**** BEGIN LOGGING AT Wed May 23 02:59:59 2012 **** ENDING LOGGING AT Wed May 23 04:35:22 2012 **** BEGIN LOGGING AT Wed May 23 04:36:15 2012 May 23 05:34:29 I think the volume is low when calling, although I've set the volume to the maximum May 23 05:34:54 I set it by the volume button on the top left side of N900 May 23 05:35:30 but do you set it DURING a call? May 23 05:35:32 or before a call? May 23 05:35:58 because they are different May 23 05:36:55 Sc0rpius, I set it during the call May 23 05:37:40 Actually, the volume is big enough, but I think it can be higher at the maxmum level May 23 05:47:16 are there any phones available or being made that would suit as a n900 replacement? May 23 05:59:59 basically no May 23 07:33:17 think i need to coax one guy who bought a n9 to sell me his old n900 May 23 07:44:30 why? May 23 07:44:33 jaska? May 23 07:45:10 I haven't bought N9, but I got N900 in ideal condition and reinforced SB port to sale May 23 07:55:59 who remember name of fm radio driver module for N900? May 23 08:21:01 javispedro, nice to know, but why haven't You released update sources to repos? May 23 08:33:53 Estel_: there is no capital on "You" or "Your" in English. May 23 08:33:57 it's not German :P May 23 08:34:05 it's just "you" or "your" May 23 08:42:18 teotwaki, know that already ;) Let's just assume it's a folklore of mine May 23 08:42:45 s/folklore/quirk/ May 23 08:42:53 whatever May 23 08:43:58 If it makes you feel better, You may assume I'm using elisabethian type of english ;) May 23 08:44:18 s/You/you/ May 23 08:44:19 Estel_ meant: If it makes you feel better, you may assume I'm using elisabethian type of english ;) May 23 08:44:20 ;) May 23 08:47:26 actually, even 5 centuries ago, "you" didn't take a capital :) May 23 08:47:44 it took in letters May 23 08:47:49 or other forms of respect May 23 08:48:07 strangely enough, "I" too, but I live still May 23 08:48:13 well, self respect isn't bad thing ;) May 23 08:48:36 FYI, it was also the case in most languages, and, it's still up to this date May 23 08:48:48 not only german, same for Polish, for example May 23 08:49:02 although,. in polish, like in english, common writing doesn't include "You" as a capitalised form May 23 08:49:10 but, it still live in letters and personal documents May 23 08:49:19 so i may assume polish is somewhere in middle ;) May 23 08:50:00 I'd love a reference to "you" being capitalised in personal writings... Because even though I'm currently studying heterologies in English litt, I've never come across that. May 23 08:50:30 hm, odd enough May 23 08:51:18 well, other than when it's the first word of a sentence, obviously :) May 23 08:53:02 obviously. wait, im searching May 23 08:53:12 Estel_: this being said, when I say "I'm currently studying", that's probably pushing it -- my girlfriend is doing a masters in that, I'm just tagging along when not at work. This being said, we are currently working on Melville's Clarel for a dossier, so I guess I'm entitled to saying I do study it :P May 23 08:53:32 for sure May 23 08:53:44 (18000 lines is quite the poem :P) May 23 08:57:19 it seems that You're right, and shakespare used it only because it was reffered to King or Queen May 23 08:57:43 capitalised form of english You wasn't used since 12th century :/ I always knew I'm old fashioned ;) May 23 08:57:46 anyway May 23 08:57:58 strangely enough, it's 2nd time I see someone carrying about it, in a short time May 23 08:58:08 + never ever heard about it before May 23 08:58:26 seems like people are turning into etymology hobbyits May 23 08:58:35 hobbysts* May 23 08:58:38 anyway May 23 08:58:50 hobbyists :) May 23 08:59:09 I really like form of capitalised You, She Hew etc, as a form of showing respect, so it sounbds like a great challenge for me to remember about need to drop it ;) May 23 08:59:24 that's why I've called it a folklore of mine May 23 08:59:45 anyway, it was a nice topic for short discussion during recovering OptFS backup of 4Gb in size May 23 08:59:59 ffs, something crashed, and I got filesystem corrupted beyond repair May 23 09:00:41 device was semi-working, although, any command written in terminal left it in limbo state May 23 09:00:44 (temrinal only) May 23 09:00:47 no power menu etc May 23 09:00:55 I was able to connect through ssh, yet... May 23 09:01:29 commands, despite being "accepted" (new line started after etc) it did nothing May 23 09:01:34 same for shutdown, reboot etc May 23 09:01:40 had to kill Xorg to force shutdown May 23 09:01:50 but optfs got screwed in the process :( May 23 09:29:19 Estel_: if you need to recover files, use photorec May 23 09:29:31 Estel_: and ddrescue to dump May 23 09:30:12 fortunatelly nothiong critical, just annoyance, as recovering take time May 23 09:30:30 I had backup - not very recent ;) But I "mastereD" recovering optfs only and fixing desync issues with rootfs by hand May 23 09:34:57 damn, device doesn't boot maemo even after recovery, WTF? May 23 09:35:15 is it possible that I'll have to do full reflash for the first time even? *headache* May 23 09:36:24 Estel_: you might want to try NIN101's rescueOS May 23 09:36:48 that's if you have some idea where the problem is :\ May 23 09:37:21 thanks, but unfortunatelly, NFC, as I've restored working OptFS backup May 23 09:37:28 (without problems during restore) May 23 09:37:42 I can't imagine how rootfs could get screwed due to such random thing, but who knows May 23 09:38:01 erm, what filesystem are we using for rootfs May 23 09:38:39 ubifs, i think May 23 09:43:28 any way to recover this? it shouldn't require fsck nor have one May 23 09:44:47 framebuffe risn't help as it goes through it OK, but fail to displaye 5 dots May 23 09:44:57 just black screen with backlight on forever, not rebooting May 23 09:45:09 need to do forced power off via power button :( May 23 09:48:58 Estel_: are you still making root filesystems for ez debian chroot? May 23 09:49:09 is there any documentation anywhere about how to make those? May 23 09:49:26 hm? you just need debian partition May 23 09:49:28 image of it May 23 09:49:47 so just do something with a loopback and debootstrap? May 23 09:52:52 it depends, what do You qant to achieve? May 23 09:52:56 well May 23 09:53:08 creating images for easy debian isn't trivial, as installed components must play nicely with Maemo counterparts May 23 09:53:10 i want to achieve a very minimal debian image with testing on it. May 23 09:53:11 see pulseaudio May 23 09:53:21 hm May 23 09:53:27 i think the easiest way to do it might be to take your image, shift it to a slug or a zaurus May 23 09:53:37 mount it there and just uninstall / update stuff. May 23 09:53:46 whenever i try and do similar on my phone it ends up rebooting itself May 23 09:53:51 yea, but installed components won't play nicely with maemo ones, until You don't need sound working etc May 23 09:53:55 (even if i nice -n 19....) May 23 09:54:04 use swap on microSd *only* May 23 09:54:07 disable swpa on eMMC May 23 09:54:14 (it will help Maemo too, not only ED) May 23 09:54:24 im not interested in sound. May 23 09:54:29 i want to play with mosh. May 23 09:54:31 :P May 23 09:54:33 and see Ed thread, ther eyou have proposed settings related to swap May 23 09:54:40 swappiness proposed settings are depreciated May 23 09:54:57 best way is to use my image, then, shrink it by uninstalling everything You don't need May 23 09:55:05 yeah May 23 09:55:07 just mount image as loopback May 23 09:55:14 move content to dedicated partition May 23 09:55:20 (or work on image, although it's slower) May 23 09:55:32 and use chroot terminal or synaptic to get rid of unused stuff May 23 09:55:39 is it normal that it keeps rebooting if i do apt-get update etc? May 23 09:55:48 after that, you might want to send Your image to qole, he will probably like to make it available for download May 23 09:55:49 i read somewhere that if you nice it that doesnt happen. May 23 09:55:50 as minimalistic one May 23 09:56:02 nope, ncie is jsut a workaround, doens't work always May 23 09:56:14 silly watchdog thing. May 23 09:56:14 but it';s sign that You have suboptimal swap setup on your N900 May 23 09:56:18 aaah. May 23 09:56:21 hmmm. May 23 09:56:21 it may crash in anymemory hungry operation May 23 09:56:39 last few pages on Ed threrad were about that May 23 09:56:42 with proposed fixes May 23 09:57:05 thanks i'll take a good look at that thread. May 23 09:57:17 ive been a bit lazy - its 250 pages or so and the s/n is quite low! May 23 09:57:18 :P May 23 09:57:41 :) May 23 09:57:52 np, yw May 23 10:20:01 is there an easy way to share my 3g connection on an existing infrastructure-mode wlan? May 23 10:20:36 mobilehotspot seems to want to create a new ad-hoc... the situation is that the wired isp here sucks today May 23 10:21:12 mobilehotspot is bets You can get May 23 10:21:17 = semi/ad-hoc May 23 10:23:20 does it support connecting to an infrastructure-mode network? May 23 10:23:53 ah, Yuo want to fully re-use May 23 10:23:55 i want to get a connection on a computer with no wireless card but a wired connection to a wireless router May 23 10:23:57 hm May 23 10:23:59 it depends on AP May 23 10:24:08 it it support DUN, no problem May 23 11:50:47 Hi :) May 23 11:51:43 I have recorded that video of wifite, but it only plays on N900. May 23 11:58:18 uploading to YT now, hope it can display it May 23 12:39:00 http://www.youtube.com/watch?v=Xnb6zUw95Ew&feature=youtu.be May 23 12:39:27 It worked :) May 23 12:42:06 merlin1991: The result ^^ encoded via avimux, so it could probably smoother if I took mjpeg encoding, I guess? May 23 15:40:06 MrPingu, nice work :) May 23 16:00:00 hm, until swap fragmentatrion occur, there is only one big free space on swap partition May 23 16:00:07 (ahead of what was written) May 23 16:00:08 Har, Google's front page has a synthesizer app on. May 23 16:00:10 Estel_, there is a speciol patch in kernel block driver for swap devices,. And it is made by Nokia May 23 16:00:10 at least for swap logic May 23 16:00:18 I see May 23 16:00:37 AFAIK, it just writes sequentialy, until all swap space is or was used May 23 16:00:56 only after that it starts to salvage opreviously used blocks May 23 16:00:57 Maybe there is some logic error in it, kicking after "end" of the swap area May 23 16:01:03 quite possible May 23 16:01:18 well, it would be great, if ereswap wouldn't be needed :) May 23 16:03:05 because what I saw in kernel block driver during my times with compcache is that when there are n pages of swap to be written, the driver searches it's mapping table for continuous space of at least n clusters May 23 16:03:31 and there is a mapping table between swap page and physical cluster May 23 16:03:49 (assuming page size == cluster size) May 23 16:04:40 at the end it sounds to me like a bug, damn, I wish there were at least 2-3 more kernel developers here :( May 23 16:07:42 Yea, I wish it too May 23 16:08:00 unfortunatelly, kernel devs are quite hard to find May 23 16:08:10 at least in our little community May 23 16:08:15 Estel_, maybe I should disable this patch (Nokia one) to check how it will behave without it May 23 16:08:27 why not, but what's the thing path should do? May 23 16:08:33 exactly? May 23 16:09:06 Well, it builds a kind of scather-gather list May 23 16:09:11 i.e. May 23 16:09:13 BTW, maybe it's just that searching for n free clusters is always slower than writing blindly "forward" of written space? May 23 16:09:22 * Estel_ listens May 23 16:09:53 The nokia modified swap algorithm is: Find largest free block in swap, start writing to that block until end of block, repeat find largest free blck. Over time, as stuff is allocated and freed, the largest free block becomes small, and unaligned. For optimal speed we want somethning like 1Mbyte alignment. May 23 16:09:59 (on cssu channel) May 23 16:10:06 it maps n(non-sequental) swap pages, to n sequental clusters before writing May 23 16:11:14 ShadowJK, what do you mean by saying "unaligned" May 23 16:11:26 from the flash POV? May 23 16:11:42 yeah May 23 16:12:07 btw, ShadowJK, I hope You've seen credits I've given You on latest package ;) May 23 16:12:25 * Estel_ would not like to see, that someone things it's his concept May 23 16:12:44 hmm, but then it is obvoius where our problems lie - we should not search for the largest, but for the smaller block May 23 16:13:11 which fits our needs May 23 16:13:51 I once plotted (through block dump) byte position of writes on Y axis, and time on x axis, and it showed a steadily climbing line of writes until it reached end of swap, after which it became groups of red dots, which became random red dots later May 23 16:14:03 No, small write is very slow May 23 16:14:31 ShadowJK, i mean - tha smallest block our n swap pages fit into May 23 16:14:54 so it still will be sequental write May 23 16:15:09 (even unaligned) May 23 16:15:12 It will slow down future writes May 23 16:16:11 why, after all there always will be a large enough block to hold the whole RAM, after all there is 768MB swap agains 256 RAM. Maybe I am missing something May 23 16:16:48 Ram isn't dedicated to mirroring swap May 23 16:16:54 :) May 23 16:17:19 ShadowJK, let me explain, maybe I am wrong, but still May 23 16:17:34 freemangordon, if my poor predicting skills permit, it seems to me, that over time, free block will become smaller and smaller May 23 16:17:39 going to total entropy May 23 16:17:47 s/going/heading/ May 23 16:17:49 Estel_ meant: heading to total entropy May 23 16:18:04 (free block on swap logical mapping) May 23 16:18:11 s/block/blocks May 23 16:18:16 well, kernel does not evict RAM pages one by one, but on groups AFAIK May 23 16:18:17 The default kernel.org algorithm is something like "find free space nearest last read/write from/to swap", and this is dreadfully slow on flash :/ May 23 16:18:56 freemangordon, writing them in pages just mean it will head into entropy slower + minimum size of single free block will be single page size May 23 16:18:59 They're not freed in groups, however. That depends on what the apps do May 23 16:19:10 it makes final entroipy less painful, yet, still slowing down May 23 16:19:28 Write to an area of memory that was swapped out: kernel reads in page from swap and marks it free May 23 16:19:31 exactly. That is why over time, red dots on ShadowJK's alghoritm become random May 23 16:19:35 yes, and here is where big enough swap would kick in May 23 16:20:32 we will always have big enough sequental swap space available for the next swapping, if swap space is big enough. Or at least my six sense tells me that :D May 23 16:21:00 hm, on heavy swap usage, it would fail May 23 16:21:01 Btw, if I accidentally load something gigantic, like fennec, swap is filled with a big chunk of what's almost exclusively fennec, and when I close it, that area of swap is almost entirely free, and the onset of swap fragmentation is delayed May 23 16:21:07 i.e. have 200-300 MB of swap used May 23 16:21:16 ...which is not uncommon for browser or EasyDebian things May 23 16:21:24 and it becomes a problem, IIRC May 23 16:21:39 sure, but we can use more that 768, can we? May 23 16:21:41 even when we would search first "big enough" free space May 23 16:21:48 sure thing May 23 16:22:15 maybe it is a matter of finding the optimum swap size then? May 23 16:22:20 well, but, it would surely limit clutterness, yet, searching for first enough swap space still would be little slower than "just write sequentialy", yes? May 23 16:22:34 in my experience "big enough" only delays the slowdown. I have 2gig swapspace, and I usually notice slowdown once about 2.3 gig has been written :) May 23 16:23:21 ShadowJK, that is inevitable with the current algorith May 23 16:23:30 ShadowJK, freeman means optimum swpa size if ekrnel would search for first big enough place in swap May 23 16:23:40 instead of "biggest" free swap May 23 16:23:56 freemangordon, but swap sometimes also write more than 1 page etc May 23 16:24:07 so searching for every single writew wouldn't slow things down? May 23 16:24:19 freemangordon; i think it's inevitable with any algorithm that can't reallocate pages that already exist in swap. I once had an idea of doing: May 23 16:24:22 you know, those 3 pages goes there, those 5 there, and this 1 there... May 23 16:24:42 hmm, the other option would be swap de-fragmentation worker thread May 23 16:24:55 bug realocating pages already sitting in swap would require re-writing already written things, + writing new ones = more slowdown on already slow swap May 23 16:25:28 honestly, as we have hardware wear-leveling anyway, the fastest thing would be to remap their logic location, without moving them physically on swap partition. But is it possible? May 23 16:25:36 divide swap into 1 meg blocks. Write sequentially filling blocks. If no fully free block is found, find n most free blocks, collect used pages from those together into a fully filled block, hopefully gaining many more totally free blocks May 23 16:25:45 a garbage collect May 23 16:25:58 Estel_; not possible May 23 16:26:31 ShadowJK, yes, something like that May 23 16:26:38 If I would have to choose between swap of maximum allowable speed but requiring "refreshinbg" via ereswap, or swap not needing ereswap at all, but slower all-together May 23 16:26:44 due to some worker re-writing it... May 23 16:26:52 I would preffer faster swap needing reflash May 23 16:27:02 s/reflash/refresh/ May 23 16:27:02 Estel_ meant: I would preffer faster swap needing refresh May 23 16:27:30 Estel_, a well written garbage collector would be unnoticable from th user POV May 23 16:27:49 if that's achievable, why not? May 23 16:27:51 swapoff/swapon is the quick and easy sh script defragmenter, it essentially reads in all pages and writes it to another fresh swap :D May 23 16:28:06 but you need a second swap space May 23 16:28:09 but, it need moving them to ram first May 23 16:28:12 and moving them to swap later May 23 16:28:19 both require time if You have 200 MB of swpa used May 23 16:28:19 * Woody14619 feels like a damn yo-yo... stupid network. May 23 16:28:24 which means you need to close everything May 23 16:28:28 and do reswap May 23 16:28:36 or it take ages May 23 16:28:40 * ShadowJK usually runs swap on microsd May 23 16:28:46 same here May 23 16:28:54 but it still take time and effort to do reswap May 23 16:28:57 me too, but that is not a general solution for stick n900 May 23 16:29:04 because it reduces the amount of random I/O to emmc May 23 16:29:05 *stock May 23 16:29:14 IMO, swap on eMMC is depreciated May 23 16:29:25 * Woody14619 runs swap on sd, then runs a script via cron at 3am that swaps to internal then back to declutter. :) May 23 16:29:27 I educate people whenever I can, and it seems to make good effects May 23 16:29:39 Woody14619, sometimes I need to refresh swap few times a day May 23 16:29:51 or few times a hour in extreme cases :D May 23 16:30:00 Estel_, after all I am a part of CSSU team :P May 23 16:30:18 :P May 23 16:30:24 btw: http://lwn.net/Articles/428584/ May 23 16:30:25 I generally have a moderate usage pattern, so for me once a day is plenty. May 23 16:31:13 BTw I noticed that poorly written html page with BIG pictures is able to use whole swap oin 10 minutes May 23 16:31:13 and there are not only hackers/gurus in CSSU users May 23 16:31:26 freemangordon,that why I've put ereswap into repos May 23 16:31:45 ...and spend 7 freaking hours to write proper script that automagically uipdates user rcS-late May 23 16:31:56 raping sed few times in the process May 23 16:31:57 yeah, sure, I am joust trying to gather as much information as possible May 23 16:32:01 of course May 23 16:32:22 fikrst of all, swap on microSd = I/O conflicts in eMMC and more wear and tear to non-replaceable medium May 23 16:32:26 IMO it's totally depreciated May 23 16:32:43 damn May 23 16:32:46 I mean swap on eMMC May 23 16:32:50 not microSd ;) May 23 16:32:54 microSd is the opposite May 23 16:33:06 BTW I thing simulationg the fragmentation should be a relatively easy task May 23 16:33:32 so one can check an algorithm performance May 23 16:33:53 damn, a typo day again :D May 23 16:34:04 bTW, freemangordon, what do You t2hink about DocScrutinizer's idea of attaching jt6ag cable and gaining some more debug info through testpads? May 23 16:34:29 if doable, it should provide us really great core dumps May 23 16:34:39 I am all with him :D May 23 16:34:50 If he can make it that will be great May 23 16:34:56 I hope he will take this project to results stage May 23 16:35:12 I would do this too, if I would understand what he means completely ;) May 23 16:35:37 Woody14619, are You still here? May 23 16:35:51 BTW there is rich-core... stuff in repos perfectly suitable to gather coredumps May 23 16:35:55 * Woody14619 was off in another window... May 23 16:36:09 Woody14619, you are using windows? May 23 16:36:11 :P May 23 16:36:21 Woody14619, as election is ending today, and, during candidate's numbers etc, we're sure that "someone" will be elected May 23 16:36:22 I'm at work... so yes. :) May 23 16:36:33 what about setting up a hyphotetical Council Meeting tommorow? May 23 16:36:54 I mean rock solid council meeting with hyphotetical list of peoiple ;0 May 23 16:37:10 You assume we're gonig to get results in a timely mannor? ;) May 23 16:37:21 Estel_, that ^^^ May 23 16:37:25 yea, as machine is doing it ;P May 23 16:37:43 Considering it took 3 days to start the election, I'm not quite certain we'll get a result at a given time. May 23 16:37:45 and have in mind that all of the candidates shoudl agree on that May 23 16:37:54 freemangordon, sure thing May 23 16:37:58 that's why I'm catching them May 23 16:38:02 I bet the machine wills stop taking votes at a given time, but seeing the results? Dunno. :) May 23 16:38:03 as noone knows (i hope) in advance who will be elected May 23 16:38:06 although some, like jrpchak, are quite hard to catch May 23 16:38:09 ...anywhere... May 23 16:38:24 freemangordon, hahaha, it see,ms X-Fade know it already, probably May 23 16:38:33 considering most votes were allocated May 23 16:38:44 seriously though May 23 16:38:48 well, besides him May 23 16:38:50 I will be on holiday Friday through Monday, so if it's going to be any time, it's tomorrow or Tuesday on. May 23 16:39:20 god damn thing about time differencews between continent and US May 23 16:39:21 Monday is a US holiday (Memorial Day). May 23 16:39:35 Woody14619, what is Your's tommorow availability? May 23 16:40:20 today is our protest day, it seems... Taxi drivers protest vs deregulation of their professions, so they've blocked country capital May 23 16:40:42 people are already thinking about revenge, like payiung with pouches of 1 gr (our equivalent of 1 cent) May 23 16:41:15 Most of the day, 14-19 UTC and 20-23 UTC. May 23 16:42:15 .oO(I think I got that right? May be an hour off from daylight saving time) May 23 16:43:05 Yup.. my bad... 13-18, and 19-22. :P May 23 16:43:07 first idea is 15 UTC May 23 16:43:14 so 14 UTC May 23 16:43:20 will check with others May 23 16:44:06 : nods, that generally works for most people. Morning for US, evening for EU. May 23 16:44:26 * Woody14619 sigh... IRC != Talker. May 23 16:45:00 erm, which UTC hour is now? May 23 16:45:07 what UTC time do we have now May 23 16:45:13 (daylight saving time problems) May 23 16:45:34 16:45 right now. May 23 16:45:42 yea, checked that too in web... so it's all ok May 23 16:46:40 I know it's always +5 or +4 (EST->UTC or EDT->UTC), just forget which is which sometimes. May 23 16:51:15 hmm, maybe 15:47 May 23 16:51:22 nah, 16... May 23 16:53:03 god bless zerobin May 23 16:53:16 today I had to recover from backup, as my optfs *and* rootfs got screwed... May 23 16:53:19 wtf??? May 23 16:53:26 and I've lsot some on-device work that I've made on eject May 23 16:53:35 Estel_: I never suggested that May 23 16:53:38 fortunatelly, I've found zerobin of latest version, that I posted here for revision May 23 16:53:49 DocScrutinizer, that = what? May 23 16:54:19 attacj JTAG May 23 16:55:36 hm? May 23 16:55:40 sorry, i must have missunderstood May 23 16:56:04 but You generally know which idea I'm reffering to? May 23 16:56:59 I talked with two or 3 colleagues, all think ETB should be usable from maemo userland or at least kefrnel May 23 16:57:31 no need jtag May 23 16:57:49 I'm almsot sure I remember that someone asked You about jtag and You replied yes May 23 16:57:51 but, whatever May 23 16:57:56 maybe it was 2 topics merge May 23 16:58:30 chck chanlog, I said works via both May 23 16:58:50 * Estel_ nods May 23 17:04:15 Estel_: [2012-05-22 13:02:29] freemangordon: IHI0014Q_etm_architecture_spec.pdf p129 "3.5 detailed register descriptions" ETMACTR [13:10] May 23 17:06:23 http://infocenter.arm.com/help/topic/com.arm.doc.ihi0014q/IHI0014Q_etm_architecture_spec.pdf May 23 17:08:10 State and mode comparison control. The assignment of these bits is: May 23 17:08:11 Bit [13, 11] Non-secure state comparison control. May 23 17:08:13 Bit [12, 10] Secure state comparison control. May 23 17:08:38 thanks a lot, although i dont understand single part of it ;) May 23 17:09:19 If the processor does not implement the Security Extensions, bits [13, 11] are reserved, May 23 17:10:19 tjhis clearly suggests that ETB is designed to work on platforms that use TrustZone Security Extensions (as well as on those who don't use or have it) May 23 17:13:30 erm, above sentence it means that trustzone security blabla doesn't matter... May 23 17:13:35 cause it works with or without it? May 23 17:27:47 yes May 23 17:28:14 you know what's ETB/ETM aka CoreSight? May 23 17:29:22 wel, obviously " Embedded Trace Macrocell ™ " according to the pdf link I pasted above ;-) May 23 17:29:48 freemangordon; in your simulaion, randomly free 4k chunks of swap May 23 17:29:52 it's basically a (iirc) 6.8K size FIFO May 23 17:30:26 that traces all instructions execured, and optionally also data of accessed registers May 23 17:31:28 and you can read out that FIFO by CPU May 23 17:32:58 ShadowJK, :nod:, along with random allocation May 23 17:38:51 >> An Embedded Trace Macrocell (ETM) is a real-time trace module providing instruction and data tracing of a processor. An ETM is an integral part of an ARM RealView® debug solution.<< May 23 17:39:52 freemangordon; and then a non-sequential write has penalty of .5M read-modify-write cycle :) May 23 17:40:59 well, that's flash May 23 17:41:05 * ShadowJK finds his trace May 23 17:41:06 http://enivax.net/jk/kswapd.png May 23 17:41:20 NAND flash, to be precise May 23 17:41:24 y is multiples of 512 or something May 23 17:42:18 notice right after linear part ended, there's short tiny patch of linear in middle of y approx May 23 17:42:23 but then that too is spent May 23 17:42:45 funny tiny scratch May 23 17:43:28 and shit literally explodes right after that May 23 17:44:19 ShadowJK: I guess that tiny thingie is the fuse of the bomb you see exploding in right half May 23 17:44:20 the dots get smaller and smaller May 23 17:47:02 sure as hell painting a 20 MB picture onto flying dust makes your kernel kswapd busy for 30min or an hour May 23 17:49:18 http://enivax.net/jk/n900/swap2.png also with reads, ends before fragmentation onset it seems May 23 17:50:43 Some data seems to stick around, and most reads are of recently written data. tiny semi-permanent chunks are sprinkled like pixie dust in the wake of the swapmonster :) May 23 17:53:08 * DocScrutinizer cleans dust off his LCD screen May 23 17:55:55 now if you'd color the dots modulo of PID of process owning the page... :-D May 23 17:56:05 Speaking of dust, it's amazing how big fuses you can trip with a few teaspoons of metal dust inside a schuko May 23 17:56:32 (and undetectable with multimeter) May 23 17:56:38 hehe May 23 17:56:53 Unfortunately kswapd does most of accesses :) May 23 17:57:26 DocScrutinizer, sh expertise needed again, I've screwed something (what a surprise :P) May 23 17:57:35 lets say I have simple line of code May 23 17:57:56 unlikely ;-P May 23 17:58:11 [ ! -e /dev/sr* ];e=1 && banner May 23 17:58:21 bad May 23 17:58:21 banner is a function described properly, it function ok so lets skip that May 23 17:58:29 e is declared as 0 before May 23 17:58:30 wil blow up if there is more than one match May 23 17:58:34 e+0 then comes our line May 23 17:58:40 oh May 23 17:58:44 good point, but anbother problem is there May 23 17:58:54 it set e to 1 even when /dev/sr0 is present May 23 17:58:58 (and sole) May 23 17:59:06 becuase you have a ; there May 23 17:59:23 with exectures the next command unconditionally May 23 17:59:24 yeah, completely messed up May 23 17:59:30 I see May 23 17:59:38 you want && between ] and the assignment May 23 17:59:55 erm, so what should be there to execute next command (and banner) only if /dev/sr* is non-existing? May 23 18:00:13 [ ! -e /dev/sr* ] && { e=1 ; banner ; } May 23 18:00:13 [ ! -e /dev/sr* ] && e=1 && banner May 23 18:00:14 ?? May 23 18:00:18 or just do it like normal people: May 23 18:00:19 ah May 23 18:00:32 well, my only excuse is that I'm debugging blue_led's code :P May 23 18:00:34 if [ ! -e /dev/sr* ] ; then e=1 ; banner ; fi May 23 18:00:54 still ignoring your own valid concern, r00t|home ? May 23 18:00:57 but it will still blow up if there is more than one match to the glob May 23 18:01:02 i'm not sure if "if" (:P) is suitable here, due to this shit being part of complex code May 23 18:01:04 gimme sex May 23 18:01:11 DENIED May 23 18:01:20 use find May 23 18:01:42 i'd rather use ls i think, find a little overkill May 23 18:01:51 or that May 23 18:02:07 also, why use that glob anyway? May 23 18:02:22 but how you instruct ls to throw error on invokation w/o argument? May 23 18:02:31 doesn't make much sense May 23 18:02:44 aaah wait, it has invalid argument then May 23 18:02:45 DocScrutinizer: the shell does NOT remove non-matching globs, unless you set a special option May 23 18:02:53 yep May 23 18:02:55 wait a second, I'll show You context May 23 18:02:55 shopt nullglob May 23 18:03:00 where it's written May 23 18:03:02 Estel_: NOOOOOO.... May 23 18:03:07 ? May 23 18:03:11 pretty please mot May 23 18:03:13 not May 23 18:03:14 Estel_: we've seen enough of it by now ;) May 23 18:03:27 to know we don't want to see more May 23 18:03:50 honestly, without seeing at least full function, this is academic discussion May 23 18:03:58 as I can't apply Your suggestion May 23 18:04:00 ls /foo/bar/xy* && echo "there's some xy*" May 23 18:04:11 without You considering context and adjusting suggestions to it May 23 18:04:15 ls /foo/bar/xy* 2>/dev/null && echo "there's some xy*" May 23 18:04:15 DocScrutinizer: &>/dev/null May 23 18:04:29 DocScrutinizer: you want to surpress normal output aswell May 23 18:04:30 err May 23 18:04:31 yep May 23 18:04:40 whatever, if You two doesn't want to see it, maybe someone else will like to help May 23 18:04:53 overall, it's simple code, simple small script May 23 18:04:54 didn't we help already? May 23 18:05:04 * Estel_ nods May 23 18:05:09 dang, s-off with shellscripts, I learnt about caif in andridiot whole day May 23 18:05:09 of course and zI'm grateful May 23 18:05:25 yet not to the point, when noob like me would be able to fix it properly May 23 18:05:51 ls /foo/bar/xy* &>/dev/null && echo "there's some xy*" May 23 18:06:02 HELL, FIRE & BRIMSTONE May 23 18:07:02 echo will set variable for me that I can use later to decide something? May 23 18:07:13 *sigh* May 23 18:07:15 * GeneralAntilles finally voted. May 23 18:07:36 GeneralAntilles, thanks a lot for You participating in election May 23 18:07:47 if you dunno how to replace "echo TheFuck" by any other cmd, I think you shouldn't try shellscripting May 23 18:08:05 Estel_, well, I probably should. I was partly responsible for the existence of the council in the first place. :P May 23 18:08:30 ;) May 23 18:08:32 http://sebsauvage.net/paste/?715bc0aa4fd6ec1a#jUMAZFyEh+dgn97YDkyrODE0gQBWkTAV88OHtAixX/c= May 23 18:08:41 here is current state of this simple script May 23 18:09:00 set -xv tells me, obviously, that it fail.s due to ;e=1 May 23 18:09:15 any way of fixing it *without* breaking whole stuff? May 23 18:09:19 * DocScrutinizer surfs to youporn instead ;-P May 23 18:09:43 It have been written by blue_led and contained multiple mistakes, I've fixed them and want to publish it May 23 18:09:52 quite useful for optical drives users May 23 18:10:02 DocScrutinizer, this script is more hardcore porn :P May 23 18:10:07 * Estel_ teases May 23 18:10:14 that's why May 23 18:10:24 or softcore, it's like foo May 23 18:10:33 it's whatever You ever wanted to look at May 23 18:11:04 seriously, it's few liner, and I'm to noob to fix it ;) anyone could help? May 23 18:11:26 line: May 23 18:11:27 *Host*)[ ! -e /dev/sr* ];e=1 && ( banner May 23 18:11:32 is responsible for fail May 23 18:11:52 as it should set 1 only if /dev/sr* (anything) doesn't exist May 23 18:12:07 but obviously it sets 1 all the time May 23 18:12:22 now, I would like to also take r00t|home valid comment May 23 18:12:33 about it failing when more than 1 matches found May 23 18:13:09 r00t|home, teased enough to look at it? :P May 23 18:13:34 maybe some good looking if would fit there? May 23 18:14:51 normally, it looked like that: May 23 18:15:08 *Host*)[ ! -e /dev/sr* ];e=1 && ( banner "Host mode detected, but no optical drive attached. Exiting...";e=1);; May 23 18:15:14 no wait May 23 18:15:24 *Host*)[ ! -e /dev/sr* ] && ( banner "Host mode detected, but no optical drive attached. Exiting...";e=1);; May 23 18:15:31 ^^^like that May 23 18:15:51 but it had bad habit of setting it to 0, even if no /dev/sr* found May 23 18:16:02 + exploded, when more than 1 matching May 23 18:16:12 idea how to fix it, anyone? May 23 18:16:36 *Host*)[ ! ls /dev/sr* ] && ( banner "Host mode detected, but no optical drive attached. Exiting...";e=1);; May 23 18:16:37 ? May 23 18:18:38 HEEEEEEELLLLLL May 23 18:19:53 if ls /dev/sr* 6>/dev/null; then e=1; banner; fi; May 23 18:20:03 s/6/&/ May 23 18:20:03 DocScrutinizer meant: if ls /dev/sr* &>/dev/null; then e=1; banner; fi; May 23 18:20:46 or May 23 18:20:57 if ! ls /dev/sr* &>/dev/null; then e=1; banner; fi; May 23 18:21:19 or May 23 18:21:47 if ls /dev/sr* &>/dev/null; then :; else e=1; banner; fi; May 23 18:22:07 or May 23 18:22:21 ls /dev/sr* May 23 18:23:08 if [ 0 -eq $?] ; then e=1; banner; fi; May 23 18:23:29 if [ 0 -ne $?] ; then e=1; banner; fi; May 23 18:24:12 hehe thanks for plentor of solutions May 23 18:24:24 trying them now... May 23 18:24:53 and those () present in original script, not needed? May 23 18:26:24 you still think I'll look at that script again? May 23 18:26:49 no :P will try everything myself May 23 18:26:58 and ask if all effort ,to no avail May 23 18:27:10 thanks a loor checking it May 23 18:34:24 DocScrutinizer, ok, now my question not related to above code, at least directly May 23 18:34:49 when I try to mount filesystem for optical drive with no disc inside, ZI get, hopefully, such output: May 23 18:34:51 mount: mounting /dev/sr0 on /media/cdrom failed: No medium found May 23 18:35:02 I would like to write simple if May 23 18:35:13 making banner output different for such situation May 23 18:35:24 than situation, when disc is present but mounting failed May 23 18:35:38 any idea how to identify former case in sh script? May 23 18:35:59 i.e. how to make sh script "catch" this "no medium found" reply May 23 18:36:13 my idea is simple: May 23 18:37:00 if ; banner "message1": else banner "messagea2"; fi; May 23 18:37:14 the problem is with method to caught this output properly May 23 18:37:21 I'm sure there is a way May 23 18:38:06 exit code of mount is same (failure) on both no medium found and other problems, yep? May 23 18:38:21 so can't use simple || May 23 18:42:13 you're sure there's such an output? May 23 18:46:20 yes May 23 18:46:28 DocScrutinizer ^^^ May 23 18:46:35 I mean, output in terminal May 23 18:46:51 it's different for other problems May 23 18:47:23 I wonder how mount know that sr0 is something that contains (should) media May 23 18:47:26 but it does May 23 18:47:54 sr0 exist, yet, trying to mount it without CD or DVD inside gives 'essage as I've pasted before May 23 18:47:57 srX is supposed to be sth like that May 23 18:48:04 of course May 23 18:48:43 I already adjusted this script to give proper answer when mounting fail instead of dying silently, but, it would be nice to make different output banner for no disc present May 23 18:48:56 and "mounting failed" overall error with mounting May 23 18:49:13 if [ "some message" == $(command if [ "some message" == "$(command even May 23 18:49:52 ah, so it will know answer... lovely May 23 18:50:14 else, just command May 23 18:50:20 but but May 23 18:50:55 wont if [ "some message" ==$(command already mount it in case of disc present? May 23 18:51:31 then trying to mount it again after "else" would give false error mesage about device in use May 23 18:52:10 if [ "root" == $(sudo whoami) ]; then echo "test was true"; fi May 23 18:52:20 generally, it must either mount it with success, and give banner1, or fail to mount it due to lack of disc and give banner2, or fail due to other reasons and give banner3 May 23 18:52:53 I see, but how it can see reply to mount command without trying? May 23 18:53:53 it's just simulating command and getting proper output? 0_o May 23 18:54:05 even with mount related to hardware devices? May 23 18:54:45 (that could have different state totally independently of N900 and without informing the latter about status change, i.e. disc present or not) May 23 18:54:55 (unless N900 actually tries to mount it) May 23 18:55:11 oh, got idea May 23 18:55:12 case "$(sudo whoami)" in \n "root") echo "root";; \n user) echo user; \n beep;; \n *) echo "neither root nor user"; \n reboot;; \n esac May 23 18:55:34 hehe May 23 18:55:37 what about May 23 18:57:50 mount /dev/sr0 /media/cdrom || if [ "no media" == "$(mount /dev/sr0 /media/cdrom)" ]; then banner "no disc"; else banner "generic error"; fi; May 23 18:58:03 DocScrutinizer, ^^^, should work I think? May 23 18:58:28 no, you invoke it twice May 23 18:58:41 ??? May 23 18:58:49 2nd time only if 1st fail May 23 18:58:52 || May 23 18:59:00 so?? is that sane? May 23 18:59:21 but it won't mess with situation when mounting was ok May 23 18:59:30 so wont give false error for success May 23 18:59:44 and if it fails, it just check why and give answer based on that May 23 18:59:51 = no false answer in any condition May 23 19:00:10 calling it twice isn't a problem, problem was with false errors or positives May 23 19:00:21 ohwell May 23 19:00:24 but of course I'll gladly read better solution and learn analysing it May 23 19:02:07 no, you'll answer hehe and then ignore any suggestion of better alternatives May 23 19:02:56 DocScrutinizer, what's the difference between CSOC and RSOC in bq27k queries? May 23 19:03:11 Sorry to burst in asking dumb things. May 23 19:03:15 wtf? May 23 19:03:24 how those with root was related? May 23 19:03:35 the one is adjusted to take into account the temperature etc of cell May 23 19:03:37 whoiam doesn't interfere like checking mount reply May 23 19:04:31 Doc, thanks. :) May 23 19:04:49 case "$(sudo whoami)" in \n "root") echo "root";; \n user) echo user; \n beep;; \n *) echo "neither root nor user"; \n reboot;; \n esac May 23 19:04:55 here You also call it twice May 23 19:05:07 o.O May 23 19:06:43 case "$(mount /dev/sr0 /media/cdrom)" in \n "no media") echo "disc not in tray";; \n generic_error) echo failed_totally; \n mounted_properly;; \n *) echo "mounted properly"; \n mount /dev/sr0 /media/cdrom;; \n esac May 23 19:06:46 or May 23 19:07:02 case "$(mount /dev/sr0 /media/cdrom) mount it already for good? May 23 19:07:19 wtf? May 23 19:07:24 I think I get it May 23 19:07:52 if ! mount cdrom ; then echo an error occured ; else echo ok ; fi May 23 19:08:10 r00t|home, You haven't read whole story May 23 19:08:18 there could be 2 reasons of error May 23 19:08:30 and I want to give different messages on them May 23 19:08:38 one is for disc not in tray May 23 19:08:44 another one for generic error May 23 19:08:58 but, it can be checked *only* during real mounting attempt May 23 19:09:07 and what makes you think the error msg of mount cmd isn't ok for your banner text? May 23 19:09:22 not verbose enough :P May 23 19:09:49 someting like "mount: no media found" is ugly May 23 19:09:50 Estel_: do a May 23 19:09:51 ;) May 23 19:10:02 or May 23 19:10:31 and it's GUI tool, so I needd answer via banner May 23 19:10:35 result=$( mount dev/srwhatnot 2>&1 ) May 23 19:10:40 or user won't see it May 23 19:10:48 then test for $result May 23 19:11:06 good idea, but result mount it already May 23 19:11:10 if it's OK May 23 19:11:37 and if result is wrong May 23 19:11:48 I need to try mounting agaion to probe for answer content May 23 19:11:59 WAAAAAAAAAAAAAAAAAAAAAHHHHHH! May 23 19:12:01 suboptimal I guess, like my previous suggestion with || May 23 19:12:26 or result will be direct answer? May 23 19:13:00 call me noob but I don't know what 2>$1 exactly does May 23 19:13:11 although I've seen it many times before May 23 19:14:29 ignore it, it's fixing a problem you haven't even detected yet May 23 19:15:13 and &1, NOT $1 May 23 19:15:27 don't do it like that May 23 19:15:37 do it like hmm May 23 19:15:46 res=`mount /dev/sr0 /media/cdrom` May 23 19:16:07 if [ $? != 0 ]; then echo this failed. May 23 19:16:26 mount will always return -1 if it fails. May 23 19:16:28 *SIGH* May 23 19:16:45 and if you somehow want the output anyway you have it in $ret May 23 19:16:49 * DocScrutinizer heads out, as that's too much May 23 19:17:00 NO, YOU DON'T May 23 19:17:07 as that's stderr output May 23 19:17:09 fool May 23 19:17:55 if [ 0 -ne $?] ; then e=1; banner; fi; May 23 19:17:59 basically same thing I just said. May 23 19:18:14 waaaaaaaaaAAAAAAAAAaaAAAAAaaAAAaAAAAaaa May 23 19:19:03 res=`mount /dev/sr0 /media/cdrom 2>&1 ` <-- big deal. May 23 19:19:06 jr@halley:~> ls -l xxxdfd May 23 19:19:07 ls: Zugriff auf xxxdfd nicht möglich: Datei oder Verzeichnis nicht gefunden May 23 19:19:09 jr@halley:~> x=$(ls -l xxxdfd) May 23 19:19:10 ls: Zugriff auf xxxdfd nicht möglich: Datei oder Verzeichnis nicht gefunden May 23 19:19:11 that's just in case I want that output. May 23 19:19:12 jr@halley:~> echo $x May 23 19:19:13 jr@halley:~> May 23 19:19:24 wtf language is that May 23 19:19:31 norwegian? May 23 19:19:51 x=$(ls -l xxxdfd 2>&1) May 23 19:20:18 naikel@ubuntu:~$ x=$(ls -l xxxdfd 2>&1) May 23 19:20:19 naikel@ubuntu:~$ echo $x May 23 19:20:19 ls: cannot access xxxdfd: No such file or directory May 23 19:20:22 that's "never compare output strings, fool! unless you've set LANG=C" language May 23 19:20:31 in that I agree. May 23 19:20:43 and not only that May 23 19:20:57 obviously the OS language also matters May 23 19:21:04 everybody should have their OSes in english anyway May 23 19:21:09 ok, now you agree May 23 19:21:13 NO MATTER what native language you have May 23 19:21:15 * NIN101 grabs some popcorn. May 23 19:21:18 OS should be installed IN ENGLISH May 23 19:21:46 but that's not the topic here May 23 19:21:51 when I suggested exactly same shite as you with>>naikel@ubuntu:~$ x=$(ls -l xxxdfd 2>&1)<< you said "no, don't do it like that" May 23 19:21:55 anyway my main point was "use $?" May 23 19:22:12 that was just demonstration you can capture stderr May 23 19:22:21 not a solution of the problem. May 23 19:22:32 *plonk* May 23 19:22:33 I insist, my main point was "use $?" May 23 19:24:00 and May 23 19:24:02 about this: May 23 19:24:13 there could be 2 reasons of error May 23 19:24:16 man mount May 23 19:24:24 /RETURN CODES May 23 19:24:32 sure May 23 19:24:33 there are different return codes for $? May 23 19:24:39 you can check against them May 23 19:24:43 and give different messages May 23 19:24:49 the point is Estel_ doesn't even understand the first approach May 23 19:25:14 and I'm out May 23 19:25:25 that's really too much for my mood May 23 19:25:30 bye ;) anyway, thanks for edu material for analysis May 23 19:25:38 :/ May 23 19:25:43 patience my friend :) May 23 19:25:50 don't be sad, remember when You've lost credit card? May 23 19:25:57 I lost a credit card once May 23 19:26:11 and then I found it like 30 minutes later in a parking lot I already left May 23 19:26:14 :) May 23 19:26:15 DocScrutinizert told funny story about how he lost it May 23 19:26:19 :) May 23 19:26:30 he was buying something in vendor machine May 23 19:26:56 thinking about how poorly it's developed, because it gives goods before asking to take credit card May 23 19:27:07 thought about it still 20 meters away May 23 19:27:22 and after 30 min, realized he left his card there May 23 19:27:41 ...which just makes his reasoning about poor design valid ;) May 23 19:28:01 except for part when he though "it could make idiots into losing card" May 23 19:28:17 hahaha May 23 19:28:28 that's funny May 23 19:28:33 karma's a bitch, you're gonna lose it in the same machine some day May 23 19:28:36 I hope DocScrutinizer doesnt mind I told this story again May 23 19:29:03 karma? TMO one? :P May 23 19:29:21 s/tmo/maemo.org/ May 23 19:29:31 s/TMO/maemo.org/ May 23 19:29:31 Estel_ meant: karma? maemo.org one? :P May 23 19:30:50 Sc0rpius, You're also experienced @ sh script, yes? May 23 19:31:05 who isn't :p May 23 19:31:13 yeah May 23 19:31:19 but mostly bash May 23 19:31:29 though I have made some scripts in ksh and other weird shells May 23 19:34:06 I see May 23 19:34:34 Sc0rpius, honestly, what 2>&1 exactly does May 23 19:35:02 MohammadAG, I'm not (who isn't) May 23 19:35:19 I would call myself beginner, kinda May 23 19:35:25 it means the second standard output is redirected to the first one May 23 19:35:30 (beginner > just user) May 23 19:35:33 the second standard output is stderr May 23 19:35:39 the first one is stdout May 23 19:35:53 > == 1> May 23 19:36:00 I see. What is that supposed to fix? May 23 19:36:01 > outputfile == 1> outputfile May 23 19:36:15 it just redirects stderr (where errors are printed) to the stdout May 23 19:36:26 all right May 23 19:36:27 when you type a command and it shows an error, it's shown in stderr May 23 19:36:40 without it, it isn't? May 23 19:36:50 that's why a $(mount something) that fails is always empty May 23 19:37:01 roger that May 23 19:37:03 because there was no output to stdout May 23 19:37:06 anyway like DocScrutinizer May 23 19:37:07 said May 23 19:37:12 you shouldn't compare strings May 23 19:37:24 because? May 23 19:37:40 because that will only work in the language of your OS, if the script is for the masses, and people has their Maemo in spanish, it'll fail May 23 19:37:57 roger roger... May 23 19:38:09 so exit code approach is only one possible? May 23 19:38:32 it's the best approach May 23 19:38:34 and if mount give only 1 error code, but different string output? May 23 19:38:47 there are several return codes for mount, you can check against them and show different messages May 23 19:39:03 there's a RETURN CODES section when you type: man mount May 23 19:39:08 the point is, as You're probably aware, that it's different situation when no media found May 23 19:39:11 I know May 23 19:39:25 I'm asking just in case writing wrappers for poorlhy wrfitten program May 23 19:39:38 that can have different string output yet same error code May 23 19:43:29 how to act in such situation? May 23 19:44:10 what are the different errors you get? May 23 19:44:46 in this case, probably error codes work, but: May 23 19:45:39 mount: mounting /dev/sr0 on /media/drom failed: no media found May 23 19:45:58 no media found is key part May 23 19:46:09 but, it may be different in other languages May 23 19:52:04 Estel_: May 23 19:52:21 LANG=C ls -l XXXxXxx May 23 19:52:39 nice May 23 19:53:03 thanks May 23 19:56:11 Sc0rpius, do You have ereswap installed, in any case? May 23 19:56:51 or, Your rcS-late is vanilla one? May 23 19:57:25 DocScrutinizer, You have some vanilla N900, yep? Could You pastebin me content of vanilla rcS-late? May 23 19:57:36 /etc/rcS-late May 23 19:57:39 sorry May 23 19:57:41 fuck no May 23 19:57:44 it will be different in other languages. May 23 19:57:52 /etc/event.d/rcS-late May 23 19:57:59 I think this one isn't exactly vanilla May 23 19:58:14 in fact I need just vanilla part enabling swap May 23 19:58:22 + one line above and one under May 23 19:59:07 Estel_: mine is unmodified, unless KP & CSSU modify it :\ May 23 19:59:14 no, its ok May 23 19:59:33 you lost the one i gave before? ;) May 23 19:59:49 BTW DocScrutinizer 2nd questiob - I agree that shell operating language should be english, but how do You change it in N900? May 23 19:59:54 Sicelo, yes May 23 20:00:06 dpkg-reconfigure locales doesn't work May 23 20:00:17 LANG=C ls -l XXXxXxx May 23 20:00:24 generally speking: LANG=C May 23 20:00:24 Sicelo, I have had to restore rootfs and optfs today May 23 20:00:29 sets $LANG to 'C' May 23 20:00:35 for that command, temporarily May 23 20:00:48 yea. But for good? May 23 20:01:39 you finally managed without reflash? May 23 20:01:41 LANG=cz_CZ ;in .profile or whatever initfile May 23 20:01:57 Sicelo, sure thing, thanks to backupmenu May 23 20:02:01 DocScrutinizer, thanks May 23 20:02:02 even May 23 20:02:15 export LANG=cz_CZ ;in .profile or whatever initfile May 23 20:02:17 Sicelo, strangely enough, my rootfs content was corrupted May 23 20:02:31 http://paste.debian.net/170841/ May 23 20:02:36 thanks May 23 20:03:05 DocScrutinizer, last question, when I want to set it for this terminal session only, not single command or all? May 23 20:03:07 yw. when you say rootfs content, you mean on N900, or the backupmenu-saved one? May 23 20:03:14 N900 May 23 20:03:25 had to reestore both optfs and rootfs May 23 20:03:31 export LANG=cz_CZ May 23 20:03:34 normally, optfs would be enough May 23 20:03:41 DocScrutinizer, I see May 23 20:04:28 without export it won't take effect even for shellscripts you start from that shell May 23 20:05:31 I see May 23 20:06:00 odd enough, export LANG=pl_PL doesn't work for maemo May 23 20:06:20 or it does May 23 20:06:31 but special chars are screwed May 23 20:07:01 pl_PL.utf8 fix that May 23 20:07:32 Estel_: man -a locale May 23 20:07:51 Maemo got mans crippled May 23 20:08:09 I'm on device, remember that. Will search in net May 23 20:08:11 we got it uncrippled :-P May 23 20:08:22 by MAG's package? May 23 20:08:35 (we = MohammadAG , me) May 23 20:08:58 I see. Will install it May 23 20:09:14 noob noobish question, how to increase backscroll lines in osso-xterm May 23 20:09:23 longer man's are cut in middle May 23 20:09:28 even tar one May 23 20:10:47 not here May 23 20:12:29 ^^^ ??? May 23 20:12:44 just pipe it to less May 23 20:12:53 hm? May 23 20:13:28 instead of scrolling the window May 23 20:13:46 "pipe it to less" what does it mean? May 23 20:14:07 isn't less a dependency in mandb-n900?? May 23 20:14:24 actually yes for man it should already be called, hmm May 23 20:14:45 it means: command | less May 23 20:14:52 yes, sorry May 23 20:15:00 ok. what it does, actually? May 23 20:15:40 DocScrutinizer51, I think not May 23 20:15:49 it should be installed, but I think I haven't depended on it May 23 20:15:50 anyway May 23 20:15:52 lets you navigate thru a longer text, space one page forward, return one line, b one page back, / search, etc May 23 20:16:00 to increase backscroll, edit osso-xterm's gconf May 23 20:16:05 Estel_: apt-get install less May 23 20:18:25 off, bye May 23 20:22:22 nox-, thanks May 23 20:22:28 MohammadAG, thanks a lot May 23 20:22:49 Estel_: h gor help May 23 20:23:10 and path to osso-xterm's gconf is? or best way to find it myself? May 23 20:23:12 for* May 23 20:23:28 /apps/osso/xterm/ May 23 20:23:29 forget that May 23 20:23:39 Sicelo, thanks May 23 20:23:47 DocScrutinizer, ??? May 23 20:27:29 Sicelo, gconftool-2 --all-dirs /apps/osso/xterm fails May 23 20:28:35 gconftool-2 -R /apps/oso/xterm May 23 20:32:45 *osso May 23 20:34:48 thanks, my mistake May 23 20:48:39 on transreflective scren (N900), does drawing white or black generates different battery usage? May 23 20:48:54 I know it's marginal, as backlight take most, but May 23 21:17:21 Estel_: yes May 23 21:17:27 + May 23 21:17:46 in Neo Freerunner it was ~30mA IIRC May 23 21:17:50 dark = lower usage on transleflective? May 23 21:17:56 can't be! May 23 21:18:04 so high value? May 23 21:18:11 between white and black? May 23 21:18:37 I wonder if it's same for N900, going to check that now May 23 21:18:42 I suspected it's been the data bus May 23 21:19:59 I = U / R May 23 21:20:27 ~ 3.3 / ( 50 / 4 ) May 23 21:20:27 0.264 May 23 21:21:33 ~ 4 * (1.8 / 1000) May 23 21:21:34 0.0072 May 23 21:22:08 hm, no measureable change on N900 May 23 21:22:22 it must be within normal fluctuation range = <10 mA May 23 21:23:38 ping X-Fade May 23 21:34:16 wtf May 23 21:34:25 my KB LEDs have ceased to function May 23 21:34:35 hmm May 23 21:34:36 how can I test them? May 23 21:35:00 is there a KB led script? May 23 21:35:05 yep May 23 21:35:27 first: stop mce May 23 21:35:33 can I enable R&D LED flashy mode... May 23 21:35:40 DocScrutinizer: ok May 23 21:36:07 2nd: find /sys -name brightness May 23 21:36:08 DocScrutinizer: done May 23 21:36:40 vi__, happened to me once, led indicator stop to work and kb light too May 23 21:36:49 3rd: echo 250 >/sys/class/leds/lp5523:kb1/brightness May 23 21:36:57 do You happen to keep something under backover? May 23 21:37:10 2nd sim, sd card, money :P, whatever? May 23 21:37:23 0.o May 23 21:37:35 Sicelo, it kwas my reaction to May 23 21:37:36 Estel_: ?? no May 23 21:38:01 pressure on certain part od N900 (under backcover) makes led driver to, erm, sleep May 23 21:38:38 when I kept 2nd sim card there, it wasn't working. I was thinking WTF and haven't foun issue case May 23 21:38:47 DocScrutinizer: I do not have an lp5523 /sys entry May 23 21:38:56 then, when 2nd card wasn't neededd and I took it out, it started to work May 23 21:38:56 for led in /sys/class/leds/lp5523:kb?/brightness; do echo 100 >$led; done May 23 21:39:04 HAHA May 23 21:39:14 vi__, same was here, no lp5523 May 23 21:39:34 even on backupmenu it haven't worked May 23 21:39:51 only a twl14030 May 23 21:39:53 one of strangest thing I've ever seen on N900 May 23 21:39:55 what surprises me is the keeping of stuff there :P May 23 21:40:02 Estel_: did you need to reboot? May 23 21:40:06 :P I got mugen cover May 23 21:40:09 I heard the chip sometimes locks up - remove battery for some minutes May 23 21:40:31 Sicelo: there was some guy who keeped monies in his! May 23 21:40:35 vi__, yea. But rebooting with someting between cover even reboots was for no use May 23 21:40:47 DocScrutinizer: May 23 21:40:51 DocScrutinizer: ok May 23 21:40:53 even during turning N900 on and off it haven't worked May 23 21:40:55 brb May 23 21:41:00 it was scary :P May 23 21:41:40 DocScrutinizer, hardware wise, any idea why pressure above camera slider sensor could make it to "lock"? May 23 21:42:01 fscked FPC May 23 21:42:07 Sicelo, with mugen cover it was best place for 2nd sim May 23 21:42:09 FPC? May 23 21:42:19 flex plastic cable May 23 21:42:27 flat plastic connector May 23 21:42:38 I see May 23 21:42:41 flexible printed circuit May 23 21:42:47 :P May 23 21:42:59 fuckin phone control May 23 21:43:09 not talking about N900, obviously May 23 21:43:27 btw, those with modem issues.. i wonder if they ever tried the other OSes to see if it still doesnamt work there (Nemo,MeeGo,Debian,Ubuntu, etc) May 23 21:43:34 hm, but what secondary board fpc does have to do with leds driver... camera leds aren't connected there May 23 21:43:41 (to driver) May 23 21:44:09 nor it's on secondary board May 23 21:44:28 mysteries :P May 23 21:45:03 Sicelo, most of the time modem issue are due to breakage in balls between sim reader May 23 21:45:05 sim slot May 23 21:45:25 or smth like that. One guy "fixed" it with nail and hammer May 23 21:45:29 literally May 23 21:45:36 w00t, meh LEDs are back. May 23 21:45:42 I suspect reflow would fix that May 23 21:45:43 thanks May 23 21:45:54 no problem user vi ;) May 23 21:45:58 citizien kane May 23 21:46:50 vi___, those guy who keeps money under N900 cover do it on purpose May 23 21:47:00 if You put ten pounds there at evening May 23 21:47:13 You find 20 pounds at the morning! May 23 21:47:31 how do I disable rotation completly without having to use rotation lock applet? like stock n900. May 23 21:47:44 Estel_: n900 is that good? May 23 21:47:53 but never put more, as You'll find used beer capsles instead May 23 21:47:56 ;) May 23 21:48:21 having tried portrait I can honestly say, I do not know what all the fuss is about. May 23 21:48:23 vi___, check what rotation lock applet does and do it manually? May 23 21:48:35 hehe I got it enabled but never use on home May 23 21:48:44 although, it doesn't hurt to have May 23 21:49:03 well, I put icons in order etc... can remember if ever used it May 23 21:49:13 Estel_: MOAR FFETURES==MOAR BETTER May 23 21:49:39 i think you can find that in ~jr-tools May 23 21:49:46 well, for people using it one-handedly, writting on screen via vkb without sliding keyb out... May 23 21:49:52 it's probably useful May 23 21:50:14 pfft, what do they think this is? May 23 21:50:19 ...a phone? May 23 21:50:24 ;) May 23 21:50:54 btw vi, using ereswap already? May 23 21:51:01 so are nokia really giving away a new pile of n950s? May 23 21:51:07 yea May 23 21:51:07 Estel_: no. May 23 21:51:21 Estel_: I literally have not had time May 23 21:51:31 40 N9 60 N900's May 23 21:51:35 * Estel_ nods May 23 21:51:40 fu May 23 21:51:44 60 N9 May 23 21:52:06 s/N900's/N9's/ May 23 21:52:30 fu again May 23 21:52:42 40 N950's and 60 N9's May 23 21:52:48 thats what I mean May 23 21:52:51 Estel_: I imagine I will download the deb, steal the main script then just run it as a cronjob every ~oouple of hours. May 23 21:53:01 vi, bad idea May 23 21:53:11 could do reswap while You're on heavy usage May 23 21:53:19 heavy ram or heavy CPU or both May 23 21:53:39 Estel_: so are the n950s returned by developers or are they just been found? May 23 21:53:49 = lock Your phone for 30 minutes or crash some apps without warning or reboot phone and corrupt fs May 23 21:53:55 "just found" May 23 21:54:09 vi=#-#- May 23 21:54:09 Estel_: well of course there will be some checking to tell if the screen is off and phone is locked etc. May 23 21:54:12 #*'" May 23 21:54:24 damn my alt_gr just stuck May 23 21:54:27 for a while May 23 21:54:32 software only May 23 21:54:36 harhar May 23 21:54:42 scarredk me that backspace stopped working May 23 21:55:09 IroN900:~/bin# cat no-portrait-gconfkey May 23 21:55:10 #!/bin/sh May 23 21:55:12 gconftool-2 -s /apps/osso/hildon-desktop/ui_can_rotate -t bool false May 23 21:55:17 vi___, I'm working on a way that it will inform user about swap usage automagically May 23 21:55:33 Estel_: cool May 23 21:55:45 Estel_: not my cup of tea thpugh. May 23 21:56:03 usually, You won't hit swap fragmentation with device locked and screen off May 23 21:56:30 and when You want to use it You'll be irritatedd when You hit fragm. and do it manually anyway May 23 21:56:48 so no much job for cron esxept for wasting cycles for checking May 23 21:57:33 btw freemangordon is checking if it's possible to get rid of this problem all-together May 23 21:57:50 by altering how kernel allocate swap on free fragments May 23 21:58:09 although I doubt 100% effective solution can be found on flash medium May 23 21:58:23 unless it awakens, sees swap useage is >50% then defrags May 23 21:58:47 Estel_: marvellous, that guy is a prodige. May 23 21:59:06 porridge May 23 21:59:08 ? May 23 21:59:12 vi, quite a solution May 23 21:59:19 prodigy? May 23 21:59:34 yea, but he's worried a little, that no more kernel devs than he and pali May 23 21:59:50 haha May 23 22:00:01 I would also *love* to se another talented dev working on kp May 23 22:00:10 Estel_: thats because there are no more kernel devs than just them May 23 22:00:14 s/se/see/ May 23 22:00:14 Estel_ meant: I would also *love* to see another talented dev working on kp May 23 22:00:36 Estel_: it is a 2 man show and they are a hard act to follow. May 23 22:00:49 :) May 23 22:00:55 and a quite autistic act May 23 22:01:04 DocScrutinizer, wtf? May 23 22:01:13 DocScrutinizer: >:( May 23 22:02:15 what's kp? May 23 22:02:36 kernel-power May 23 22:03:02 oh ok May 23 22:03:10 I really like my stock kernel :) May 23 22:03:19 You shouldn't ;) May 23 22:03:27 kp=kangaroo penis. May 23 22:03:29 it's bugged and lack basic things May 23 22:03:33 why not? overclocking is bad for your health May 23 22:03:41 like what? May 23 22:03:44 I don't talk about oc this time May 23 22:03:45 no ipv6 in the stock kernel May 23 22:03:48 I've been using it for years with no probs May 23 22:03:50 Sc0rpius: so is bacon May 23 22:03:54 no ext4 May 23 22:03:59 there's no IPv6 around me May 23 22:04:08 there's ext3. May 23 22:04:11 o_O May 23 22:04:12 oh, that's geting funny now, debating which kernel has more flaws and bugs, stock or power May 23 22:04:15 Kangaroo... penis. May 23 22:04:17 ext3 is worse ;) May 23 22:04:18 Sc0rpius: and drugs May 23 22:04:28 And drugs. May 23 22:04:30 Good night. :D May 23 22:04:38 no crypt modules for truecrypt May 23 22:04:40 what's a good advantage of an ext4 fs on a phone? May 23 22:04:54 no vdd=1 powersaving May 23 22:05:00 tuneability May 23 22:05:01 Sc0rpius: it is faster May 23 22:05:06 speed May 23 22:05:10 DocScrutinizer: which regressions are there actually in the power kernel? the only one I noticed was broken reboot, though that is fixed with the latest version May 23 22:05:15 better journal May 23 22:05:28 ahem... May 23 22:05:36 DocScrutinizer, is well known kp negatron :D May 23 22:05:36 I don't know why in my mind everything you said translate to less battery May 23 22:05:37 ...skype audio is broken May 23 22:05:43 and battery is the most precious value in a phone May 23 22:05:47 Sc0rpius, bullshitl;) May 23 22:05:49 Estel_: suuure May 23 22:05:51 BS May 23 22:05:52 Sc0rpius: lol wut May 23 22:06:01 if you tell me May 23 22:06:10 fixed SR values = 1/3 more battery May 23 22:06:13 "upgrade to kernel-power and your battety will last a week in 3G" May 23 22:06:16 then we're talking May 23 22:06:22 "battery" May 23 22:06:25 Sc0rpius: not quite May 23 22:06:32 buy dual scud from me and it will ;) May 23 22:06:50 Sc0rpius: however smart reflex has improved battery shitloads May 23 22:06:55 Estel_: I'm a patch-oile'o'crap hater, and a git fan though I dunno much about it May 23 22:06:57 (vdd1) May 23 22:07:11 Sc0rpius, I'm using 3000 mAh battery May 23 22:07:12 * Sicelo has KP for the sole purpose of using hostmode May 23 22:07:20 and will use 6000 mAh soon May 23 22:07:41 Estel_: ha, you have more battery than n900? May 23 22:07:42 I think I'm gonna install Nemo soon May 23 22:07:44 honestly, it's hard to mention all kp benefits May 23 22:07:52 Sicelo: the you found the reason why Estel_ 's assumption regarding me and KP was terribly off May 23 22:07:54 Sc0rpius, nemo won't be usable for months or years May 23 22:08:10 by the way in another topic I installed Lubuntu in a Netbook and that's amazing. It even has Office suite and everything and it's just 3 GB on the HDD May 23 22:08:20 can't I make a call with current Nemo? May 23 22:08:24 yeah DocScrutinizer. i was about to mention that to him :P May 23 22:08:32 Sc0rpius, using stock kernel is like sitting on pr 1.0 or 1.1 May 23 22:08:54 haters gonna hate=docsrutinisers gonna be May 23 22:08:55 Sc0rpius, you can install lubuntu on N900 May 23 22:08:55 I don't agree. May 23 22:09:03 yeah but it will stop being a phone May 23 22:09:06 and I want a phone May 23 22:09:16 it's my main phone if I can't make a call is useless May 23 22:09:17 Sc0rpius, why You've installedd pr1.3? May 23 22:09:24 just because Nokika released it May 23 22:09:30 after all, it's bugfixes May 23 22:09:30 well I'm in PR 1.3 + CSSU May 23 22:09:32 kp too May 23 22:09:38 gtg May 23 22:09:40 bb May 23 22:09:45 why, You havent had phone on pr 1.1? May 23 22:09:52 see ya Vvi May 23 22:09:54 I don't remember May 23 22:09:59 * merlin1991 is about to bid on n900 on ebay May 23 22:09:59 one can never have enough xD May 23 22:09:59 or rather I'm tired of my cssu-testing device burning its battery in 10 hrs May 23 22:10:08 I think my phone came with 1.2 pre-installed when I bought it May 23 22:10:25 so why 1.3 Sc0rpius May 23 22:10:30 what I notoriously hate though are patches introduced to kernel by fools who got no clue, like e.g. that bq27200.ko and the I2C locking fuckup it brought with it May 23 22:10:35 or not downgraded, 1.1 uses less space ;) May 23 22:10:38 Estel_: he actually meant on ubuntu.. that he wont' have a phone May 23 22:10:50 DocScrutinizer, but you still use kernel-power nonetheless, right? May 23 22:10:55 DocScrutinizer, bq i2c id dedcade old thing May 23 22:11:07 oh yeah if I install Lubuntu on my N900 it stops being a phone, that's what I meant. May 23 22:11:07 fixed when dinosaurs walked on earth May 23 22:11:21 bwahaha May 23 22:11:39 DocScrutinizer, now every path must be first submitted on tmo thread and explained May 23 22:11:49 if those dinosaurs hadn't kicke3d some asses, it wasn't fixed til now yet May 23 22:12:00 ;) May 23 22:12:30 probably because it wouldn't hurt anyone until sun becomes supernova (although I agree it was wrong) May 23 22:12:44 oh yeah tmo, home of the *real* kernel experts May 23 22:12:47 btw it was patch from titan's times May 23 22:13:00 DocScrutinizer is so filled of hate today :) May 23 22:13:01 DocScrutinizer, maintainers read patches May 23 22:13:20 so, can I make a call with current Nemo? May 23 22:13:31 during Pali and freeman times every patch is reviewed, that's why luke-jr thing was kickedd out May 23 22:13:45 Sc0rpius, would love to answer, althoughm no idea May 23 22:13:52 what luke-jr thing? May 23 22:14:14 there was some patch enabling low frequency cutoff via smth May 23 22:14:26 but, it would also affect headphones May 23 22:14:50 gdenerally, it would be nice thing to have, despite fact that we have speaker's protection mechanism already May 23 22:14:51 and how T F is that kernel now? May 23 22:14:52 well there's only one way to know May 23 22:15:01 but in gstreamer pulseaudio or somewhere May 23 22:15:11 but his kernel patch did it in kernel May 23 22:15:28 if code would be better and adapt itself to situtionp, it would be ok May 23 22:15:35 DocScrutinizer, we have some hw thing May 23 22:15:41 controllable by kernel May 23 22:15:42 Sc0rpius: report back, please :) May 23 22:15:48 that was meant to do filtefring May 23 22:15:56 * Sc0rpius is installing U-Boot, first step of the installation May 23 22:16:05 anyway, probldem was that luke-jr just used some code May 23 22:16:12 and wouldn't be able to maintain it May 23 22:16:16 btw we got a nice gifted guy named HNT or sth like that, who implemented that stuff into ALSA like 18 months ago May 23 22:16:17 in case of bugs etc May 23 22:16:26 so patch was kicked out in a second May 23 22:17:13 with a lot of help from my side May 23 22:17:15 I just mean that whinning about quality of patches in KP is wrong since maaany months May 23 22:17:23 DocScrutinizer, where is his patch? May 23 22:17:28 implemented anywhere? May 23 22:19:26 just for the record: May 23 22:19:29 IroN900:~/bin# uname -a May 23 22:19:30 Linux IroN900 2.6.28-omap1 #1 PREEMPT Fri Aug 6 11:50:00 EEST 2010 armv7l GNU/Linux May 23 22:21:44 well, so it's not iron anymore ;) May 23 22:21:58 I would rather say mercurium May 23 22:22:04 or graphit May 23 22:22:09 :P May 23 22:23:19 as much as I would be afraid of using 2.6.28 on desktop due to kernel bugs, I would be afraid of using it on N900 without all KP patches, making it no longer 2.6.28 in reality (ammount of upstream patches) May 23 22:23:43 at least, I would not call it "more stable" than patched counterpart May 23 22:25:36 ping X-Fade May 23 23:16:56 merlin1991, I lost refference to wiki article, where it was explained with details how to properly encode icon for debian/control file May 23 23:17:11 8 know it's base64, but exact command and tool was given there May 23 23:17:25 do You remember proper way of doing so? I can't find this document May 23 23:17:39 and wiki got plentora of less or more verbose docs on packaging May 23 23:17:51 (none for now got detailed info on icon) May 23 23:19:38 nvm, found it finally May 23 23:48:21 we,ll I'm booting Nemo May 23 23:50:53 nice May 23 23:51:17 I can make a call May 23 23:51:31 but it's funny, I call my voicemail and it says the password is incorrect May 23 23:54:21 well, I blame voicemail May 23 23:54:32 I pressed the numbers very slowly and it worked May 23 23:54:41 the portrait virtual keyboard works VERY well May 23 23:55:13 I can't find an SMS app :( May 23 23:56:31 virtual leyboard on N900 sounds almost like an regression :P May 23 23:56:47 a/leyboard/keyboard/ May 24 00:11:28 I wonder how am I supposed to zoom in Nemo with an N900 if it's not multitouch :) May 24 01:03:29 Estel_: shrug, I'd rather miss some freqs in headphone, than my N900 speakers be destroyed **** ENDING LOGGING AT Thu May 24 02:59:58 2012