**** BEGIN LOGGING AT Mon Nov 19 02:59:59 2012 Nov 19 03:07:44 RST38h: you see how sometimes cat food can nice looks? Nov 19 03:11:32 Doc: Honestly, I do not know Nov 19 03:12:17 Some of the cat stuff does smell suspiciously delicious Nov 19 03:20:49 yes confirm... Nov 19 03:21:37 sometimes pets dont want eat thats specific food... Nov 19 07:31:56 Pali: congrats, bq.. and rx51_battery are haeding for 3.8 :) Nov 19 07:32:16 hi Nov 19 07:32:30 now I'm back from Kraków and reading mails... Nov 19 07:32:44 http://git.infradead.org/battery-2.6.git Nov 19 07:33:09 I see :-) (I got mails) Nov 19 07:33:24 :) Nov 19 07:33:35 Pali: BTW who maintains that page? http://elinux.org/N900 Nov 19 07:34:12 Some edits are my and some by Sre Nov 19 07:34:47 (Sebastian Reichel) Nov 19 07:34:59 aah, ok. As I was wondering how is that possibel the page is updated even before the commits on the battery tree :D Nov 19 07:35:44 BTW there IS SGX driver for 3.6 Nov 19 07:36:10 well, fremantle one forward-ported Nov 19 07:38:10 both bq2415x and rx51_battery needs board data Nov 19 07:38:20 Pali: yeah Nov 19 07:38:21 I will write small patches for this Nov 19 07:39:11 Pali: BTW did you follow alarmUI and tklock and EDS conversations? Nov 19 07:39:42 I was not read anything after 16.11 Nov 19 07:40:00 ok. In short (if you are curious): Nov 19 07:40:06 I was in Krakow on CERC ACM ICPC Nov 19 07:40:31 there were 3 memory leaks in alarmUI. so far :) Nov 19 07:40:46 tklock leaks like mad on every swipe to unlock Nov 19 07:42:19 ahh.... Nov 19 07:42:28 every libedataserver.so client process uses a huuuge amount of heap (>2MB) Nov 19 07:42:51 funny, ain't :D Nov 19 07:44:30 fortunately the last one is open source, and I hope luf knows WTF is going on Nov 19 07:44:55 (though he is mia right now :I ) Nov 19 07:45:50 Pali: "CERC ACM ICPC" ? Nov 19 07:47:15 Pali: BTW bqxxx battery driver is missing board data too Nov 19 07:48:54 hmm, scratch that Nov 19 07:55:26 Central Europe Regional Contest ACM - competition in programming (for students) Nov 19 07:57:00 wow Nov 19 07:58:27 hehe http://git.infradead.org/battery-2.6.git/commit/6c47a3e00c6e4f3cdac7566c1480de34d9e32e07 Nov 19 07:58:53 seems we are lucky having upstream maintainer owning n900 :D Nov 19 07:59:40 Pali: please, submit my thumb2 patches :P Nov 19 08:09:40 freemangordon, try to send thumb2 patches yourself - you are author :-) Nov 19 08:10:52 Pali: I know I am the author, I just can't put myself together to do that. And as you are riding the upstreaming wave... Nov 19 08:11:26 Pali: it is not laziness, see ^^^ . Too much other useful stuff I can do Nov 19 08:11:44 ^^^ is for the memory leaks and such Nov 19 08:12:37 I'll waste too much time to figure out the correct channels, mails, procedure, tools, etc Nov 19 08:13:06 and you have all this already set up Nov 19 08:29:08 Pali: will you be able to add stuff to the bq* drivers and have them merged upstream? Nov 19 08:29:20 having single-file flags would be neat :3 Nov 19 08:29:33 kerio: what? Nov 19 08:29:47 bq drivers are in queue for 3.8 Nov 19 08:29:56 freemangordon: /sys/class/power_supply/bq27200-0/flags/VDQ for instance Nov 19 08:29:57 read the backscroll Nov 19 08:30:10 freemangordon: i am reading the backscroll Nov 19 08:30:14 that's why i said *add* Nov 19 08:30:33 ok Nov 19 10:13:55 yo. Nov 19 10:17:41 Vi_: Howdy! Nov 19 10:23:05 MrPingu: Hey man. Nov 19 10:23:47 MrPingu: What do you know about making ovi maps work? Nov 19 10:24:31 let me tell you the puzzle. Nov 19 10:24:48 I have 1 n900 that I am using to create the ultimate reflash image. Nov 19 10:25:03 That is ALL the shite stripped out and some very basic essentials loaded in. Nov 19 10:25:24 Seeems nice ;) Nov 19 10:25:45 I don't know anything about ovi maps, sorry Nov 19 10:25:50 I have almost finished. Nov 19 10:26:13 So I have attempted to clone the stripped n900 to another n900. Nov 19 10:26:30 But? Ovi maps doesn't work? Nov 19 10:26:46 And to all purposes it looks like it has worked. One device is the absolute clone of the other. Nov 19 10:27:27 ovi maps does work on the device, however when I connect to nokia suite it says the clone device does not have maps installed! Nov 19 10:27:33 I just do not get it. Nov 19 10:27:37 How can that happen? Nov 19 10:27:38 For one thing I know, ovi maps is NOT optified sits totally in rootfs so Nov 19 10:27:44 One device is IDENTICAL to the other. Nov 19 10:28:13 Tried reinstall the maps package? Nov 19 10:28:19 Not yet. Nov 19 10:28:44 I am not sure which packages I need to purge&binge. Nov 19 10:28:54 Let me see :P Nov 19 10:29:01 MrPingu: how Do you make conversations rotate? Nov 19 10:29:29 MrPingu: I found your account on symbianfreak. Nov 19 10:29:44 I know I knew your name from somewhere! Nov 19 10:30:01 The legendary pingu. Nov 19 10:30:21 Adding it to the whitelist in transitions ini, ditch the whole blacklist concept Nov 19 10:31:11 vi_: should I know you too? Or were you a silent freak? :P Nov 19 10:31:24 I came too late. Nov 19 10:31:39 The whole forum was winding down before I got started. Nov 19 10:31:53 Also I did not like the some of the culture of the place. Nov 19 10:32:02 All moved to dailymobile Nov 19 10:32:13 It's a bit personal there ;) Nov 19 10:32:36 Most of the people I have on FB too Nov 19 10:32:40 anyway Nov 19 10:32:49 The only thing I contributed was modifying a user agent-switcher program that someone wrote to be 'iphone' instead n97. Nov 19 10:32:58 whitelist = rtcom-messaging-ui Nov 19 10:33:20 voila, conversations rotate. Need to have a modified css file though Nov 19 10:33:29 MrPingu: Do I have to modify my conversation theme? Nov 19 10:33:34 There we go! Nov 19 10:33:44 vi_: yes Nov 19 10:34:29 Let me find the default theme + fixed portrait Nov 19 10:35:33 vi_: http://talk.maemo.org/showpost.php?p=995119 Nov 19 10:36:03 use the display-outside, it looks like default + portrait Nov 19 10:36:42 but I like the small compact themes. Nov 19 10:36:50 No avatar. Nov 19 10:36:54 small text. Nov 19 10:37:00 no wasted space. Nov 19 10:37:50 http://wiki.maemo.org/Conversation_Mods Nov 19 10:38:08 Hope there's a good mod for you that has portrait support Nov 19 10:38:49 vi_: what do you know bout left over files after purging testserver? Nov 19 10:39:25 MrPingu: There is some files left in event.d Nov 19 10:39:27 1 file. Nov 19 10:39:41 and 1 in /etc/init.d Nov 19 10:40:17 You are building the crap cleaner script. Nov 19 10:40:28 You are doing god's work son. Nov 19 10:41:55 vi_: but's fucking dangerous I mean the N900 is my only device :P So I am looking forward to your updated rootfs :P Nov 19 10:47:03 MrPingu: heh. Nov 19 10:47:22 MrPingu: It is not so easy. Nov 19 10:47:38 :( Nov 19 10:48:00 fiasco-gen layout file for emmc is completely undocumented so I am unable to generate an emmc image. Nov 19 10:48:22 Which is a complete ballbreaker. Nov 19 10:48:27 However there is hope. Nov 19 10:48:54 There is ~40m-50b free in the root partition. Nov 19 10:49:29 All I have to do is hijack the 'first boot&optify' script that gets run at first boot after a reflash. Nov 19 10:49:59 That way I can sneak in a bunch of stuff to the rootfs and have it copied to the mmc at boot time. Nov 19 10:50:21 Regarding that optify-runonce, can it be purged after first boot? Nov 19 10:50:40 MrPingu: I guess so. Nov 19 10:50:59 However if I were you I would be a little more cautious. Nov 19 10:51:17 You do not want to hose your n900! Nov 19 10:52:14 I don't purge anything, I am only adding it to my scrap-paper for investigation Nov 19 10:52:22 where is optify-runonce kept again? Nov 19 10:53:05 /usr/sbin/ Nov 19 10:56:01 I guess it can't be purged, at the first glance it seems looking at the files it auto-optify not only after first boot Nov 19 10:57:34 I do not seem to have an optify-runonce in /sbin. Nov 19 10:57:36 W Nov 19 10:57:37 T Nov 19 10:57:37 F Nov 19 10:58:17 its a bunch of scripts: /usr/sbin/maemo-optify-* Nov 19 10:58:55 MrPingu: and I do not have them. Nov 19 10:58:57 mmm Nov 19 10:59:03 I must start from a fresh flash! Nov 19 11:00:06 Didn't you purge it already? Nov 19 11:00:32 I am looking on my 'normal' n900. Nov 19 11:00:42 It desperatley needs a reflash. Nov 19 11:00:56 I cannot remember what I have hacked/bodged in the last year. Nov 19 11:01:33 That is the stimulus for creating a replacement flashable root image. Nov 19 11:02:07 The question is, what crap can I add to this replacement fiasco image? Nov 19 11:02:27 obviously strip out most of the stuff as mentioned in your script. Nov 19 11:02:46 Good xbindkeys, proper keymap Nov 19 11:02:53 but would it be desirable to add flash10 and 720p HD codecs for example? Nov 19 11:03:03 For me, defenitely Nov 19 11:03:06 Busybox power? Nov 19 11:03:32 bb-power, flash10, 720p, shipped with kp50 or even newer Nov 19 11:03:48 that is easy enough, what else? Nov 19 11:03:57 Let me think Nov 19 11:04:11 A script to wipe the emmc. Nov 19 11:04:32 the uber-usd arrived Nov 19 11:04:36 You're russian roulette script, one of the six boots fail? Nov 19 11:04:41 So it is like flashing emmc but without adding 200MB of videos+mp3 junk. Nov 19 11:04:48 * ShadowJK wont have time to properly benchmark before work :( Nov 19 11:04:49 ShadowJK: How big? Nov 19 11:04:55 16gig Nov 19 11:05:04 ...do not say ~5mm Nov 19 11:05:15 16GB? Nov 19 11:05:26 THat is not so big by todays standards! Nov 19 11:05:45 but uber because it promised 150 write iops (compared to 2-4 of other cards) Nov 19 11:05:59 ShadowJK: Super fast write speeds? Nov 19 11:06:10 super fast random read/write Nov 19 11:06:16 ShadowJK: Who advertises such features on their cards? Nov 19 11:06:40 ShadowJK: LINK PLOX. Nov 19 11:06:43 well, 45M sequential read and 40M sequential write too, but I don't give a shit about sequential Nov 19 11:06:54 ShadowJK: HOLY SHIT. Nov 19 11:06:56 shortcutd? Getting too personal? Nov 19 11:07:08 MrPingu: yes, I think so. Nov 19 11:07:20 Oh vi_ remember your xbindkeys depends on some lib Nov 19 11:07:22 http://www.adata-group.com/index.php?action=product_feature&cid=7&piid=178 Nov 19 11:07:31 MrPingu: Although I cannot live without it, I would not force it on another person. Nov 19 11:07:38 http://www.adata-group.com/upload/ProductFeature/productImage3281.jpg Nov 19 11:08:10 I have got a real nice keymap, derived from you Nov 19 11:08:14 ive yet to test whether uhs-i cards work with sdhc deviced at all tho lol Nov 19 11:08:30 random write speeds are significantly improved by 99% Nov 19 11:08:41 but "more" default Nov 19 11:08:52 ^holy shit. Nov 19 11:09:12 So random write speed increased by 100% will make it ~8kB/S write? Nov 19 11:09:21 MrPingu: How do you mean? Nov 19 11:10:37 if it holds what it promises, random write should come in at about 400-600 kbyte/s Nov 19 11:10:38 ShadowJK: wow, you even get a free 60-day trial of norton internet security! Nov 19 11:10:47 ShadowJK: That is insane. Nov 19 11:11:04 finally usd card makers are getting their shit together. Nov 19 11:12:27 ShadowJK: so uh...how much monies and do they come in 32GB? Nov 19 11:13:23 Vi_: € still there, ? still on fn + . Nov 19 11:13:54 plus added `´ on fourth level F + G Nov 19 11:14:21 Are those danish characters? Nov 19 11:14:32 no, french + dutch Nov 19 11:14:39 ShadowJK: I am EXTREMELY interested in your results. Nov 19 11:14:51 ShadowJK: when will you have done your testing? Nov 19 11:15:06 Yeah they come in 32G, about 80E for that, iirc Nov 19 11:15:08 vi_: and fourth level . = ¨ Nov 19 11:15:20 used in german Nov 19 11:15:21 vi_, I started random write test now, so first glimpse in 5 minutes Nov 19 11:15:27 * ShadowJK goes change clothes before work in the meanwhile Nov 19 11:16:20 All the characters are mentioned are dead Nov 19 11:16:26 51 ops/sec, 4K random write Nov 19 11:16:38 s/are/I Nov 19 11:16:40 (12-25X improvement over fastest cards I've benched so far) Nov 19 11:16:43 ShadowJK: forget work, you have more important issues to settle! Nov 19 11:16:58 from an oinline review: Nov 19 11:17:00 Not as advertised. Not fast. Just a normal class 4 product. after many many many benchmarks the highest write speed was 4.3MBs the slowest 3.77MBs. Sucks. Nov 19 11:17:56 MrPingu: My irc client does not display utf8 characters properly so I cannot see what you are typing. Nov 19 11:18:14 Aah. That is too specific. Nov 19 11:18:18 English only! Nov 19 11:18:53 But there again, the RX-51 file DOES contain a DE specific layout. Nov 19 11:19:06 Perhaps it should be modified. Nov 19 11:19:56 vi_: true, but the english keyboard = dutch keyboard Nov 19 11:20:23 and these sign are used in dutch too, not often but they are Nov 19 11:20:47 16k: 24 (.38M/s), 32k: 24 (.78M/s), 64k: 11 (.75M/s), 128k: 6(.81M/s) Nov 19 11:20:48 MrPingu: Å¡ is not used in German :) Nov 19 11:21:23 ShadowJK: I'll wait for sandisk to answer ;) Nov 19 11:21:29 kolp: but ¨ + letter so ë ü ö Nov 19 11:21:31 adata is crap Nov 19 11:21:55 ShadowJK: But can an n900 make use of the uhs-i features? Nov 19 11:22:20 vi_: what features exactly? proprietary fs? Nov 19 11:22:43 freemangordon__: The 'magic' what makes it go fast. Nov 19 11:22:48 I dunno. Nov 19 11:23:07 vi_: the "magic" is in the card itself AFAIK Nov 19 11:23:18 The litrature suggests it has some kind of 'backwards' compatability. Nov 19 11:23:39 That implies you need special moder hardware to make it work fully. Nov 19 11:23:58 and i really doubt a card can saturate host controller io on n900 anytime soon Nov 19 11:24:04 iirc UHS-I specifies new bus mode Nov 19 11:24:21 I know nothing about uhs-i cards. I only just doscovered them about 20 minutes ago. Nov 19 11:24:35 ShadowJK: sure. But anyway we don;t really care about sequental access Nov 19 11:24:40 yes Nov 19 11:25:21 "UHS-I cards, specified in SD Version 3.01, support a clock frequency of 100 MHz (a quadrupling of the original "Default Speed"), which in four-bit transfer mode could transfer 50 MB/s." Nov 19 11:26:11 Isn't there something about the n900 usd slot running at only half speed? Nov 19 11:26:12 hmm, I think host controller on n900 is clocked at 80Mhz Nov 19 11:26:23 Because nokia were to pussy to crank it to the max? Nov 19 11:26:32 but might be wrong, need to check in the sources Nov 19 11:26:34 Or am I thinking of something else? Nov 19 11:26:55 vi_: iirc this was on n8x0 iirc Nov 19 11:27:02 freemangordon__: oh well. Nov 19 11:27:05 You're thinking of N800/N810 maybe, the TI controller isn't specced to run at max Nov 19 11:27:16 * ShadowJK was actually thinking about putting this card in his N800 Nov 19 11:27:26 iirc it runs full speed on n900 Nov 19 11:27:32 ok Nov 19 11:27:57 but whether it was 50MHz or 80MHz i just can't remember Nov 19 11:28:44 but even @50MHz, the max theoretical throughput is 20 MB/s Nov 19 11:28:45 which I am pretty ok with :D Nov 19 11:29:50 testing sequential write now Nov 19 11:30:10 ShadowJK: do you test it in n900? Nov 19 11:30:11 the card itself advertises a preferred erase size of 4 Megs Nov 19 11:30:16 No I'm testing it in sheevaplug Nov 19 11:30:21 aah,ok Nov 19 11:30:37 it will be ineresting to check how it will behave in n900 Nov 19 11:30:46 freemangordon__: did you ever finish your defragging ramzswap? Nov 19 11:30:46 * ShadowJK wonders if the binary compiled on sheevaplug would work on N900 Nov 19 11:31:10 ShadowJK: connected via USB should be ok afaik Nov 19 11:32:10 vi_: I am fighting memory leaks, no matter how good your swap algo is, it will became ineffective at some point if your system leaks memory like a botomless basket Nov 19 11:32:15 at 64k block size sequential write, I'm already getting 11.7 Mbyte/s Nov 19 11:32:40 freemangordon, not if you have 2G swap :))) Nov 19 11:32:47 hehe Nov 19 11:32:56 and 4G of RAM :P Nov 19 11:33:31 I'm actually at 256M swap used on my N800 now Nov 19 11:33:37 a little bit sluggish, since it doesn't have nokia swap algo Nov 19 11:33:45 which is why I think I'll put card in that Nov 19 11:33:57 ah, have to run Nov 19 11:34:08 will log into sheevaplug from work later nad check sequential write speeds Nov 19 11:34:21 vi_: Most probably I will finish that someday. But have in mind my queue - mem leaks, kernel in cssu, 720p stuff, etc. Nov 19 11:35:11 vi_: or better said - I will winish it, but it will take time Nov 19 11:35:14 *finish Nov 19 11:35:29 freemangordon__: you will *finnish* it. Nov 19 11:35:32 Like nokia. Nov 19 11:35:36 is finnish. Nov 19 11:35:42 aaah. Nov 19 11:36:22 hehe Nov 19 11:36:37 If I rolled a 'community' firmware that contained 720p codecs and flash10, whould I be allowed to talk about it on tmo or would that count as advertising warez? Nov 19 11:36:39 back, had to run to get my train xD Nov 19 11:37:04 chem|st: If I rolled a 'community' firmware that contained 720p codecs and flash10, whould I be allowed to talk about it on tmo or would that count as advertising warez? Nov 19 11:37:04 vi_: I suspect nobody on earth would care Nov 19 11:37:19 vi_: I don;t see a reason why not talk about it Nov 19 11:37:32 vi_: I mean other than those wanting to install Nov 19 11:37:40 So linkinking to distributing nokia binarys is not considered bad? Nov 19 11:37:58 Refer to qgil, statement Nov 19 11:38:26 We were allowd to distribute anything if it would aid the communiy, right? Nov 19 11:38:40 MrPingu: like sygic mobile maps? Nov 19 11:38:47 flash isn't part of that, I suspect Nov 19 11:38:56 vi_: angry birds! Nov 19 11:39:28 vi_: sygic is not nokia, so that would be applicable to qgil statement Nov 19 11:39:40 would NOT Nov 19 11:40:02 I suppose qgil statement applies to binaries (nokia/non nokia) already freely distributed by nokia Nov 19 11:40:38 MrPingu: cssu-s+flash 10+720p codecs+some vm tweaks (done the right way, not swappolube)+power kernel+fapman+rootsh-all preinstalled crap. Nov 19 11:40:45 And I'm not sure it continues to apply if nokia stops distributing them Nov 19 11:41:25 vi_: fapman you sure? Nov 19 11:41:44 yeah, too personal. Nov 19 11:42:20 also +script to wipe mmc just like emmc flash. Nov 19 11:43:00 vi_: That script must explicitly be called by user Nov 19 11:43:02 what's bad about swappolube? Nov 19 11:43:35 MrPingu: Then you would have /opt full of junk! Nov 19 11:43:36 I don't want to my EMMC be wiped when I flash combined ;& Nov 19 11:43:53 MrPingu: It can leave myDocs Nov 19 11:44:35 tadzik: the defaults iirc are outdated Nov 19 11:44:41 Vi_: that seems good Nov 19 11:45:23 Vi_: Add Docs root on password with some default passwd, that can be altered later Nov 19 11:45:45 g2g, Nov 19 11:45:50 bb Nov 19 11:46:30 VM tweaks for contiguous swap writing would be awesome. Nov 19 11:47:25 SpeedEvil: I mean simple stuff like dirty backround, etc & nr_requests. Nov 19 11:47:39 vi_: you are not allowed to distribute nokia binaries afaik so "yes that is warez" Nov 19 11:47:57 chem|st: Thankyou for the clarification. Nov 19 11:47:58 SpeedEvil: nokia algo works pretty well, I suspect memory leaks are the reason for such a bad swap fragmentation Nov 19 11:48:16 chem|st: those are not Nokia's :P Nov 19 11:50:09 vi_: preinstalling fapman does not sound like a good idea to me Nov 19 11:50:18 freemangordon__: Yeah. Nov 19 11:51:00 freemangordon__: Ah, it is just an idea. Nov 19 11:51:06 Luck, next train had wifi too :) Nov 19 11:51:13 MrPingu: heh. Nov 19 11:51:24 MrPingu: you have wifi in trains? Nov 19 11:51:33 freemangordon__: Not being able to put it on TMO would somewhat undermine the point of it as well. Nov 19 11:52:04 freemangordon__: Yes, most of them Nov 19 11:52:31 getting about 12M/s write seq Nov 19 11:52:33 vi_: well, I don;t think you can host it on TMO, if not for legal, then for size reasons Nov 19 11:52:49 freemangordon__: Not host, talk. Nov 19 11:52:56 aah Nov 19 11:53:11 But if the link is to something that is deemed questionable then it is a no go. Nov 19 11:53:25 ShadowJK: toldya, adata si crap Nov 19 11:53:27 *is Nov 19 11:53:36 I do not want to taste the wrath of chem|st mighty ban hammer! Nov 19 11:53:44 :D:D:D Nov 19 11:53:44 fastest card I've ever tested :) Nov 19 11:53:57 had another a-data card oncea. it died in a month Nov 19 11:54:05 so yeah I know adata Nov 19 11:55:02 ShadowJK: I have a really old 1g kingsaton in mine that gets the same sequentiaomwrite Nov 19 11:55:11 well, 11.5m IIRC Nov 19 11:55:19 ShadowJK: 12MB/s is just slightly over 10MB/s a cheap class 10 adata does. I own one of those, so I know what I am talking about. Nov 19 11:55:37 8 G class 10 Nov 19 11:55:38 class 10 is a massive bullshit. Nov 19 11:55:48 It is just more bigger numbers==better. Nov 19 11:55:56 but it gives i/o timeout errors avary now and then Nov 19 11:56:04 *every Nov 19 11:57:43 I wonder what is the random io of sandisk uhs 64G cards, I am thinking to buy one if it makes sense Nov 19 12:30:32 anybody has a debian chroot, and chromium gives segmentation fault when started? Nov 19 12:49:24 vi_: talking about that stuff would not cause a ban, I'm sure Nov 19 12:50:18 vi_: it's N900 related so the larger definition doesn't fit. Yet you shouldn't even ask ;-) Nov 19 12:52:03 vi_: applying common sense is always a good rule Nov 19 12:52:46 common sense says there's nothing terribly fishy in talking about a fiasco.bin Nov 19 12:54:53 maybe I should release some beta version of new 0xFFFF :-) Nov 19 12:59:09 vi_: so you found some more info about fiasco-gen? Nov 19 13:00:07 DocScrutinizer05, I think I understand full fiasco format, I also updated fiasco documentation Nov 19 13:00:44 Pali: that's fine, but we don't know how to operate fiasco-gen Nov 19 13:01:10 0xFFFF has implemeted both unpacking & generating Nov 19 13:01:16 OOOH Nov 19 13:01:23 so 0xFFFF can be used instead fiasco-gen Nov 19 13:02:59 Pali: that's extremely nice news Nov 19 13:04:26 GeneralAntilles: ^^^ how about MWKN? worth a note I think Nov 19 13:04:59 I will release some beta version of new 0xFFFF Nov 19 13:05:43 Pali: could you paste a good link to a tmo thread or whatever, which GeneralAntilles could use for MWKN? Nov 19 13:06:24 ok, when I drop tarball of new beta version, I will create new tmo thread Nov 19 13:06:44 :-) Nov 19 13:31:00 DocScrutinizer05, is wiki and extras-devel working? Nov 19 13:31:04 or still not? Nov 19 13:31:20 just tested wiki, borked as usual, no edit Nov 19 13:31:31 no idea about e-dev Nov 19 13:31:47 any other news? Nov 19 13:31:54 Pali: 0xFFFF can be used as a replacement for fiasco-gen? Nov 19 13:31:56 nope Nov 19 13:32:08 vi_, yes Nov 19 13:32:10 Pali: No bullshit with layout files? Nov 19 13:32:14 vi_: that's what he said :-) Nov 19 13:32:19 where?? Nov 19 13:32:25 Do I get it from nopcode? Nov 19 13:32:29 vi_, you need to create layou file still Nov 19 13:32:43 Pali: But no one knows how to make the layout files. Nov 19 13:32:48 old version on nopcode support some fiasco format Nov 19 13:32:56 ~pali Nov 19 13:32:57 extra, extra, read all about it, pali is http://atrey.karlin.mff.cuni.cz/~pali/ Nov 19 13:33:11 new version will have full fiasco support Nov 19 13:33:19 hrm. Any idea what's /usr/lib/locale/locale-archive? It currently eats almost 23 megs on my rootfs Nov 19 13:33:32 vi_ last time when I looked at that layout file it was plain text file Nov 19 13:33:35 tadzik: a mega abomination Nov 19 13:33:46 somewhere on tmo MAG wrote info about layout Nov 19 13:33:49 tadzik: It contains your locales. Nov 19 13:33:50 try to find it Nov 19 13:33:54 it's the *closed source* file of all Nokia translations/locales Nov 19 13:34:06 tadzik: You can move it to /opt and symlink it back with no trouble. Nov 19 13:34:15 Pali: I found it. Nov 19 13:34:18 really? Nov 19 13:34:19 It is not correct. Nov 19 13:34:51 The layout example on TMO is as it appears in the .bin header. Nov 19 13:35:02 not that I lack space, I just released ncdu on / and it takes 10% of it :) Nov 19 13:35:07 It is not what is required to generate a flashable image. Nov 19 13:35:26 I think moving /usr/lib/locale/locale-archive to /opt will render some binaries used during early boot extremely unhappy, no? Nov 19 13:35:44 DocScrutinizer05: Like what? Nov 19 13:35:50 nfc Nov 19 13:36:02 MMC is mounted well before any locale stuff is done. Nov 19 13:36:03 the friggin player for shake-hands Nov 19 13:36:25 The shake hands player occours after x has started. Nov 19 13:36:31 ^X11 Nov 19 13:36:43 locales 'n shit is before that. Nov 19 13:37:21 Pali: I EAGERLY await your release of 0xFFFF Nov 19 13:37:22 then why does PA break when you optify the codecs? due to that bootvideo shit Nov 19 13:37:43 DocScrutinizer05: dunno. Nov 19 13:37:52 tadzik: You must try this, for science. Nov 19 13:38:54 when my rootfs is 95% filled rather than 59%, I'll do that and see what happens :) Nov 19 13:39:28 vi_: please verify your claim that nothing needs /usr/lib/locale/locale-archive during early boot, before /home mounted Nov 19 13:39:39 DocScrutinizer05: I will. This evening. Nov 19 13:39:46 thanks Nov 19 13:39:57 I will symlink it on my stunt n900. Nov 19 13:40:26 ...right after pali releases 0xFFFF 0.6b Nov 19 13:41:49 * DocScrutinizer05 idly wonders why Nokia hasn't optified /usr/lib/locale/locale-archive if it's really not needed during early boot Nov 19 13:42:41 friggin abomination, a 21MB file with locales you wouldn't even know how to spell them, and no way to shrink it to the locales you really need Nov 19 13:43:36 well, nokia didn't optify much at all, did they Nov 19 13:45:10 Why didn't they optify maps either Nov 19 13:45:25 uh? Nov 19 13:45:56 maps? Nov 19 13:46:18 Most of nokia maps is in rootfs Nov 19 13:46:19 yeah, or microb stuff Nov 19 13:46:24 please give FQN of any maps related file that could use optification Nov 19 13:47:23 /usr/share/nokia-maps Nov 19 13:47:46 ncdu -x / is insightful :) Nov 19 13:49:02 MrPingu: right, that's terribly odd I'd say Nov 19 13:50:39 tadzik: yes that run ncdu -x /usr/share/ , be prepared to find nokia maps listed as third Nov 19 13:50:59 yeah, something of this sort Nov 19 13:51:24 Whole 7.3 MB wasted rootfs space Nov 19 13:51:28 . o O ( does anyone still use this? ) Nov 19 13:53:50 Nokia maps is not that bad. Nov 19 13:53:56 at least they still update the maps. Nov 19 13:55:18 but but, nothing bad happends when symlinked to opt Nov 19 13:55:34 this few times I tried to use them ("tell me how to go from here to this bar") it worked quite fine, but the overall UX was *horrible* Nov 19 13:55:46 "wtf do I click now" every few seconds Nov 19 14:03:54 so let's start a hardcore optification effort? strace -ff -e trace=file (pre)init -o /optify-trace.log; and see which files are actually used until /home gets mounted Nov 19 14:04:33 s/-ff/-f/. Nov 19 14:15:14 never had such a boring lecture Nov 19 14:16:18 Normally I am not at IRC when I am ay university but this is extreme ;o Nov 19 14:16:23 what lecture is that? Nov 19 14:16:46 Economics and infrastructures Nov 19 14:17:19 I remeber a Compilers lecture few weeks ago, monday, 8 AM. We watched Star Trek with a friend and he still got asleep. Nov 19 14:17:46 I'm fascinated by the topic, frankly, but the lecturer is a genius in making people asleep Nov 19 14:18:43 Next lecture will be javascript, that's getting my attention, my first programming language :P Nov 19 14:19:23 Yes lecturers are good at that, even if the topic is interesting they still can make boring Nov 19 14:19:25 MrPingu: Dont let those JS heretics get to you first. Nov 19 14:19:42 well, is there a single language that does not suck as a first language? ;) Nov 19 14:19:50 C Nov 19 14:19:58 I thought that to be C, but there are still people who say it's bad Nov 19 14:20:00 C is the only language Nov 19 14:20:04 C is the mother lanuage Nov 19 14:20:19 C...bad?? Nov 19 14:20:20 I've only seen sandisk UHS in huge sizes like 64g :( Nov 19 14:20:25 well, not that I agree ;) Nov 19 14:20:41 the a-data was smaller so not too much money to spend on benchmarking/testing Nov 19 14:20:44 ShadowJK: so sandisk do offer an alternative. Nov 19 14:20:49 I sometimes think that if someone can't understand C, they should probably stay out of programming anyway Nov 19 14:20:52 vi_: I can't help if I do, JS is forced on me Nov 19 14:21:35 Well we don't know if sandisk uhs has any better random write than their other cards :) Nov 19 14:21:36 MrPingu: May you pass through the shadow of the valley of JS and fear no inhetitance. Nov 19 14:21:36 vi_: think MIT and their Scheme Nov 19 14:21:56 ^fuck y spelling. Nov 19 14:22:02 although their choice doesn't neceserilly mean "C is bad for teaching" Nov 19 14:22:50 Don't worry, I probably only get 1 period of programming. It's not the main aim of my study ;) Nov 19 14:23:41 BRB, got to work for a living. Nov 19 14:26:53 MrPingu: the more sense it would make to use a real programming language instead of JS Nov 19 14:27:54 unless your studies are for "Media designer" or whatver else there might be that creates "web programmers" Nov 19 14:28:13 DocScrutinizer05: It's only used to crunch to some serials of numbers, and do some calculations Nov 19 14:28:41 calculations in JS, sure ;-P Nov 19 14:28:56 yeah, that's what fortran's for, right? ;) Nov 19 14:29:29 Well, don't ask me, I am not the teacher :P Nov 19 14:30:47 honestly, if I were do pick a language for teaching a semester of programming course, I'd probably use a proper assembler (IBM, whatever) for that Nov 19 14:42:02 macro assembler: http://en.wikipedia.org/wiki/Macro_assembler#Macros Nov 19 14:43:49 Back, stupid eduroam went down :P Nov 19 14:44:20 help!! i have esy debian on partition of 6gb but stl its sayin no space in /var/cache Nov 19 14:44:32 help plzz. Nov 19 14:44:53 Yeah would have liked that too, maybe I will learn qt as hobby, only maybe Nov 19 14:45:10 devilinside: mkdir /var/cache Nov 19 14:45:53 IIRC, that folder is not present on some images Nov 19 14:46:31 but my easy deb is already on 6gb partition, copied by dd command ... it has enough free space Nov 19 14:47:03 Are you sure that folder is created? Nov 19 14:47:29 its there.. i have already installed many apps but now its saying not wnough space in /var/cache Nov 19 14:47:50 devilinside: WHAT is saying it? Nov 19 14:49:09 commended packages:commended packages:commended packages:commended packages: Nov 19 14:49:58 etrying to install jdk in easy deb Nov 19 14:50:07 I'm out, next lecture is started Nov 19 14:50:45 its saying - need to get 39.2 mb of archives. not enuf space in /var/cache/* Nov 19 14:51:48 who is saying that? Nov 19 14:52:08 easy-deb-chroot Nov 19 14:52:48 sorry, can't reproduce, no such command on my device Nov 19 14:53:26 I guess you're executing a command that wants to download and install stuff *outside* of chroot Nov 19 14:53:41 no thats not a command. i tried installing app in easy-deb. its xterm is saying it Nov 19 14:53:47 nope.. Nov 19 14:54:21 df -h /var/cache/ Nov 19 14:54:23 its working fine expect for this space problem. i had shifted the image to partition by dd command.. Nov 19 14:55:49 its saying 3.0gb used out 0f 3.0gb . 32 mb free.. Nov 19 14:56:15 but its on a partition of 6gb, copied from image of 3gb! Nov 19 14:56:15 echo test >/var/cache/delete-me Nov 19 14:56:46 scratch that Nov 19 14:56:57 you messed it up by dd Nov 19 14:57:10 dd is NOT meant to get used for that Nov 19 14:57:57 you created an image, including size and all, of original Nov 19 14:58:09 then?? what is the procedure?? it was given in tutorial to shift the img to partition. Nov 19 14:58:42 devilinside: resize2fs (if it's ext2) Nov 19 14:59:35 i didn create any image .. i used debian-wheezy-sulu.img.ext2 and used this command dd if=pathtoimg of=/dev/mmcblk1p2 Nov 19 14:59:37 devilinside: the filesystem does not know/care it was moved to a larger device, unless you tell it to Nov 19 14:59:51 *sigh* Nov 19 15:00:01 yes its ext2 Nov 19 15:00:22 resize2fs /dev/mmcblk1p2 Nov 19 15:00:55 did he say he moved it to mmcblk1p2?? Nov 19 15:01:07 i didn create any image .. i used debian-wheezy-sulu.img.ext2 and used this command dd if=pathtoimg of=/dev/mmcblk1p2 Nov 19 15:01:11 aah he did Nov 19 15:01:36 DocScrutinizer05: and his device still works? Nov 19 15:01:45 yeah, amazing Nov 19 15:01:51 yess! Nov 19 15:01:53 ~dd Nov 19 15:01:54 [dd] `dd': Convert and copy a file `dd' copies a file (from standard input to standard output, by default) with a changeable I/O block size, while optionally performing conversions on it. 'info dd' for more info Nov 19 15:02:13 ~dd is GNU Disk Destroyer Nov 19 15:02:14 ...but dd is already something else... Nov 19 15:02:26 vi_: don't dare! Nov 19 15:02:42 DocScrutinizer05: I would'nt. I can see it is allready defined. Nov 19 15:02:51 ~tm0 Nov 19 15:02:52 extra, extra, read all about it, tm0 is trolls, mooses and orangutans Nov 19 15:04:07 btw wat to do wid resize2fs or something likethat... and where to use it?? easy debian or xterm?? Nov 19 15:05:47 r00t|home: ^^^ ;-P Nov 19 15:06:19 r00t|home: resize2fs is a nifty advice, but not exactly self explanatory Nov 19 15:06:48 why not? Nov 19 15:07:07 the default action is to enlarge the filesystem to fill the device Nov 19 15:07:27 tbh I wonder which genius created that tutorial and pkg, I'd probably prefer to extraxt a tarball Nov 19 15:07:56 some idiot who does not understand that chroot can be used on any directory, not just mountpoints Nov 19 15:08:09 :nod: Nov 19 15:09:26 devilinside: so, according to r00t|home fine advice, you should type `resize2fs /dev/mmcblk1p2` in your hist environment's root shell Nov 19 15:09:30 host* Nov 19 15:10:28 you maybe need to umount /dev/mmcblk1p2 prior to that Nov 19 15:10:48 or rather, 'stop' the chroot Nov 19 15:11:17 by whatever means the easy-debian provides for that - - or simply reboot Nov 19 15:14:02 didn work! its still 3gb. Nov 19 15:19:27 devilinside: you got a spare location with >3GB free storage? Nov 19 15:19:39 DocScrutinizer05: resise2fs works on both not mounted and mounted Nov 19 15:19:55 r00t|home: thanks, wasn't sure Nov 19 15:20:00 DocScrutinizer05: actually keeping it mounted has the advantage that it doesn't demand it being fscked first Nov 19 15:20:15 mhm Nov 19 15:20:22 DocScrutinizer05: in the mounted case, the resize is performed by the kernel driver, else by the userspace program Nov 19 15:20:37 r00t|home: could you talk devilinside thru that lil inconvenience? Nov 19 15:21:05 didn work! its still 3gb. Nov 19 15:21:13 what did you do, what was the output? Nov 19 15:22:49 r00t|home: I'm not sure our kernel on fremantle provides that function, so - according to your explanation - it might work better when umount'ed Nov 19 15:23:19 DocScrutinizer05: the kernel would have to be damn old not to support it Nov 19 15:23:29 run e2fsck first Nov 19 15:23:30 it *is* damn old ;-) Nov 19 15:23:40 devilinside: WELL, READ... Nov 19 15:23:42 maybe not THAT old Nov 19 15:23:45 which gave a segmentation fault Nov 19 15:24:04 well, your filesystem is totally broken, congrats, use your backups Nov 19 15:24:25 r00t|home: lol Nov 19 15:24:28 also, file a bugreport Nov 19 15:24:31 devilinside: you created that stuff on uSD, no? Nov 19 15:24:41 yup Nov 19 15:24:46 given that it's on the sd card, put it into a pc and run fsck there Nov 19 15:24:52 insert uSD to your linux PC Nov 19 15:25:24 not near any pc now :( Nov 19 15:25:44 will get to you when done .. thnx a lot guys.. Nov 19 15:25:57 devilinside: try fsck on the image file... does that also crash? Nov 19 15:26:47 then copy the whole chroot content to some folder on your linux-PC, mkfs-ext2 on that partition, and copy content back from PC to uSD Nov 19 15:27:12 hmm, probably better use tar Nov 19 15:27:15 DocScrutinizer05: better use tar... or make sure you use cp -a... Nov 19 15:27:20 heh Nov 19 15:27:23 or simply try what r00t|home suggested Nov 19 15:27:51 i've seen fsck crash before... but only on badly corrupted filesystems... Nov 19 15:28:09 BTW in subsequent tests I find that my SIP quality is almost as bad via usb net as via wifi (which I find oddly encouraging, as it means I can do packet traces on my laptop without risking changing things on the n900) Nov 19 15:28:47 fil: indeed this finding has good as well as bad aspects Nov 19 15:29:02 so the one clean call I got after changing to usb net was a fluke Nov 19 15:30:25 a) it proves that WiFi isn't the culprit and b) it enables proper tracing by full wireshark, on PC Nov 19 15:31:21 still the basic rationale aplies: if inbound RTP stream is clean, then the problem is on N900 side. Otherwise between PC and far and Nov 19 15:32:05 and vice versa for outbound stream Nov 19 15:34:15 #define SPI0_SCR *((volatile word *)0x0000F2A0) Nov 19 15:34:24 #define SPI0_SCR *((volatile word *)0x0000F2A0) Nov 19 15:34:30 ^what EXACTLy this means? Nov 19 15:34:43 specifically the second part. Nov 19 15:35:04 What are the asterisks for? Nov 19 15:35:35 tries to read a word from memory addr 0xf2a0 Nov 19 15:36:17 (volatile word *)0x0000F2A0 = a pointer to a volatile word at 0xf2a0 Nov 19 15:36:30 *((volatile word *)0x0000F2A0) dereference said pointer. Nov 19 15:36:33 vi_: (volatile word *) casts the integer a memory address, and the * in front makes it a read from that address... Nov 19 15:37:01 welcome to the departmant of redundancy department Nov 19 15:37:29 which, whatever it means, sounds pretty odd, for fiasco Nov 19 15:38:06 DocScrutinizer05: I am working on a gas sensor right now. Nov 19 15:38:11 Not fiasco! Nov 19 15:38:11 aaah Nov 19 15:38:28 then it might make sense, for a weird way to do IO Nov 19 15:38:40 DocScrutinizer05: I have been away from embedded c coding for some years. I am still trying to get back into it. Nov 19 15:38:44 memory mapped i/o Nov 19 15:39:15 ooh, it's a #define, so yes. Probably clean Nov 19 15:39:51 so it is saying: Nov 19 15:40:31 SPIO_SCR is the volatile_word found at 0x0000F2A0? Nov 19 15:41:24 SPI0 Nov 19 15:41:24 volatile tries to make it so that itll reread the contents of that memory location Nov 19 15:41:32 instead of stuffing it into a register Nov 19 15:41:35 I know what it means more or less, I am just trying to understand why it is written in this way. Nov 19 15:41:38 vi_: volatile forbids the compiler to optimize access, always reading and writing the actual memory when the variable is read/written... Nov 19 15:41:42 redundancy... Nov 19 15:41:54 department! Nov 19 15:42:03 r00t|home: aaah, I see. Nov 19 15:42:07 vi_: there is no other way to write it... Nov 19 15:42:42 r00t|home: there is: inline assembler ;-) Nov 19 15:42:54 DocScrutinizer05: damn you... Nov 19 15:43:02 vi_: there is no other way to write it IN C... Nov 19 15:43:11 r00t|home: ty Nov 19 15:43:29 vi_: jaska provided the same information, no? Nov 19 15:46:10 jaska: thank you. Nov 19 15:50:45 Is there a word for when a programmer is lazy and uses some variable as a temporary container . A variable that is actually defined for something else? Nov 19 15:51:35 in embedded that's not even necessarily bad style... Nov 19 15:52:35 ok. Nov 19 15:54:46 well, still bad style, but might be justified as an optimization method Nov 19 15:55:08 declare the variable as a union for cleaner style ;) Nov 19 15:58:55 yep, good advice Nov 19 16:02:50 ty Nov 19 16:06:38 DocScrutinizer05: and i thought i was just being silly... ;) Nov 19 16:21:43 ShadowJK: so... how's this adata awesome card? Nov 19 16:22:16 vi_: it's a bunch of words actually Nov 19 16:22:22 "programmer with a really awful compiler" Nov 19 16:33:03 fil: any further results with your SIP woes? Nov 19 16:36:13 kerio; well, it's not good, but first impression is best microsd I've tested. Nov 19 16:36:29 what do you mean "not good"? :( Nov 19 16:38:18 not as good as advertised Nov 19 16:38:25 :( :( :( Nov 19 16:40:02 unfortunately i can't access results from other cards ive tested Nov 19 16:40:16 it's in a ooo spreadsheet at home :/ Nov 19 16:41:08 hmm, aren't 150 4k IOPS just 600kB/s? Nov 19 16:43:34 yeah Nov 19 16:43:44 that's why HDDs are so slow Nov 19 16:44:18 ShadowJK: i bought very good SD card patriot EP pro Nov 19 16:44:30 ShadowJK: half price of sandisk and it's faster Nov 19 16:44:45 >good Nov 19 16:44:48 >not sandisk Nov 19 16:44:52 DOES NOT COMPUTE Nov 19 16:46:30 http://www.tomshardware.com/charts/2011-sd-cards/CrystalDiskMark-3.0-x64,2712.html Nov 19 16:47:26 i don't get why the sequential write speed would matter Nov 19 16:47:40 it does matter for cameras Nov 19 16:48:07 (also we need uSDs for the n900) Nov 19 16:48:35 it's first card that i've tried that is fast enough for my camera Nov 19 16:48:49 i can shoot at full 6fps and buffer will not fill up Nov 19 16:49:59 I have had 2 patriot uSD. Nov 19 16:50:09 Both broke within 1 month. Nov 19 16:50:20 i've had no problems with mine Nov 19 16:50:33 jacekowski: well...make sure you keep backups. Nov 19 16:50:37 and well, those cards get some heavy sequential use Nov 19 16:50:38 jacekowski: we need it for the purpose of n900 swap, we need random read-writes Nov 19 16:50:54 well no uSD card will live long as swap Nov 19 16:51:03 why is that? Nov 19 16:51:14 a lot of writes Nov 19 16:53:09 and that is why women are no good at swap. Nov 19 16:53:37 Too many rights. Nov 19 17:00:15 I will show myself out/ Nov 19 17:00:17 :Q Nov 19 17:35:19 okay!! SOLVED :) thank u !! :) Nov 19 17:48:59 dafuq Nov 19 17:50:42 jacekowski: SDs are perfect block devices in theory Nov 19 17:50:47 they remap bad sectors automatically Nov 19 17:50:50 and stuff Nov 19 18:02:45 DocScrutinizer05: ping Nov 19 18:06:52 DocScrutinizer05: Confirming: 21MB /usr/lib/locales CAN be moved to /opt/usr/lib/locales and symlinked back. Nov 19 18:07:12 MrPingu: Confirming: 21MB /usr/lib/locales CAN be moved to /opt/usr/lib/locales and symlinked back. Nov 19 18:09:17 vi___: Awesome :) Nov 19 18:10:09 MrPingu: in fact it is a simple case of adding the path to the 'auto optification' paths list to be done on first boot for it to be rolled into a new firmware. Nov 19 18:13:10 vi__: Maps(usr/share/nokia-maps/) can be added to that list too Nov 19 18:15:13 vi___: no more info yet -- I've been fiddling with pulseaudio which was upset during an upgrade, in the hope that twinkle might work if I fixed unrelated sound issues -- seems not :-/ Nov 19 18:15:35 should have time for another play shortly Nov 19 18:17:22 i had moved many things to /opt. !ike gcc , perl , python, many more except for microb-engine, the device seemed to work fine Nov 19 18:17:58 moving microb-engine causes problems! Nov 19 18:18:58 fil: good. I am interested in your results. Nov 19 18:20:26 do you happen to know what one needs to do to make i.e. Speex 16k work with freeswitch -- setting that as my only codec results in calls no longer working, and I'm not sure which end is upset Nov 19 18:21:28 (there was a forum post that claimed that 16k helped them - I guess that the problem is a bit random, and changing packet sizes is enough to make a difference, but it might have been more than that) Nov 19 18:21:28 Devilinside: python is already optified, no? Nov 19 18:22:19 there was something re!ated to python in /usr/share.. i kicked that out too Nov 19 18:22:35 yeah it is optified Nov 19 18:24:25 sorry in /usr/lib/pymodules Nov 19 18:27:32 the fuck, ltrace traces nothing on the device :( Nov 19 18:32:11 how abut jar support? any alternative to microemulator? Nov 19 18:34:41 vi___: i never had any luck getting the autooptifier to work Nov 19 18:35:20 vi___: i'm also getting "initctl: MOUNTS_OK event failed" in its log Nov 19 18:38:57 heh, I've written to a Sandisk Cruzer 16GB over 20G daily for about two years now. Endurance seems like a non issue.. Nov 19 18:42:39 * ShadowJK wishes some major benchmark site benchmarked generally available cards, with useful tests, or atleast tests described what they do Nov 19 18:43:10 although, this looks vaguely useful: http://www.tomshardware.com/charts/2011-sd-cards/IOMeter-2008.08.18,2708.html Nov 19 18:44:59 ShadowJK: will a uSD grow slower after you've wrote random data over it once? Nov 19 18:45:21 they're black boxes, you dont know Nov 19 18:48:11 http://www.tomshardware.com/charts/2011-micro-sdhc-charts/Random-Write-4-KB-QD-1-MB-s,2750.html Nov 19 18:48:14 ooh Nov 19 18:49:41 holy shit the first one is powerful Nov 19 18:51:26 holy shit, 1.4MB/s@4k blocks! Nov 19 18:51:40 vi___: my ssd can do that Nov 19 18:51:48 vi___: which one?!? Nov 19 18:52:37 freemangordon: a sandisk :D Nov 19 18:52:44 freemangordon: sandisk uhs-i card Nov 19 18:52:46 sandisk extreme pro UHS-I 16GB Nov 19 18:52:51 SDSDQXP-016G Nov 19 18:53:08 hmm, lets see if there is at least 32G Nov 19 18:53:58 unfortunately there isn't Nov 19 18:54:30 they only come in 8,16 Nov 19 18:55:01 16 is enough, to be honest Nov 19 18:55:02 It is good to see the regular sandisk 8gb class 4 up there with the best, shitting on those class 10 monsters. Nov 19 18:55:19 i doubt my sandisk 32gb is good enough to do that :c Nov 19 18:55:21 how would i test it Nov 19 18:55:22 ? Nov 19 18:56:42 Goddamit, wtf cant nokia suite find 'maps' on this n900? Nov 19 18:56:56 It is a goddamn identical CLONE of one that works. Nov 19 18:56:58 WTF. Nov 19 18:57:11 vi___: still fighting with that? Nov 19 18:57:23 MrPingu: Jesus christ it does not make any sense. Nov 19 18:57:41 vi___: make full BM backups, make a diff of the tarballs Nov 19 18:57:45 rootfs+kernel+mydocs+/opt is a clone. Nov 19 18:58:05 kerio: I just took the backup off a working device. Nov 19 18:58:08 kerio: He took another approach Nov 19 18:58:31 The fucking fuck. Nov 19 18:58:36 I just dont get it. Nov 19 18:58:56 vi___: you said nokia maps wasn't installed? Nov 19 18:59:07 clearly it's not a perfect clone then Nov 19 18:59:09 Nokia maps IS installed. Nov 19 18:59:23 just like it is installed on the originator. Nov 19 18:59:26 vi__: You reinstalled the package now? Nov 19 18:59:29 have you tried copying CAL too? Nov 19 18:59:30 It is a motherfucking clone. Nov 19 18:59:44 (note: don't do that) Nov 19 18:59:45 kerio: That is a one way ticket to a dead n900. Nov 19 19:00:54 I have even purged maps, reinstalled them Nov 19 19:02:07 all nokia suite does is look for the 'cities' folder. Nov 19 19:02:22 If it cannot find it, it says you have no maps or you need to run it once. Nov 19 19:02:28 The cities folder is there. Nov 19 19:02:47 If I rm it and rerun maps, it is recreated freshly. Nov 19 19:02:57 \_o_/ Nov 19 19:03:06 i don't even know how pc suite works Nov 19 19:03:13 what does it use to access files? Nov 19 19:03:36 kerio: pc_suite mode Nov 19 19:03:47 vi___: yeah but how, exactly? Nov 19 19:04:04 I imagine networking over usb Nov 19 19:04:55 pc-suite.... Wasn't maps only ovi-suite? Nov 19 19:05:14 oh wait your other n900 Nov 19 19:05:21 Nevermind.... Nov 19 19:05:24 You are not getting it. It is a CLONE of a working n900. Nov 19 19:05:56 vi___: clearly it's not Nov 19 19:06:17 I won't be saying anything anymore, seems I start to talk crap :P Nov 19 19:11:16 How to tell the n900 is in act_dead mode? Is it when the backlight is slightly lit? Nov 19 19:13:30 kolp: pushing the power key gives a different LED light than when it's off Nov 19 19:13:34 and turning it on doesn't give you a vibration Nov 19 19:13:41 instead, you get a reboot Nov 19 19:14:25 Mkay, thanks Nov 19 19:19:43 can someone pastebin their /var/log/*-optify-* ? Nov 19 19:25:41 kerio http://pastebin.com/iFU64ARB Nov 19 19:35:26 mm Nov 19 19:35:45 it would seem they generate different config.cfg files inside the maps folder Nov 19 19:36:05 which makes no sense. Nov 19 19:38:59 Anyone know what sapwood-server is? Nov 19 19:48:58 any idea about using gdebi with easy debian? need to use ubuntu facebook app but the dependency requires python2.7-ubuntu something something! Nov 19 19:50:23 This makes no damn sense. Nov 19 19:50:37 Every n900 I clone from this working n900 does not work. Nov 19 19:56:43 well fuck me. Nov 19 19:56:58 If I manually load g_nokia, it works. Nov 19 19:57:04 which leads me to ask: Nov 19 19:57:07 THE FUCK. Nov 19 19:58:27 Devilinside: what's wrong with dpkg -i --force-depends ? Nov 19 20:03:24 man, gdebi is a silly program... a dpkg frontend written in python... with it's own wikipedia page... Nov 19 20:04:42 not working! i have python 2.7 in debian but its not opening! Nov 19 20:05:10 trying once again!! Nov 19 20:06:02 well, what's the error you're getting... Nov 19 20:06:14 (it MIGHT help to start it from a terminal) Nov 19 20:06:51 debian repositories won't contain packages with "ubuntu" in their name, so i don't see how any automatic dependancy resolution is supposed to help you Nov 19 20:08:12 thats the problem. debian repo wont contain dubuntuin their name. Nov 19 20:08:29 you can just apt-get install python-2.7 and you'll have a python 2.7... Nov 19 20:09:45 but that would be cheating! Nov 19 20:14:05 doing dpkg and then apt-get -f Nov 19 20:17:05 is it possible to chroot into a copy of the maemo 5 filesystem and run dpkg within it. Nov 19 20:17:18 In the fremantle sdk for example (armel). Nov 19 20:18:28 Devilinside: apt-get will just remove the package again Nov 19 20:19:05 Devilinside: PACKAGE DEPENDENCIES DO NOT WORK OUTSIDE OF A DISTRIBUTION Nov 19 20:21:56 it worked! got some of the missing packages too.. Nov 19 20:22:15 but i cant find the facebook in lxde Nov 19 20:22:50 lol Nov 19 20:23:16 you should check out microsoft's products, they have this "windows" thing, you will like it Nov 19 20:23:36 now now, everyone has to start somswhere. Nov 19 21:04:53 I just stumbled across a forum post stating that Nokis didn't implement clock-rates other than 8k on the n900 -- is that right? (would ceratinly explain why calls were failing to establish when I tried 16k) Nov 19 21:05:19 Nokia, even Nov 19 21:05:24 fil: regarding what? Nov 19 21:07:42 kerio: speex clock-rate=16000 vs. clock-rate=8000 -- is that what you're querying? Nov 19 21:07:55 oh, for... huh Nov 19 21:08:01 SIP phone calls? Nov 19 21:09:46 SIP, yes Nov 19 21:10:27 (sorry, I've been wibbling on about sip since last night) Nov 19 21:42:10 wow smplayer vo fbdev source 720p slowly but works... Nov 19 21:42:21 without sound Nov 19 21:42:56 screen quality is awesome... Nov 19 21:43:49 ittle powerfull n900... hrhr Nov 19 21:50:18 well, the screen is a bit too small for 720p to make sense, no? Nov 19 21:50:48 can somone explain this command to me: Nov 19 21:50:52 mknod /dev/mtdblock0 b 31 0 Nov 19 21:51:35 vi_: that creates a block device major 31 minor 0 in /dev Nov 19 21:51:48 named mtdblock0 :P Nov 19 21:52:33 vi_: where did you get that from? /sbin/preinit? Nov 19 21:53:17 I was gonna say, if it looks familiar to me, with my limited experience, I'd have to guess that script as well. Nov 19 21:54:01 tadzik: is no so bsd quality is amazing... Nov 19 21:54:17 But a quick grep indicates that the stock /sbin/preinit does not have that command exactly, at least. Nov 19 21:54:49 but in 720p works only fbdev and x11 strange is why xv show green screen... Nov 19 21:55:08 Drathir: what kernel? Nov 19 21:55:58 tadzik: the point is you don;t have to reencode ;) Nov 19 21:56:21 (I presume we're talking about N900's /sbin/preinit)Though it does something that might effectively do that, since it does a bunch of command-substitution to craft the string used in the final mknod command. Nov 19 21:56:34 freemangordon: No, I found it in an online guide to mounting a jffs2 file system. Nov 19 21:56:52 I get as far as: dd if=./rootfs.jffs2 of=/dev/mtd0 Nov 19 21:56:59 vi_: hmm, modprobe mtdblock should do that for you Nov 19 21:57:03 and I am told no space left on device. Nov 19 21:57:14 vi_: you need nandsim Nov 19 21:57:16 iirc Nov 19 21:57:31 well the method on the maemo wiki is complete bullshit. Nov 19 21:58:38 vi_: you'd better ask pali, he was the one who helped me to mount the CAL partition while I was REing libcal Nov 19 21:59:22 vi_: BTW where do you try to mount that, on your PC? Nov 19 21:59:29 1920x800 also x11 and fbdev better works than xv green show video wheen seek Nov 19 21:59:43 freemangordon: power Nov 19 22:00:03 hmm, 720p xv should work with KP Nov 19 22:00:08 freemangordon: ver 51 Nov 19 22:00:13 yep Nov 19 22:00:16 freemangordon: can you suggest a way to extract the pr1.3 firmware filesystem from the fiasco bin BEFORE optification? Nov 19 22:00:42 vi_: flasher -x -F blah-blah Nov 19 22:01:02 freemangordon: yeah, that gets me a jffs2 Nov 19 22:01:10 that I cannot seem to mount anywhere. Nov 19 22:01:13 vi_: you'll get root.jffs2 Nov 19 22:01:20 a-data card throws input/output error during benchmarking :) Nov 19 22:01:33 freemangordon: kp? Nov 19 22:01:34 vi_: they lie to you, it is not jffs2 :P Nov 19 22:01:41 Drathir: kernel-power Nov 19 22:01:45 the one you use Nov 19 22:02:06 freemangordon: ...wut? Nov 19 22:02:11 freemangordon: open payer say 720p unrecognozed file Nov 19 22:02:30 yep, it is not jffs2, but, umm..., lemme check Nov 19 22:02:48 vi_: ubifs ;) Nov 19 22:03:16 Drathir: yeah, you need 720p codecs installed to be able to play 720p with mediaplayer Nov 19 22:03:22 then wtf is it called rootfs.jffs2??? Nov 19 22:03:24 afaik Nov 19 22:03:34 vi_: to make you mad Nov 19 22:03:47 freemangordon: well, I am quite mad. Nov 19 22:03:53 but amazing is that the responsive of system is quite good with that high resolution play... thats is interesting... Nov 19 22:04:17 vi_: well, they have reached their goal :D:D:D Nov 19 22:04:33 freemangordon: ok i will be looks that codec... Nov 19 22:04:56 Drathir: talk.maemo.org/showthread.php?t=77695 Nov 19 22:05:34 vi_: you need ubimount and stuff like that Nov 19 22:06:49 Drathir: though have in mind you need to overclock the DSP, so think twice Nov 19 22:07:12 to play 720p using DSP that is Nov 19 22:07:14 freemangordon: thanks... when more discover more things suprised and shocking me... positive way ofc... Nov 19 22:07:59 freemangordon: but without oc this codec also will work? Nov 19 22:08:10 yes Nov 19 22:08:20 thats good Nov 19 22:08:21 but 720p will stutter Nov 19 22:09:29 Drathir: though even without it you should be able to see 720p using xv. HW downsampler in KP >= 47 supports that resolution Nov 19 22:09:37 so how do I mount a ubifs fs then? Nov 19 22:09:44 ubimount Nov 19 22:09:52 one thing sad that core player from symbian isnt ported to maemo... this will be powerfull mashine... Nov 19 22:10:20 vi_: ubitools or something, not sure about the name , Nov 19 22:10:56 Drathir: never used symbian so NFC what core player is :) Nov 19 22:11:26 vi_: http://www.linux-mtd.infradead.org/faq/ubifs.html Nov 19 22:11:42 though get a bber and some snacks Nov 19 22:11:45 *beer Nov 19 22:12:32 but working 1920x800 slowly but working is amazing power of n900 Nov 19 22:12:35 vi_: but you deffinitely need nandsim. And you need Pali to tell you the parameters to pass to nandsim module Nov 19 22:13:08 freemangordon: What a complete and total utter fsck ballache. Nov 19 22:13:34 vi_: though I might be wrong about nandsim, maybe a simple dd will do the job Nov 19 22:14:39 now I have an n900 that will not flash. Nov 19 22:14:49 Wtf is going wrong with my n900 stash? Nov 19 22:15:04 freemangordon: core player is amasing thing in symbian s60v3 eg. n80 phone able to play all audio and video formats with mkv included... also network stream works great... Nov 19 22:15:12 Correction, it will flash, it just boot loops. Nov 19 22:15:34 Drathir: n80? Nov 19 22:15:59 or n8? Nov 19 22:16:41 vi_: read that howto, there is a guide how to make ubgs images Nov 19 22:16:47 *ubifs Nov 19 22:17:05 vi_: why hot flash? power lost? Nov 19 22:17:24 Drathir: he is trying to make n900 firmware image Nov 19 22:17:25 freemangordon: yes back to the past - N80 Nov 19 22:17:54 Drathir: hmm, i doubt it will play many formats, but... if you say so :P Nov 19 22:18:30 Drathir: for how long do you own your n900? Nov 19 22:18:51 freemangordon: couple of days.... Nov 19 22:19:09 aah, ok Nov 19 22:19:31 Playing different formats isn't that hard Nov 19 22:19:57 ShadowJK: on phone like n80? Nov 19 22:20:04 It's once the content encoded with the codecs inside the format require more processing power than you have, that you get issues :) Nov 19 22:20:45 ShadowJK: I don't think a device like n80 have DSP Nov 19 22:20:55 probably Nov 19 22:21:04 on n900 yes its easy but n80 with this time when popular flac and mkv play was somthng big... Nov 19 22:21:29 and with its (maybe) 300 MHz arm11 or something like that... Nov 19 22:23:21 Without dedicated decoder hw, it'd just about, maybe, manage h264 baseline profile, at almost its screen resolution? :) Nov 19 22:23:42 oh, 220MHz Nov 19 22:23:58 220 MHz Dual ARM 9 Nov 19 22:24:07 what is that supposed to mean ?!? Nov 19 22:24:09 15mb ram and 206mhz Nov 19 22:24:14 I get as far as: dd if=./rootfs.jffs2 of=/dev/mtd0 <--- mtd devices aren't block devices normal programs can use... you need to write to mtdblock0 Nov 19 22:25:09 r00t|home: not to say he was trying to mount it as jffs2 :) Nov 19 22:26:45 freemangordon, modem is the other arm Nov 19 22:27:05 ShadowJK: thanks. makes sense Nov 19 22:27:24 Later symbian gained realtime features and could run on same cpu as modem Nov 19 22:32:03 for me the s60v3 is the last worth somthing edition s60v5 not so good... Nov 19 22:32:06 r00t|home: I am as dumb as that guy you were helping earlier when it comes to ubifs. Nov 19 22:32:34 too candy way... Nov 19 22:32:53 heh Nov 19 22:33:21 i don't know ubifs ;) Nov 19 22:33:38 vi_: the link I posted ^^^ is the way to go Nov 19 22:34:18 vi_: though you may try to dd to /dev/mtdblock0 Nov 19 22:34:58 I am almost sure it will not work, as this is supposed to be nand image, i.e. mtd char, not block device Nov 19 22:35:15 vi_: mtd "memory technology device" is raw flash, without wear leveling... which is used by jffs2 (and ubifs i guess)... the mtdblock driver implements the wear-leveling part so you can access it as a normal block device Nov 19 22:36:27 r00t|home: afaik it is not so somple, one needs nandsim with image as backing file Nov 19 22:36:34 *simple Nov 19 22:36:37 freemangordon: i know from both my ipaq pda and motorola phones that you can use mtdblock to grab and restore images of mtd devices... and i don't see why an image file would be in another format Nov 19 22:37:24 freemangordon: nandsim sounds to me like it would provide an mtd device backed by a normal block devices, for testing drivers? Nov 19 22:37:41 http://www.linux-mtd.infradead.org/faq/nand.html#L_nand_nandsim Nov 19 22:37:52 r00t|home: It could be I am talking nonsense :) Nov 19 22:39:05 r00t|home: but deffinitely I needed nandsim to read/write /dev/mtdblock1 (CAL partition) dd-ed from n900 to a file and then back from file to /dev/mtdblock1 Nov 19 22:39:26 or was it mtdblock2? Nov 19 22:39:41 makes no difference Nov 19 22:39:50 this says you can write directly to /dev/mtdX: http://www.linux-mtd.infradead.org/faq/nand.html#L_nand_copyimg Nov 19 22:40:08 we shouldn't talk about stuff we don't know jack about ;) Nov 19 22:40:41 hehe Nov 19 22:40:44 freemangordon: but the tool for flashing is nandwrite, not nandsim Nov 19 22:41:06 r00t|home: ok, but vi wants to mount it Nov 19 22:41:34 either by flashing and then mounting, or with the sim, you're right then i guess Nov 19 22:42:02 freemangordon: I give up for tonight, it is getting late and I still need to fix mrs vi_ n900. Nov 19 22:42:14 afaik nandwrite will take care of bad block in nand Nov 19 22:42:28 exactly Nov 19 22:43:54 well, lets wait Pali to appear, he knows a bit about that Nov 19 22:44:17 let's hope he won't read the backlog ;) Nov 19 22:44:35 hehe, why not? Nov 19 22:45:00 he'll be aware that he explains to noobs :D Nov 19 22:49:36 ((233416 vi_: though you may try to dd to /dev/mtdblock0)) NOOOO! this doesn't work! see BM-v1 Nov 19 22:50:02 ~bm Nov 19 22:50:03 somebody said backupmenu was http://talk.maemo.org/showthread.php?t=63975 Nov 19 22:50:25 DocScrutinizer51: well, I don;t think it will work, but costs nothing Nov 19 22:50:28 to write to nand you need nandwrite Nov 19 22:50:57 mtd doesn't handle bad blocks Nov 19 22:51:57 DocScrutinizer51: hmm, we are not talking nand here, but nandsim Nov 19 22:52:10 ooh Nov 19 22:52:19 i.e. mounting rootfs in a PC Nov 19 22:52:34 :nod: Nov 19 22:52:55 I suppose vi wants to chroot and do his nice'n'cool stuff :D Nov 19 22:52:55 loopback mount Nov 19 22:53:10 mtd does not support loopback ;) Nov 19 22:53:41 the only way is nandsim with or without cache_file Nov 19 22:53:54 afaik Nov 19 22:53:58 we'rnot talking mtd here? Nov 19 22:54:15 mtd is ont top of nand Nov 19 22:54:25 ubifs is on top of mtd Nov 19 22:54:26 we're talking ubi image Nov 19 22:54:32 yes Nov 19 22:54:50 ubi supports loopmnt afaik Nov 19 22:54:57 you think ubi can mount a loopback? Nov 19 22:55:05 aah. well, NFC Nov 19 22:55:45 nandsim prolly for simulating real nand, with all quirks Nov 19 22:56:15 http://www.linux-mtd.infradead.org/doc/ubifs.html#L_rednote Nov 19 22:56:16 to run stuff like nandstress against it Nov 19 22:56:22 loopback is a nogo Nov 19 22:56:28 ok Nov 19 22:56:48 I just know jffs works Nov 19 22:57:24 since we built images for freerunner that way Nov 19 22:57:25 yeah, jffs does, but ubifs it totally diferent afaik Nov 19 22:57:32 *is Nov 19 23:10:46 I wonder if it's been vi or freemangordon who "DNGAF about what a ***** (like kerio) found out on #maemo" - anyway *he* knows how to write to mtdblk Nov 19 23:10:56 after I tought him Nov 19 23:11:13 freemangordon: why is used that ? Nov 19 23:11:16 apt-get remove gstreamer0.10-openmax Nov 19 23:13:06 Drathir: you don;t need it when you install updated gst-dsp Nov 19 23:13:14 sorry for that i ask but mostly i like to know what im doing? Nov 19 23:13:20 ok thanks Nov 19 23:13:27 you know gstreamer? Nov 19 23:13:44 that mean installed packages is replacment for that? Nov 19 23:14:01 not exactly. but it provides the same codecs Nov 19 23:14:30 Drathir: take care to follow howto exactly Nov 19 23:14:33 as i know gstreamer is some kind of different codec liblaries... Nov 19 23:15:29 Drathir: just follow the steps and you should be ok :) Nov 19 23:15:55 read the thread for some insight (yeah, I know it is huge) Nov 19 23:16:16 however, I am going to have some sleep, night Nov 19 23:16:26 freemangordon: yes i know that... only like verbose mode... thanks again a lot... Nov 19 23:17:00 freemangordon: colorfull dreams Nov 19 23:18:12 nandsim "loopmount": http://www.linux-mtd.infradead.org/faq/ubifs.html#L_ubifs_nandsim Nov 19 23:20:03 filename: /lib/modules/3.1.10-1.16-desktop/kernel/drivers/mtd/nand/nandsim.ko Nov 19 23:20:05 description: The NAND flash simulator Nov 19 23:20:06 author: Artem B. Bityuckiy Nov 19 23:28:23 (([2012-11-19 23:03:22] then wtf is it called rootfs.jffs2???)) for hysterical raisins, vi ;-P Nov 19 23:29:34 iirc it actually been jffs2 on maemo4/diablo Nov 19 23:30:36 IroN900:~# flasher --help Nov 19 23:30:37 flasher v2.8.2 (Jan 8 2010) Nov 19 23:31:07 I guess they CBA to change the file ending, prolly hardcoded to flasher Nov 19 23:32:30 my rough estimation: 10..25% of us had that "WTF"-moment about rootfs.jffs2 Nov 19 23:33:48 either forward or reverse (== concluding we might have jffs2 fs on root) Nov 19 23:40:42 * fil just installed asterisk on his n900, and the demo call to digium sounds very clear on the wire, and only a little worse on wifi ... but I guess that'll be doing IAX rather than SIP for the remote leg Nov 19 23:42:24 I'm now wondering if I can sort out the VoIP woes by connecting to a local asterisk instance, and have that talk to the remote freeswitch -- would be mad if that makes things better Nov 19 23:45:44 freemangordon: all works perfect... Nov 20 00:02:48 nice looks when played somthing in omp go to backround with ctrl backspace to app chose screen in background could see somthing funny moving pixels... but perhaps this is connected with open media player... Nov 20 00:04:55 fil: * has a relatively nice debug print iirc Nov 20 00:20:38 DocScrutinizer05: I just need to work out the relevant incatations to get * to log into FS and let me dial in and out via there, and I can do some tests, but that can wait till tomorrow -- time for bed Nov 20 00:20:58 * fil hasn't touched * for about 6 years Nov 20 00:21:45 actually, make that 8 or 9 -- time flys when you're having fun, eh? Nov 20 00:26:29 actually my time flies when I'm bored Nov 20 00:26:40 the days are endless, the years fly by Nov 20 00:27:55 I think the subjective duration of a year is measured by the number of remarkable events in it **** ENDING LOGGING AT Tue Nov 20 02:59:57 2012