15 Gear_, eat red pill and get to know what is happening on your device. Reading wiki - starting in "beginners" section - helps a lot Jan 19 19:16:16 and i DON'T remember doing that :) Jan 19 19:16:16 then i didn't optify that Jan 19 19:16:24 http://wiki.maemo.org if it does work Jan 19 19:16:30 it works Jan 19 19:16:47 last time i've checked Jan 19 19:16:57 Estel I'll do that but not straight away, I'm just quickly seeing which repositories work and which don't anymore Jan 19 19:17:18 nice to hear, I'm out of loop what does and doesn't work in this [strike]titanic[/strike] queen mary 2 anymore Jan 19 19:17:19 anymore? Jan 19 19:17:20 and "syncing" fapman and normal app man Jan 19 19:17:28 i've got 243M used on / Jan 19 19:17:39 kerio: nice Jan 19 19:17:40 I thought nokia has discontinued any support Jan 19 19:17:49 (uncompressed 243M, i mean) Jan 19 19:18:02 kerio: you're on the edge AIUI Jan 19 19:18:16 yeah, i'd have to optify a lot more things Jan 19 19:18:22 and the nokia repositories were still enabled Jan 19 19:18:42 Gear_: Nokia support was non-existant for the last... 2 years or so Jan 19 19:18:48 freemangordon: pong Jan 19 19:18:56 alright Jan 19 19:19:00 qwazix: what is that "missing symbol"? Jan 19 19:19:08 freemangordon, s/anymore/or doesn't anymore/ Jan 19 19:19:10 freemangordon: do you reckon microb-engine can be optified? Jan 19 19:19:21 rootfs 73/227mb free opt 1240/2015 mb free Jan 19 19:19:27 some garbled things and then hildonPlugin Jan 19 19:19:36 not @pc atm Jan 19 19:19:39 is that bad? Jan 19 19:19:53 Gear_: sudo du -hsx / Jan 19 19:19:57 read that might be version mismatch Jan 19 19:20:02 ubifs compression is *magic*! Jan 19 19:20:02 qwazix: aah,ok. Jan 19 19:20:19 kerio what does that do? Jan 19 19:20:33 magic, as he said Jan 19 19:20:37 but I tried running on simulator and got the same thing so probably not Jan 19 19:21:04 Gear_: actual data on /, without displaying the subdirectories, with the appropriate unit of measure, and without crossing filesystem boundaries Jan 19 19:21:10 qwazix: well, I can try to help if you post the "garbled symbols" somewhere Jan 19 19:21:19 and also marxian has the same problem Jan 19 19:21:25 yeah, saw that Jan 19 19:21:26 I'm using fapman to try to clear out the extras testing stuff left behind Jan 19 19:22:14 and what's so bad about fapman? Jan 19 19:22:19 okay I'm at the movies now so I'll post that later. It'something like undefined symbol: Zcx_jdHildonPlugin Jan 19 19:22:31 qwazix: no hurry Jan 19 19:22:40 Gear_, honestly, nothing is bad about fapman, just don't use it for cssu updates Jan 19 19:22:48 :nod: Jan 19 19:22:50 I know that much Jan 19 19:22:51 freemangordon, thanks. Talk later. Jan 19 19:23:12 or, do like me, and use it for everything, but be prepared to know how to fix things if something goes wrong without crying for help on cssu developers Jan 19 19:23:29 :nod:^2 Jan 19 19:23:36 that sudo du -hsx / is asking for a password Jan 19 19:23:47 not that something went wrong at all in my years of using it, but it's not supported way of updaing cssu Jan 19 19:23:47 kerio: don;t tell him :P Jan 19 19:24:12 Gear_: try without sudo Jan 19 19:24:15 I never use fapman for important stuff Jan 19 19:24:15 password starts with rm -fr, I suppose... Nah, just joking Jan 19 19:24:20 freemangordon: permission error, probably Jan 19 19:24:32 kerio: missing rootsh probably Jan 19 19:24:48 can't open anything, permission denied Jan 19 19:24:48 or df is not sudoer Jan 19 19:24:55 sould I go sudo gainroot Jan 19 19:25:00 important varies - for some, cutetube for watching porn is most important stuff on device Jan 19 19:25:02 Gear_: do you have rootsh installed? Jan 19 19:25:14 I have power kernel whatever Jan 19 19:25:16 Estel_: i tried that, but i couldn't get it to work Jan 19 19:25:16 and cssu Jan 19 19:25:16 it's about not using it for system updates (cssu is considered system update) Jan 19 19:25:29 Gear_: can you get a root shell somehow? Jan 19 19:25:37 sudo gainroot? Jan 19 19:25:47 Gear_: just type "root" without quotes Jan 19 19:25:57 well, launch "du -hsx /" as root in some way Jan 19 19:26:08 and be veeery careful what you do, double-check your commands Jan 19 19:26:26 293.9M Jan 19 19:26:31 what does that mean Jan 19 19:26:43 your root filesystme usage Jan 19 19:26:50 *filesystem Jan 19 19:26:52 how much is there in total? Jan 19 19:27:04 293.9MB Jan 19 19:27:11 Gear_: that you have 294MB stored in your occupied 154MB Jan 19 19:27:26 what? Jan 19 19:27:34 rootfs uses compression Jan 19 19:27:45 uh Jan 19 19:27:53 so it is using 154MB to store 294MB Jan 19 19:28:07 ok Jan 19 19:28:09 303.7MB. He's bette than me :( Jan 19 19:28:21 also, aiui ubi volumes autoexpand a bit Jan 19 19:28:26 how much space is there in total Jan 19 19:28:27 so you might get a bit more Jan 19 19:28:39 Gear_: ~252MB Jan 19 19:28:53 how do you know it's using 154mb Jan 19 19:29:10 251.25MiB Jan 19 19:29:13 Gear_: you said so! Jan 19 19:29:19 21:16 rootfs 73/227mb free opt 1240/2015 mb free Jan 19 19:29:33 oh. right. Jan 19 19:29:34 probably ~227-73 Jan 19 19:29:37 ooh, why 227? Jan 19 19:29:46 kerio: ^^^ Jan 19 19:29:47 also, aiui ubi volumes autoexpand a bit Jan 19 19:30:02 well, in my case it's 228 Jan 19 19:30:08 I'm looking for some cool eyecandy visualisation stuff Jan 19 19:30:09 kerio: can't parse Jan 19 19:30:17 same here, honestly Jan 19 19:30:21 mine too, but that's probably due to rounding errors Jan 19 19:30:22 what autoexpand? Jan 19 19:30:30 like sigil in toment? Jan 19 19:30:34 torment* Jan 19 19:30:37 Estel_: yes, precisely like that Jan 19 19:30:44 * freemangordon is lost Jan 19 19:30:51 or maybe not Jan 19 19:31:02 ubifs will try using more space as you start running out Jan 19 19:31:14 ubisoft Jan 19 19:31:18 i noticed that on the sheevaplug too Jan 19 19:31:27 when you're constrained, try to expand. Logical. Jan 19 19:31:35 hey I have pwnie express on my spare mint phone Jan 19 19:31:38 it's especially noticeable when your rootfs fills up Jan 19 19:31:51 kerio: do you know the exact partition size? Jan 19 19:31:54 because you've added too many torrents and the fucking transmission info directory is in /var/lib Jan 19 19:31:54 freemangordon, I think both of us need to read manual on ubifs Jan 19 19:31:58 how can I format that to swap all this stuff over, this phone is falling apart and nearly dead Jan 19 19:32:27 Gear_: reflash Jan 19 19:32:34 ...as i don't get what he is saying about, too Jan 19 19:32:42 s/saying/speaking/ Jan 19 19:32:43 Estel_ meant: ...as i don't get what he is speaking about, too Jan 19 19:32:43 Gear_: install backupmenu, make a full backup, reflash your other phone, install backupmenu there, load the backup Jan 19 19:33:00 and experience attack of clones Jan 19 19:33:21 my imei number will change Jan 19 19:33:30 it wont Jan 19 19:33:36 afaik Jan 19 19:33:47 if I flash the other phone? Jan 19 19:34:02 Gear_: no Jan 19 19:34:09 and transfer everything then use that as my phone, and this as my toy? Jan 19 19:34:16 the IMEI is stored in the rapuyama CAL, so there's no way to change that Jan 19 19:34:31 but it'll go to whatever the other phone is Jan 19 19:34:33 Gear_: restoring a backupmenu backup will give you an exact copy of your current phone Jan 19 19:34:42 kinda Jan 19 19:34:48 I'm not sure I want that, there might be crap on here that I don't actually need Jan 19 19:34:49 unles connecting few cables here and there and electroducing it while submerge in liquid nitrogen Jan 19 19:35:06 Gear_, jokes aside Jan 19 19:35:15 first - go and read wiki Jan 19 19:35:17 make sure you have the correct kernel preinstalled Jan 19 19:35:18 second Jan 19 19:35:36 I wouldn't mind doing a "fresh install" unless there's no point Jan 19 19:35:43 ~flashing Jan 19 19:35:44 methinks maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware Jan 19 19:35:44 do full reflash, install kernel-power, then reformat partition layout as per wiki Jan 19 19:35:45 Secondly, did you choose a good value for vol_size or enable autoresize in your ubinize configuration? It is possible that your UBI volume is not utilising the full space of the available flash. Jan 19 19:35:54 freemangordon: do you have 252M used? Jan 19 19:35:56 er Jan 19 19:36:01 252M in your reported size Jan 19 19:36:04 no Jan 19 19:36:07 how much? Jan 19 19:36:16 Gear_, changing partition layout is best done as early as possible Jan 19 19:36:24 kerio: 305.6M Jan 19 19:36:32 then, while doing it, you may also decide to change (or not) home filesystem to ext4 Jan 19 19:36:35 changing to what Jan 19 19:36:38 i mean what df -h reports as "Size" Jan 19 19:36:40 rootfs 227.9M 163.9M 59.8M 73% / Jan 19 19:36:45 same as me Jan 19 19:36:46 hmm Jan 19 19:36:47 Gear_, again, first read wiki Jan 19 19:36:52 I'm not very good with linux Jan 19 19:37:02 it's quite useless to copy instructions from there to here, on irc Jan 19 19:37:09 thats why ik said read wiki Jan 19 19:37:15 Gear_: are you good on reading? Jan 19 19:37:15 http://wiki.maemo.org Jan 19 19:37:33 yes I am thanks for asking Jan 19 19:37:53 I have the opportunity to pick up a new in box n900 for 100 bucks Jan 19 19:38:00 there is "beginners section" exactly for "not good with linux" guys Jan 19 19:38:00 rootfs 227.6M here Jan 19 19:38:10 would you guys grab that Jan 19 19:38:28 unless it's from hong-kong, yes Jan 19 19:38:36 without second thought Jan 19 19:38:41 Gear_: like a priest would grab a choir boy's butt Jan 19 19:38:48 :D Jan 19 19:39:03 it's another phone I don't really need, but I plan on using an n900 until they're so dead they don't do anything Jan 19 19:39:18 there's no other phone I like Jan 19 19:39:26 Gear_: the you definitely should start reading ;) Jan 19 19:39:29 *then Jan 19 19:39:32 android is so shitty Jan 19 19:39:46 and I will never own an apple product Jan 19 19:39:57 Gear_, no worries, at beginning, kerio was as irritating as you Jan 19 19:40:02 but he started reading Jan 19 19:40:07 and see where he ended now Jan 19 19:40:16 talking about barriers, ubifs, and whatsnot Jan 19 19:40:23 are you sure you want end like that? :P Jan 19 19:40:26 and choir boys butts Jan 19 19:40:32 Estel_: since when kerio is not irritating? :P Jan 19 19:40:34 instead of you know, getting girls with new iphone? Jan 19 19:40:42 freemangordon, ;) Jan 19 19:40:46 i was a long-time linux user though Jan 19 19:41:14 I managed to stream video through wifi Jan 19 19:41:28 that's the most I've done with linux Jan 19 19:41:50 no worries, as next step wifi would stream you over video Jan 19 19:42:00 omg Jan 19 19:42:08 I'm making a kite that can drop things on people usi9ng HEN Jan 19 19:42:29 you want to use usb cable as kite line? Jan 19 19:42:38 no, to enable a little solenoid Jan 19 19:42:41 i wouldn't do that, the musb port is really flaky Jan 19 19:42:45 yep Jan 19 19:42:52 freemangordon, what, no one said my sense of humour is sensible today :P Jan 19 19:42:55 and to stream the video back to a netbook Jan 19 19:43:05 Gear_, for that -* ad now seriously... Jan 19 19:43:05 so I can see who I'm dropping stuff on Jan 19 19:43:11 yopu use OpenPilot Jan 19 19:43:12 Estel_: we still can't believe it :P Jan 19 19:43:20 what's openpilot Jan 19 19:43:30 and it's software is available for maemo Jan 19 19:43:35 google openpilot Jan 19 19:43:46 i.e. not that it's made by google Jan 19 19:43:53 just google for it Jan 19 19:44:12 can it enable the host mode and trigger a little 5v thing to drop stuff on people? Jan 19 19:44:19 then, you steer your airplane, zeppelin, or whatsnot over wifi or 3G or... Jan 19 19:44:27 kite, it's stationary Jan 19 19:44:31 Gear_: if you just need the 5v, then it's even easier Jan 19 19:44:42 as long as it doesn't use more than 200mA, yes Jan 19 19:44:45 well, yes Jan 19 19:44:46 do it with a battery Jan 19 19:45:02 don;t break USB port on your n900 Jan 19 19:45:07 I shouldn't need a relay Jan 19 19:45:22 sure, a simple switch will work for you Jan 19 19:45:25 DocScrutinizer05, bugs.maemo.org looks fine to me. Search works, I can access attachments, so I think it's fine. Jan 19 19:45:33 actually I might if I need to light a fuse Jan 19 19:45:40 using nichrome Jan 19 19:45:49 Gear_: do you have access to a soldering iron and a steady hand? Jan 19 19:46:11 yes I do Jan 19 19:46:37 Gear_, why doing it over hen? use bluetooth and bkuetooth-to-serial adaptor Jan 19 19:46:46 screw that Jan 19 19:46:53 ~usbfix Jan 19 19:46:53 extra, extra, read all about it, usbfix is http://talk.maemo.org/showthread.php?t=75920 - and **NEVER** use epoxy (unless you want to seal your device for underwater) Jan 19 19:46:54 I'll be logged into my netbook with the phone anyway Jan 19 19:47:03 unless your definition of kite is different than mine, I don't see usb cable hanging from it fancy Jan 19 19:47:11 so I can putty and open HEN and drop things Jan 19 19:47:35 the netbook is a wifi access point and the phone is logged in and streaming video Jan 19 19:47:43 oh you need a camera, too Jan 19 19:47:44 hm Jan 19 19:48:01 I wish I knew how to put crosshairs up Jan 19 19:48:14 kerio: wifi bandwith should allow streaming Jan 19 19:48:23 overlay something onto the display Jan 19 19:48:51 damn I want to get openpilot now Jan 19 19:49:09 Gear_: http://media.giantbomb.com/uploads/13/131403/2158022-magnavox_odyssey_overlays_large.jpg Jan 19 19:49:20 over your netbook :D Jan 19 19:49:38 heh Jan 19 19:49:46 that'd be retarded. Jan 19 19:49:52 in all seriousness, fix your n900's usb Jan 19 19:50:08 you mean beef up the ground? Jan 19 19:50:25 andre__: many thanks for confirming Jan 19 19:50:26 i mean Jan 19 19:50:27 ~usbfix Jan 19 19:50:27 somebody said usbfix was http://talk.maemo.org/showthread.php?t=75920 - and **NEVER** use epoxy (unless you want to seal your device for underwater) Jan 19 19:50:44 BTW I still don;t get the "epoxy" part Jan 19 19:50:47 to be fair, this n900's usb port seems to be resisting more Jan 19 19:50:47 you mean beef the ground up Jan 19 19:50:53 what's wrong with epoxy? Jan 19 19:51:01 Gear_: *wink wink nudge nudge saynomore* Jan 19 19:51:14 shouldn't be anything wrong with expoxy if you use it sparingly Jan 19 19:51:21 but I'd solder it on Jan 19 19:51:35 i think the issue with epoxy is that it doesn't actually fix the problem Jan 19 19:51:36 or just leave it, it's been fine since I got it Jan 19 19:51:47 epoxy would hold it there Jan 19 19:51:55 kerio: mine is 3 year old, the only thing i did is to file my charger/cable connector teeth a bit Jan 19 19:52:06 so it seems it depends on your luck Jan 19 19:52:17 hey my speaker is really quiet Jan 19 19:52:33 when I pulled that metal bit off it never properly sealed after that Jan 19 19:52:34 kerio: why epoxy should not glue metal to the board? Jan 19 19:52:41 * ShadowJK has never boken usb and hasn't done any mods either Jan 19 19:52:44 Gear_: the Marshal Jacques de la Palice would be alive, if he weren't dead Jan 19 19:53:03 I tried putting some double sided tape around the "speaker channel" but that didn't really work Jan 19 19:53:11 smart Jan 19 19:53:16 yay, a new kim dotcom video Jan 19 19:53:31 * freemangordon is off for cigarettes Jan 19 19:53:54 gordon freeman can't smoke in that hyper oxygenated suit Jan 19 19:54:21 mainyl because it has a helmet Jan 19 19:54:47 when is ep3 supposed to come out? Jan 19 19:54:53 he can't pee with it on too, so he removes it sometimes :D Jan 19 19:55:06 I thought it recycled all waste Jan 19 19:55:12 freemangordon: what do you mean he can't pee with it on? Jan 19 19:55:19 Gear_: indeed Jan 19 19:55:25 all good exo-suits do that! Jan 19 19:55:33 he can't take it on or off without help Jan 19 19:55:50 yeah, as shown in half life and in half life 2 Jan 19 19:55:58 where he needs a huge apparatus to wear the suit Jan 19 19:56:03 ...waaaaaaaaaaaaaaaaait Jan 19 19:56:11 he jsut appears with the suit Jan 19 19:56:33 "still fits like a glove! at least, the glove parts do" Jan 19 19:56:43 I'm sur ethe process of putting it on would be at least as difficult as a space suit Jan 19 19:57:04 also it should have a rebreather Jan 19 19:57:23 freemangordon: (epoxy) a) epoxy is more flexible than solder, so solder will break first. and b) epoxy would only glue to the green varnish unless you remove it and thoroughly sandpaper the copper under it Jan 19 19:57:47 it can admimister intravenous drugs and has muscle augmentation, a rebreather would be a simple addition Jan 19 19:58:02 freemangordon: and c) epoxy is nearly impossible to remove or rework when sth goes wrong Jan 19 19:58:09 freemangordon, as "some" claims, hardened epoxy is much more elastic than solder joins, so while it keeps usb in place, it doesnt protect it from damage to usb pads and it's solder Jan 19 19:58:11 I'd use solder Jan 19 19:58:21 and it's true for most consuimer grade epoxies Jan 19 19:58:34 epoxy could gum up stuff you dont' want it to Jan 19 19:58:37 there are, available for purchase, epoxies that are less elastic than metal Jan 19 19:58:40 I' duse solder Jan 19 19:58:50 but they're not as that adhesive Jan 19 19:58:51 but I haven't taken the phone that far apart Jan 19 19:59:02 soldering is better option Jan 19 19:59:43 if someone doesn't touch soldering iron with long stick and want to use epoxy no matter what, I would recommend ones used to repair hydraulic Jan 19 20:00:49 in form of 2-part "putty", seals underwater (lol@ factoid) and after sealing ressemble rock, not elastic at all. Sadly, it's still quite hard to attach it properly, it resemble toy putty in consistence Jan 19 20:01:05 frankly, i see soldering both easier and more robust Jan 19 20:01:11 Are the temp repos linked on tmo trustworthy? Jan 19 20:01:28 downside - even people experienced with soldering (like me or vi) happen to run into strage issue after soldering Jan 19 20:01:53 eyescream: is repository.maemo.org trustworthy? Jan 19 20:01:54 despite that only usb shield is soldered to shieldplane, device starts to turn on itself upon inserting battery Jan 19 20:02:25 Estel_: have you tried a Jan 19 20:02:26 and in extreme cases, turn on in act_dead mode no matter what, requiring one to set r&d flag pwr_button Jan 19 20:02:28 ~n900-full-reset Jan 19 20:02:29 n900-full-reset is, like, when the user presses the PWRON (power-on) button for 8 seconds and removes the battery in the next 8 seconds, the TPS65950 enters NO SUPPLY state instead of BACKUP state, even if a valid backup battery is present. In such a situation, the backup domain registers are also reset, along with the VRRTC domain registers. Jan 19 20:02:33 sure Jan 19 20:02:36 kerio: I've wondered that, too. My answer was more trustworthy than anything else. Jan 19 20:02:59 http://www.firenock.com/main.php?page=iBowSight Jan 19 20:03:06 personally, I suspect the following: Jan 19 20:03:11 can you emulate iphone software? Jan 19 20:03:22 quite a heat is required to attach usb shield to shieldplane Jan 19 20:03:36 eyescream: packages are still signed with the same gpg keys Jan 19 20:03:50 considering how tiny connections inside multi-layer PCB are... Jan 19 20:04:04 so even the weird repos are a lot more trustworthy than, say, packages from nokia's repos Jan 19 20:04:08 there is huge chance of burning out pcb isolating some two of them Jan 19 20:04:15 use flux and solder carefully Jan 19 20:04:15 while soldering shield to groundplane Jan 19 20:04:28 resulting in some connection, where it shouldn't be present Jan 19 20:04:47 lots of people have done it Jan 19 20:05:04 DocScrutinizer05, any comment on that from someone familliar with N900 PCB point of view? Jan 19 20:05:11 is it plausible explanation? Jan 19 20:05:16 I haven't yet because it requires taking the hell apart of the phone Jan 19 20:05:35 and the usb port is fine Jan 19 20:06:00 I though it's just me, but I've talked with at least 6 people experiencing same issue after soldering down port. Personally, I soldered it on few devices, and two of them are affected by that, immediately after soldering Jan 19 20:06:00 my brothers one broke off and I'm gonna use his as a guineapig Jan 19 20:06:27 use an adjustable heat iron and flux and do it carefully Jan 19 20:06:28 DocScrutinizer51: yep, got it now (epoxy) Jan 19 20:06:52 2nd one was soldered very carefully by me, yet, it still shows this behavior Jan 19 20:06:59 it only has to stick Jan 19 20:07:05 Estel_: gas iron? Jan 19 20:07:11 not armour the whole thing up with solder Jan 19 20:07:15 well, pwr_button r&d flag is nice way to nullify it, but... Jan 19 20:07:25 kerio: I see. Thanks. I might as well go ahead and try them out, then. Jan 19 20:07:32 freemangordon, yes, but I made sure that heat from it doesn't affect other parts via air Jan 19 20:07:55 my gas iron is brutal, I wouldn't use it on any n900 Jan 19 20:08:04 Estel_: could it be electrostatic discharge? I know you are experienced, but... Jan 19 20:08:37 freemangordon, hm, I always touch solid ground before soldering it, and gas iron, as said... don't think so Jan 19 20:08:43 you'd have to overheat it so badly, pcb isn't that heat condictive Jan 19 20:09:06 Estel_: only touch? no grounded ring on your wrist? Jan 19 20:09:07 exact same copies of software on all devices, some get affected, some not - just after soldering, immediately, si I would say "hardware" Jan 19 20:09:32 eyescream: skerion is Jan 19 20:09:36 freemangordon, gas iron handle is insulative and tip grounded, so my wrist shouldn't matter Jan 19 20:09:44 Gear_: :nod: Jan 19 20:09:50 Estel_: what? Jan 19 20:09:52 wedrop.it and merlin too, i'd say Jan 19 20:10:22 both yoiu and the iron should be grounded afaik Jan 19 20:10:25 as soon as the solder flows let that side cool Jan 19 20:10:26 I mean that grip of my gas soldering iron is made of non-conductive material, and grounded to avoid static Jan 19 20:10:48 aah, ok Jan 19 20:10:49 is it really necessary? my phone has been going since it was new Jan 19 20:11:20 freemangordon, I can't ground "whole" iron or it would take heat to the ground, still, I would rather blame heat Jan 19 20:11:36 what would you ground it to anyway Jan 19 20:11:49 Estel_: do you have an idea what material is pcb made from? Jan 19 20:11:53 frankly, it just looks like device"thinks" there is charger connected, on low lvl, but data pins are *not* short Jan 19 20:12:06 DocScrutinizer05: Thanks. And thanks for the updates on the migration process. :) Jan 19 20:12:28 no idea, but i would guess some epoxy-derrivative and glass fiber? Jan 19 20:12:47 epoxy is too gunky Jan 19 20:12:55 yw Jan 19 20:12:56 Estel_: i doubt it is glass/epoxy because of hi freq parts Jan 19 20:13:01 right Jan 19 20:13:06 so no idea, as said Jan 19 20:13:31 why asking? some idea? Jan 19 20:13:31 so if it is tephlon (or whatever the english term is) i'd say it is overheated Jan 19 20:13:35 (soldering) make sure you got no static or parasitic electricity on your solder tip Jan 19 20:13:35 can iphone software be ported? Jan 19 20:13:52 Gear_: what software? Jan 19 20:13:56 http://www.firenock.com/main.php?page=iBowSight Jan 19 20:14:33 WTF? Jan 19 20:14:59 it uses the camera and "levels" Jan 19 20:15:03 I would say "heat", as usb port is (relatively) big, and to solder it's outer side to ground plane on device (under scrapped green mask) you need quite a heat, even if you do it with small ammount of solder Jan 19 20:15:05 accelerometers Jan 19 20:15:59 Gear_: if you get the source, I guess it could be ported Jan 19 20:16:03 it's a relatively useful piece of software for bow Jan 19 20:16:05 i suspect that "something" is very fragile, be it a path inside pcb or whatnot. Maybe ikt's not direct contact, even Jan 19 20:16:07 shooting Jan 19 20:16:22 it seems to affect device only at boot time Jan 19 20:16:31 but that hardware is designed for iPhone form factor Jan 19 20:16:42 no higher power usage during runtime (by a slightest bit), no other ill effects Jan 19 20:16:43 what? Jan 19 20:16:53 usb working fine for both hostmode and client Jan 19 20:16:56 http://www.firenock.com/images/iBowSight/iPhoneBParts.jpg <- Gear_ Jan 19 20:16:57 bef0rd: he'll use duct tape :D Jan 19 20:17:07 oh you mean the chassis Jan 19 20:17:17 it could be altered Jan 19 20:17:18 duct tape should work Jan 19 20:17:21 :P Jan 19 20:17:25 dremel Jan 19 20:17:27 it works everytime :D Jan 19 20:17:29 just that during boot, in enters act_dead, and shutdowns, due to lack of charger Jan 19 20:17:36 (boots ok with charger connected) Jan 19 20:17:43 estel that sounds bad Jan 19 20:18:01 how many did this happen to? Jan 19 20:18:12 30%? Jan 19 20:18:12 r&d flag pwr_button "fixes" it and no further ill effects observed, for more than a year of usage Jan 19 20:18:29 yea, would say 30%, but it's quite random Jan 19 20:19:11 funny enough, people who doesn't care for grounding rings on wrist or even touching anything grounded before starting, doesn't seem to be afffected :P Jan 19 20:19:18 so i would say luck Jan 19 20:19:22 hehe Jan 19 20:19:32 I see no reason to really do that mod until the ground thing actually breaks off Jan 19 20:19:43 it's too late then Jan 19 20:19:44 Gear_: it will be late then Jan 19 20:19:56 it takes copper with it Jan 19 20:19:59 it'll still work, it's just soldered on anyway Jan 19 20:20:01 *the copper Jan 19 20:20:09 and you're soldering the sides Jan 19 20:20:17 data and power pads gets damaged, and even if they look all clean (rarely they do), and you're able to fix it, there is high chance it will stop working again Jan 19 20:20:19 that's all connected to the same ground isn't it? Jan 19 20:20:39 Gear_: it is not only ground on USB Jan 19 20:20:51 eh? Jan 19 20:21:22 soldering 4 pins of microusb is pita enough, and if copper pads get broken too, it's impossible unless one use under_battery pads Jan 19 20:21:30 if the connector falls of, it takes data planes off too Jan 19 20:21:32 even then, it's pita to fix Jan 19 20:21:44 i thought the under-battery pads were for rapuyama's usb Jan 19 20:21:48 my brothers one didn't take the data pins off Jan 19 20:21:53 for what? Jan 19 20:21:56 just the little shield fell out Jan 19 20:22:03 gear, wf? :P Jan 19 20:22:07 wtf* Jan 19 20:22:30 my brothers n900 the little ground shield fell out and it still worked with some wiggling Jan 19 20:23:24 can't parse, we're talking about microUSB port on N900, it can't break partially ;) Jan 19 20:23:29 it just goes off Jan 19 20:23:36 freemangordon, once, I had fun idea Jan 19 20:23:39 it's just the ground metal thing Jan 19 20:23:43 checking pcb schematics Jan 19 20:23:50 so that iphone software, could it be ported do you think? Jan 19 20:24:07 to ensure that no path cross pcb just at sides of port Jan 19 20:24:15 Gear_: already answered, gimme the source code and i guess i'll port it Jan 19 20:24:19 drilling it with 0.5 mm drill Jan 19 20:24:26 I don't have the source code Jan 19 20:24:35 what to "port" then Jan 19 20:24:46 I dunno. can it be reverse engineered? Jan 19 20:24:49 then, putting microusb port with legs soldered on the other said, through holes Jan 19 20:25:24 sure, get tools and reverse engineer it, alongside windows, and few other minor software bits Jan 19 20:25:29 Gear_: if you find me HexRays for ARM, I promise i'll RE it for you. Or will write a replacement Jan 19 20:25:46 writing replacement seems less work :P Jan 19 20:25:50 yes Jan 19 20:26:10 though I don;t have a bow here, might need to get one :D Jan 19 20:26:13 Gear_, start a kickstarted project Jan 19 20:26:35 frankly, it would be possible to do using N900 software already available Jan 19 20:26:38 look up bowtech destroyer 350 back ops Jan 19 20:26:40 Estel_: sound sane (USB holes) Jan 19 20:26:43 there is plenty of it measuring angles Jan 19 20:27:02 black ops Jan 19 20:27:08 that's my bow, it's brutal Jan 19 20:27:09 hexrays/ida is way too expensive Jan 19 20:27:09 though I don;t think we have access to PCB Jan 19 20:27:20 bef0rd: I know Jan 19 20:27:22 freemangordon, it does, and it would be rock solid thereafter, like mikniusb ports or regular ones, but that unidentified "heat" damage scarred me off Jan 19 20:27:38 heat it slower Jan 19 20:27:45 but it's a fine piece of software Jan 19 20:28:12 time for FOSS RE tool :P Jan 19 20:28:19 bef0rd: yeah. And we are lucky that at least x86 hexrays leaked :) Jan 19 20:28:39 I can't stand this afront to freedom, no foss RE thing being better than commercial ones? :) Jan 19 20:28:59 it might actually be worth just getting an iphone if I'm gonna pay for that mount Jan 19 20:29:12 Estel_: come on , writing decompiler/disassembler is huuge project Jan 19 20:29:30 actually I think that software might come installed on an iphone already, it seems to be a custom iphone Jan 19 20:29:42 custom iphone? Jan 19 20:29:47 freemangordon, so writing operating system is :P Jan 19 20:29:50 it has camo Jan 19 20:30:12 does the guy say it's an app? Jan 19 20:30:21 Gear_: maybe I dont get it but that software seems quite basic, just a camera and a level indicator? Jan 19 20:30:28 freemangordon, there are N900's sold with Faircrack preinstalled, for 2k$ Jan 19 20:30:29 and a dot in the middle of the camera viewer ? Jan 19 20:30:30 bef0rd: ;) Jan 19 20:30:36 "custom hacker phones" Jan 19 20:30:43 I don't know Jan 19 20:31:03 Estel_: what?!? 2000$ for n900? Jan 19 20:31:06 it's the first time I've seen that software and I haven't researched it yet Jan 19 20:31:09 freemangordon, I though of selling N900's like that myself :P Jan 19 20:31:14 come on Jan 19 20:31:33 I saw at least few of them sol for 1800 usd, but it was a while ago. Maybe people got smarter since... Or not Jan 19 20:31:34 I wouldn;t believe someone is stupid enough to buy that Jan 19 20:31:46 You wouldn't belive how stupid people can be Jan 19 20:31:55 though... people are stupid, they buy iphones Jan 19 20:32:00 hello! i just tried registering on the maemo.org website, but the security images are offline? Jan 19 20:32:29 stuff N900 with pentest things, make a fancy hackish theme, add smscon with pre-confogured scripts for "eavesdropping" on owner gps position and talks... Jan 19 20:32:45 and you have 5k dollars device for sale :P Jan 19 20:32:59 hacker/spy ultimate device combined Jan 19 20:33:13 recaller hidden and set up to record calls Jan 19 20:33:24 smscon to monitor gps position and take photos of user Jan 19 20:33:44 = ultimate gift for suspected husband or employe Jan 19 20:33:48 I think Estel_ meant this one http://pwnieexpress.com/products/pwnphone Jan 19 20:33:50 Estel_, sounds very cool - been thinking about similar stuff =) Jan 19 20:34:17 wonder where are they getting the new n900 from :P Jan 19 20:34:30 well, if not for my funny morals, I would already made small fortune on it, "spy phones" are expensive on local wanna-be-ebay sites Jan 19 20:34:37 hong-kong Jan 19 20:34:41 "brand new" :P Jan 19 20:35:13 hah yea Jan 19 20:35:27 995 $? this one is cheaper Jan 19 20:35:42 =) Jan 19 20:36:19 still fun way to sell devices, though Jan 19 20:36:38 freemangordon went away, I bet he is buying stock of N900's ;) Jan 19 20:36:57 to stuff them with stuff Jan 19 20:37:40 i'm going to test building a n900 cluster Jan 19 20:37:57 only have two of them currently though =P Jan 19 20:40:42 * DocScrutinizer05 throws another 5 N900 plus 14 Openmoko Freerunner GTA02 on jyrjyr Jan 19 20:40:54 =D Jan 19 20:41:25 it would make a nice little compiling farm Jan 19 20:41:33 actually I pondered that when we had a stock of a few 1000 GTA02 which were pretty unclear if they'd ever get sold Jan 19 20:41:47 oh! Jan 19 20:42:13 DocScrutinizer05, you could create darwinia with them Jan 19 20:42:20 i'd like to make that farm available in the net Jan 19 20:42:40 Estel_, exactly =) Jan 19 20:44:51 I figured how they'd look when put up to the wall in a 30*30 grid, or 30*80 Jan 19 20:46:23 what is the fastest interconnect available for the n900? Jan 19 20:46:30 usb ethernet? Jan 19 20:46:36 :nod: Jan 19 20:46:41 yes Jan 19 20:47:26 theoretically should max out at 100+some mbit Jan 19 20:48:00 maybe even 2 or 3 hundred Jan 19 20:48:02 not much there then. have to spend a long time optimizing mpi Jan 19 20:48:26 could do a dsitributed grid over wan, with a client that only activates when the phone is charging Jan 19 20:48:31 folding@go Jan 19 20:48:58 hahaha Jan 19 20:49:11 usb ethernet max out at 5Mb/s most of the time Jan 19 20:49:32 in real system yes Jan 19 20:49:38 more expensive ones can get 15 Mb/s max, if we're talking about usb to ethernet dongles Jan 19 20:49:41 * Estel_ nods Jan 19 20:50:05 it's not really about throughput though in a compute cluster Jan 19 20:50:11 true Jan 19 20:50:13 you worry more about latency Jan 19 20:50:28 and reliability, of course Jan 19 20:50:29 latency seems ok even in 2$ ethernet dongles Jan 19 20:50:46 but usually I get yelled at if average latency exceeds about 250ns Jan 19 20:51:04 I thought we're talking about ethernet-over-usb, not dongles Jan 19 20:51:08 well, in gmy limited tests el-cheapo usb->ethernet solution was quite reliable, suprisingly, but it doesn't apply to old davicom (dm96) module powered ones Jan 19 20:51:25 yea, ethernet over usb sounds more approriate Jan 19 20:51:52 still, suprisingly, hostmode and dongles are cheap and effective, if you buy el-cheapo qf9700 ones Jan 19 20:52:18 quite compatible with any network infra yoiu may have or want Jan 19 20:52:29 I want someone to buy a license from altair for the scheduler software and put down "assloads of nokias" as the environment Jan 19 20:52:34 look *that* up on your price sheet Jan 19 20:52:43 ethernet over usb for 1k devices could prove problematic on hardware level for obvious reasons Jan 19 20:53:14 :) Jan 19 20:54:49 i wonder what the real life limits of N900 wifi hardware are, when it comes to bandwidth Jan 19 20:55:08 probably limit is in spi bus Jan 19 20:55:24 any ideas how it looks there? Jan 19 20:56:47 ethernet work well through usb? Jan 19 20:57:05 estel just try torrenting something Jan 19 20:57:46 not so easy, you need to ensure stable testing enviromnent with no other limiting factors Jan 19 20:57:47 and Jan 19 20:57:53 flash speed limits you ;) Jan 19 20:58:01 that sucks Jan 19 20:58:06 flash doesn't run fast Jan 19 20:58:08 would need ramdisk Jan 19 20:58:15 something inside ram Jan 19 20:58:18 uploaded Jan 19 20:58:24 oh that flash Jan 19 20:58:25 nulldisk Jan 19 20:58:35 tadzik, true :D Jan 19 20:58:42 a fuse FS that doesn't actually store anything Jan 19 20:58:50 sending zeroes over wifi is fun, too Jan 19 21:01:49 Evening! Jan 19 21:04:23 5 left Jan 19 21:04:38 5 what Jan 19 21:04:43 thanks Jan 19 21:05:00 2995 Jan 19 21:05:09 karma whore Jan 19 21:07:55 Let's see if I can find 5 posts for you :P Jan 19 21:08:03 lol Jan 19 21:14:05 ah, it's why hildon foundation [strike]wasted[/strike] used donated money to buy vbulletin propertiary license Jan 19 21:14:22 upon migration to SMF or PHPBB DocScrutinizer05 would lose his thanks! Jan 19 21:14:24 ;) Jan 19 21:15:56 DocScrutinizer05: Congratulations on your 3k'th thanks Jan 19 21:16:51 can you cancel a "thanks"? Jan 19 21:17:44 kerio: yes Jan 19 21:18:37 :D Jan 19 21:19:03 Thanked 2,999 Times in 537 Posts Jan 19 21:23:01 I think I have more posts left to thank than you to unthank :P :P Jan 19 21:34:22 ._. Jan 19 21:34:38 well, i only had one thanked post Jan 19 21:34:42 MrPingu: \o/ :-D Jan 19 21:36:47 Hi. Anyone uses Kasvopus on n900 here? Jan 19 21:36:58 Its a facebook client. Jan 19 21:39:49 Ex-Opesa: used, but that's a long time ago. Now I use m.facebook.com in MicroB Jan 19 21:41:17 on a sidenote, FYI (to whom it may concern): Jan 19 21:41:22 > You probably know that the SSL setup on vcs.maemo.org is broken. The Jan 19 21:41:23 > certificate and the private key file do not match. Jan 19 21:41:25 > Jan 19 21:41:26 > The certificate will anyway expire on the 3rd of Jan, 2013. Hope that Jan 19 21:41:28 > the Hildon Foundation can request a new one, as soon as the domain Jan 19 21:41:29 > ownership is transferred. If somebody needs help with requesting a Jan 19 21:41:31 > certificate please contact Niels or myself. Jan 19 21:41:32 > Jan 19 21:41:44 MrPingu: Okay. I was just wondering why we can't check our messages in it. I mean that is one of the basic and most important feature of a fb client Jan 19 21:42:19 MrPingu: Thanks for replying though. I might post on forum later. Jan 19 21:43:37 Ex-Opesa: YW, I might dissapoint you but I think FB client development for N900 is dead Jan 19 21:48:50 MrPingu: Yeah. zzz Jan 19 21:49:31 I might be that fMobi from ovistore is decent but it costs you Jan 19 21:49:37 *it Jan 19 21:50:52 Yea Jan 19 21:51:05 I will check it too Jan 19 22:01:39 Btw is there a way to reset word prediction (dictionary)? Jan 19 22:02:51 i didn't know anyone used it Jan 19 22:02:52 :o Jan 19 22:04:02 kerio: sometimes I use the autocomplete but I mixed 2 languages Jan 19 22:12:57 MrPingu, on truecrypt thread, there are instructions how to nuke dictionary database Jan 19 22:18:57 MrPingu: I'm quite sure it has a file storing all the new words somewhere Jan 19 22:19:43 MrPingu: teach it a new word, then immediately check which file changed during last 5 min Jan 19 22:19:56 find cmd is your friend Jan 19 22:20:32 /home/user/.osso/dictionaries/.personal.dictionary Jan 19 22:21:16 or that :-D Jan 19 22:22:21 NIN101: thanks :) Jan 19 22:23:13 yw Jan 19 22:30:57 NIN101: though... "don't provide a solution, rather teach how to find it" Jan 19 22:31:13 OWTTE, from Jan 19 22:31:23 ~question Jan 19 22:31:24 rumour has it, question is If you have a question and want people to give useful answers, make sure you have read this first: http://www.catb.org/~esr/faqs/smart-questions.html Jan 19 22:33:47 http://www.catb.org/~esr/faqs/smart-questions.html#idp30009456 actually Jan 19 22:38:03 meh, I am not hear as a teacher. Jan 19 22:38:05 *here Jan 19 22:38:11 although of course that's reasonable, but anyway. Jan 19 22:39:10 NIN101: we did a good job complementing each other Jan 19 22:39:12 :-) Jan 19 22:40:22 and actually I just took the opportunity to explain why I didn't answer the way you did, though I could've done and it would've been less effort for me Jan 19 22:40:50 no criticism intended Jan 19 22:41:35 ok :-). Jan 19 22:46:41 This way it was less effort for me too :P Jan 19 22:51:40 this might trip the trunaround point sooner than you might have thought Jan 19 22:52:26 NIN101: any clue which format is that file ? ispell? Jan 19 22:53:55 aspell, gnome-dictionary...? Jan 19 22:54:04 Anyway I am off to bed, good night folks! Jan 19 22:54:58 n8 Jan 19 22:55:22 /home/user/.osso/dictionaries/.used.dictionary: 8086 relocatable (Microsoft); /home/user/.osso/dictionaries/.personal.dictionary: lif file; :-S Jan 19 22:55:24 DocScrutinizer05: I don't recall anymore Jan 19 22:56:24 alas my dicts are basically unused and virgin, since I never enabled this spell correction/prediction Jan 19 23:09:37 hi everyone Jan 19 23:10:04 where can I download packages for OS2008? Jan 19 23:10:26 It looks like most repositories are down Jan 19 23:10:46 at least repository.maemo.org Jan 19 23:10:59 are there any other good choices? Jan 19 23:11:02 ? Jan 19 23:13:11 anyone know? Jan 19 23:13:54 I need to install opera Jan 19 23:22:11 anyone? Jan 19 23:24:04 allotrope94: You could check the topic, discover the first link, check the last update of the OP and then follow the given link. Jan 19 23:24:51 OP? I tried loading the install files on the N810 and installation failed. Jan 19 23:25:12 eyescream: which topic? Jan 19 23:25:55 it looks like repository.maemo.org is down Jan 19 23:26:04 hi David Jan 19 23:26:29 is it still possible to install packages on the N810? Jan 19 23:26:33 /topic and Original Post. Jan 19 23:26:40 Or just this: http://talk.maemo.org/showthread.php?t=88707 Jan 19 23:26:53 oh I see :) Jan 19 23:28:27 i'll check that out Jan 19 23:31:17 cool Jan 19 23:31:36 I have an n810 btw...not an n900, eyescream Jan 19 23:31:44 should this still work? Jan 19 23:32:41 allotrope94: From a cursory look, the repos hold the files for other maemo versions, too. If you specify your version, it should work. Jan 19 23:32:49 cool beans Jan 19 23:33:18 allotrope94; i don't know about the backup repos, but, repository.maemo.org should be restored next week we think, so latest at that point normality will also return to N810 Jan 19 23:34:04 nice Jan 19 23:34:19 ShadowJK: Is there an estimate if and when they will be back up next week? Jan 19 23:36:58 eyescream; "next week" is just my own guess. Best case 9am EET monday +3hrs, worst case "longer" Jan 19 23:37:22 I see. Might as well use Valve time. :) Jan 19 23:38:32 That's based on Nokia ppl returning to work after weekend, at 9am local time monday, and immediately implementing requested changes, plus 3 hours for dns caches worldwide to expire and refresh Jan 19 23:40:23 That seems incredibly likely. Jan 19 23:40:43 lol. :P Jan 19 23:41:17 I most certainly will not be holding my breath. :P Jan 19 23:45:14 I have a N800, a N810, 1 broken N900 and 3 working N900 Jan 20 00:08:41 we got a DNS for repository.maemo.org already pointing to IP of the VM originally assigned for that purpose. We found out that akamai polling stage.maemo.org instantly killed the rack's uplink. We took down stage.m.o VM and things somewhat became normal, but Eero realized that actaully stage.m.o is our new *real* repository.m.o. Eero implemented a firewall redirection from repo.m.o IP to stage.m.o IP/VM. When we started up stage.m.o VM Jan 20 00:08:43 again, it again instantly killed *.m.o by congestion of some bottleneck in uplink. At that stage we shut down stage.m.o again and called it a week on Friday afternoon. Monday we will investigate whether maybe the akamai->stage.m.o path still was open and thus starting of stage.m.o VM caused same issue we initially seen on Wednesday Jan 20 00:12:21 you may quote me on that, whenever somebody asks what's up with repository and when it will come back to life Jan 20 00:12:51 also see last update on http://talk.maemo.org/showthread.php?p=1314588 Jan 20 00:15:42 PLEASE EVERYBODY help testing of moved maemo.org infra. We know repo doesn't work, and we know tmo can't send mails. Please everybody check everything you can and report stuff that started to fail with the move during last 7 days. THANKS Jan 20 00:24:52 allotrope94, eyescream, there you have it from the horse's mouth :-) Jan 20 00:26:52 well nice update. Jan 20 00:28:47 * DocScrutinizer05 neighs Jan 20 00:32:26 DocScrutinizer05: Thanks for the update. Jan 20 00:32:31 ShadowJK: Thanks for the highlight. :P Jan 20 00:32:32 yw Jan 20 00:36:09 stage.maemo.org. 300 IN A 188.117.59.205 Jan 20 00:36:32 when vanishing, might fix akamai DDoS attack for good Jan 20 00:37:16 blocking that path on firewall might help for an interim bandaid Jan 20 00:37:56 (that's what I suspect we might have missed to configure correctly on Friday) Jan 20 00:38:35 repository.maemo.org. 300 IN A 188.117.59.203 Jan 20 00:39:43 got redirected on FW/NAT to 188.117.59.205, but sh*t hit the fan as soon as we started the stage VM Jan 20 00:41:11 I wonder if it might be easier to change the NIC IP on stage VM Jan 20 00:42:11 AIUI stage as well as autobuilder is X-Fade's domain of expertise Jan 20 00:42:53 oooh, ps in "PLEASE EVERYBODY": autobuilder and drop also don't work yet Jan 20 00:43:52 just wondering, is there any issues with devel atm? Jan 20 00:45:51 or did I just add the repository wrongly in the ham version of sources.list (in the sources.list.d directory)? Jan 20 00:46:14 I get the following cryptic error: Jan 20 00:46:19 W: Failed to fetch http://repository.maemo.org/community/dists/fremantle/free/i18n/Translation-sv_SE.bz2 Unable to connect to repository.maemo.org http: Jan 20 00:46:54 this happens for all maemo repositories Jan 20 00:47:00 extras, testing and devel Jan 20 00:47:26 repository.maemo.org is offline. see the topic. Jan 20 00:47:38 sorry, just wondering, is there any issues with devel atm? s/devel/repository/ Jan 20 00:47:40 huh Jan 20 00:47:56 ohh, I'm very sorry Jan 20 00:48:04 I read the topic before asking but somehow missed it anyway Jan 20 00:50:15 Quite a rough few days, isn't it? DocScrutinizer05, are we just waiting for RMO's DNS entries to get updated? Jan 20 00:50:22 hmmm interesting, the link which is told to show a hosts file with the correct IP is dead Jan 20 00:50:33 (@ http://mwkn.net/2013/01/community.html ) Jan 20 00:50:57 FIQ: http://talk.maemo.org/showthread.php?t=88707 Jan 20 00:51:09 ty Jan 20 00:51:15 You're welcome. Jan 20 01:02:05 Hurrian: I really don't know how to explain situation more verbosely than what I did ending 11 postlines above your initial post Jan 20 01:02:22 mmm. Jan 20 01:02:29 freemangordon, [code](/opt/QtSDK/Simulator/Qt/gcc/imports/org/hildon/components/libhildonplugin.so: undefined symbol: _ZTV12HildonPlugin) [/code] Jan 20 01:02:29 The same appears on device too. Jan 20 01:02:54 oops, forgot to remove bbtags after copying from tmo Jan 20 01:03:43 qwazix, _ZTV12HildonPlugin is manged Jan 20 01:03:56 AFAIK hildon library is C Jan 20 01:04:12 ~#maemo repo-down is http://mg.pov.lt/maemo-irclog/%23maemo.2013-01-20.log.html#t2013-01-20T02:08:42 Jan 20 01:04:12 okay, DocScrutinizer05 Jan 20 01:04:27 Pali, HildonPlugin is just a name of a qml plugin by Marxian Jan 20 01:04:36 not related to real hildon Jan 20 01:04:47 FIQ: http://mg.pov.lt/maemo-irclog/%23maemo.2013-01-20.log.html#t2013-01-20T02:08:42 Jan 20 01:04:55 Hurrian: http://mg.pov.lt/maemo-irclog/%23maemo.2013-01-20.log.html#t2013-01-20T02:08:42 Jan 20 01:05:03 qwazix, ok Jan 20 01:05:10 Pali, see http://talk.maemo.org/showthread.php?t=85955 Jan 20 01:06:07 lol, these temporary mirrors give bizarre errors Jan 20 01:06:43 _ZTV12HildonPlugin <- likely missing extern "C" Jan 20 01:07:11 wait, c++filt says it's a vtable Jan 20 01:07:25 oh, one of them was actually due to a typo Jan 20 01:07:26 jon_y, it's a qt lib Jan 20 01:07:32 otherwise they're still strange Jan 20 01:08:02 one of them had bz2 files which bzip refused to open ("returned with an error code (1)"), another one just give hash sum mismatch errors Jan 20 01:08:14 qwazix: isn't QT C++? Jan 20 01:08:34 "one of them" being skeiron, the other one merlin1991's Jan 20 01:08:38 FIQ, I get the same with gzip Jan 20 01:08:43 Qt is C++ in the same sense that C++ is C Jan 20 01:08:54 the file contains 'no' if you open it with a browser Jan 20 01:09:04 now let's try the third provided Jan 20 01:09:14 from webdrop.it Jan 20 01:09:25 jon_y, qt is C++ plus some things Jan 20 01:09:43 C++++ :P Jan 20 01:09:47 I don't get it, you mixed C with C++, obviously you get mangled symbols Jan 20 01:09:53 *wedrop.it Jan 20 01:10:11 jon_y I didn't mix C with C++ Jan 20 01:10:15 khm: what does that supposed to mean? Jan 20 01:10:24 jon_y: Qt is a superset of C++. Jan 20 01:10:37 ok Jan 20 01:10:58 It's qt code (not written by me) but it looks like valid Qt code. Jan 20 01:11:47 well, whatever it is, your vtable export is missing Jan 20 01:12:04 let's see if this one works Jan 20 01:12:29 * qwazix is looking up vtable export Jan 20 01:12:46 qwazix: symbol export Jan 20 01:13:12 vtables are compiler generated Jan 20 01:13:36 jon_y I think I found it Jan 20 01:13:43 let me make sure Jan 20 01:13:54 likely present in the object that has the HildonPlugin defintion Jan 20 01:14:18 hmmmmmm Jan 20 01:14:30 both wedrop and merlin1991's repositories gave hash mismatch errors Jan 20 01:14:37 the third one gave a bzip error Jan 20 01:14:39 w/e Jan 20 01:14:50 nah, it wasn't that... Jan 20 01:15:09 does this line seem ok? Q_EXPORT_PLUGIN2(hildonplugin, HildonPlugin) Jan 20 01:15:26 check your compiled objects Jan 20 01:15:50 no amount of syntax checking will tell you if it was actually exported Jan 20 01:15:57 so the repositories don't work? darn Jan 20 01:16:01 how do I check them? Jan 20 01:16:42 compile your code? Jan 20 01:16:55 * jon_y don't use IDEs Jan 20 01:16:57 Yes I got the compiled objects Jan 20 01:17:03 I have no idea how you do that Jan 20 01:17:14 what can I do with them? Jan 20 01:17:14 now use nm or strings to find it Jan 20 01:17:38 nm .o | grep HildonPlugin Jan 20 01:17:43 ok Jan 20 01:21:17 only _ZTV12HildonPlugin Jan 20 01:26:56 qwazix: T 00000000 _ZTV12HildonPlugin? Jan 20 01:27:15 whole line: U _ZTV12HildonPlugin Jan 20 01:27:25 that means it is missing Jan 20 01:27:29 with some leading whitespace Jan 20 01:27:51 are you sure this is the object that holds the definition? Jan 20 01:28:07 I grepped all the files to make sure Jan 20 01:28:17 but this should be the one, yes Jan 20 01:28:33 I think i've got some studying to do Jan 20 01:28:38 hmm, not sure, but I'll be soing some tests Jan 20 01:29:08 jon_y, source is on tmo link I pasted to pali above Jan 20 01:29:14 page 2 Jan 20 01:29:21 if you want to have a look at it Jan 20 01:29:37 I'll check the qt plugin example to hunt for differences Jan 20 01:36:45 qwazix: maybe http://gcc.gnu.org/faq.html#vtables can help Jan 20 01:36:56 I have no idea how to do this in QT Jan 20 01:37:05 continuously getting the error http://pastebin.com/3QcsShwx . thru syslog . Jan 20 01:37:14 jon_y, thanks I'll check it out Jan 20 01:37:17 any help Jan 20 01:37:40 cpu % is 100 Jan 20 01:38:39 jon_y, qt generating some special variables for plugins Jan 20 01:38:58 there is qt macro for plugin exported by qt code Jan 20 01:39:24 so hyw is the class missing the vtable? Jan 20 01:39:26 or that vtable error can happen also when you forgot to generate and link qt MOC files Jan 20 01:39:26 *why Jan 20 01:39:34 ok Jan 20 01:39:47 see moc program Jan 20 01:39:53 UberNeo, maybe messaging ui theme is to blame? Jan 20 01:40:28 what should I do .. in this case ? Jan 20 01:40:31 Pali, moc_* files are there in the build folder if that's of any use Jan 20 01:40:39 jon_y: each class in Qt should have Q_OBJECT macro in declaration Jan 20 01:40:46 and that macro defining some functions... Jan 20 01:40:50 UberNeo, revert to default theme if you changed it Jan 20 01:41:01 and moc generating that functions Jan 20 01:41:17 and if you forgot to link moc cpp files, then you will see linker errors Jan 20 01:41:23 UberNeo, I also see messaging failing to get a handle name Jan 20 01:41:28 in Q_OBJECT are only function headers Jan 20 01:41:30 maybe deleting the messages would help Jan 20 01:41:44 Not sure, somebody more experienced should see it Jan 20 01:41:56 qwazix: see Pali's message Jan 20 01:42:23 changed theme .. no effect Jan 20 01:42:33 jon_y, ok Jan 20 01:42:57 actually .. I have my /opt and /users directory on the uSD card .. Jan 20 01:43:31 UberNeo, I'd say to purge rtcom-messaging-ui and reinstall but repos are down Jan 20 01:43:45 was working perfectly fine . I just copied the whole stuff in another uSd card .. so it starts showing these errors Jan 20 01:43:55 UberNeo: why? uSD card might be faster in itsef, but it is still connected by only 2 links vs 4 for eMMC Jan 20 01:44:10 Maybe borked messaging db Jan 20 01:44:25 my eMMC is corrupted .. Hardware fault Jan 20 01:44:35 oh by the way - copying R/W mounted FS isn't best idea Jan 20 01:45:04 so most of the stuff on uSd Jan 20 01:45:40 I would try to get new one - there's loads of it on market at the moment - all you need is look for it (you might have to look for it in broken devices, but still you can find one quite easily) Jan 20 01:46:36 UberNeo - how heavy corruption is it? Jan 20 01:46:50 enough to involve Interpol :) Jan 20 01:47:13 WizardNumberNext: i havent copied mounted FS .. removed the usd . copied the contents on my usd and then copied it on another usd Jan 20 01:47:18 maybe you would be able to use ext4 on it created with -ccc Jan 20 01:47:41 I mean /home and stuff Jan 20 01:48:48 UberNeo: what about swap? Jan 20 01:49:12 I would move swap as this would create loads of problems if it would stay on corrupted device Jan 20 01:49:42 maybe you can try RescueOS and export the partitions to a PC Jan 20 01:49:56 Nokia-N900:/etc/systemui# cat /proc/swaps Filename Type Size Used Priority /dev/mmcblk0p3 partition 1804280 20444 -1 Jan 20 01:49:57 and attempt to recover via dd or something Jan 20 01:50:08 haha It worked :) Jan 20 01:50:36 I just needed to move the declaration of the plugin class to a header file instead of a source file Jan 20 01:50:38 swap seems to be on uSd Jan 20 01:50:56 get swap on uSD, if your eMMC is corrupted you have more problems with swap, then without (or on uSD) Jan 20 01:50:56 jon_y, Pali, thanks for the pointers Jan 20 01:51:16 mmc0 is eMMc Jan 20 01:52:02 do you run GNU/Linux on PC? Jan 20 01:52:07 this app http://www.cgsecurity.org/wiki/TestDisk_Download is pretty good at recovering stuff Jan 20 01:52:16 runs on Linux AND Windows Jan 20 01:52:20 I am on MAc Jan 20 01:52:43 here is my mount and free logs Jan 20 01:52:44 http://pastebin.com/vPbxSrcm Jan 20 01:53:09 if you do, then using pipe should enable you to move whole mmc0 to PC (partition by partition is easier to cope with) Jan 20 01:53:21 then you have SSH Jan 20 01:53:31 you can use pipe on N900 Jan 20 01:53:41 why yes TestDisk runs on Mac too Jan 20 01:54:08 I used g_file_stuff to export the partition over USB Jan 20 01:54:22 or was it g_file_storage or something Jan 20 01:54:29 mkfifo /some_pipe; scp /some_pipe user@host:/file; dd if=/dev/mmcblk0p1 of=/some_pipe Jan 20 01:54:30 anyway, mass storage mode Jan 20 01:54:36 whole mmc0 to pc?? Jan 20 01:55:04 g_file_storage is gadget which makes pendrive Jan 20 01:55:09 that would work as well Jan 20 01:55:32 yes, but do it partition by partition - splitting might be tricky Jan 20 01:55:34 WizardNumberNext: ? Jan 20 01:55:53 WizardNumberNext: can't UberNeo just export /dev/mtdblock0? Jan 20 01:56:06 and you can safely omit swap - you do not need outdated memory contents ;) Jan 20 01:56:08 *mmcblk0 Jan 20 01:56:50 jon_y and UberNeo: yes, he can export mtdblock0 as well in rescueOS Jan 20 01:57:15 via pendrive facility (uses g_file_storage) Jan 20 01:57:37 I just did that when my n900 screen died Jan 20 01:58:25 it does work well - I unbricked my N900 twice using rescueOS and USB-storage (g_file_storage on N900, but it doesn't matter, because afair it was called there usb-storage) Jan 20 01:58:33 modprobe g_file_storage file=/dev/ Jan 20 01:58:38 not sure then .. howcome . the things were working fine on my old usd Jan 20 01:59:08 UberNeo: you can have another copy Jan 20 01:59:19 how to check . which Swap its is using ? Jan 20 02:00:06 but on PC/MAC try to use dd_rescue for getting data of it - dd_rescue would get more data in better state and it does create sparse file and doesn't fill holes, where it cannot read source Jan 20 02:00:16 cat /proc/swaps Jan 20 02:00:23 if I am not wrong Jan 20 02:00:26 second Jan 20 02:00:46 yes thats the file Jan 20 02:01:13 WizardNumberNext: (( mmc0 is eMMc)) no, not always Jan 20 02:01:19 didn't change from time of 2.6.2, so definitely on N900 would be same Jan 20 02:02:11 DocScrutinizer05: I never have seen eMMC on mmcblk1, but yes - it can be in different place Jan 20 02:02:26 rescueOS has it swapped backwards, at least in my case Jan 20 02:02:36 if you would detect uSD first, then uSD would be 0 and eMMC would any next Jan 20 02:02:51 WizardNumberNext: please stop talking UberNeo thru an adventure that's unrelated to the problem he came here for Jan 20 02:03:11 jon_y and UberNeo: afaik RescueOS does swap it around Jan 20 02:04:09 OK, anyway I still think he should make proper backup and get rid of swap from his eMMC Jan 20 02:04:20 WizardNumberNext: in early boot mmc-port0 is called mmc0 and internal port is mmc1. Later during boot it gets renamed, but not everywhere Jan 20 02:05:10 doc is it up to udev? just being courious and lazy to check it out by myself Jan 20 02:05:26 yes, it's a binary called by udev Jan 20 02:06:20 I was actuallythinking they did it easier way (something like with ethX and such stuff) Jan 20 02:06:48 they like to complicate their lives, I see Jan 20 02:07:34 the complicator's gloves! Jan 20 02:08:04 unfortunately I belong to them in a way - I like to complicate my life as well Jan 20 02:08:17 any final suggestion . confused here .. how to debug http://pastebin.com/3QcsShwx Jan 20 02:08:47 like, when I changed partition type of first partition on eMMC to 83 - endless loop of booting Jan 20 02:09:05 UberNeo: how did you manage to move /home to uSD? Jan 20 02:09:44 I am interested as well, because fstab is dynamic and it is created every boot Jan 20 02:10:03 at least looks like Jan 20 02:10:36 WizardNumberNext: yes, and some 5 poeple talked one user with similar problem thru such switch here on this very channel a whole night long, several months ago Jan 20 02:11:24 doc what would happen if I would make fstab immutable? Jan 20 02:11:29 it been not exactly trivial, and in the end I doubt we missed to switch swap Jan 20 02:11:38 DocScrutinizer05: its u . who have helped me .. more than a year back . 31ss Oct 2011 Jan 20 02:13:13 http://mg.pov.lt/maemo-irclog/%23maemo.2011-10-31.log.html Jan 20 02:13:17 here are the logs Jan 20 02:13:22 DocScrutinizer05: Jan 20 02:14:23 DocScrutinizer05: this is magic line . which u gave me "[03:31] umount /media/mmc1; sfdisk /dev/mmcblk1 < ok, I been wrong, it's not a binary Jan 20 02:21:03 WizardNumberNext: read those chanlogs and see how to *really* swap eMMC with uSD Jan 20 02:21:14 for *whole* system, incl swap Jan 20 02:21:40 UberNeo: I suggest you blame your new uSD, get a better one and do the copying again Jan 20 02:22:04 copy of *whole* physical uSD Jan 20 02:22:09 DocScrutinizer05: you mean those logs? I am just reading them - it is quite interesting Jan 20 02:22:23 incl bootsector/MBR and all Jan 20 02:22:25 DocScrutinizer05: .. i think .. i have done something wrong . while copying /opt and /users from my old uSd to new uSd . i did it using the Konqueror Jan 20 02:22:51 UberNeo: what about hidden files? Jan 20 02:23:19 if you don't do a physical diskimage copy, you need to create partitions on new uSD exactly same way they were on old one Jan 20 02:23:41 and I am not sure if konqueror keeps permissions (user on N900 have uid of 29999, Linux normaly created users from 1000) Jan 20 02:24:08 nah, copy by konqueror is prone to fail in many ways Jan 20 02:24:21 for such stuff Jan 20 02:24:45 then hw shud I copy Jan 20 02:24:54 yep Jan 20 02:25:21 dd if=/dev/mmc24 of=/dev/mmc26 Jan 20 02:25:31 on your linux desktop pc Jan 20 02:25:39 something similar to that Jan 20 02:26:30 here mmc24 is my old usd and mmc26 is the new usd .. Jan 20 02:26:43 yep Jan 20 02:27:29 or you use an intermediate storage on your HDD: dd if=/dev/mmc7 of=/tmp/myuSDimage Jan 20 02:27:52 dd if=/tmp/myuSDimage of=/dev/mmc7 Jan 20 02:28:03 i only have one card reader .. and can read/write only one usd at a time . can i save it on my Desktop and then transfer to new usd Jan 20 02:28:09 and of course swap uSD in between the two cmds Jan 20 02:28:23 and don't forget to sync after second cmd Jan 20 02:28:47 sync? Jan 20 02:28:47 since driver will write buffers for almost one minute after dd finished Jan 20 02:28:57 the command "sync" Jan 20 02:29:05 ok Jan 20 02:29:24 UberNeo: just type sync after those commands, then wait till you will get prompt Jan 20 02:29:36 only after sync returns, you may remove uSD Jan 20 02:29:46 I do not remember if sync works from user or you have to do it as root Jan 20 02:29:55 root Jan 20 02:30:13 just like dd Jan 20 02:30:30 do i need the sync command . after both dd statements Jan 20 02:30:47 you can actually make yourself extra sure - press ALT+PrintScreen+S, then dmesg as see if Emergency Sync is complete Jan 20 02:30:56 after second Jan 20 02:31:40 this one which I told is sure as hell, as it is done from kernel and kernel won't say it synced until it is Jan 20 02:32:09 it may take a while to sync and it depends on size of your RAM, so be patient Jan 20 02:33:03 okk . doing that Jan 20 02:34:16 and check, if card gets mounted automatically, if so unmount it before you start to copy! Jan 20 02:35:39 UberNeo: forget alt-prtscr-S, it most probably isn't even enabled on your standard linux PC. You can use `sync` after writing to uSD before removing it. Works reliably Jan 20 02:36:13 yep, that is actually an extremely important advice Jan 20 02:36:18 it does work reliably - it safed my day many times Jan 20 02:37:35 UberNeo: after inserting uSD, do a `mount` and see if anything points to the card. If so, umount it Jan 20 02:37:44 : I just use SysRq as it is much faster to do (by keyboard hitting), so no wonder I always like to say people about this one Jan 20 02:38:02 there mustn't any mounts on the card before you dd to it Jan 20 02:38:29 SysReq has to get enabled. Usually it isn't Jan 20 02:38:46 ^^ or your copy might get highly inconsistent and even fsck could have big problem to fix it Jan 20 02:39:34 DocScrutinizer05: actually I even do not know if it is enabled in Debian, as I have it enabled in kernel configuration (my own kernel) Jan 20 02:39:57 uhhh Jan 20 02:39:59 really Jan 20 02:40:43 *stuff updates* "error, couldn't do some thing as repo is down, do you want to check errormsg?" *checks, and as I exit the UI, it updates AGAIN* Jan 20 02:40:49 ham :( Jan 20 02:40:57 FIQ: are you asking me? Yes, for last few years I use my own kernel (I configure it, I compile it) - from time 2.6.2 Jan 20 02:41:15 HAM is cute, no? X-P Jan 20 02:41:20 WizardNumberNext: nah, "really" was @ ham stupidity Jan 20 02:41:26 DocScrutinizer05: yaay Jan 20 02:41:39 my new usd card is of higher capacity so I have given more blocks to it .. Jan 20 02:41:40 I see - I finished typing and read your message Jan 20 02:42:01 do it make any effect if the source and destnation are of different sizes Jan 20 02:42:03 errr Jan 20 02:42:18 let's NOT press "information" this time Jan 20 02:42:19 UberNeo: what do you mean? To any particular partitions? Jan 20 02:42:40 Target -- /dev/sdd2 21506 28674 7341056 83 Linux Jan 20 02:43:06 actually makeing bigger partition which contains DDed FS wouldn't make FS bigger Jan 20 02:43:09 apparently I need PC Suite to install latest thumb update Jan 20 02:43:10 Source -- /dev/sdd2 277 965 5254314 83 Linux Jan 20 02:43:15 * FIQ doesn't believe that Jan 20 02:43:42 FIQ: that's really stupid, I agree Jan 20 02:43:57 though repo@maemo is down so that is probably the problem Jan 20 02:45:24 UberNeo: then you should go the "hard" way, create proper partitions similar to old card, just extending the "/home" partiton or "MyDocs" partition, whatever you like. Then mkfs on each partition, and finally cp -A stuff from old to new card, partition by partition Jan 20 02:45:42 UberNeo, if you want to use bigger patition, then you need to mount your dump (if you did it partition by partition) and make new FS in place of bigger partition, then copy ALL files (cp -a) Jan 20 02:46:02 overlap Jan 20 02:47:34 actually, if you use PowerKernel you can use ext4 (quite fast I think) Jan 20 02:47:49 mkfs.ext4 /device Jan 20 02:48:20 not sure if jfs is more appropriate for flash devices Jan 20 02:48:24 but actually you can safe some space by: 'mkfs.ext4 -J size=16 /device' Jan 20 02:48:46 jon_y: do you mean JFS or JJFS2 Jan 20 02:48:54 or whatever it is called Jan 20 02:49:02 or you can try to make it *work* by not doing any such tweaks, first Jan 20 02:49:50 yeah, maybe I'm confusing the 2 Jan 20 02:49:52 jfs is for NAND, not for mmc Jan 20 02:49:52 that is just tweak of size of journal, but if partition is smaller then 1GiB, then anyway he would get 16MiB journal Jan 20 02:50:15 jjfs2 actually Jan 20 02:50:38 DocScrutinizer05: if i have to use the same partition size . then there is no point to move it to new usd card . Jan 20 02:50:46 jon_y: you cannot use JJFS2 or UBIFS directly on block device - it have to be flash or be an emulated flash Jan 20 02:50:57 UberNeo: please reread my instruction Jan 20 02:51:40 >>*similar* to old card, just *extending* the "/home" partiton or "MyDocs" partition<< Jan 20 02:52:31 UberNeo: just do DD dumps of partitions (skip if you did it partition by partition already), make new FSs on new card and 'cp -a' for each FS Jan 20 02:53:02 huh? Jan 20 02:53:49 cp -a a dd dump? I guess he had to loopmount the dump before, which might already be way more cumbersome than a proper tarball Jan 20 02:54:21 oh sorry I think with shortcuts Jan 20 02:54:23 so i can cp -a the dumps created by dd as single file .. Jan 20 02:54:28 no! Jan 20 02:54:42 WizardNumberNext: ok, JFS Jan 20 02:54:44 sorry it is fault of my shortcut thinking Jan 20 02:54:58 you have to moun t those dumps Jan 20 02:55:21 or you create tar archives of your old card Jan 20 02:55:22 mount -o loop,ro /dump_file /mount_point Jan 20 02:55:31 this would work as well Jan 20 02:55:47 and might be faster Jan 20 02:55:55 depends on computer Jan 20 02:56:17 or you have two sd card slots on your PC and simply cp -a, like you did with konqueror before Jan 20 02:56:35 on mine it is definitely faster - no ordinary HDD is able to write so damn fast Jan 20 02:57:18 but still highly advisable to mount it read-only Jan 20 02:57:34 loopmount is not exactly trivial stuff for beginners Jan 20 02:57:38 nothing can touch anything on it Jan 20 02:57:55 mostly works out of the box Jan 20 02:58:06 never had any problem Jan 20 02:58:17 still my old . dd command is ruuning .. which copying to bigger size usd partion (as compared to source partion size) Jan 20 02:58:35 your problem however is talking beginners thru procedures Jan 20 02:58:39 just kill it CTRL+C Jan 20 02:59:03 or let it finish Jan 20 02:59:17 the use gparted to extend the MyDocs partition Jan 20 02:59:24 then* Jan 20 02:59:32 might work Jan 20 02:59:36 fdisk will work for extending ?? Jan 20 02:59:41 no Jan 20 02:59:48 and there is backup, so it is anyway perfectly safe **** ENDING LOGGING AT Sun Jan 20 02:59:58 2013