**** BEGIN LOGGING AT Sun Nov 04 03:00:02 2012 Nov 04 05:25:40 Hello world Nov 04 05:26:10 http://web.archive.org/web/20110703120502/http://wiki.meego.com/ARM/N900/Install/Dual_Boot -- Is it still safe to use the 21.2011.38-1 kernel instead of the PR1.3? Nov 04 09:13:00 archey: why would you wanna do this? but probably the answer is a "yes, but..." Nov 04 09:23:27 DocScrutinizer05: Why not dualboot maemo with meego? Nov 04 09:24:24 meego isn't supported anymore, but what I meant is: why not use uBoot as of now? why not use current pr1.3 kernel? Nov 04 09:24:56 btw I'm not sure there ever been a 21.2011.38-1 kernel Nov 04 09:36:43 archey: you nmight wanna look there: https://wiki.merproject.org/wiki/Nemo/Installing#Nokia_N900 Nov 04 10:38:51 spoofy: yo Nov 04 10:39:16 Is there any chance I can have your nmap 6 deb again? Nov 04 11:00:10 what's the issue with 'current' nmap? Nov 04 11:01:44 you'd think that except for capture-all function it's pretty OS agnostic a tool, no? Nov 04 11:29:40 DocScrutinizer05: the current nmap is epicly broken. Nov 04 11:30:09 when you exit the program by ctrl+c it brakes the shell. Nov 04 11:30:36 nothing appears when you type until you have entered redet twice. Nov 04 11:30:36 mmmpf Nov 04 11:30:40 :-S Nov 04 11:30:46 ^reset Nov 04 11:31:03 furthermore it cannot be run as root. Nov 04 11:31:10 WTF??? Nov 04 11:31:30 it freaks out trying to figure out which interface to use. Nov 04 11:31:32 what else if not root could nmap run under anyway? Nov 04 11:31:50 if you specify an interface it still freaks out. Nov 04 11:31:58 sudo Nov 04 11:32:11 yeah, interfaces been what I already mentioned (implicitly) above Nov 04 11:32:23 the only way to make it run is as sudo. Nov 04 11:32:30 o.O Nov 04 11:32:45 even then only somje of the features work. Nov 04 11:32:54 as I said, utterly broken. Nov 04 11:32:58 sudo id tells me that any program run this way has user: "root" Nov 04 11:33:34 the user spoofy compiled a non optified version of nmap 6 and made a deb. Nov 04 11:33:45 it works perfectly. Nov 04 11:33:59 wtf wrong with optification?? Nov 04 11:34:08 but he took it down because he wanted to make zenmap work as well. Nov 04 11:34:23 sounds like a fix applied due to not understanding the problem Nov 04 11:34:27 apart from being a kludge? Nov 04 11:35:16 I wouldn't want a non-optified nmap just for the sake of somebody not getting the fundamental unix concepts used by optification Nov 04 11:35:19 spoofys nmap 6 worked fine, you just needed to copy+symlink to get your rootfs space back. Nov 04 11:35:49 DocScrutinizer05: it was just a quick compile. Nov 04 11:36:01 he had not gotten round to it yet. Nov 04 11:36:04 well move and symlink is exactly what maemo-optify* does Nov 04 11:36:12 yup. Nov 04 11:36:53 and for the life of me I can't see why an optified nmap would break while non-optified it worked Nov 04 11:36:58 anyway, instead of waiting for zenmap to happen (pah! gui nonsense!) I am bugging him to let me have the deb again. Nov 04 11:37:30 DocScrutinizer05: I dont think optification had anything to do with the breakage. Nov 04 11:37:30 I'd bug nmap maintainer to update the friggin pkg Nov 04 11:37:49 DocScrutinizer05: I highly doubt he is still around. Nov 04 11:38:10 then reclaim the pkg maintainership and do it :-D Nov 04 11:38:26 fame basically for free Nov 04 11:38:35 anyway, this evening I have set aside some time to try and set up SB on my new PC. Nov 04 11:38:41 I hate fame. Nov 04 11:38:50 hehe Nov 04 11:39:22 when I get SB set up I intend to start updating some packages that I like to use. Nov 04 11:39:22 my empathy on your SB install adventure Nov 04 11:39:54 DocScrutinizer05: once you have trie to set up bitbake, scratchbox is a holiday. Nov 04 11:40:04 but the rest of your plan is well worth the effort to install SB Nov 04 11:40:13 indeed Nov 04 11:40:24 I had it running before. Nov 04 11:40:38 I "know" bitbake from openembedded and SHR/Openmoko Nov 04 11:40:40 anyways I have to get to fixing my motorcycle. Nov 04 11:40:50 bitbake... Nov 04 11:40:55 * vi__ shudder Nov 04 11:41:04 *scream* Nov 04 11:41:04 the horror. Nov 04 11:41:35 bitbake has literally turned me bald and ground potential years off my lifespan. Nov 04 11:42:01 I happily take that for true Nov 04 11:42:45 bitbake sounds like something used to make bitcookies, the new bitcoin flavour. Nov 04 11:43:05 EHLO teotwaki Nov 04 11:43:15 Host DocScrutinizer05 Nov 04 12:32:15 dang, learning something old every other day: http://wiki.maemo.org/Opt_Problem obviously there been pretty comprehensive reasonings regarding optification - the more I wonder why we ended with what we got now Nov 04 12:32:48 anyway seems in Nov 04 12:32:52 ~optification Nov 04 12:32:53 optification is a inventive duct tape workaround to reclaim space in fs root, done due to the fact the systeminit *and* partitioning is FUBAR, http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging,_Deploying_and_Distributing/Installing_under_opt_and_MyDocs, or ""OMG - I wish they looked into FHS and moved /usr to eMMC"", http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 bullet1,2 and fhs-2.3.html#PURPOSE16 dot3" Nov 04 12:33:38 there's a misconception: they *did* look at FHS, and for some reason rejected or ignored the resulting design decisions Nov 04 12:36:44 >>You know, historically /usr exists entirely for the purpose of separating the application partition from the underlying system, which is exactly our problem. -> Make /usr another partition << LOL! Nov 04 12:37:44 I suspect the issue been that this point is all down at bottom of this list, so everybody already picked his favorite before reaching it Nov 04 12:38:06 (just kidding - maybe) Nov 04 12:38:30 hard to tell without insight into the actually process Nov 04 12:39:16 "All the core nokia software and libraries should be relocated to / or maybe a /maemo." Nokia: "WHAT THE HELL? Not gonna happen!" Nov 04 12:41:22 also see recent ~poettering systemd-introduced fuckup of definition of /usr Nov 04 12:42:24 Lennart: "FHS? Pah, obsolete bullshit!" Nov 04 12:43:05 ~poettering Nov 04 12:43:05 'sth is poettering' means it acts invasive, possessive, destructive, and generally in an egocentric exacerbating negative way. ``this cancer is extremely poettering'' Nov 04 12:45:43 SpeedEvil: actually it seems to me that they boggled from the hard cut they would've had to introduce with PR1.2 to get this flying Nov 04 12:47:04 there's hardly any sane way to *upgrade* from a non-FHS optified system like PR1.1.1(?) and before to something obeying FHS and moving /usr to a separate partition mounted 'late' Nov 04 12:49:39 so better perpetuate bugs introduced in the past rather than make a clear cut and start all new while resetting current system config during the process. And then there's all the packages that might already rely on mess regarding /usr supposed to me part of kernel rootfs in early boot, and all those pkgs would break on the new system, unless fixed Nov 04 12:50:07 all the packages *in repo* Nov 04 12:51:14 hard to tell if *any* of those packages would notice /usr missing during early boot - after all those 3rd-party pkgs are not exactly supposed to get started early Nov 04 12:52:29 http://wiki.maemo.org/Opt_Problem has a 'running gag' which is "speed? hard numbers?" Nov 04 12:53:30 * ShadowJK thinks extras should've been placed in /usr/extras/ or similar Nov 04 12:55:30 ROTFL @ >>OneNAND performance figures: Sustained read performance: 108MB/s. Sustained write performance: up to 17MB/s<< Nov 04 12:56:06 unless that's Mbit, they're delusional Nov 04 12:56:45 read sounds high Nov 04 12:56:51 indeed Nov 04 12:59:08 * ShadowJK gets 20M/s Nov 04 13:01:09 It's pretty pointless anyways Nov 04 13:01:32 if it had 1M/s write speed it'd still be faster than emmc and all sd cards Nov 04 13:01:39 because it can be managed intelligently Nov 04 13:08:20 t900:~# time dd if=/dev/mtd5ro of=/dev/null bs=1M count=100 100+0 records in 100+0 records out real 0m5.818s user 0m0.008s sys 0m5.250s Nov 04 13:08:55 -> 20MB/s as well Nov 04 13:09:32 though: t900:~# time cp /usr/lib/locale/locale-archive /dev/null real 0m2.459s user 0m0.016s sys 0m1.984s Nov 04 13:10:06 -rw-r--r-- 1 root root 29763056 Jan 22 2010 /usr/lib/locale/locale-archive Nov 04 13:10:36 NB this is valid only for very first run, since after that it comes from buffer Nov 04 13:12:14 t900:~# time dd if=/dev/mmcblk0p1 of=/dev/null bs=1M count=100 100+0 records in 100+0 records out real 0m6.722s user 0m0.000s sys 0m2.594s Nov 04 13:12:21 20% slower Nov 04 13:13:28 watch the diffs in sys usage though Nov 04 13:14:36 while for mmc it seems iowait is the bottleneck, for OneNAND that's CPU performance Nov 04 13:14:42 l Nov 04 13:16:16 yes Nov 04 13:16:55 so it might easily turn out that overall system performance is better on mmc Nov 04 13:19:51 ShadowJK: (1MB/s faster than...) well, that's for sure true for *writing*, just writing is maybe 1% of IO activity, possibly less Nov 04 13:20:48 definitely for topic of a separate /usr partition on eMMC the writing should be next to zero Nov 04 13:21:49 Yes, but writing a single file touches a minimum of two erase blocks, which in worst case creates 8M of read-modify-write cycles, which in class 10 card is almost a second's worth of latency, which blocks all pending (and future) reads for that duration Nov 04 13:22:13 so all the "we might move files to OneNAND aka /-fs if they are frequently read, to speed up stuff" reasoning seems rather moot Nov 04 13:23:26 ShadowJK: yes, sure. But that's irrelevant for evaluation pros and cons of proper separate /usr partition Nov 04 13:25:22 actually keeping stuff that gets written to, like /var/log, on rootfs while moving stuff that needs frequent reads like /usr to eMMC would _cure_ that blocking issue Nov 04 13:25:34 Well yes, entire speed discussion is irrelevant to that. :P Nov 04 13:26:08 also moving /home (excluding MyDocs and opt) would help Nov 04 13:26:25 hmm Nov 04 13:26:29 moving to NAND? Nov 04 13:26:37 yes Nov 04 13:26:43 yeah, sure Nov 04 13:26:55 but not without sane quota ;-) Nov 04 13:27:00 I get much better performance from gpodder after moving its dir to uSD :-) Nov 04 13:28:19 and of course only after killing trackerd/thumbnailerd with fire Nov 04 13:31:50 * DocScrutinizer05 ponders keeping ext3-journal on a different storage and forbid *any* writeback unless device locked and thus not interactive Nov 04 13:32:45 what about using ext2? :) Nov 04 13:32:52 horrible idea that Nov 04 13:33:17 how's that helping to avoid 1second+ write lockups? Nov 04 13:33:32 ext2 performs worse in my experience Nov 04 13:34:25 the journal of ext3 acts a bit like a write cache, which helps reduce the stalls a bit Nov 04 13:34:31 well, ext2 has no journaling Nov 04 13:34:53 if you keep your journal on NAND or uSD, your eMMC will never get locked by write events in the way ShadowJK explained above, until you allow journal writeback to ext3 Nov 04 13:36:34 Writing *to* the journal on emmc/usd is almost "free", you take the hit once it starts putting stuff from journal to the correct places, but then you gain a little bit because it can benefit more from the ioscheduler getting fed more stuff at once Nov 04 13:37:20 so while interactive, the device can read from eMMC full speed and doesn't see 'stalls' from write events, and when screen gets blanked (device locks, non-interactive period) the journal can get played back and stuff gets written to eMMC Nov 04 13:37:26 I played with the idea of using a *gigantic* journal, and gigantic writeback delay Nov 04 13:37:59 that's exactly what I sketched above Nov 04 13:38:22 virtually infinite writeback-delay Nov 04 13:38:37 until user goes inactive Nov 04 13:38:56 right moment to have some inevitable 'freezes' Nov 04 13:39:08 even better would be filesystems that only have journal, kinda like logfs ;p Nov 04 13:39:09 nobody will notice Nov 04 13:39:12 what happens if user did not shut down properly? Nov 04 13:39:30 battery too low, user unplug batter etc Nov 04 13:39:43 then journal gets replayed on next boot Nov 04 13:39:44 that's the point of journal Nov 04 13:40:21 so normally the data gets written twice? Nov 04 13:40:28 yes Nov 04 13:40:42 ok, makes sense Nov 04 13:41:19 I hope so, otherwise God wasted energy when giving me my brain ;-D Nov 04 13:41:39 it's mounted with writeback option, metadata is written twice, filedata once Nov 04 13:42:29 sure you need to tweak mount writeback options - among other stuff Nov 04 13:48:00 ~ask does anyone know how to change the histogram region in FCam API? Nov 04 13:50:51 eh? Nov 04 13:51:00 what's that ~ask ?? Nov 04 13:51:53 ~ ~ Nov 04 13:52:01 ~~ Nov 04 13:52:12 ~key Nov 04 13:52:19 wtf? Nov 04 13:52:23 ~botsnack Nov 04 13:52:24 DocScrutinizer05: aw, gee Nov 04 13:52:49 ~ask Nov 04 13:52:49 Questions in the channel should be specific, informative, complete, concise, and on-topic. Don't ask if you can ask a question first. Don't ask if a person is there; just ask what you intended to ask them. Better questions more frequently yield better answers. We are all here voluntarily or against our will. Nov 04 14:45:13 where can i get the kernel for cssu manually? Nov 04 14:46:13 hm? Nov 04 14:46:30 timmy: there's not yet any such thing like a kernel for CSSU Nov 04 14:47:12 there is, of course for cssu-thumb2 or something like this Nov 04 14:47:17 CSSU is (and will remain to be) supposed to work with all kernels Nov 04 14:47:48 i had to install kernel kernel power 51 or cssu kernel for cssu thumb2 Nov 04 14:47:52 though recently we're working to ship a kernel with bugfixes via CSSU, however optional Nov 04 14:48:11 cssu thumb is not same as cssu Nov 04 14:48:32 for thumb you need a bugfix in kernel Nov 04 14:49:27 ok, then where can i get that? Nov 04 14:51:08 see #maemo-ssu Nov 04 17:13:48 Ugh.. Hi / Hey / Hello :) Nov 04 17:15:01 Nmap and working version of PPP (PolishPwniePhone) will be ready atthe end of the month Nov 04 17:15:46 We made a newest metasploit with newest ruby1.9.2 working with postgres Nov 04 17:15:53 looking forward :-) Nov 04 17:16:55 Now I'm using N900 as my primary device to work with my server's Nov 04 17:19:18 So when I'm done this my n900 will be changed to dev device and I'll finish stuff ;) Nov 04 17:30:20 spoofy: isn't the n900 easy to exploit too? Nov 04 17:30:31 spoofy: (because of old kernel, etc.) Nov 04 18:20:35 fasta: yeah.. that's why u should always use iptables Nov 04 18:20:55 I'll not compile snort, psad and other things to maemo Nov 04 18:21:16 but at least I made a easy front-end script to iptables Nov 04 18:21:28 It's called Bulletproof firewall Nov 04 18:21:38 Will be released with PPP ;) Nov 04 18:22:22 spoofy: acronym being? Nov 04 18:23:44 fasta: PPP? Nov 04 18:23:51 spoofy: yes Nov 04 18:24:00 PolishPwniePhone ;) Nov 04 18:24:19 Polish - couse I"m polish :P That's why me english is not so well Nov 04 18:24:33 PwniePhone - couse inspiration to made it was PwniePhone Nov 04 18:24:39 spoofy: I only would need something to semi-protect me from outside traffic, and preferably from browser exploits. Nov 04 18:24:45 But It's clean image Nov 04 18:25:13 fasta: I'm working on compiling new webkit and midori Nov 04 18:25:15 Too bad security still comes at the 100th place. Nov 04 18:25:31 spoofy: have you overclocked your phone? Nov 04 18:25:49 fasta: N900 non oc. is like a bike without wheels.. Nov 04 18:26:14 spoofy: so, how did you overclock it? Nov 04 18:26:28 fasta: power-kernel-settings? Nov 04 18:26:38 U mean the profile or limits? Nov 04 18:26:44 spoofy: frankly, I don't really have time to read through thousands of posts on t.m.o. Nov 04 18:26:53 fasta: Yeah.. me too Nov 04 18:26:54 spoofy: some people had widgets to control this stuff. Nov 04 18:27:20 I always update searea - the siri clone but I don't have a time to check updates and progress Nov 04 18:27:21 spoofy: it looked really cool and I was like: ok, I will have one of those, but then there are no clear instructions. Nov 04 18:27:37 fasta: like vi said - gui is for n00bs :P Nov 04 18:28:09 spoofy: I am more of an Emacs person Nov 04 18:28:15 spoofy: but I can use both Nov 04 18:28:24 spoofy: in fact, I like both. Nov 04 18:28:30 Isn't that insanity!? Nov 04 18:28:41 nope :) Nov 04 18:28:52 I only think the VIM extension language is rather bad. Nov 04 18:29:04 Emacs at least will some day become Guile based. Nov 04 18:29:14 I don't use any gui stuff on my desktop but one of my servers have a centos with fluxbox Nov 04 18:29:21 Isn't that insanity?! Nov 04 18:29:25 No GUI stuff? Nov 04 18:29:30 I mean Nov 04 18:29:36 I like compiz and xfce for example Nov 04 18:29:47 Those are certainly GUIs. Nov 04 18:30:00 I know people who run without X. Nov 04 18:30:05 but it's only for making transparent terminals look more fancy :) Nov 04 18:30:06 I think they are insane, though. Nov 04 18:30:25 spoofy: I use gestures too. Nov 04 18:30:33 spoofy: for opening terminals of course. Nov 04 18:31:25 fasta: u mean - mouse/pointer gestures? Nov 04 18:31:34 spoofy: yes Nov 04 18:31:35 I prefer keyboard shortcuts :) Nov 04 18:31:40 spoofy: I also have those. Nov 04 18:31:46 spoofy: depends a bit on where my hand is. Nov 04 18:32:10 fasta: that's why I love ibm/lenovo T series notebooks :) Nov 04 18:32:23 my hands are always on the keyboard Nov 04 18:33:49 fasta: so.. You want to overclock your n900, right? Nov 04 18:34:01 You have any SSU or CSSU? Nov 04 18:34:08 spoofy: yes, because by default it's rather sluggish. Nov 04 18:34:12 spoofy: CSSU Nov 04 18:34:29 fasta: what kinda cssu? Stable, testing, thumb? Nov 04 18:34:43 spoofy: where can I see that? Nov 04 18:35:02 spoofy: I have ssh on the phone. Nov 04 18:35:04 First of all I suggest you to read t.m.o ;) Nov 04 18:35:12 spoofy: I just forgot. Nov 04 18:35:24 spoofy: I do know where I can see the version. Nov 04 18:35:53 spoofy: stable Nov 04 18:35:56 Settings > About Community SSU Nov 04 18:35:57 good old uname? Nov 04 18:36:07 Ok.. stable Nov 04 18:36:30 Hmm.. It's like a redhat stuff.. I prefer some bleeding edge feel with thumb Nov 04 18:37:20 spoofy: I do prefer things that work. Nov 04 18:37:41 I think you must install a powerkernel with power-kernel-settings Nov 04 18:37:58 reboot and load it from terminal like > kernel-power load Nov 04 18:38:47 that how I did it last time when I had stable cssu Nov 04 19:42:46 yo Nov 04 19:43:19 is it possible to have 2 scratchbox targets? Nov 04 19:43:24 Hi vi_ Nov 04 19:43:28 pingu! Nov 04 19:43:42 thumb2 and regular maemo for example? Nov 04 19:43:48 Yes ;) Nov 04 19:45:13 After default sb-setup you have i386 and armel target Nov 04 19:45:27 You can add thumb to that list Nov 04 19:46:16 tonight I shall set up scratchbox. Nov 04 19:46:37 while listening to an audiobook of dune with mrs vi_ Nov 04 19:46:44 What a glorious evening! Nov 04 19:57:12 sounds like mere fun Nov 04 19:57:18 what does one need to do to install the T-maemo versions later than T-maemo3.1 (which is what my n900 seems to have settled on, rather than 5.1 which I see mentioned in the CSSU changelog) ... and is doing so a good idea Nov 04 19:58:09 it's a good idea, and usually updates should come as SSU (you see, the name ;-D ) Nov 04 19:58:22 well, quite Nov 04 19:58:37 hence the question Nov 04 19:59:12 if update doesn't show up in HAM for you, there's sth wrong with either your MP-community-PR pkg, or with your vcatalogs, or with HAM in general Nov 04 20:00:44 I normally use the Faster Aplication Manager, but I'll give HAM a kick to see if it does anything different ... Nov 04 20:00:50 when however you installed CSSU via apt, fapman, or other unsupported method, then ... well, no support, since we got no idea what's the state of your system Nov 04 20:01:41 fapman is *deprecated* and known to not work and even fsckup CSSU (as it does with all maemo installations sooner or later) Nov 04 20:02:02 I _think_ I did all CSSU updates when prompted by HAM, but not 100% sure Nov 04 20:02:24 fapman doesn't need to deal with CSSU update to mess up CSSU Nov 04 20:02:27 ah, that'll be it then :-) Nov 04 20:02:40 it does autoremove on arbitrary occasions Nov 04 20:04:43 * fil wonders why it was necessary for Nokia's devs to pretty much ignore Debian's dependency system when cobbling Maemo together, but that's probably my Debian background showing through ;-) Nov 04 20:05:17 I guess it's been driven by mere "we don't need to deal with that cruft" Nov 04 20:06:05 it would be nice to be able to do a "apt-get dist-upgrade" without it totally killing the device, for instance (which is a mistake I made almost immediately after getting my first n900) Nov 04 20:06:21 :nod: Nov 04 20:06:29 say "thank you Nokia" Nov 04 20:07:30 I'm struggling hard to convince everybody involved into CSSU development that we need to get rid of this MP abomination and re-establish sane dependencies Nov 04 20:07:59 some "lazy" devels prefer to stick with the Nokia "way" Nov 04 20:08:26 so, the answer to my question is: flash your spare n900 down to the ground, install CSSU, and never touch fapman again ? Nov 04 20:08:28 and Nokia philosophy to forcefeed users an all-or-nothing package Nov 04 20:08:49 DocScrutinizer05: that would be a very good thing to do -- good luck Nov 04 20:09:23 fil: I'd try to apt-get install --force-reinstall(?) MP-community-PR Nov 04 20:09:40 something along that line (the names are all wrong though) Nov 04 20:09:56 just --reinstall I think Nov 04 20:10:03 :nod: Nov 04 20:10:08 ok, I'll give that a go Nov 04 20:10:23 and MP-community-PR is also slightly differentöly named Nov 04 20:10:43 apt-cache search community should help Nov 04 20:10:51 sorry, tired Nov 04 20:12:43 mp-fremantle-community-pr Nov 04 20:12:46 apt-get --reinstall install mp-fremantle-community-pr Nov 04 20:12:48 yup Nov 04 20:13:26 good luck Nov 04 20:13:45 if that fails, you still could try to simply install CSSU-T again Nov 04 20:13:52 without prior reflashing Nov 04 20:15:32 would it not be a good idea to flush fapman off of people's systems by having one of these CSSU packages conflict with it? (or has that already happened, and I missed it because I'm not getting updates?) Nov 04 20:16:28 hey, nice idea Nov 04 20:16:54 we probably should make improved HAM conflict with fapman Nov 04 20:17:36 vi_: read channel logs for my answer about nmap ;) Nov 04 20:17:49 the problem though is, iirc, that HAM doesn't resolve conflicts (yet) Nov 04 20:18:09 spoofy: I do not have the channel logs. Nov 04 20:18:15 care to repeat the answer? Nov 04 20:18:38 as soon as anything conflicts, it just throws that annoying "use PC-Suite" error Nov 04 20:18:48 DocScrutinizer05: oh, of course it doesn't (that makes so much sense ;-) ) Nov 04 20:19:32 at least we nuked that idiotic advice regarding "use PCsuite" now, along with some minor bugfixes and speedups, in CSSU HAM Nov 04 20:21:00 the wiki instructions on setting up scratchbox are wrong. Nov 04 20:21:12 vi_: http://mg.pov.lt/maemo-irclog/latest.log.html ;) Nov 04 20:21:12 plus CSSU HAM re-introduced red-pill mode Nov 04 20:21:13 BTW that reminds me, is there a way of shutting up the "Do you agree to these terms" that happen on upgrades? (of course I agree with the sodding terms, at least for the Free stuff) Nov 04 20:21:21 5. Since Scratchbox Hathor is based on Debian Squeeze, and the SDK script assumes etch is to be used, we need to make a small change to it: Nov 04 20:21:42 fil: HAM is opensource Nov 04 20:21:45 ;-) Nov 04 20:21:46 The instruction should say something a bit more generic. Nov 04 20:23:54 DocScrutinizer05: yes, I know -- although I _was_ thinking more in terms of a configuration option than: Step 1, set up your cross compliation environment... Nov 04 20:29:24 DocScrutinizer05: Hmm, no change there -- when you say "simply install CSSU-T again" what do you mean? (sorry about this, but I always feel like I missed the induction session when reading the CSSU docs) Nov 04 20:34:05 DocScrutinizer05: BTW I do have 2 n900's at least in part so that I can flash one back to the metal while still being able to make phone calls, so if the only sure route to a sane phone is by doing that, it's no great hardship Nov 04 20:34:50 fil: seriously, you're missing instructions on http://wiki.maemo.org/Community_SSU ? Nov 04 20:35:18 fil: I meant: click that icon on that page again Nov 04 20:39:23 fil: btw what happened in detail when -reinstall-ing mp-fremantle-community-pr ? Nov 04 20:39:52 also, did you update before doing so? Nov 04 20:40:31 fil: what's the result of `apt-cache policy mp-fremantle-community-pr` ? Nov 04 20:41:20 nope, I'm not missing instructions on that page ... but I seem to have been missing that page entierly of late. Is it linked to from anywhere obvious on the maemo.org site? Nov 04 20:41:44 err, I think so, yes Nov 04 20:41:55 I didn't do an update immediately prior to the reinstall, so I'll just try that Nov 04 20:41:56 we had that advertisment banner upper right Nov 04 20:42:15 (not here, since I'm using a different skin/theme) Nov 04 20:42:33 do an update, then a apt-cache policy Nov 04 20:43:04 see what's your installed mp-fremantle-community-pr, and what's the latest available version Nov 04 20:47:43 Ah, I see -- seems that HAM thinks I have "no catalogues in use" which explains why it's no longer doing any updates Nov 04 20:48:56 lxterm fucking piece of fucking shit. Nov 04 20:49:39 apt-get is offering to do a fairly substantial update now -- am I to believe that HAM fapman and apt are all using independent package lists locally? Nov 04 20:50:04 Scratchbox login executable... no │20:48 [ ZogG_laptop ] Nov 04 20:50:07 E: Scratchbox login found but not executable by user. Nov 04 20:50:12 wtf this means Nov 04 20:50:13 ? Nov 04 20:53:00 anyone? Nov 04 20:53:03 ? Nov 04 20:53:17 what is the problem with SB? Nov 04 20:53:37 did you logoff/logon after installing SB? Nov 04 20:53:54 vi__: ^^^ Nov 04 20:54:06 yes, I even went as far as closing the shell. Nov 04 20:54:27 unless of course you mean I should go init 2 Nov 04 20:54:31 on "scratchobox/login" does not work? Nov 04 20:54:33 vi__: looks like it's owned by root, and not executable by all, so you (as the user "user") are not alloed to execute it Nov 04 20:55:03 fil: but SB is not supposed to be run as root Nov 04 20:55:40 all the executables in /scratchbox/sbin have executable flag set for everyone. Nov 04 20:55:53 ah, it's trying to run "login" is it? (I know nothing about maemo really, so am just falling back on *nix experience BTW) Nov 04 20:56:16 vi__: is the above the result of "/scratchbox/login"? Nov 04 20:56:57 ... Nov 04 20:57:05 executable flag not set for user. Nov 04 20:57:10 fil: kinda different lists, yeah Nov 04 20:57:14 hur dur. Nov 04 20:57:28 what distro is that? Nov 04 20:57:31 though they update one the other, or rather supplement Nov 04 20:57:42 debian testing Nov 04 20:57:54 fil: you should start HAM and see if it offers an update now Nov 04 20:58:00 DocScrutinizer05: genius Nov 04 20:58:27 fil: wait for several minutes, HAM might still need quite a while until the update shows up Nov 04 20:58:31 debian wheezy Nov 04 20:58:36 vi__: hmm, I think someone had problems with newer distros Nov 04 20:58:55 DocScrutinizer05: BTW, what I did to fix things (which I seem to have done) is: apt-get update ; apt-get install hildon-application-manager Nov 04 20:59:14 jesus christ, dont tell me even our compiler environment is now obsolete. Nov 04 20:59:17 hey :-) Nov 04 20:59:37 vi__: why not just download vmware image and give it a shot Nov 04 20:59:49 yeah, it's now offering me ....5.1 Nov 04 20:59:50 where? Nov 04 20:59:54 that is what I use, NP at all Nov 04 21:00:05 freemangordon: I would like that. Nov 04 21:00:06 on Nokia's maemo SDK page Nov 04 21:00:07 linky? Nov 04 21:00:17 is that the maemo wiki Nov 04 21:00:19 ? Nov 04 21:00:28 just a minute Nov 04 21:00:42 I have found it. Nov 04 21:01:08 be careful to download PR1.3 version Nov 04 21:01:27 ah, I see, now HAM dies because I didn't let get apt to deal with the dependencies, so let's do that now... Nov 04 21:01:42 noo Nov 04 21:01:55 That image is an wonderful SDK, everything you need without hassling with setting-up Nov 04 21:01:56 wait, HAM most probably doesn't 'die' Nov 04 21:02:09 Maemo_Ubuntu_Intrepid_Server_SDK_Virtual_Image.zip 232057407 Maemo Ubuntu Intrepid Server virtual SDK image Nov 04 21:02:12 it just takes loooong Nov 04 21:02:12 Maemo_Ubuntu_Lucid_Desktop_SDK_Virtual_Image_Final.7z 1457450621 Maemo Ubuntu Lucid Desktop virtual SDK image (Fremantle PR1.3) Nov 04 21:02:15 Maemo_Ubuntu_Lucid_Server_SDK_Virtual_Image_Final.zip 591051761 Maemo Ubuntu Lucid Server virtual SDK image Nov 04 21:02:18 ^which? Nov 04 21:02:38 hold on till I check Nov 04 21:02:39 fil: check your catalog list Nov 04 21:02:56 DocScrutinizer05: no, sorry, let's say "chokes" :-) Nov 04 21:03:09 though, that should be ok, just make sure toi disable extras-devel and -testing, as they make HAM take forever Nov 04 21:03:21 vi__: Maemo_Ubuntu_Lucid_Desktop_SDK_Virtual_Image_Final.7z Nov 04 21:03:33 this is doing the job though: apt-get install mp-fremantle-community-pr Nov 04 21:03:38 vi__: 1457450621 Maemo Ubuntu Lucid Desktop virtual SDK image (Fremantle PR1.3) Nov 04 21:04:59 at least, it's dragging in a load of packages, which seemed to be the ones that made HAM choke and mutter about PC Suites Nov 04 21:05:49 mmm Nov 04 21:06:04 only 10 hours remaining at 31 kb/s. Nov 04 21:06:08 -_- Nov 04 21:06:38 * fil boggles at the effort that seems to have been applied by Nokia to subtly break things that work fine in Debian --- Microsoft are welcome to them ;-) Nov 04 21:07:11 vi__: WHAAAT? Nov 04 21:07:52 anyway, once you have it up and running - do not upgrade in guest OS Nov 04 21:08:20 fil: should work as well, if you got the proper catalog/repo (which you obviously got since otherwise you couldn't install the MP) Nov 04 21:09:01 Ok Nov 04 21:09:25 I have altered the script to think I have debian squeeze. Nov 04 21:14:40 does not including diff in busybox actually save any worthwhile amount of space? (it's a bit tiresome not being able to diff configuation file changes) Nov 04 21:15:32 fil: Nothing about busybox is worth it. Nov 04 21:15:48 The crippled tools do not justify the small space saving. Nov 04 21:16:05 At least not an a maemo device anyway. Nov 04 21:18:52 vi__: having spent years sodding about with d-i, I somehow have a soft spot for busybox, but it seems unlikely that leaving diff out is saving anything noticeable -- and as you say, the n900 is perfectly capable of running the real tools, I'd think -- smacks of premature optimisation to me, presumably by embedded engineers who _knew_ _best_ Nov 04 21:22:45 vi__: the rule to not upgrade is that at some point kernel upgrade breaks qemu and dh_helper is not able to build some packages anymore Nov 04 21:23:16 If those maemo-bootscripts weren't that messy, we probably would running full utils already? Nov 04 21:23:31 fil: indeed Nov 04 21:23:39 fil: anyway, get bash plus proper tools Nov 04 21:23:44 since: Nov 04 21:23:47 ~messybox Nov 04 21:23:47 messy... err busybox is meant for lean scripting. Regarding all the missing options and immanent limitations (see su, passwd) it's not really the interactive shell of choice. A lot of people hate busybox because a lot of system integrators don't understand the difference between busybox and a decent user interactive shell plus unix utils Nov 04 21:24:02 also see: Nov 04 21:24:05 ~jrtools Nov 04 21:24:05 [jrtools] http://wiki.maemo.org/User:Joerg_rw/tools Nov 04 21:25:12 fil: nota bene I don't suggest to replace busybox by anything, au contraire: keep stock busybox for boot, use proper bash and unixtools in your interactive shell Nov 04 21:25:51 fil: I bet you're aware there's no limitations of any kind regarding concurrent installation of different shells Nov 04 21:26:09 busy box is measurably faster in some cases Nov 04 21:26:15 just get your $PATH right in your interactive shells Nov 04 21:26:30 when it comes to forking Nov 04 21:26:50 internal tool startup is very light Nov 04 21:27:01 SpeedEvil: so what? if I'd bother about that, I still could use `busybox cmd` isntead of `cmd` Nov 04 21:27:22 I mean for boot Nov 04 21:27:27 in my interactive shell and scripts Nov 04 21:27:31 vi was arguing there is no reason Nov 04 21:27:39 should have clarifies Nov 04 21:27:41 d Nov 04 21:27:43 oh yeah, for boot messybox is absolutely fine Nov 04 21:31:17 does it win over e.g. dash? (I'd guess it does for the builtin stuff) Nov 04 21:37:03 I think would sacrafice slower boot for bash + full gnu tools :P Nov 04 21:42:18 me too. Nov 04 21:47:58 Pali: do you know, is it possible for OSS radeon driver to mimic closed source behaviour when there is second display attached Nov 04 21:48:24 i.e. to play videos fulscreen on second display Nov 04 21:48:33 *fullscreen Nov 04 21:48:51 freemangordon, do you want to play fullscreen video on other display? Nov 04 21:48:56 yep Nov 04 21:49:02 this is not driver, but application problem Nov 04 21:49:25 but closed source driver does it without application support AFAIK Nov 04 21:49:47 at least under windows it behaves like that Nov 04 21:50:02 I do not understand now... you can move any window to any from randr Nov 04 21:50:03 and according to google the same happens under linux Nov 04 21:50:07 naah Nov 04 21:50:18 my second display is actually a tv-set Nov 04 21:50:29 did you started second Xserver for second display? Nov 04 21:50:34 no Nov 04 21:50:42 or you configured output via xrandr? Nov 04 21:50:57 i configured it via "displays" in ubuntu :D Nov 04 21:51:11 NFC what it does Nov 04 21:51:21 and it works, np with that Nov 04 21:51:26 can you move that window to second display? Nov 04 21:51:37 (excluding that id did not recognoze the supported resolutions) Nov 04 21:51:38 s/display/output/ Nov 04 21:51:38 Pali meant: can you move that window to second output? Nov 04 21:51:40 yes Nov 04 21:52:00 but I don;t want that Nov 04 21:52:15 hm, then problem is in application I think Nov 04 21:52:18 i want overlay cloned fullscreen on the second display Nov 04 21:52:36 NFC if overlays are used at all Nov 04 21:53:03 do you have second output clone of first? Nov 04 21:53:08 no Nov 04 21:53:27 and I don;t want to Nov 04 21:53:54 ok, so you want 2 outputs: 1. monitor and 2. tv, right? Nov 04 21:53:59 as I am seeing only the back of the second monitor (tv-set) Nov 04 21:54:03 yep Nov 04 21:54:14 and I have them Nov 04 21:54:23 bot heads work Nov 04 21:54:26 *both Nov 04 21:54:59 and problem is that if you change in your app mode from window to fullscreen it move window to first output? Nov 04 21:54:59 second one shows only background, but that is normal Nov 04 21:55:21 no, lemme explain again Nov 04 21:56:12 under XP, when I play a video, I can setup CCC to output the overlay fullscreen on the second display. without it being fullscreen/maximised on the first Nov 04 21:56:34 it is called "theater mode" or something Nov 04 21:56:59 is it clear now? Nov 04 21:57:35 actually it is the same on n900 :D Nov 04 21:57:54 you play video in mediaplayer and have the same video cloned fullscreen on TV-out Nov 04 21:58:19 what is CCC? Nov 04 21:58:27 catalyst control center Nov 04 21:58:37 driver UI Nov 04 21:59:56 I do not know how it worked on winxp, but is this correct: you want to tell mplayer to play fullscreen video on tvout output? Nov 04 22:00:25 Pali: something like that. but it is not TV-out, it is DVI Nov 04 22:00:36 it just happens I have TV-set connected there Nov 04 22:00:54 video-card is dual-headed Nov 04 22:01:26 BTW i know you don;t know how it works under XP, that is why I explained :D Nov 04 22:01:51 if you can move mplayer window to second output, then press here F and it will be in fullscreen Nov 04 22:01:56 (on second output) Nov 04 22:02:59 Pali: sure, but that is just a workaround. And this is more or less unusable as I see only the back of the second display Nov 04 22:03:43 back of second display? Nov 04 22:03:44 anyway, thanks, will continue my fight :D Nov 04 22:03:47 yep Nov 04 22:03:56 I do not understand... Nov 04 22:04:19 it is rotated in such a way that I see only its back Nov 04 22:04:37 mrs vi_ just got a nexus 7... Nov 04 22:04:42 with host mode! Nov 04 22:04:54 I don't see the image when I am in front of my primary display Nov 04 22:04:58 (that does not support writing to external media) Nov 04 22:05:05 wtf. Nov 04 22:05:20 yes. Nov 04 22:05:38 the ability to mount external filesystems without the option to write to them. Nov 04 22:05:46 oh google such a faux pas! Nov 04 22:06:12 Pali: I use the second display to watch videos. Or better said my daugther watches videos while I use my PC :D Nov 04 22:06:30 anyway, thanks for the effort :) Nov 04 22:06:37 bbl Nov 04 22:07:07 ~bbl Nov 04 22:07:08 rumour has it, bbl is be back later Nov 04 22:08:22 freemangordon, as I understand correctly, you want to tell mplayer to start fullscreen playback on specified output? Nov 04 22:14:39 Pali: no, to clone the video which plays on the primary as fullscreen on the secondary Nov 04 22:16:06 so you want to clone video window (which is on first output) to fullscreen second Nov 04 22:17:54 fakeroot apt-get install maemo-sdk-debug Nov 04 22:17:58 fails with: Nov 04 22:18:07 fakeroot: preload library `libfakeroot-sysv.so' not found, aborting. Nov 04 22:21:59 fakeroot: preload library `libfakeroot-sysv.so' not found, aborting. Nov 04 22:22:08 ^the fuck this means? Nov 04 22:30:28 wow, all I had to do was wget libfakeroot from merlin1991 server Nov 04 22:30:36 how...cryptic. Nov 04 22:45:05 freemangordon: AV/TV-out as on N900 is not app nor is it a generic display setup, there's a 'daemon' mirroring the video framebuffer to second display Nov 04 22:48:56 ok night guys Nov 04 22:49:19 usual "config display" multihead setup tools just offer either the most commonly used multidisplay (N adjacent displays form one large screen), or a complete mirroring of 1st display to secondary Nov 04 22:49:27 night ZogG_laptop Nov 04 22:51:19 ... Nov 04 22:51:34 I may now have a scratchbox environment... Nov 04 22:51:46 freemangordon: ping. Nov 04 22:52:43 That was a slice of velvety cake compared to bitbakes punch in the dick with a cheesegrater. Nov 04 22:52:44 vi__: \o/ Nov 04 22:53:28 now all I need is a spare n900 or 2 to test shit on. Nov 04 22:53:43 first you need shit to test ;-P Nov 04 22:53:54 packages I want to update: Nov 04 22:53:57 mplayer Nov 04 22:53:59 nmap Nov 04 22:54:07 great Nov 04 22:54:15 some kind of xterm that does not blow goats. Nov 04 22:54:15 start with "hello world" ;-) Nov 04 22:54:27 DocScrutinizer05: I was thinking gnu fortune... Nov 04 22:54:37 ...WITH FULL NEON OPTIMISATION. Nov 04 22:54:44 -O3 Nov 04 22:54:54 EAT IT COWSAY. Nov 04 22:56:55 xterm blows goats? Nov 04 22:57:19 DocScrutinizer05: 2 pixel border making it 79 chars wide. Nov 04 22:57:24 no middle click paste. Nov 04 22:57:27 Hello. I'm trying to install hamachi on n900 but I'm getting this error "getopts: not found". Is there any fix for that? or maybe I can simply link getopts to getopt? Nov 04 22:57:42 middle click paste ? :-o LOL Nov 04 22:57:47 ok ok ok Nov 04 22:58:05 great to hear you're using h-e-n Nov 04 22:58:20 lol all you like. taking my fingures off the KB just to open the silly menu to get to copy and paste is a PITA. Nov 04 22:58:32 DocScrutinizer05: me? Nov 04 22:58:56 DocScrutinizer05: no, middle click paste from the keyboard. Nov 04 22:59:02 what else would you do do issue a middle click, than use a 3+ button USB mouse? Nov 04 22:59:17 shift+touch screen. Nov 04 22:59:25 ouch Nov 04 22:59:26 shift+touch screen==middle click Nov 04 22:59:27 yep Nov 04 22:59:40 function+touch screen==right click Nov 04 23:00:06 mrxvt is almost perfect but you cannot change the font size. Nov 04 23:00:07 I'm pretty sure that shift+touch is a right click Nov 04 23:00:24 ~vi-kbd Nov 04 23:00:24 extra, extra, read all about it, vi-kbd is http://wiki.maemo.org/Remapping_keyboard/user_vis_awesome_kbdmapping Nov 04 23:00:48 Sysaxed: tbh I cannot remember. Nov 04 23:01:13 the point is osso-xterm ignores middle click. Nov 04 23:01:19 and it is slow. Nov 04 23:01:24 vi__, yep, I'm using your keyboard. Nov 04 23:01:37 Sysaxed: it is everyones keyboard! Nov 04 23:01:45 :D Nov 04 23:01:53 Sysaxed: if you have suggestions let me know. Nov 04 23:02:38 vi__: I'd rather try to follow the idea of "shift+ctrl+v" - at least this works on my desktop xterm Nov 04 23:03:10 even in ssh seeion to N900, so it's obviously something with X and local xterm Nov 04 23:03:22 DocScrutinizer05: generally yes, however ctrl+shift is a pain in the ass. Nov 04 23:03:57 yes, osso-xterm is the POS in the equation. Nov 04 23:04:12 once it works, it's for sure no problem to remap it to arbitrary other key cobos. Particularly for master of awesome kbd mapping Nov 04 23:04:13 if you load up xterm or mrxvt middle click paste works. Nov 04 23:04:21 ctrl+shift is fine :) Nov 04 23:05:27 DocScrutinizer05: hey, you came up with the title of the page. I merely filled it in! Nov 04 23:05:38 vi__, I have a question... how come shift+fn+j,k,z or x types two symbols instead of one? Nov 04 23:06:14 Sysaxed: http://wiki.maemo.org/N900_Hardware_Subsystems#Keyboard Nov 04 23:06:21 written by DocScrutinizer05 Nov 04 23:06:45 vi__: you're aware you as well could've 'filled in' lkSAoueQU:edit page on wiki? Nov 04 23:07:25 I did'nt want to spoil the fun. Nov 04 23:07:46 Sysaxed: I hope it's somehow expplaining stuff so normal users can understand Nov 04 23:08:10 I can change the wiki if people feel the title is too obnoxious. Nov 04 23:08:23 nah Nov 04 23:08:35 :-) Nov 04 23:08:44 anyway, gtg Nov 04 23:08:45 gn Nov 04 23:08:47 :q Nov 04 23:08:49 n8 Nov 04 23:09:12 :q? Nov 04 23:09:24 errr vi grrr Nov 04 23:09:31 DocScrutinizer05, it's fine :) Nov 04 23:09:34 thanks Nov 04 23:09:36 vi Nov 04 23:09:37 yw Nov 04 23:10:48 vi_, as a suggestion - set Delete key as fn+backspace Nov 04 23:11:25 it's not going to work in xterm because it's shit, but anywhere elese it works Nov 04 23:14:43 oh lol, it's too late, he left... Nov 04 23:15:15 anyway, is there anything I can do with my getopts problem? Nov 04 23:15:58 DocScrutinizer05: av/tv on n900 is a HW feature Nov 04 23:16:09 in DSS Nov 04 23:17:17 it is an overlay clone, video overlay goes through the resizer pipeline and ends in TV encoder Nov 04 23:24:51 * DocScrutinizer05 just looked at that hooker800, wondered what strange ideas made Nokia think it was a good idea to donate a windows phone to each of the maemo devie program members, and laughs out loud Nov 04 23:25:13 haahahahha Nov 04 23:25:20 really???? Nov 04 23:25:24 yep Nov 04 23:25:36 hahahahaha Nov 04 23:26:21 everyone who got a N950 meego developer device eventually got a hooker800 for free Nov 04 23:26:55 now try to get maemo working on it :DDD Nov 04 23:27:06 not worth the effort Nov 04 23:27:19 maemo anyway isn't exactly portable Nov 04 23:30:33 I wish they stopped doing phones for dumb asses and start making ones for human beings Nov 04 23:32:01 I really should sell this lumia800 device, while it's not rotten and forgotten Nov 04 23:32:33 if it earns me 20 bucks, that's already a pizza and beer Nov 04 23:33:27 oh, that's why they sent phones... Nov 04 23:34:00 to pay for the maemo effort :) Nov 05 02:25:59 I wouldnt mind someone sending me free phones :| **** ENDING LOGGING AT Mon Nov 05 02:59:59 2012