**** BEGIN LOGGING AT Tue Jun 05 02:59:58 2012 Jun 05 09:53:40 can anyone point me towards what the difference between the qt-components and qt-components-10 packages? Which should I be depending on for maximum compatibility? Jun 05 09:58:34 ah, looks like qt-components-10 is what I want Jun 05 11:55:04 Elleo: you are mike? Jun 05 11:55:14 chem|st: yep Jun 05 11:55:24 ah ok, Jun 05 11:57:19 you can have the testing live... you brake all dependencies and push to testing a bit fast Jun 05 11:57:55 I just changed the dependency from qt-components to qt-components-10 Jun 05 11:58:01 nothing else about the package has changed Jun 05 11:58:23 since qt-components doesn't appear to contain harmattan compatible components, whereas qt-components-10 does Jun 05 11:59:38 and you cannot install qt-components-10 with qt-components installed... Jun 05 11:59:42 so update brakes Jun 05 11:59:52 not much I can do about that Jun 05 12:00:25 lol Jun 05 12:00:49 fsck the rest, as long as my app runs? Jun 05 12:01:07 it's not at all clear what I should be doing here Jun 05 12:01:14 Elleo: make eyrie force uninstall of qt-components first... Jun 05 12:01:19 chem|st: how? Jun 05 12:01:54 the procedure for porting from harmattan to the n900 seems very unclear to me Jun 05 12:02:04 with multiple incompatible versions of qt-components around Jun 05 12:02:28 hell, an app that's going to replace qt on my system... You know what... Jun 05 12:02:35 DocScrutinizer51: it doesn't replace qt Jun 05 12:02:51 it replaces qt-components, the com.nokia.meego qml elements Jun 05 12:03:05 the I'm mistaken Jun 05 12:03:14 DocScrutinizer51: qt-com10 is not able to overwrite needed stuff of qt-comp so it stalles and dpkg extis 1 Jun 05 12:03:23 and it's not the only app that depends on qt-components-10, it was a mistake to depend on qt-components in the first place Jun 05 12:03:35 surely qt-components-10 should be marked as replacing qt-components? Jun 05 12:03:37 Elleo: and it does not replace it tries to install at the same time Jun 05 12:03:50 well isn't that a bug in qt-components-10? Jun 05 12:04:21 for now a workaround is to have your install replace it by hand and note it to maintainers of qt-components Jun 05 12:05:02 yes but you cannot push things to testing if installing buggy stuff is required with out doing a workaround yourself Jun 05 12:05:54 or it will brake when updating... Jun 05 12:06:13 what it does Jun 05 12:06:33 0.2-1 updating to 0.2-2 dies Jun 05 12:06:49 qt-components-10 is marked with Conflicts: qt-components (<< 1.0~20111000), the latest version of qt-components is 1.0~20111000-1, presumably that's the real issue Jun 05 12:07:16 since qt-components-10 actually conflicts with qt-components 1.0~20111000 as well Jun 05 12:09:08 Elleo: 0.2-2 has just a nice and lit black screen... Jun 05 12:09:32 chem|st: with qt-components-10? Jun 05 12:09:39 yes Jun 05 12:10:07 1.0~git20111014-1+0m5 Jun 05 12:10:17 try doing an "apt-get --reinstall install qt-components-10" maybe qt-components removed some parts of that package if they conflict Jun 05 12:10:54 I uninstalled qt-components and all things eyrie pulled and started from the beginning Jun 05 12:11:15 including qt-components-10? Jun 05 12:11:44 because on my system which had qt-components-10 from the start it works fine Jun 05 12:12:20 Elleo: eyrie pulls qt-components-10 (1.0~git20111014-1+0m5) Jun 05 12:12:36 is there another version in devel? Jun 05 12:13:34 yeah, looks like there is Jun 05 12:13:42 how do I unpublish a package? Jun 05 12:14:05 I think I'll just leave this until qt-components is sorted in testing Jun 05 12:14:18 leav it in devel for now... Jun 05 12:14:26 how do I revert it to devel? Jun 05 12:16:00 I have no idea, I am no maintainer Jun 05 12:16:51 what else is different versions Jun 05 12:19:44 well I've filed a bug for the qt-components-10 issue Jun 05 12:19:49 not sure there's really much else I can do for now Jun 05 12:20:56 make it work with the version in testing? Jun 05 12:23:40 Elleo: I am pulling in from devel now, anything to know? Jun 05 12:24:11 chem|st: pass Jun 05 12:24:28 I haven't used my n900 in over a year, this is my first attempt at doing anything with QT on it Jun 05 12:24:52 qt-components-base-theme gets an upgrade too Jun 05 12:25:13 :) no worries, I am sure you will get it sorted over time! Jun 05 12:26:47 hmm, what should I be specifying with -t to get packages from extras-testing? simply doing apt-get install -t extras-testing qt-components-10 still seems to pull from extras-devel Jun 05 12:27:33 oh, I guess they all end up with the release name of "fremantle" or something Jun 05 12:28:13 I'll just disable devel temporarily Jun 05 12:28:23 for instance yes Jun 05 12:29:11 I'd try to replace qt-components-base-theme first, see if that brakes it already Jun 05 12:30:25 qt-components-base-theme 0.1.90-1+0m5.1 got upgraded with qt-components-base-theme 0.8.0-1+0m5 Jun 05 12:31:10 okay, I'll try with the various combinations of versions of both packages until I track it down Jun 05 12:31:45 seems to be qt-components-10 doesn't have com.nokia.meego Jun 05 12:32:01 Elleo: just to tell.... for the records... awesome! battery; cell strength; network; clock displayed^^ Jun 05 12:32:15 (in testing) Jun 05 12:32:27 hopefully it's just a change in the package naming Jun 05 12:32:43 because there's no way I'm rewriting the entire UI with new components ;) Jun 05 12:33:33 Elleo: ehrm how do you get the minimize and close buttons on n900? Jun 05 12:34:20 chem|st: just tap where they would otherwise be Jun 05 12:34:38 it should automatically show a message at the top immediately after start mentioning that Jun 05 12:34:52 (handled by qt-components on the n900, not by me) Jun 05 12:37:03 hmm Jun 05 12:37:07 the message does not show for very long though Jun 05 12:37:12 looks like I might have to depend on qt-components for now after all Jun 05 12:37:31 since they do provide com.nokia.meego components (even if they have all the wrong dimensions) Jun 05 12:37:43 and just hack the QML a bit for the n900 to change the sizes of things Jun 05 12:39:38 it looks like it's probably mostly the toolbar component in qt-components that's broken Jun 05 12:40:54 and it does not work... Jun 05 12:41:10 chem|st: make sure you're using a song it actually knows about Jun 05 12:41:16 that fingerprint database is quite small at the moment Jun 05 12:41:23 tried world charts... Jun 05 12:41:54 http://www.youtube.com/watch?v=UGuOm7_yOL8 <-- is known to work for example Jun 05 12:42:14 alternatively http://echoprint-data.s3.amazonaws.com/echoprint-list-of-tracks.txt has a big list of all the supported tracks Jun 05 12:43:13 nope still Jun 05 12:43:13 hi, i am having problems installing ubuntu-mobile on my n900 Jun 05 12:45:06 int_ua: are you here? Jun 05 12:45:28 michaaa62: then you may go to #ubuntu-mobile Jun 05 12:45:36 ~seen int_ua Jun 05 12:45:37 int_ua is currently on #maemo-ssu (1h 48m 30s), last said: 'freemangordon: Hi, we are talking about the 3.0 kernel with mpoirier on the #linaro channel if you are interested :)'. Jun 05 12:46:11 thanks, i try. Jun 05 12:48:25 Elleo: does the app need to find a fingerprint or does it create one? Jun 05 12:48:45 echonest claims 30M songs Jun 05 12:48:49 chem|st: it creates a fingerprint then looks it up Jun 05 12:49:01 chem|st: that's for their file fingerprinting Jun 05 12:49:08 ok Jun 05 12:49:18 chem|st: for echoprint (their open source over the air algorithm) there's only about 200,000 Jun 05 12:49:35 a shame Jun 05 12:55:52 right, I've managed to cobble together a version that works with qt-components in testing Jun 05 12:56:46 although you need to use the power button menu to close it since that version doesn't support any of the top bar stuff Jun 05 12:57:13 good... note that in the screen somewhere maybe°° Jun 05 12:58:01 maybe in the next version, I'm just going to push this out to get rid of the broken version in testing then get some sleep Jun 05 12:58:14 although hopefully by then qt-components-10 will work properly Jun 05 12:58:33 since I don't really want to have to maintain a bunch of extra hacks for the n900 Jun 05 12:59:05 Elleo: sleep well! Jun 05 12:59:21 thanks Jun 05 13:10:51 keep in mind powerbutton menu might not offer the needed functions Jun 05 13:11:33 user is free to reconfigure that menu to their liking Jun 05 13:12:20 DocScrutinizer51: well I'll add a quit button or something in a future release Jun 05 13:12:35 for now I just want to get a version in testing that doesn't use qt-components-10 since that seems broken Jun 05 13:14:09 Elleo: there's always ctrl-bs ;) Jun 05 13:14:29 heh, I guess Jun 05 14:03:44 hmmm, why can't I have my micro-sd formatted to ext3? isn't there support for that fs? Jun 05 14:04:14 dreamer: you can, but Maemo stupidly hard-coded vfat expectations some places Jun 05 14:04:19 meh Jun 05 14:06:47 luke-jr: so the mediaplayer and stuff will only read media on a vfat card? Jun 05 14:07:00 nah, you just need to mount ext3 manually Jun 05 14:07:11 also, ext3 is a poor choice for flash Jun 05 14:07:15 might consider ext2 Jun 05 14:07:29 ok. is that because of the journalling? (read something about it) Jun 05 14:07:34 ext3/ext4 assume magnetic driver behaviours Jun 05 14:07:37 yeah Jun 05 14:07:59 hmm, I just copied over 8Gb from my previous card .. don't really feel like re-formatting :P Jun 05 14:08:16 your risk :P Jun 05 14:08:27 ext3/4 are more likely to corrupt data, and destroy the card Jun 05 14:08:37 well it's a brand new class 6 card Jun 05 14:08:48 personally I'm more worried about data corruption Jun 05 14:08:56 and isn't journalling supposed to prevent data corruption? :) Jun 05 14:09:07 it's supposed to, but it assumes magnetic behaviours ;) Jun 05 14:09:40 MicroSD often write things out of order, which makes those assumptions bad Jun 05 14:09:55 because ext3 won't even do an integrity check later if it thinks it knows what's going on Jun 05 14:10:29 at least ext2 will scan the filesystem for errors Jun 05 14:10:57 luke-jr, latest CSSU-testing automounts all partitions on SD card Jun 05 14:11:05 freemangordon: cool Jun 05 14:11:10 (assuming fs is supported) Jun 05 14:11:27 * luke-jr makes note not to consider CSSU until he has a long day free Jun 05 14:11:35 (that'd mess up my own mounting scripts) Jun 05 14:11:41 :D Jun 05 14:11:50 yeah, you should remove them Jun 05 14:12:01 freemangordon: not practical! Jun 05 14:12:09 I arrange them for other OS's Jun 05 14:12:18 freemangordon: does it auto-enable swap too? Jun 05 14:12:29 would be nice if it disabled the internal swap if it did that <.< Jun 05 14:12:33 so, bla. reformat to ext2? is that the concensus? (still have to manually mount I guess?) Jun 05 14:13:07 dreamer: yes Jun 05 14:13:21 dreamer: unless you need permissions/owners, you might also consider vfat <.< Jun 05 14:13:21 luke-jr: no, there is a code for auto-enabing swap in ke-recv (legacy from maemo4 I suppose), but it is commented out Jun 05 14:13:52 luke-jr: or filesnames with things like ? or ! Jun 05 14:13:57 freemangordon: probably something to do with my experience that Maemo crashes with uSD swap Jun 05 14:14:11 luke-jr: well, I don't need it particularly. I just may want to have files >4G on it some day .. Jun 05 14:14:30 dreamer: or that Jun 05 14:14:42 you planning to run VMs on N900? XD Jun 05 14:14:47 luke-jr: It is you the only one having problems with swap on uSD then. Jun 05 14:15:08 for others it leads to a massive performance gain Jun 05 14:15:29 BTW i think flash wearing is a bit overrated Jun 05 14:15:59 O.o Jun 05 14:16:48 even an average TtF of 100000 hours means 700 years according to a document vi_ posted somewhere (compcache thread?) Jun 05 14:43:44 Anyone knows anyting about getting PEAP+mschapv2 working on maemo? Jun 05 14:50:22 dreamer: you may mount ext3 as ext2... Jun 05 14:51:41 i've been using ext3 for a couple of yrs without problem. there is some fix in a couple of files somewhere you have to make to get it to automount in maemo, but it does work. Jun 05 14:52:40 cbeau: we do not want a sd-card mounted as ext3 but if you do not want to remake-fs you can just mount an ext3 as ext2 and journaling is off Jun 05 14:53:20 cbeau: what is it, eduroam with peap mschapv2 seems to work... Jun 05 14:53:53 chem|st: I'm tempted, but I may just reformat Jun 05 14:56:05 dreamer: you do not need to reformat ext3, if you mount it as ext2 it is just ext2... Jun 05 14:56:29 ext3 is ext2+journal afair Jun 05 14:56:41 think that never changed?! Jun 05 14:57:55 chem|st: well, eduroam or not, it doesn't work for me. I think it doesn't work for anyone... these maemo forum/wiki posts are from before the last update I think... I think that's what broke my mschapv2 Jun 05 14:58:41 chem|st: do you use peap + mschapv2 yourself? Jun 05 14:59:41 nope Jun 05 14:59:49 cbeau: told you yesterday Jun 05 15:00:20 hey everyone Jun 05 15:00:39 but your outer tunnel seems to work, but your mschap doesn't Jun 05 15:00:47 chem|st: oh, i have terrible memory. see, i'd really like to find someone who is using it and it's working fine for them. or i'd like to get help from anyone to get it working. Jun 05 15:00:54 i'm having this weird problem .. Jun 05 15:01:26 chem|st: that's right. the outer tunnel is correctly communicated as PEAP, but the inner tunnel is not communicated to be mschapv2 so when I send my password, it doesn't know what to do with it. Jun 05 15:01:28 cbeau: certificate? anonymous username? Jun 05 15:02:08 mlwane: elaborate Jun 05 15:02:11 when i try to install libqtm-12-contacts i get this error: "libqtm-12-contacts: Depends: libqtm-12 but it is not going to be installed" Jun 05 15:02:23 chem|st: certificate is up-to-date and working (no warning, no issue) and the username is not annonymous and completely well received (acknowledged) as per our computer people's logs. Jun 05 15:02:35 but libqtm-12 is perfectly installed ! Jun 05 15:03:25 there is a 2nd username for mschap that has to be set afair Jun 05 15:03:48 mlwane: that may be a version problem Jun 05 15:04:00 manualy Jun 05 15:04:12 chem|st: yes, in the advanced menu. And I've done that. And I am able to verify through our computer people that it has been properly received. Jun 05 15:04:18 what network is it you are connecting to? Jun 05 15:04:36 eduroam? Jun 05 15:05:01 "The following packages have unmet dependencies: libqtm-12-contacts: Depends: libqtm-12 but it is not going to be installed E: Broken packages" that is what i get exactly Jun 05 15:05:05 you must not set a password there Jun 05 15:05:16 chem|st: my university's wireless network. it uses the same infrastructure for its eduroam ssid as for the others (multiple ssid mapping to same network) Jun 05 15:05:29 mlwane: lock into package which version it depends on Jun 05 15:05:40 chem|st: you must not set a password where? In the advance menu? I couldn't even if I wanted to: no field for that there. Jun 05 15:06:03 ok then I have no further suggestions Jun 05 15:06:09 :( Jun 05 15:06:53 yes, I know. I've tried all the obvious, read all the forums and tried everything everyone suggested. At this point I have tried all the combinations of all the options in all the menus, still no luck. Argh! Jun 05 15:07:11 I think javispedro could help me, but he's not around. Jun 05 15:07:15 cbeau: ask on the mailinglist if someone from your university or someone with eduroam setup may have a suggestion Jun 05 15:08:34 chem|st: I've done one better: I've sat down with the wireless guy and by tracking my MAC address, he could look at my requests while they were coming in and failing. It is through him that I found out the inner tunnel protocol was not being communicated to the system. Jun 05 15:09:45 chem|st: and even he has no suggestion for me at this point. There are other protocols opened (like GTC and WEP), but the former requires that I keep retyping my password everytime I move a little bit and the latter requires that I authenticate through a website every time :( Jun 05 15:09:53 but that is eap-mschap and that is working for my inner tunnel Jun 05 15:10:02 wep? Jun 05 15:10:33 horray Jun 05 15:13:28 I mean peap+gtc and wep. yeah, wep has been kept around for a few very old machines which required it. Jun 05 15:14:57 chem|st: what do you mean that is eap-mschap and that is working for my inner tunnel? Do you mean that you use gtc and if that's used than mschapv2 should work too? Jun 05 15:18:02 wait, no, that makes no sense: what doesn't work for me is PEAP+mschapv2 but what does work is PEAP+gtc but the latter method does not allow you to cache a password (in the maemo interface) so it's not practical (i have to type password everytime I move a little) Jun 05 15:18:24 chem|st: what are you saying works for you? Jun 05 15:19:51 ttls outer mschap inner tunnel Jun 05 15:20:47 chem|st: oh, i see. so i perhaps it is the peap+mschapv2 that messes up on communicating the mschapv2 part of things...? Jun 05 15:22:26 yes and that may be not an issue for you but the guys at uni... Jun 05 15:24:08 chem|st: i checked .. i have libqtm-12 version 1.2 Jun 05 15:25:03 chem|st: and libqtm-12-contacts doesn't specify any version it just says depends on libqtm-12 Jun 05 15:25:51 chem|st: do you have any idea what package on maemo would be responsible for this problem? Jun 05 15:44:45 cbeau: no but if, I am pretty sure it is closed source Jun 05 15:47:37 mlwane: do you have that problem when installing cssu? Jun 05 15:48:44 mlwane: on tmo there is a thread about having those troubles when installing cssu if PR1.3.1 is present Jun 05 15:49:53 http://talk.maemo.org/showthread.php?t=79713 Jun 05 15:57:16 ~seen ascipa Jun 05 15:57:22 i haven't seen 'ascipa', freemangordon Jun 05 15:57:35 ~seen acsipa Jun 05 15:57:36 freemangordon: i haven't seen 'acsipa' Jun 05 15:57:47 ~seen achipa Jun 05 15:57:48 achipa <~attila@a91-155-190-80.elisa-laajakaista.fi> was last seen on IRC in channel #maemo, 204d 6h 6m 56s ago, saying: 'also libqtm-12 seems to have lost the promotion block so it's borked in general as it propagated to testing'. Jun 05 15:57:49 infobot: you're blind Jun 05 15:58:19 hehe exactly what we are after but 204d ago^^ Jun 05 15:58:55 yeah :D Jun 05 16:00:27 chem|st: I already have CSSU installed Jun 05 16:00:38 chem|st: depending packages: "mbarcode-plugin-qrcode libqtm-12-dev libqtm-12 libqtm-12-declarative " Jun 05 16:02:34 hmm, maybe a new upload with "replaces: libqtm12-versit" will do the job Jun 05 16:04:07 anyway, someone should report that problem to Attila Csipa Jun 05 16:06:03 chem|st: thanks for the help, I guess it was a version conflict since it installed smoothly after disabling dev, testing and ovi repositories Jun 05 16:12:19 mlwane: good to know o_O Jun 05 17:20:17 Does anyone know how to get geolocation working with a QTWebView? Jun 05 18:07:36 helo Jun 05 18:09:26 seems like I have a broken rootfs memory. Can somebody guide me how to mark the bad blocks? Jun 05 18:13:56 anybody here to help? Jun 05 18:16:37 is it possible to clone maemo5 to an sdcard ? Jun 05 18:16:57 mlwane, yes Jun 05 18:17:22 hmm I would be also interested in that. seems my memory for rootfs is corrupted Jun 05 18:18:05 one moment, let me find the page again Jun 05 18:18:35 ok, see http://talk.maemo.org/showthread.php?t=83719 Jun 05 18:18:42 thx Jun 05 18:20:45 ok, that will not work for me, as my rootfs is borked, i/o errors even after fresh reflashing. Jun 05 18:21:08 thanks a lot taixzo Jun 05 18:21:33 so I cannot copy it over, I need to fix the filesystem / flash memoy first Jun 05 18:21:53 *memory* Jun 05 18:23:40 c3sso, perhaps you could copy from a emmc image? Jun 05 18:24:25 not from emmc Jun 05 18:26:58 I mean for flashing. Jun 05 18:32:38 seems that I could use backupmenu to fsck the root partition - does anybody know about that? Jun 05 18:33:25 do you already have bm installed? Jun 05 18:33:55 nope Jun 05 18:34:13 trying to install now Jun 05 18:34:22 hmm. I don't know if you can install it if you can't boot. Jun 05 18:36:30 I CAN boot. but the flash mamory is corrupted. So it works after first boot, but on second, I e.g. use the images, some programs to not behave correctlz, apt fails to write to /var/cache/apt/archives and so on. But on fresh flasheing, it could work.maybe I am lucky. Jun 05 18:40:18 c3sso: the guide for cloning maemo5 may help you, there is a step in which you mount the rootfs on a pc Jun 05 18:40:49 ok I will check it out Jun 05 18:51:29 ok, installing backupmenu fails - proram manger spits out "not enough memory left" Jun 05 18:54:30 seems I am stuck, as I would have to have cloned the phone before.. also how to boot into other os if I cannot install bootloader? Jun 05 19:06:10 c3sso, you can flash a kernel with uboot Jun 05 19:08:33 ok, seems I have to read the wiki more thouroughly, thanks! Jun 05 20:08:08 c3sso: you can try booting meego from an sdcard, you may be able to mount rootfs there and maube use some tools to check the system Jun 05 20:09:18 is meego more stable than Nemo? Jun 05 20:13:35 Can I get wpa2 support on an n900? Jun 05 20:15:23 taixzo, MeeGo is defunct. Jun 05 20:15:26 taixzo, so, yes. Jun 05 20:16:12 GeneralAntilles, ok. I was wondering because it seemed to run smoothly, whereas nemo always hangs after a few minutes of use. Jun 05 20:18:45 to mount maemo's rootfs 'externally', you may find NIN101's RescueOS unbeatable, among other methods Jun 05 20:22:17 about Nemo, eew, i was hoping things would get better Jun 05 21:20:42 how can I update the library of the default mediaplayer? I just added my new card Jun 05 21:21:15 fasta: it should already be supported Jun 05 21:21:30 i mean, it would be weird if it wasn't, my wifi network is wpa2/aes only, and it works Jun 05 21:22:29 kerio: also in the GUI? Jun 05 21:22:35 yeah, sure Jun 05 21:22:43 kerio: it certainly isn't in my version. Since when? Jun 05 21:23:07 kerio: It's not in the default Nokia release. Jun 05 21:23:15 kerio: is there some wiki page? Jun 05 21:23:22 * kerio just tapped on his wifi network Jun 05 21:23:31 i'm on PR1.2 fwiw Jun 05 21:23:35 which is actually a bit old Jun 05 21:23:54 kerio: are you sure your network is not just using fallback? Jun 05 21:24:10 kerio: i.e. did you explicitly configure or see the string wpa2 anywhere on your phone? Jun 05 21:24:20 kerio: I think I also have PR1.2. Jun 05 21:25:11 So, not wpa, but wpa2. Jun 05 21:25:14 pretty sure, yeah Jun 05 21:29:43 DocScrutinizer: YT? Jun 05 21:31:03 not really Jun 05 22:09:04 fasta: I use wpa2 at home, no fall-back, and have since at least PR1.2. There are tons of things that can bork wpa2 though. Search TMO forums for wpa2 setup problems and you'll find manythreads, almost all with a solution. Jun 05 22:09:30 Woody14619: and it is shown in the GUI? Jun 05 22:10:54 fasta: yes and no. There's a bug that can cause it to not show up depending on how you add the network. If you add it the normal way (hit top bar, network, and select it) then it shows up fine. If you try to use settings, internet connections, add, it doesn't. Jun 05 22:11:29 fasta: I think that particular bug was fixed in PR1.3, but I'm not sure of it, as I tend to use the former method, not the later. Jun 05 22:11:30 Is there also some way to add mediaformats to play? Jun 05 22:11:45 media formats? Jun 05 22:11:54 I.e. those that come from more modern phones like the Samsung Galaxy S2? Jun 05 22:12:33 I have no idea, what it is, quite likely it's playable with vlc on a desktop system. Jun 05 22:12:59 Still not sure what you mean... but... mplayer and vlc are available on N900. Jun 05 22:13:52 But if a particular phone is using some proprietary codec, good odds unless you're on that particular device it won't work. Jun 05 22:27:09 Woody14619: it's just that it doesn't play with the built-in media player. Jun 05 22:27:29 Woody14619: so, perhaps vlc on the phone just plays it. vlc on the desktop surely plays it. Jun 05 22:28:07 The thing which is great about this channel is that the signal to noise level is relatively high (if you don't count me ;) ) . Jun 05 22:28:15 fasta: vlc has a lot of extra codecs in it, and is activly still developed. media player has a nice base set. Jun 05 22:28:50 Woody14619: is there some way to tell the media player to take a hike and to let vlc take over? Jun 05 22:30:02 fasta: there are app reassigners, not sure what the name of the package is though. So yes. But vlc is much heavier then media player, and not quite as stable. Jun 05 22:30:23 K, am off... have to get out of here. :) Good luck fasta Jun 05 22:31:08 Woody14619: thanks anyway. I will lookin to the whole wpa2 thing. Jun 05 23:09:16 ffplay applehttp+http://www.nasa.gov/multimedia/nasatv/NTV-Public-IPS.m3u8 - venus transit, 720 Jun 05 23:09:17 p Jun 05 23:09:25 (ot :) Jun 05 23:24:57 nox-, thanks for non flash link! Jun 05 23:25:10 np Jun 05 23:25:14 :) Jun 06 00:59:55 hello Jun 06 01:00:54 i connected to my n900 with "ssh -X hostname" and did "startx" but the command is not found. is there a way to make this work? Jun 06 01:01:26 you're trying to do what? Jun 06 01:42:29 i am trying to run X applications on the n900 and display on a separate machine Jun 06 01:46:32 you don't generally want to run startx then - that will start an X server, which won't work at all remotely. Jun 06 01:46:52 Simply running an X application should forward over the X link Jun 06 01:46:57 the ssh link Jun 06 01:48:17 good point Jun 06 01:49:16 mtnman, You may want to read: Jun 06 01:49:17 http://talk.maemo.org/showthread.php?t=81408 Jun 06 01:49:27 it is about different thing, but problems with X forwarding are mentioned there Jun 06 01:49:47 thank you i will read it. thanks. Jun 06 01:50:09 SpeedEvil, what kind of X applications on N900 are suited to run over remote X? Jun 06 01:50:12 isn't it broken? Jun 06 01:50:28 many don't work Jun 06 01:50:35 do we know cause? Jun 06 01:50:45 They require local shared memory Jun 06 01:51:04 and proper way to fix it would be?... Jun 06 01:51:19 mount /dev/shm ? Jun 06 01:51:23 Rewrite the app to not require local shared memory Jun 06 01:51:31 (.e. tweaking some configs, Or recompiling something, or RE some closed bit in Maemo...) Jun 06 01:51:57 Hurrian, on N900's side? and why it won't work?:P Jun 06 01:52:27 (won't work - assumption based on common opinion that it won't work without rewriting) Jun 06 01:52:29 hmm, tmpfs is already mounted on /dev/shm Jun 06 01:52:37 yep Jun 06 01:52:41 tmpfs is irrelevant Jun 06 01:52:42 perhaps we should make it bigger? Jun 06 01:52:48 it's only 64M Jun 06 01:52:52 It requires shared memory access with th evideo buffer Jun 06 01:53:00 which is on another computer Jun 06 01:53:04 ah Jun 06 01:53:04 which will be ... trick Jun 06 01:53:08 y Jun 06 01:53:11 SpeedEvil, why is it problem on N900, and isn't in generic? linuxbox Jun 06 01:53:19 It is. Jun 06 01:53:24 Not all stuff works remotely. Jun 06 01:53:27 normally Jun 06 01:53:32 understood. Jun 06 01:53:36 More stuff is designed to fallback gracefully Jun 06 01:54:28 SpeedEvil, the link I mentioned is a WIP to make N900 connected with Raspberry Pi via USB networking, into using Pi's HDMI-out Jun 06 01:54:34 it already works for easy debian... Jun 06 01:54:54 forfwarding whole thing to different video output... (easy debian processed on N900) Jun 06 01:55:07 but, still can't make it work properly for Maemo itself Jun 06 01:55:17 Estel^, easy-debian uses xephyr Jun 06 01:55:22 what would be, in easy steps a way to forward out little Maemo? Jun 06 01:55:27 Hurrian, yep Jun 06 01:55:51 SpeedEvil, the point is to sue different resolutions, and to not use PAL video out ;) Jun 06 01:56:15 = I would eb glad to even use real 800x480 via dhmi on external screen, but of course, other, higher resolutions would be great to Jun 06 01:56:16 Estel^, vncserver -geometry 800x480 -nolisten tcp :1 Jun 06 01:56:21 Then you need to startup a whole new desktop Jun 06 01:56:26 Estel^, vncserver -geometry 1920x1080 -nolisten tcp :1 Jun 06 01:56:27 no no no, vnc is too slow Jun 06 01:56:30 If you don't want to use the same resolution. Jun 06 01:56:35 And almost nothing at all will work Jun 06 01:56:41 As it's all one resolution only Jun 06 01:56:54 SpeedEvil, ok, lets start with desktop @ 800x480 Jun 06 01:57:02 Generally only minimally ported stuff without hildon tweaks works properly Jun 06 01:57:10 yea Jun 06 01:57:16 higher resolutions would be mean to run with only such stuff Jun 06 01:57:19 (at other res) Jun 06 01:57:53 Hurrian, VNC is too slow in such setup, and we need at least good 10 FPS :P Jun 06 01:58:22 SpeedEvil, could You, if got some spare time anytime soon, look @ http://talk.maemo.org/showthread.php?t=81408 and problems mentioned there? Jun 06 01:58:30 would be evry rgateful for Your input :) Jun 06 01:58:45 s/evry rgateful/very grateful/ Jun 06 01:58:46 Estel^ meant: would be very grateful for Your input :) Jun 06 01:59:02 as You sounds like someone who know how to deal with this stuff Jun 06 01:59:15 Hurrian, You're also welcomed there, but You're there already, so I haven't invited You :P Jun 06 01:59:38 Estel^, DISPLAY=192.168.1.x:0.0 startx, i guess Jun 06 01:59:43 with an X client listening Jun 06 02:00:07 In principle, you can start the maemo X stack, pointed to another X server. Jun 06 02:00:08 one caveat - N900's things are, many times, hardcoded to Display Jun 06 02:00:14 However. Jun 06 02:00:27 you still don't have opengl anything though Jun 06 02:00:28 Most stuff - possibly including the window manager - won't actually work like this. Jun 06 02:00:30 which means hildon breaks Jun 06 02:00:43 hm... Jun 06 02:00:44 So, you won't end up with a functioning system at all. Jun 06 02:00:52 yea. Jun 06 02:01:05 If you do manage to patch it into sort-of-working - none of the 'main' applications actually works. Jun 06 02:01:19 (I think I found a few which did - but they were few and far between) Jun 06 02:01:32 any way to overcome this problem? Jun 06 02:01:37 And then you run into the issue that doing it this way - you don't have a working normal display. Jun 06 02:01:55 Or alternatively, you have two copies of X and the whole stack in RAM. In 256M Jun 06 02:02:01 :D Jun 06 02:02:05 Which is a good benchmark of swap. Jun 06 02:02:08 But little else Jun 06 02:02:33 In principle, something faster than VNC could be done to clone the desktop. Jun 06 02:02:36 lets say that switching between EITHER normal display and remote isn't a problem, but, any way to actually make it *work*, not proof-of-concept? Jun 06 02:02:37 More is challenging. Jun 06 02:02:43 yea Jun 06 02:02:54 I've not seen any clean way to switch programs between X servers Jun 06 02:03:04 cloning desktop could be ok, if we would have something that doens't care about bandwidth Jun 06 02:03:19 and, for example, tries to maintain constant FPS Jun 06 02:03:26 In principle, you would need to fork the X server, and make an X server with an integrated optimised VNC client. Jun 06 02:03:27 0 compression, etc, to not add much overhead Jun 06 02:03:30 SpeedEvil, such a thing is not at all good. app will need to be relaunched with DISPLAY variable Jun 06 02:03:33 This could be quite a lot faster. Jun 06 02:03:45 ouh Jun 06 02:03:55 Hurrian: I don't think it's in principle impossible, but it would need a 'shim' with lots of state Jun 06 02:04:03 ok, so now, why Easy Debian works liek a charm without additional overhead? due to xephyr? Jun 06 02:04:10 Hurrian: the patched X server is possibly the easier way. Jun 06 02:04:12 Estel^, local networking Jun 06 02:04:15 yes. Jun 06 02:04:30 and no way to add xephyr layer to Maemo or part of it, I suppose? Jun 06 02:04:59 Estel^, you'd have to start all of hildon-desktop again Jun 06 02:05:14 Hurrian, I'm talking about local networking as whole backend - USB networking between raspberry Pi and N900 Jun 06 02:05:20 In principle, you could instead of forking the X server, make Xephyr understand multiple displays Jun 06 02:05:46 SpeedEvil, basically, making the raspi look like a second screen under the current xserver Jun 06 02:05:55 And then start the whole stack under xephyr Jun 06 02:06:09 it's a pity that no existing tools could be used... or I'm wrong? I'm quite not experienced enough to modify code of existing things that much. Jun 06 02:06:11 Hurrian: Or that. Jun 06 02:06:23 As far as I'm aware, there are no suitable tools. Jun 06 02:06:39 Is there an ##x? Jun 06 02:06:40 SpeedEvil, actually, we should look into the kernel driver for the TV-Out, and see how it clones the display, scales and sends Jun 06 02:06:52 Hurrian, making things to see RPi as second screen, would include any very hard to do modifications, and/or closed bits roadblocks? Jun 06 02:07:04 Estel^, closed bits shouldn't be roadblocks Jun 06 02:07:06 hm, kernel-driver for tv-out? I wonder if it's open Jun 06 02:07:11 It is Jun 06 02:07:23 Estel^, the kernel patch is open, so I assume it is Jun 06 02:07:36 Hurrian, using it to send non-scaled output would be great Jun 06 02:08:15 will try to buggy freemangordon or Pali about it - it is, for sure, much more benefit than messing for ages with facebook upload plugin (without any disrespect to work being done to RE the latter) Jun 06 02:08:18 Estel^, yeah, create a fake framebuffer device to send output of any resolution Jun 06 02:08:25 if no rocket-science involved, it could be best option Jun 06 02:08:55 Hurrian, fake framebuffer is re modified tv-out kernel module, yep? Jun 06 02:09:39 Stuff like video is unlikely to work. Jun 06 02:10:00 no problem, it's not meant for that - ho ever, if You think about it, video works via tv-out, so why not? Jun 06 02:10:08 (using 900x480) Jun 06 02:10:18 s/900/800/ Jun 06 02:10:18 Estel^ meant: (using 800x480) Jun 06 02:10:21 Estel^: It will work to the framebuffer Jun 06 02:10:44 you can't expect to be able to either transport or compress 800*480*3*30fps however. Jun 06 02:11:23 probably, some optimizations to send it with respect to USB networking bandwidth could be used Jun 06 02:11:29 heck, it could be even scaled down to WiFi Jun 06 02:11:39 Optimisations = CPU Jun 06 02:11:49 and allow to either send via tv-out, OR fast usb networking OR slower WiFi Jun 06 02:11:57 no no, I mean optimisation = lower output Jun 06 02:12:00 for example, lower fps Jun 06 02:12:31 it works with tv-out with "transparent" fps (i.e. looks like native), with all those resizing and PAL compression... Jun 06 02:12:42 That's all hardware Jun 06 02:12:50 why should it over USb networking? Jun 06 02:12:53 understand Jun 06 02:12:57 There is no PAL compression Jun 06 02:13:06 used wrong word Jun 06 02:13:08 scaling Jun 06 02:13:15 probably that's what I mean ;) Jun 06 02:13:20 It's simply another DAC reading the framebuffer out line by line, and outputting it in the correct format. Jun 06 02:13:31 As a hardware unit, using no CPU or DSP bandwidth at all. Jun 06 02:14:05 Ok... I wonder, how much CPu (or maybe DSP?) would use reading framebuffer and sending it @ 900x480 (or other resolutions) at usable FPS Jun 06 02:14:15 800x480, heck Jun 06 02:14:21 CPU - unless you can find someone to code DSP Jun 06 02:14:34 Plus - if you have DSP decoding video, it can't be doing other things Jun 06 02:14:37 we have some people udnerstanding openGLES, but I think it's not related? Jun 06 02:14:51 reasonable. Jun 06 02:15:11 hey Jun 06 02:15:20 would it be possible to use raspberry CPU to read from N900's framebuffer? Jun 06 02:15:23 Going to sleep - sorry :) Jun 06 02:15:28 USB is a bus Jun 06 02:15:34 It has to have CPUs at both ends Jun 06 02:15:42 The way you're using it Jun 06 02:15:47 yea Jun 06 02:15:56 Firewire in principle can do that I think Jun 06 02:16:02 so I wodne rif it's possible to utilise Pi to process N900's framebuffer, instead of doing everything on N900 Jun 06 02:16:33 no problem, good night :) Jun 06 02:16:35 800*600*3*30 >>480MBPs Jun 06 02:16:49 formk which those *3 comes from? Jun 06 02:16:52 from where* Jun 06 02:16:54 coloru Jun 06 02:17:16 ouh Jun 06 02:18:05 800*480*3*10 = 11,5 Mb/s Jun 06 02:18:09 if I count correctly Jun 06 02:18:32 that erm, 11,5 MB/s actually Jun 06 02:18:52 that should work flawlessly over USB Jun 06 02:19:00 even 20 fps Jun 06 02:19:30 Hurrian, Your thoughts about it? isn't my math screwed?:D Jun 06 02:26:22 * Sc0rpius yawns Jun 06 02:33:01 hmm Jun 06 02:33:23 800x480=384000 pixels on the screen. Jun 06 02:34:19 384000*3=1152000b color maximum Jun 06 02:35:20 works out to 1.1 Mb/frame **** ENDING LOGGING AT Wed Jun 06 02:59:58 2012