**** BEGIN LOGGING AT Fri Sep 21 02:59:57 2007 Sep 21 03:28:53 hey Sep 21 03:29:41 on the wiki, the pdf file reported to be the manual for the cpu wont open for me in any pdf viewer, Neo1973_Hardware#Processor Sep 21 03:29:57 can anybody actually get this file open? Sep 21 03:30:10 how many bytes should it be? Sep 21 03:30:35 I'm speaking of the "2410UserManual.pdf" Sep 21 03:30:47 about 2.6mbyte says my browser Sep 21 03:30:53 and it opens fine in evince Sep 21 03:31:07 and xpdf Sep 21 03:31:25 ok, well i tried both of those on my 2.6M file Sep 21 03:31:27 dunno. dont use that Sep 21 03:31:37 and acrobat reader as well Sep 21 03:31:50 2807559 bytes um_s3c2410a_rev11.pdf Sep 21 03:31:54 ah Sep 21 03:32:01 md5 is 109268084905f13233521ff7df387364 Sep 21 03:32:23 i have 2695885 bytes Sep 21 03:32:38 it is bork Sep 21 03:32:53 well wget wont let me continue Sep 21 03:33:23 orzo samsungs httpd does not support 206 partial content Sep 21 03:33:30 that sucks Sep 21 03:33:43 can one of you send me the end of the file? Sep 21 03:33:46 heh Sep 21 03:33:56 thats also why firefox does not give a reasonable progressbar and does not calculate the time Sep 21 03:34:25 i suppose it's not legal to host the file on openmoko? Sep 21 03:34:31 can't you redownload it? or is your bandwith limited so hard? Sep 21 03:34:39 low bandwidth Sep 21 03:34:42 it comes real slow Sep 21 03:34:53 and pauses a lot Sep 21 03:35:01 orzo i dunno what the license on that file is Sep 21 03:40:52 hm Sep 21 03:41:54 i think it is just my connection to samsung that is so slow Sep 21 03:41:57 it's stalled now Sep 21 03:42:23 either of you mind chopping the last chunk off the file and emailing it to me? Sep 21 03:42:50 http://195.160.173.6/tmp/um_s3c2410a_rev11.pdf Sep 21 03:43:34 thanks, i got it now Sep 21 03:45:24 hm, my md5 is not the same tho Sep 21 03:45:37 but it opens **** BEGIN LOGGING AT Fri Sep 21 04:34:50 2007 Sep 21 04:57:56 moin Sep 21 05:12:36 CM: you're not here are you? Sep 21 05:18:07 ecloud: I'm always here ;) Sep 21 05:19:15 ecloud: Was just about to go to work though.. Sep 21 05:19:40 * CM is back in a while when changing offices is done Sep 21 05:50:29 the mailing lists are quiet as of late Sep 21 05:52:17 Sure are. The planet too Sep 21 05:53:59 everyones busy using their qtopia-neo's? Sep 21 05:55:01 or hacking after OSiM? Sep 21 05:55:31 ljp: I am. Really nice work, I'm impressed :) Sep 21 05:55:56 Even managed to use up my prepaid sim credits.. ;) Sep 21 05:56:07 i fixed a few bugs today, so I think i will update the image next week sometime Sep 21 05:56:18 Nice Sep 21 05:56:41 I have one annoying bug that locks the touchscreen Sep 21 05:56:41 bugs and workarounds for things that got broke Sep 21 05:56:47 Anywhere I can report it? Sep 21 05:56:51 It's very simple Sep 21 05:56:53 #qtopia Sep 21 06:10:32 i find it ironic and funny that you wrote that trash talking blog post and now hang on #openmoko Sep 21 06:12:27 notserpe: Honestly, qtopia now works better than OpenMoko ever have for me Sep 21 06:13:07 I'll probably go back to using om as soon as gsmd works, but I don't mind some competition. :) Sep 21 06:14:47 (script) wiki RecentChanges: [[Community_nominees_for_GTA02_P0]] Sep 21 06:14:49 it was only your perception that thought it was trash talking Sep 21 06:15:22 backpeddle all you want. "and does not look like a hockey puck!" Sep 21 06:15:37 if you actually read the blog I said the neo is not the first open source phone Sep 21 06:16:05 I'd say that depends on how you define open Sep 21 06:16:15 yeah. exactly. Sep 21 06:16:18 and if i said the greenphone looks like a brick would you be happy? Sep 21 06:17:08 so I take it you are against the use of the LGPL Sep 21 06:17:21 that was a simple counter point to you suggesting it was a misperception on my part. Sep 21 06:17:33 ljp i think its really only a question how people define their 'open' .. for most it does not mean that you can run opensource on it, but that its parts are all opensource. Sep 21 06:17:49 roh: exactly Sep 21 06:18:12 lgpl is good Sep 21 06:18:21 sure, but the LGPL allows closed source products, and that is not very free Sep 21 06:18:30 besides that it does not matter who was first. i think everybody should have their fair share in contributing his perception of how it has to look like and let the people choose which open is the one they want. Sep 21 06:18:38 it allows both, which is freedom of choice Sep 21 06:18:40 ok, yousay lgpl is good, so you must be ok with closed source Sep 21 06:18:59 but you then have a problem with closed source, which the lgpl allows Sep 21 06:19:37 I'm ok with closed source sucking and existing only in userspace. Sep 21 06:19:55 but the kernel also allows closed drivers Sep 21 06:20:12 which sometimes people get stuck with Sep 21 06:20:20 ljp lgpl is not the idealistic form. but sometimes its a neccessary one to be able to create certain software or a product at all. not really in our case neccessarly Sep 21 06:20:23 I'm using vmware-server and the nvidia-drives, and they're both working fine Sep 21 06:20:55 * CM prefers to practical instead of a zealot Sep 21 06:21:06 ljp: better do some research on that. Sep 21 06:21:40 will catch you another time, gotta get some sleep. Sep 21 06:23:04 roh: sure, but the Neo is touted as a 'free' phone, why allow closed source, then Sep 21 06:23:29 and then bag the people that do have closed sources Sep 21 06:23:41 (not saying openmoko does,) Sep 21 06:24:48 If there was a closed source gps navigation program that was way better than anything open and having good maps, I would be ok with buying it. Sep 21 06:25:08 I don't see that any differently than using a TomTom Sep 21 06:25:34 Ment using it on the Neo of course Sep 21 06:26:48 ljp the important point is that all stuff which is needed is free. thats also why we are not too happy about gllin Sep 21 06:28:29 ljp the other thing is to lock in users/applications to certain licenses. we think thats not our part to decide. it e.g. some company would want to release a navigation-software for it and can (due to restrictions in the license) not release the code to parse the map data they license, they wouldnt be able to code and release that app with gpl-ed libs only beneath Sep 21 06:29:41 That's exactly the way I see it too Sep 21 06:30:00 CM: I just found OE bug 2928, which includes a patch and it takes care of the libiberty.a problem Sep 21 06:30:13 ecloud: Ah, ok. :) Sep 21 06:30:26 of course i'd like to see a free version of all the software, but sometimes when 3rd partys and their licenses are involved it locks you in. we want to go a pragmatic way. we don't lock people in by licenses, we have a trademark which still can be used to ensure what happens under the openmoko label Sep 21 06:30:27 so now the build is proceeding further. We'll see... Sep 21 06:31:50 besides that, ljp you should be happy.. that gpl-lockin gets people the interrest to buy licenses from trolltech, doesnt it? Sep 21 06:32:47 roh: Couldn't resist a little jab could you? ;) Sep 21 06:33:02 which is totally ok, don't misunderstand me there. its just another business-model than ours Sep 21 06:33:28 CM ;) Sep 21 06:34:01 sure, but the use of lgpl promotes non free software, which most open source users despise Sep 21 06:34:47 ljp well, then i have to assume these users never had the pain to live in the real world, where sometimes you get into conflict with 3rd party licenses. Sep 21 06:36:02 ljp: I think that closed source apps in general gets developed faster, but eventually open source alternatives will catch up when the market is more mature Sep 21 06:36:55 besides that lgpl gives us the feature not to need people to sign away their copyright to somebody else to allow for dual-licensing. less paperwork and hassle (or howdoes troll solve the problem of the copyright which a patch submitted by the community has) Sep 21 06:38:17 atleast digium does require you to sign some papers before accepting patches (for asterisk) .. which is part of why it got forked into openpbx (i think they renamed it again or so) Sep 21 06:38:49 ah.. callweaver it is named now Sep 21 06:39:27 openpbx? Sep 21 06:41:14 CM basically people wanted to change stuff on asterisk, but digium did not want these changes since the would link against gpl projects, which would kill their businesscase of dual-licensing it. Sep 21 06:41:23 I modified gtk+-directfb_2.10.14.bb and its related file about pango & cairo for directfb supporting. Sep 21 06:41:29 when do_rootfs is processing, the bb report that: Sep 21 06:41:36 ERROR: Cannot satisfy the following dependencies for gtk+-directfb: Sep 21 06:41:42 libx11-6 (>= 1.1.2) Sep 21 06:41:48 i checked the status file in rootfs dir, there is no libx11-6 in it. i wonder where the libx11 demanding is come from. Sep 21 06:41:52 roh: Ah, ok. Didn't know anything about that fork Sep 21 06:42:06 and after some piles of changes got bigger and people were refusing to sign over these rights to digium they forked the gpl2-ed release of asterisk and renamed it openpbx. then some trademark issue came up so its callweaver now Sep 21 06:42:45 and can to new fancy stuff like fax, proper sip afaik and some things like a real config parser, not fiddled stuff Sep 21 06:43:03 Ah, nice. I like sip :) Sep 21 06:49:10 roh: Ah, interesting. I'll have to see what advantages this might have over my current asterisk installation. Sep 21 06:50:33 cworth i think they are still hacking hard on it, but in the end they have already removed some code which was only there because it was from digium, and not because it was good and there was no alternative.. so i hope it gets more stable in the end Sep 21 06:51:26 roh: The couple of times I've dared to look into the asterisk implementation I've been quite appalled. There's some dire code in there. Sep 21 06:54:08 cworth heh... yep.. but thats why opensource persists. as long as somebody is there who really wants to do better he/she can fork Sep 21 07:01:16 good morning Sep 21 07:01:29 morning Hopscotch Sep 21 07:06:00 hi Sep 21 07:09:30 hello all Sep 21 07:09:52 kristian-m: all you online? Sep 21 07:12:00 s/all/are/ Sep 21 07:12:00 zdanek meant: kristian-m: are you online? Sep 21 07:13:12 hrw: shoudn't you shutdown computer? Sep 21 07:13:27 your wife will be mad :) Sep 21 07:16:21 zdanek: hrw|honeymoon uses an IRC bouncer. Sep 21 07:16:31 hmm Sep 21 07:16:44 I've never used such thing Sep 21 07:17:09 is it some kind of bot/voice mail with proper message? Sep 21 07:17:24 does it filter and send sms to the neo? ;) Sep 21 07:17:35 it's a proxy server which stays connected to IRC for you, and gives you the backlog when you connect to it Sep 21 07:17:41 I use dircproxy myself. Sep 21 07:18:46 so, hrw still wanted to connected while honeymooning" Sep 21 07:18:56 my wife would be mad Sep 21 07:19:15 on vacations there's no such device as computer :) Sep 21 07:19:18 only phone Sep 21 07:19:20 I thought openpbx was dead and everyone was using freeswitch Sep 21 07:19:36 but... neo is a phone Sep 21 07:19:38 :D Sep 21 07:20:13 zdanek: hm, are u shure? mine is a webserver already... ;) Sep 21 07:20:38 I need brave man to make a package with latest SettingsGUI Sep 21 07:20:53 emdete: which soft are you using? Sep 21 07:20:58 i'm interested Sep 21 07:21:02 zdanek: webpy... Sep 21 07:21:28 that was already done. it reads contacts, appointment and such Sep 21 07:21:34 where is it available? Sep 21 07:22:06 webpy.org - it a web devlopment platform Sep 21 07:22:14 ..there is a bb for it ;) Sep 21 07:23:41 my app has all data in sqlite3 but i thnk i have to switch it to eds soon Sep 21 07:24:38 I don't have dev. environment set yet so I cannot bitbake it, could you do it for me? Sep 21 07:25:15 zdanek: u need python for that an some packages... and u need a app using it. it's more or less a lib Sep 21 07:25:42 oh, understand Sep 21 07:57:36 openmoko: 03thomas * r3022 10/trunk/src/target/OM-2007.2/libraries/moko-gtk-engine/ (ChangeLog data/gtkrc): * data/gtkrc: Define some standard icon sizes Sep 21 07:58:41 hi...were can I change the font size for the apps (today2, contacts, ...)? Sep 21 08:00:38 in openmoko-dates i have the normal font size but in all other apps the font size is _very_ small Sep 21 08:00:49 (i build om for Palm TX) Sep 21 08:14:49 (script) wiki RecentChanges: [[Business_Development/it]] [[Main_Page/it]] [[Buying_Interest_List]] [[Python]] Sep 21 08:25:02 how can i restart the top panel? Sep 21 08:26:58 zefanja: restarting X restart that panel as well. I don't know how to restart only that Sep 21 08:30:16 Ghiottone, thx. do you know how to change the font size? Sep 21 08:34:43 Ghiottone Sep 21 08:34:55 hi, will you make SettingsGUI package? Sep 21 08:35:05 Kriss has released 0,7v Sep 21 08:36:03 zefanja: sorry, but I don't know how to change the font, other than that of the terminal, for which there is something in the wiki Sep 21 08:36:51 zdanek: sure. I'm looking at the sources right now. Sep 21 08:37:17 zdanek: it could take a bit because we decided to switch do distutils packaging Sep 21 08:37:23 grazie :) there are some new improvments including mine with proper connection AT command Sep 21 08:40:23 zdanek: :) Sep 21 08:53:18 new day, same question: any updates on the shipments? Sep 21 08:53:23 hi all ) Sep 21 08:54:07 nope. if i knew i would have posted it. Sep 21 09:05:24 Ghiottone: please drop me a line when package is ready, I have to disconnect, ok? Sep 21 09:08:44 my mail is bartek.zdanowski[]gmail.com Sep 21 09:08:47 thank you. Sep 21 09:18:00 anyone fixed libsdl-net yet? Sep 21 09:27:54 roh: do you post updates somewhere on the wiki? Sep 21 09:28:56 Anyone that knows how long time to expect from order to delivery on GTA01Bv4 (Neo Advanced)? Sep 21 09:28:57 esben_: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 21 09:30:07 XorA, don't suppose you're about? Sep 21 09:30:17 quinton: no Sep 21 09:30:36 =p Sep 21 09:31:02 wazzzzaaaap? Sep 21 09:31:34 sorry to bug you, but did you get it working with a UK t-mobile sim? Sep 21 09:31:46 quinton: yes, both PAYG and contract Sep 21 09:31:56 was it a 2G or a 3G sim? Sep 21 09:32:02 openmoko: 03thomas * r3023 10/trunk/src/target/OM-2007.2/libraries/libmokoui2/ (ChangeLog libmokoui/moko-stock.c): Sep 21 09:32:02 openmoko: * libmokoui/moko-stock.c: (_moko_stock_add_icon): Sep 21 09:32:02 openmoko: Correctly load all the available sizes for additional stock icons Sep 21 09:32:16 quinton: AFAIK t-mobile dont issue 3g sims in UK Sep 21 09:32:52 really? Sep 21 09:33:09 i think the guy said it was 3g, and it's got 8 contacts rather than 6 Sep 21 09:33:30 quinton: last company I worked for used to do GSM testing, they went to all networks and asked for 3g sims to test, they kept getting told that UK was using 2G sims in 3g handsets for cheapness Sep 21 09:33:50 oh, i didn't know that Sep 21 09:34:12 quinton: the form factor doesnt change for a USIM Sep 21 09:34:23 quinton: they are both ISO7816 compatible cards Sep 21 09:34:50 quinton: my info might be out of date, but I only ordered SIM from t-mobile 2 weeks ago and it works fine Sep 21 09:35:14 hmm, i've been having problems getting the phone to detect sim cards and i just found out about the problems with 3g Sep 21 09:35:36 okay, thanks very much for the info, XorA Sep 21 09:36:27 quinton: the number on my SIM are 147095/278458 Sep 21 09:36:38 quinton: they might be some form of version number :-) Sep 21 09:38:39 mine says 147096/623954 Sep 21 09:39:29 i have a GPRS service here in austria .. and i'd like to test it on the neo .. anyone know if its worth the effort yet? Sep 21 09:39:38 i've tried the qtopia image, but it doesn't come with a browser onboard. Sep 21 09:39:52 was i wrong to think that 3g were the ones with two extra contacts? Sep 21 09:40:22 does qtopia support gps Sep 21 09:40:23 ? Sep 21 09:41:00 quinton: yes, 3G USIM has no extra contacts, but it looks like t-mobile are using the metal space for art, mine looks like something klingon Sep 21 09:41:26 heh Sep 21 09:41:41 okay, thanks a lot, i don't know how that idea got in my head Sep 21 09:41:52 well they do look like contacts Sep 21 09:42:11 but ISO7816 is quite clear on the allowed shape of a card :-) Sep 21 09:45:31 qtopia looks really good, do you think openmoko will be able to catch up on time? Sep 21 09:46:47 edistar: I'm sure it will :) Sep 21 09:55:03 edistar: i'm a bit concerned to be honest. the progress on openmoko has been abysmal. Sep 21 09:55:28 the more weeks go by with buggy build systems and no public releases, the more it looks to me like openmoko is a red herring designed to deflect real development. Sep 21 09:56:41 i wish there where secret development anywhere. but i don't think so. Sep 21 09:57:18 my only point is that i wish there was a visible, public person, making the openmoko results more widely known to those who wish to contribute. Sep 21 09:57:48 its terribly frustrating to have to learn that the build system is borked for a few weeks because nobody has bothered to tell anyone, and then watch 9 or 10 new coders come along and stumble upon the same problems, individually. Sep 21 09:57:56 this to me is a terrible way to organize things. Sep 21 09:58:30 torpor: rumour has it OM has hired a full time OE developer, we are just waiting to find out who Sep 21 09:59:30 well whoever he is, i wish he'd stop gallavanting around and get on with the job Sep 21 09:59:55 i'm trying to push code into openmoko for months now, and for the last 3 weeks, just when i feel like i've fixed things so i can reliably code for it, the build system gets borked. Sep 21 10:00:10 i'm getting frustrated. Sep 21 10:00:51 torpor: he is probably on notice at another company, some companies have long notice periods :-( Sep 21 10:01:05 gah. Sep 21 10:01:11 torpor well.. be sure some of the team are too.. thats why we now finally have atleast one fulltime admin, taking a first peek this week and shopefully starting fully next week Sep 21 10:01:40 i think there is alot to do for gta2... that is nothing usable for me now... Sep 21 10:01:44 torpor getting qualified people seems to be extremely problematic nowadays... Sep 21 10:01:45 i never realy tried hard. but the fact that no build ever compiled for me and i was never abel to make a single phone call via gui set my entusiasiasm a bit back. So it's easy to say i have no time at the moment. :-/ Sep 21 10:02:02 and: try to develop for the greenphone and you will be fine with whatever happens here with the neo Sep 21 10:02:06 roh: i'm qualified and i'd love to work full time on openmoko Sep 21 10:02:14 cudos for the browser btw. who ever did that. impressing Sep 21 10:02:35 e.g. we are still missing one admin for the taipeh office who can handle all the stuff we use Sep 21 10:02:45 so you have to have someone in taipeh/ Sep 21 10:02:46 ? Sep 21 10:03:22 torpor we'd like to. we now got the best ip you can get there it seems, but still, >300msec is not fun when you need to work on such a machine remote. Sep 21 10:03:23 what does "handle all the stuff we use" mean? Sep 21 10:03:44 i'd be quite happy to work at >300ms if it was actually useful. Sep 21 10:03:49 i've done far worse. Sep 21 10:03:59 <-- used to run an ISP with a 14.4k PPP line .. Sep 21 10:04:49 torpor heh.. we know it works, but it does not help if somebody in a building on the other side of the planet has a problem you need to debug. but hey.. why don't you just send us a mail ;) Sep 21 10:05:07 roh: address? Sep 21 10:05:24 i'd prefer to work on code/development full time, personally, but maybe you'll have such a position open in the future? Sep 21 10:05:41 also, i'd be very happy to do developer support issues, documentation, assistance, etc. Sep 21 10:06:32 torpor there should be a jobs@openmoko, but i'm not 100% sure. it if bounces mail to coreteam@ Sep 21 10:06:45 ok thanks roh Sep 21 10:07:03 i will send my details off, and at least they'll be on file. and in the meantime i'll soldier on with my new apps for openmoko Sep 21 10:08:39 torpor we have som queue for hr in rt, so if we get something it should land there Sep 21 10:10:29 anyone using automatic networking with the openmoko and fedora Sep 21 10:12:10 roh: you are from fic / om? Sep 21 10:13:25 morning Sep 21 10:13:38 emdete: he is. Sep 21 10:14:11 (but i'm in his personal ignore filter :D ) Sep 21 10:14:24 emdete i work for om Sep 21 10:14:50 (script) wiki RecentChanges: [[Vertical_applications/it]] Sep 21 10:14:51 is it possible to discuss some generel disign ideas somewhere Sep 21 10:14:52 ? Sep 21 10:15:10 emdete depends what you have in mind Sep 21 10:15:54 in short: i think of using webkit as general gui renderer also for phone functionality. apps would be written in js. Sep 21 10:16:13 one way to do that would be to put the webserver on board and run all the hard stuff through it .. Sep 21 10:16:19 cgi scripts to dial phone, etc. Sep 21 10:16:26 this would include much more developer as before Sep 21 10:16:31 i thought of doing that too, but it really depends on how heavy/light the server setup would be .. Sep 21 10:16:36 torpor: no Sep 21 10:16:46 torpor: i would like a api from js Sep 21 10:16:46 emdete hm.. nice idea.. seen that concept some times on different stb devices. but it does not really fit very well to a phone and would require a lot of ajax hacks Sep 21 10:16:59 roh: no ajax Sep 21 10:17:04 <_buz> google gears maybe Sep 21 10:17:04 roh: js hacks Sep 21 10:17:07 i rather prefer that it just get easier and easier to write normal C/GTK apps for the phone. Sep 21 10:17:13 <_buz> ajax IS a js hack Sep 21 10:17:28 _buz: yes, but i have no server in mind Sep 21 10:17:37 emdete thats the same. you need to reload parts of the gui to get any kind of interactivity. and that slows the whole thing down Sep 21 10:17:38 i rather prefer that it just gets even easier to write a normal Qt/opia apps for the phone Sep 21 10:17:39 <_buz> frankly, i rather have python+gtk than js Sep 21 10:17:47 roh: no Sep 21 10:17:52 you change the dom Sep 21 10:17:53 <_buz> or qtopia, indeed Sep 21 10:17:54 on the stb they often used html4 engines and frames to do that Sep 21 10:18:14 roh: no frames. may tabs to run different 'apps' Sep 21 10:18:15 <_buz> do you get the qt4 "css" stuff in qtopia4? Sep 21 10:18:18 emdete still needs a redraw and possibly rerendering of the whole page Sep 21 10:18:37 <_buz> manipulating dom is horribly inefficient Sep 21 10:18:42 roh: depends on the intelligence of the renderer ;) Sep 21 10:18:50 _buz: as gtk is ;) Sep 21 10:19:02 <_buz> no thats a few orders of magnitude more efficient Sep 21 10:19:11 <_buz> just look at your average AJAX app Sep 21 10:19:21 <_buz> usually it's not slow because of the server round trips Sep 21 10:19:22 sry, i have to leave for lunch. but is it interesting enough to discuss that further? Sep 21 10:19:22 emdete i think nothing is holding you back trying it out. i just think it does not give us alot of features, but will make changes to the gui a horrible html and js experience Sep 21 10:19:25 _buz: hmm i will look Sep 21 10:19:57 <_buz> emdete: you could always stick gecko+xul on the phone Sep 21 10:19:59 roh: there are better screen dsigners out there that are capable of js and xhtml Sep 21 10:20:18 <_buz> frankly, i dont believe we need screen designers as much as usability people Sep 21 10:20:27 emdete you could just use the gtk-webkit component and run that standalone as fullscreen x11 app and start hacking some object-gui onto the js to get access to the 'middleware' Sep 21 10:20:56 _buz i think when we start using xul i should buy a firearm Sep 21 10:21:15 <_buz> i can see uses for xul, but not on a phone Sep 21 10:21:30 just kidding.. but i really look forward to having a reasonable browser also on my workstation, instead of firefox Sep 21 10:21:40 <_buz> i have konqueror ;) Sep 21 10:22:11 _buz does not help. its still slow and does not integrate reasonable into gnome Sep 21 10:22:16 <_buz> now if the konqueror devs finally figured that you may actually NEED a wysiwyg text area thingy in 2007 i could retire firefox Sep 21 10:22:34 <_buz> roh: oh, i dont care for gnome :) Sep 21 10:22:49 <_buz> actually, it's about twice as fast as firefox Sep 21 10:22:56 <_buz> no match for opera, but good enough Sep 21 10:22:57 what i would find fascinating is some html-engine which is NOT written in c++. Sep 21 10:23:07 <_buz> opera mini ;9 Sep 21 10:23:09 _buz: seems like css should be supported Sep 21 10:23:21 roh: gtkhtml? ;-) Sep 21 10:23:29 <_buz> emdete: so there's your easy way to style apps Sep 21 10:23:44 _buz but its incompatible with big parts of the web, renders bullshit on complicated ajax sites and is the same buggy when i comes to 3rd party plugins like flash -> does not get me further Sep 21 10:23:59 <_buz> that's true Sep 21 10:24:19 browser world is doomed. html is last century technology. Sep 21 10:24:20 thos gtkhtml is an insult to some degree ;) but aslong as only gtkdoc uses it... Sep 21 10:24:20 <_buz> for all its faults, gecko at least renders very gracefully Sep 21 10:24:37 roh: hm, don't think gtk doc uses it Sep 21 10:26:01 in the end i just hope that webkit-gtk will be reasonably integrated with some 'browser' around it which gets the basic features of ff while using the gnome facilitys like the keyring.. and eventually then also gnomevfs will be reasonably debugged and fixed Sep 21 10:26:45 heh. using reasonably debugged and working kde components is a big nono i presume Sep 21 10:27:06 i'd really like some kind of single-sign-on via the gnome-keyring. i'm fed up with different keyrings and auth-methods Sep 21 10:27:21 <_buz> THAT actually works very well in kde with kwallet Sep 21 10:27:32 <_buz> except for well firefox ;) Sep 21 10:27:35 berkus i don't care about c++-code. i find it bad enough webkit also is written in it Sep 21 10:27:46 roh: right. the supreme C Sep 21 10:27:56 kwallet works very well Sep 21 10:28:08 i don't remember any of my passwords except my master password to the wallet Sep 21 10:28:12 C++, that nasty hack written by a guy who couldn't program in C Sep 21 10:28:17 <_buz> berkus: just like i Sep 21 10:28:26 XorA: yeah! that ugly beast! Sep 21 10:28:32 berkus i think thats personal preference. cÃ++ just looks ugly to me, and i never can accept that some tab renders slower the first time just because of some crazy language and its linking Sep 21 10:28:40 hah Sep 21 10:28:41 * berkus uses C++ without too much feelings Sep 21 10:28:44 * XorA wishes sun had made Java compilable by default Sep 21 10:28:57 <--- C and C++ programmer here. got no problems with either. Sep 21 10:30:11 * _buz wishes java didnt have all that classpath weirdness Sep 21 10:30:21 java is overdone. Sep 21 10:30:22 i really think lowlevel-stuff in object-oriented C and dynamic languages for non-performance-eating stuff like userinterfaces and logic is the way. like pygtk for the gui and doing all the hard work in optimized C Sep 21 10:30:45 or replace python by vala for example.. also looks very nice Sep 21 10:30:48 i don't see why you need to bung a whole script interpreter onto a C/C++ project just to do user interfaces Sep 21 10:30:50 (hint: C++ is converted to C) Sep 21 10:31:02 there are plenty of good ways to do user interfaces in C/C++ .. Sep 21 10:31:03 berkus: erm, not true Sep 21 10:31:13 <_buz> i can only think of one, really ;) Sep 21 10:31:13 it started off that way Sep 21 10:31:15 XorA: well, to assembly Sep 21 10:31:25 berkus: C++ used to be converted to C, but as of C99 that was no longer possible Sep 21 10:31:32 XorA: but still. there are C++-to-C converters like llvm Sep 21 10:31:51 anyway meeting time Sep 21 10:32:25 roh: this vala? http://www.vala.org.au/ Sep 21 10:32:55 http://live.gnome.org/Vala this one Sep 21 10:34:59 ew Sep 21 10:35:37 roh: oh Sep 21 10:36:43 basically a 'looks like c crossed with python, java and mono'-but-translates to gobject C without looking as ugly as doing that by hand Sep 21 10:37:24 still relatively new, but it smells like potential for gui and plugin stuff to me Sep 21 10:37:33 ooh, Vala seems impressively perverted Sep 21 10:38:28 and hey, Qt doesn't settle for standard C++, so why should we settle for standard C! Sep 21 10:38:33 .oO(somehow funny that you find vala malderan from the stargate universe with a similar prominent google-rank) Sep 21 10:38:57 mjr who cares about standards if the code looks ugly? *duck* Sep 21 10:39:19 yeah, everyone knows pretty code works better Sep 21 10:39:50 (and that other Vala too was perverted, but not that impressively, really) Sep 21 10:39:57 roh: its really a subset of c# Sep 21 10:40:17 ScaredyCat sounds funny, but it is, really. people tend to read nice code and learn from it. they also do from bad code, but that does not help them write good code. Sep 21 10:40:39 robtaylor somehow, jap. Sep 21 10:40:53 i think they are really proud of the lambda ;) Sep 21 10:41:56 roh though it is quite powerfull Sep 21 10:43:09 vala sounds nice. probably a better solution than python for making gtk apps for embedded systems. Sep 21 10:43:41 DPThought ack Sep 21 10:43:43 DPThought i don't think its quite ready yet for that.. but we'll see... Sep 21 10:44:33 i just called it up since most people tend only to know java, cü++,python, ruby, perl and some more.. even if there are really gazillions of sick ideas Sep 21 10:44:51 yea, i wonder how bindings to other libs work. python offers a incredible protfolio for almost everything Sep 21 10:45:16 python for example only makes sense if you let the interpreter in memory and do not reload it all the time. Sep 21 10:45:28 ack Sep 21 10:46:02 and with jit and bytecode caching... Sep 21 10:46:33 *grmbl* Verzögerung bei der Verzollung Sep 21 10:47:49 hm.. does dhl deliver on saturdays? Sep 21 10:48:04 roh: i belive so Sep 21 10:48:10 if the sender has paid the extra for saturday delivery Sep 21 10:48:28 by default, no Sep 21 10:48:59 cjb_ie its an international corp. shipment.. so it can be anything ;) Sep 21 10:49:13 but i'm in germany, so you never know Sep 21 10:49:22 btw. .njet/mono sound also very convincing to me. It is somehow evil, but you have nice languages and really good performance Sep 21 10:49:26 Hello, everyone. Sep 21 10:49:55 * pavelm3 is trying to make qtopia play nice with X, but no luck so far. Sep 21 10:50:19 ...the only thing I was able to do is to make LCD flicker like an very bad TV. Not sure what is mechanism behind that. Sep 21 10:50:28 hi pavelm3 Sep 21 10:50:57 hi just found mokomenu - pity that I havn't found it earlier - nice tool Sep 21 10:51:09 openmoko: 03chris * r3024 10/trunk/src/target/OM-2007.2/libraries/libmokoui2/ (ChangeLog libmokoui/moko-finger-scroll.c): Sep 21 10:51:09 openmoko: Don't ignore 'mode' in moko_finger_scroll_new_full, fix documentation Sep 21 10:51:09 openmoko: for "velocity_max" property Sep 21 10:52:40 ah, i see. vala binds against everything that uses gobject. Well, that could be quite something. *make note to try this out* Sep 21 10:52:52 kristian: There are more nice tools around mokomenu. See pygps :-). Sep 21 10:53:25 pavelm3 i will :-) Sep 21 10:54:38 problem is I got used to mokomenu/pygps and would now want to make it somehow coexist with qtopia Sep 21 10:55:02 * kristian-m heads to work Sep 21 10:55:26 ls -l Sep 21 10:55:46 the mtdblocks are resizeable right ? Sep 21 10:56:36 openmoko: 03thomas * r3025 10/trunk/src/target/OM-2007.2/libraries/libmokoui2/ (ChangeLog libmokoui/moko-finger-scroll.c): Sep 21 10:56:36 openmoko: * libmokoui/moko-finger-scroll.c: (moko_finger_scroll_init), Sep 21 10:56:36 openmoko: (moko_finger_scroll_new), (moko_finger_scroll_new_full): Sep 21 10:56:36 openmoko: Complete the documentation for MokoFingerScroll Sep 21 10:56:37 what recipe rebuilds the kernel image file, anyone know? Sep 21 10:59:35 torpor linux-openmoko Sep 21 11:00:06 torpor eh.. that pulls it in.. also Sep 21 11:00:14 roh: thanks Sep 21 11:00:56 all i want to do is rebuild the final image file (coz i deleted the contents of my build/tmp/glibc/images dir) Sep 21 11:01:06 openmoko-devel-image doesn't seem to do this Sep 21 11:01:46 linux-openmoko: downloading gcc-4.1.2.tar.bz2 .. wtf? Sep 21 11:02:17 it looks for its deps, and builds them if they are not there yet Sep 21 11:02:47 but i've already completed make all and make openmoko-devel-image numbers of times.. why's it thinking the deps are not set up? could it be my use of rm_work? Sep 21 11:04:55 Hello, scaredycat. Sep 21 11:05:02 And thanks a lot for your images :-) Sep 21 11:05:23 lo .. no probs... working ok for you ? Sep 21 11:08:32 scaredy: Yes, they were the best I could find. Sep 21 11:09:40 :) Sep 21 11:10:38 ScaredyCat: whats the best image from your set to run? Sep 21 11:10:40 i wanna try it out Sep 21 11:11:31 scaredycat: I'd love to combine them with qtopia, so I could get _both_ usable phone software _and_ usable X system (with pygps etc). But it is not easy :-( Sep 21 11:20:02 openmoko: 03thomas * r3026 10/trunk/src/target/OM-2007.2/libraries/libmokoui2/ (ChangeLog libmokoui/moko-hint-entry.c): Sep 21 11:20:02 openmoko: * libmokoui/moko-hint-entry.c: (moko_hint_entry_new), Sep 21 11:20:02 openmoko: (moko_hint_entry_clear): Correct and complete documentation for Sep 21 11:20:02 openmoko: MokoHintEntry Sep 21 11:36:05 hi Sep 21 11:36:23 does any body know when the openmoko gta02 releases Sep 21 11:36:36 frapzzt: latest official word is december Sep 21 11:36:53 okay Sep 21 11:36:59 and unofficial ? Sep 21 11:37:00 frapzzt: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 21 11:37:22 np: 'In the year 2525'. Sep 21 11:37:34 I've always said Nov for P0, Jan for P1 and who knows for P2. Sep 21 11:38:16 I saw somewhere a "we're working also on some new models", any word on that? Sep 21 11:38:31 no. Sep 21 11:38:43 thanks. Sep 21 11:40:19 hn Sep 21 11:40:44 in the german wikipedia is 23th octobre the release date Sep 21 11:41:26 viq till we don't have gta02 completed and working i don't think we will have ressources to start development on later models Sep 21 11:44:37 roh: ah, ok. From what i remembered from the message I was under an impression that there is going to be some announcement ("hey, here's what we're thinking of doing next") somewhere around gta02 release. Sep 21 11:45:18 viq well.. we could do that.. but were not there yet. Sep 21 11:46:17 roh: yeah, I know. Sep 21 11:46:22 to make that more understandable.. when i'm lucky i'll get my gta02v2 preseries prototype the next few days... but since we already have a v4 in planning it still takes some time Sep 21 11:46:27 roh: re, do u have some more time to talk about this gui stuff? Sep 21 11:46:36 roh: so it would be reasonable to rewrite the page as 'after GTA02 goes on sale, there will be announcements of further devices'? Sep 21 11:47:28 SpeedEvil i think so. atleast about the next one. the far future is still a bit more in a nebula Sep 21 11:47:30 hm, the build of theme.edj on the feed is dying with a segmentation fault... Sep 21 11:47:32 viq: there is, however, a HXD8 device in the svn respository. Sep 21 11:47:36 * cesarb glances sideways at raster Sep 21 11:47:48 Just, well, this toy is somewhat expensive, so I am thinking whether to get it ASAP, or wait a short while and see what's planned next, and then decide whether gta02 is interesting enough, or should I wait for next model. Sep 21 11:47:49 SpeedEvil what is there now/which wikipage you mean? Sep 21 11:48:36 The one that was saying 'in october, we will give more information about new models' Sep 21 11:48:39 * SpeedEvil forgets. Sep 21 11:49:01 rwhitby: as in system/environment, or as in description of hardware? Sep 21 11:49:06 viq, I think you won't be waiting a short while if you're planning to wait past GTA02... Sep 21 11:49:28 viq that purely depends on your cash-for-toys/need-for toys ratio ;) if you can wait, wait.. but there will always be a cooler device in the queue that what is buyable Sep 21 11:49:39 mjr: short while with decision, not necessarily with buying Sep 21 11:49:42 s/that/than Sep 21 11:49:55 yeah, what roh said Sep 21 11:51:00 roh: yes, certainly ;) It's just that at the current exchange rate 600USD is pretty much exactly half my salary - and that's before government appropriates 1/3 of it. Sep 21 11:51:04 currently gta01 was mostly interresting for people who don't care about the 300$ but want to hack right away and need real hardware. thats why we released it at all. Sep 21 11:51:09 viq: http://svn.openmoko.org/trunk/src/target/kernel/patches/hxd8-core.patch Sep 21 11:51:49 * SpeedEvil cares deeply about the $300 - it's a large chunk of change for me too. Sep 21 11:51:54 And as to $600. Sep 21 11:52:06 for a development board it was cheap, but since its not really a phone to use right now, we were totally buzzed by the interrest Sep 21 11:52:10 You don't need the debug board for 99.99% of stuff Sep 21 11:52:26 you can never brick the GTA02 without the debug board. Sep 21 11:52:31 (barring hardware bugs) Sep 21 11:52:48 speed: mis-flashing uboot seems to be easy enough. Sep 21 11:52:51 Which seem unlikely. Sep 21 11:52:55 SpeedEvil exactly thats the goal. making the debugboard only needed as less as possible Sep 21 11:52:59 there is a magic recovery flash Sep 21 11:53:12 with a sane version of uboot that you can't update without the debug boatd Sep 21 11:53:18 roh: Well... actually it is really possible to use neo since qtopia is out. Sep 21 11:53:21 SpeedEvil: hmm, I guess. And it's somewhat likely one of my friends will have the debug board, and there are already 4 people listed in the wiki in my town that have gta01, some with the debug board already ;) Sep 21 11:53:39 speed: Magic recovery flash? I did not know about that way. Is there wiki entry somewhere? Sep 21 11:53:44 roh: re, do u have some more time to talk about this gui stuff? i already play around with webkit. i would like to pimp js to access phone functions. i think the design of this api is importand and could be similar to the design openmoko uses. is there such a thing? Sep 21 11:54:01 SpeedEvil: you can easily brick things by adding bogus things to the environment too, does that magic recovery flash avoid that? Sep 21 11:54:04 pavelm3 gta02 will have an additional nor flash which holds an known-to-be-working uboot to recover even from a blanked nand. normal boot goes directly to nand Sep 21 11:54:52 pavelm3 and the write enable line goes to the debug port and not the app-cpu, thus without the board and without soldering around you should have no possible way to brick an gta02 Sep 21 11:54:53 roh: can it recover from a valid environment partition (valid CRC) with evil commands? (i.e. does it proprerly ignore the uboot environment partition?) Sep 21 11:55:01 any developers here? http://folks.o-hand.com/thomas/libmokoui2/ Sep 21 11:55:58 emdete i have time, yes, but i dunno what exactly you want to discuss? Sep 21 11:56:22 roh: aha, okay, so magic flash is gta02 thingie. Yes, that makes sense. Sep 21 11:56:37 cesarb: it was said that there would be a way via buttons to make it go to the magic flash Sep 21 11:56:55 roh: :D an api for telefony. java does that (see jsr), qtopia has it. i see a need for a unified api to phone functionality so that thrd parties can build apps and rely on a stable api Sep 21 11:56:57 in which case it would ignore all normal flash contents Sep 21 11:57:51 roh: js has something called LiveConnect which would make access to jsr possible. but that would need java. my plan is to provide an api following the jsr without Sep 21 11:58:08 emdete hm. stable apis only make sense if youre sure its 'right' the way it is designed. since we don't have a stable api in the rest of the middleware yet that api also would be in flux, atleast for some time Sep 21 11:58:35 emdete you should be able to build a prototype with what we have right now, i think.. Sep 21 11:58:57 Is there place where to download recent builds? Sep 21 11:59:01 roh: we have designers here that are eager to start developing of apps but are bound to html/svg/wcd/js by their skills Sep 21 11:59:20 emdete like using the webkit widget in fullscreen with a default url as target and start hacking some demo-api into the js-aÃpi Sep 21 12:00:02 roh: yes, that's what i want. problem is that i can't even use gsmd in a stable fashion right now (i have a pure-c app already) Sep 21 12:00:07 rwhitby: thanks, that allowed me to find http://wiki.openmoko.org/wiki/HXD8 Sep 21 12:02:11 yep, we know exactly what the internal hardware is, but have no idea about the external case or the intended use. Sep 21 12:02:44 that's what I find so much fun about openmoko future devices:-) Sep 21 12:03:53 hehe Sep 21 12:03:57 roh: is there any possibility of poking whoever's in charge of the online shop - at least to update the frontpage to reflect current dates? Sep 21 12:04:07 And the smaller screen somewhat suggests a real keyboard... ;) Sep 21 12:05:01 roh: is there any interest in wworking together on that issue? Sep 21 12:06:21 emdete i dunno. i think yet another potential gui is nice, but i don't think that the current plan to code the apps in gtk will be changed Sep 21 12:06:45 roh: shure. i asked for working together on an api Sep 21 12:06:55 SpeedEvil good question.. i can bring that up in my next phone-meeting Sep 21 12:07:09 Thanks. Sep 21 12:07:51 emdete since the is only a very basic one yet, i think youre free to propose what you think is best to glue what we have together. just don't think anything is final yet Sep 21 12:08:25 such stuff is only part design and quite a big part evolution by trying it out and making it better in a progress Sep 21 12:08:42 (atleast if it should be a nice to use and extensible api at some point) Sep 21 12:14:24 Weather is just too nice outside, I have to go. Sep 21 12:14:33 :) Sep 21 12:14:34 Bye! Sep 21 12:15:30 (script) wiki RecentChanges: [[User:ChrisLord]] [[Neo1973_compatible_cellphone_providers]] Sep 21 12:22:35 roh: good stuff. so i think i continue fighting with webkit. anyway i urgently need hw-acceleration for graphic... Sep 21 12:23:59 There is sharply limited accelleration that can be done with existing hardware. Sep 21 12:24:10 SpeedEvil: what do u mean? Sep 21 12:24:13 The DMA engines can in principle be used to do some moves. Sep 21 12:24:23 and the virtual screen feature could be used. Sep 21 12:24:36 SpeedEvil: if no alpha channel is used - i love alpha channel & use it alot... :( Sep 21 12:24:39 but the current hardware has _no_ dedicated acceleleeration functions at all. Sep 21 12:24:47 it's all done in the CPU Sep 21 12:25:07 SpeedEvil: yes, that's my prob. i dont need wlan at all, but graphic accel Sep 21 12:25:30 what do you need accel for? Sep 21 12:25:49 SpeedEvil: screen animations? Sep 21 12:26:06 You mean stuff like window wipes? Sep 21 12:26:19 SpeedEvil: my scrolling is much to slow. oon the greenphone it was just okay but screenresolution on the neo killed it Sep 21 12:26:45 SpeedEvil: a screenupdate takes about half a second Sep 21 12:26:54 horizontal scrolling can be done easily - with the virtual screen feature, or omnidirectional scrolling, with the virtual screen enabled Sep 21 12:27:24 SpeedEvil: i scroll elements with alpha channel - no panning possible Sep 21 12:27:27 horizontal scrolling can be done easily - with the virtual screen feature, or omnidirectional scrolling, with no toolbars Sep 21 12:27:28 I mean Sep 21 12:28:06 Ah - if you want transparancy, that gets expensive Sep 21 12:28:47 but it is very nice and a main thing in my gui... perhaps /that/ was a wrong decision ;) Sep 21 12:29:42 Personally I don't think transparency has any place on a phone. Sep 21 12:30:07 Simply as it makes the transparent info harder to see. Sep 21 12:31:03 SpeedEvil: :D you haven't seen my app. it makes thinks transparent if they are not active Sep 21 12:31:04 * XorA needs transparency for Fog of War Sep 21 12:31:14 Hehe Sep 21 12:31:38 XorA: spray some milk on your glasses ;) Sep 21 12:31:56 viq: I don't want to think what that would look like... Sep 21 12:32:17 hmm, with three-four neos I can make a Gaming hub Sep 21 12:32:40 XorA: like me and my 4 gp2x's .. ;) Sep 21 12:32:40 has someone thought about using the neo as remote via bluetooth? Sep 21 12:32:45 yes. Sep 21 12:32:49 emdete: yes. Sep 21 12:32:57 * zash` wants a ir-remote thingy Sep 21 12:32:58 i'd like to turn the neo into a bluetooth mouse. ;) Sep 21 12:33:21 torpor: its missing a camera for that gps is much too far to walk Sep 21 12:33:21 emdete: is the background static? Sep 21 12:33:21 map on one, character stats on another, dice roller on another, combat manager on another Sep 21 12:33:22 torpor: Heh, that one was new though :-) Sep 21 12:33:32 all linked in a scatternet cluster Sep 21 12:33:35 XorA: What game btw? Sep 21 12:33:41 CM: WFRP Sep 21 12:33:42 accelerometer should work well. Sep 21 12:33:55 For the mouse. Sep 21 12:33:59 :D Sep 21 12:34:02 that would be cool Sep 21 12:34:06 The Wiimote is BT, right? Emulation ftw. Sep 21 12:34:13 Can't do wiimote Sep 21 12:34:17 XorA: Ah, ok. Never tried that Sep 21 12:34:23 mouse buttons otoh Sep 21 12:34:24 the 'sensor bar' is picked up with a little camera in the wiimote Sep 21 12:34:33 CM: also play ADND and Earthdawn Sep 21 12:34:47 SpeedEvil: yes, a tiled image Sep 21 12:35:07 I've tried ADND, but mostly used swedish games. Sep 21 12:35:17 emdete: it's probably best to fake that, rather than using generic stuff. Sep 21 12:35:59 SpeedEvil: there is not much to fake i'm afraid, alpha is changed with the animation... Sep 21 12:36:02 emdete: simply write a pixmap with the transparancy computed, rather than writing a pixmap with alpha, and then having the software have to lookup the base image, and do the transparency for you Sep 21 12:36:05 CM: as I dont speak Swedish I guess never played any of them Sep 21 12:36:19 emdete: no using the touchscreen of course... Sep 21 12:36:35 XorA: You've missed out on some fun then. ;) Sep 21 12:36:44 * CM takes a coffee break Sep 21 12:36:51 SpeedEvil: the elements are faded out... Sep 21 12:37:03 torpor: i like the gps idea better... Sep 21 12:37:52 emdete: my point is that if you know the base pixmap, you compute the resultant image directly, and write that. Rather than writing a pixmap+alpha, and letting the underlying libraries read the base image and add transparency Sep 21 12:39:26 SpeedEvil: so you need a pixmap for every y-pos.... because the alpha channel is a function of the y-pos :/ Sep 21 12:39:53 pretty much Sep 21 12:40:16 it's just you're doing it yourself, and not going through what may be several layers of code to write the image Sep 21 12:40:18 question is if smedia supports alpha channel... do you know? Sep 21 12:40:37 I _think_ it does. Sep 21 12:40:49 There is a very basic datasheet linked from the wiki Sep 21 12:40:57 what about turning the neo into a poormans optimus mini 3 Sep 21 12:40:59 ... Sep 21 12:41:18 Optimus Prime! Sep 21 12:41:50 CM: Sep 21 12:42:10 good luck turning the Neo into a fucking TRUCK Sep 21 12:42:21 Willittruck.com Sep 21 12:42:24 mjr: with enough of them... ;) Sep 21 12:42:36 http://www.artlebedev.com/everything/optimus-mini/ Sep 21 12:42:38 ^^ this hting Sep 21 12:42:44 you spoon-fed American consumer drones. Sep 21 12:42:59 torpor, yeah that, shouldn't be a problem... Sep 21 12:43:07 little gadget driver love Sep 21 12:43:37 I'm not sure gadget is the way to go (at least if using a Linux desktop), just some client/server doodah Sep 21 12:43:52 using gadget drivers would make it more compatible. Sep 21 12:44:04 true, hence the parenthesis Sep 21 12:44:09 torpor: isn't the optimus mini charper than a neo anyway? Sep 21 12:44:24 cheaper Sep 21 12:44:34 i suppose it is but you can't make a phone call with it. ;) Sep 21 12:44:44 * ScaredyCat coughs Sep 21 12:45:28 go for "Optimus Neo" Sep 21 12:47:02 Kinda doubt it. Isn't all the lebedev stuff in the silly price range? Sep 21 12:47:47 or is it the "for bill gates only" price range? Sep 21 12:49:07 Bill "has no taste" G is probably not in the target demographic. ;-) Sep 21 12:51:11 torpor: it'd be a hi-res optimus mini - 200x200 instead of 96x96 Sep 21 12:51:18 (per screen) Sep 21 12:51:30 rwhitby: i'm tempted to just go ahead and write a gadget driver for it. ;) Sep 21 12:51:31 and you could have six screens Sep 21 12:51:35 yup Sep 21 12:51:59 brb Sep 21 12:52:00 that'd be so cool to plug into an NSLU2 :-) Sep 21 12:52:38 or anything for that matter. Sep 21 12:53:01 yeah, but most other things already have a display of some sort. that's what makes plugging it into an nslu2 so cool :-) Sep 21 12:54:39 i'm personally looking at a 19" rack full of synthesizers and other music gear that i'd relish the thought of controlling with a bluetooth remote control that has a builtin screen Sep 21 12:54:43 ;) Sep 21 13:00:04 whats the name of the package manager GUI on moko again, i forget .. ? Sep 21 13:00:40 ipkg Sep 21 13:00:46 oh, gui. Sep 21 13:00:52 dunno. Sep 21 13:01:03 openmoko-appmanager ? Sep 21 13:02:16 thats the one thanks Sep 21 13:02:25 <-- really away now, bbiab Sep 21 13:07:12 hi all Sep 21 13:08:07 ScaredyCat: Yo :) Sep 21 13:08:55 Was it about that wireless optimus-mini-neo stuff? Sep 21 13:11:28 good morning Sep 21 13:14:46 no, nm too late anywy Sep 21 13:38:57 anyone found out how to delete text in qtopia? Sep 21 13:40:16 Sup3rkiddo: slide right->left Sep 21 13:40:49 ah ok..works, funny it didnt last time Sep 21 13:41:49 Sup3rkiddo: The keyboard is a pain to use, at least for a newbie Sep 21 13:42:26 you couls say that....i sent some garbage to my girl and now shes mad at me Sep 21 13:43:26 Sup3rkiddo: name completion did Cinthia instead of Cindy? :) Sep 21 13:44:10 blindcoder, hehe yeah something like that Sep 21 13:46:44 My name became Claws instead of Claes.. Kind of sounds the same if you pronounce it in swedish at least Sep 21 13:47:00 hi Sep 21 13:58:33 can anyone tell me how to force bitbake to make modules for kernel 2.6.21.6 ? (default is 2.6.22.5) Sep 21 14:02:36 hey can anyone point me to the images which has zero problems with making a call Sep 21 14:03:04 do you have a time machine? :P Sep 21 14:03:06 You mean ones that won't boot? :) Sep 21 14:04:31 ok atleast "calling sorta works" images, qtopia is too blocky Sep 21 14:04:32 jpcass_: PREFERRED_VERSION_linux-openmoko = "2.6.21.6" Sep 21 14:04:46 jpcass_: but this will only work if there is a kernel recipe with that version Sep 21 14:05:29 XorA: thanks, I think that recipe is gone... i only see 2.6.22.5 in packages/kernel Sep 21 14:05:46 jpcass_: update the kernel? Sep 21 14:05:53 jpcass_: easier than going backwards Sep 21 14:06:10 jpcass_: Maybe there's still some old modules left on the buildhost? Sep 21 14:06:53 XorA: yes i guess you're right: qtopia works with the new kernel, doesnt it? Sep 21 14:07:02 Cm: thanks for the tip, I'll have a look... Sep 21 14:07:03 jpcass_: Or use these: http://buildhost.automated.it/OM2007.2/modules-2.6.21.6-moko11-moko11-r2-fic-gta01.tgz Sep 21 14:07:22 jpcass_: I'm using the latest moko kernel with qtopia Sep 21 14:08:05 CM: ok that sounds like the best option. wonder why they shipped an older kernel Sep 21 14:08:41 Known to work I guess Sep 21 14:08:57 Lorn said he'll release a new image next week or so with some fixes Sep 21 14:09:21 for 'make build-package-webkit', what is 'ERROR: No providers of build target webkit (for [])' telling me? Sep 21 14:10:10 ok, good. Sep 21 14:10:28 hello. Someone used "ipkg update && ipkg upgrade" succesfully? I can ssh into openmoko but i can't run internet Sep 21 14:10:28 vciaglia: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 21 14:10:35 DukeOfURL: bb file is webkit-gtk, try make build-package-webkit-gtk Sep 21 14:10:54 thanks Sep 21 14:12:19 vciaglia: have you made a /etc/resolv.conf with your nameservers? without, moko can't resolve hostnames Sep 21 14:12:50 try ping 84.92.40.53 (known good IP address) - if that works you can access the net but need nameserver resolution Sep 21 14:12:55 jpcass_: of course Sep 21 14:13:06 jpcass_: ok, i'm going to try Sep 21 14:13:06 nameserver 4.2.2.1 should be enough for the resolv.conf right? Sep 21 14:13:48 * ScaredyCat uses a proxy on the usb0 Sep 21 14:14:47 basically means you set it to nameserver 192.168.0.200 on the neo Sep 21 14:14:55 very useful on a lappy Sep 21 14:15:31 (script) wiki RecentChanges: [[User_talk:Emreturkay]] [[User:Emreturkay]] Sep 21 14:15:32 http://wiki.openmoko.org/wiki/USB_Networking#Proxying_DNS_requests Sep 21 14:15:40 I even provide the script :) Sep 21 14:16:29 apt-get install dnsmasq for debian people :-) Sep 21 14:16:51 * ScaredyCat avoids debian where possible Sep 21 14:17:38 * CM suggests emerge dnsmasq ;) Sep 21 14:17:40 w00t! ScaredyCat you avoid debian Sep 21 14:17:56 it's too easy Sep 21 14:18:15 a little pain never hurt anyone ;P Sep 21 14:18:15 ScaredyCat: Heh, so you're on lfs? Sep 21 14:18:42 That must be for masochists only Sep 21 14:18:46 hehehe Sep 21 14:19:53 tbh debian pisses me off, at least the installer dis last time I tried to build for a very small disk... Sep 21 14:20:31 I selected what I wanted and stated the install and it happily went on without actually checking the size of what was requested compared ot the disk Sep 21 14:21:05 45 mins later if fails because it runs out of disk space... but it doesn;t actually stop on the first failure... Sep 21 14:21:23 ScaredyCat: high five mate, i avoid debian whenever possible too Sep 21 14:21:25 it keeps going until it's 'completed' it's list of things to install Sep 21 14:23:22 qtopia-phone takes forever to build.. Sep 21 14:23:24 I think 'ipkg install webkit-gdklauncher' was successful because I see files /usr/lib/ipkg/info/webkit-gdklauncher.control and .list. Sep 21 14:23:28 * ScaredyCat needs a faster laptop Sep 21 14:23:34 how do I start the browser? Sep 21 14:24:10 openmoko: 03thomas * r3027 10/trunk/src/target/OM-2007.2/libraries/moko-gtk-engine/ (ChangeLog src/moko-style.c src/moko-style.h): Sep 21 14:24:10 openmoko: * src/moko-style.c: (moko_rc_style_merge), (moko_rc_style_parse): Sep 21 14:24:10 openmoko: * src/moko-style.h: Sep 21 14:24:10 openmoko: Use bit flag to indicate which engine options have been set (fixes any Sep 21 14:24:10 openmoko: options not set in the first matched style) Sep 21 14:25:39 jpcass_: "4 packets transmitted, 0 packets received, 100% packet loss" :/ Sep 21 14:26:07 :o Sep 21 14:26:18 mwester: http://wiki.openmoko.org/wiki/Community_nominees_for_GTA02_P0 :D Sep 21 14:26:20 alienware are selling laptops with solid state disks Sep 21 14:26:47 vciaglia: sounds like you need to set up ip forwarding on the machine your neo is plugged into: http://wiki.openmoko.org/wiki/USB_Networking Sep 21 14:28:45 jpcass_: i'll try to proxying dns requests Sep 21 14:30:54 vciaglia: That ping looks like a rounting problem, not a dns problem. Sep 21 14:30:58 routing Sep 21 14:32:03 Lynet: i've a router-adsl on 192.168.0.1 so i tried to add "nameserver 192.168.0.1" to /etc/resolv.conf and "route add defaultgw 192.168.0.1 netmask 255.255.255.0" but isn't working. I still ssh but i can't upgrade. Sep 21 14:32:54 someone have the same environment? Sep 21 14:34:14 vciaglia: For the phone, the PC is the default gw. The PC needs to do NAT. Sep 21 14:36:38 Lynet: you're right thanks! Sep 21 14:36:41 or you need to set up a bridge on the pc Sep 21 14:37:24 jpcass_: now i'm pinging the known host "4 packets transmitted, 4 packets received, 0% packet loss" Sep 21 14:37:47 but "get: buildhost.openmoko.org: Host name lookup failure" Sep 21 14:38:54 thats a nameserver issue, your resolv.conf needs to point to a valid nameserver (your broadband provider usually have several) Sep 21 14:39:37 try 212.159.13.49 Sep 21 14:40:22 nameserver 212.159.13.49 in /etc/resolv.conf you mean? Sep 21 14:40:28 yes Sep 21 14:40:35 ok, i try Sep 21 14:41:32 nothing, still the same. i'm trying with the dns of my broadband provider Sep 21 14:43:18 yeah. It's working! :) Sep 21 14:43:20 As someone else mentioned here earlier, the default network config should be dhcp client with locallink fallback. Sep 21 14:43:48 good well done! Sep 21 14:44:08 so i'm upgrading the moko system Sep 21 14:45:23 but i've another problem. When the moko boots i can automatically register with my GSM network. Sep 21 14:46:36 *i can't Sep 21 14:49:19 sometimes the status bar and that icons news "TODAY" disappears Sep 21 14:49:29 morning all Sep 21 14:50:26 *near Sep 21 14:50:55 hey C7 ! Sep 21 14:51:09 ~praise C7 for organizing the OpenMoko booth extraordinairly well Sep 21 14:51:14 All hail C7 for organizing the OpenMoko booth extraordinairly well! Sep 21 14:51:27 vciaglia: there are many problems in the UI and gsmd... Sep 21 14:51:42 jpcass_: i see...unfortunately Sep 21 14:51:42 hey hi mickey :) thanks! Sep 21 14:52:07 gah. forgot my openmoko at home today. :/ Sep 21 14:52:19 C7: anything new from your backpack? Sep 21 14:52:24 e.g.: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 Sep 21 14:53:06 mickeyl: unfortunatly not :( Sep 21 14:53:42 does the media player work? Sep 21 14:58:41 Sup3rkiddo: kind of ;) Sep 21 14:59:08 jpcass_: patching the kernel will work? http://bugzilla.openmoko.org/cgi-bin/bugzilla/attachment.cgi?id=313&action=view Sep 21 14:59:15 C7, alright, i will reframe the question to, "Does it play sound"... Sep 21 14:59:32 * Sup3rkiddo knows how smartass people in here are =P Sep 21 14:59:44 Sup3rkiddo: yes, if you kill pulseaudio it will play sound Sep 21 14:59:51 vciaglia: :-( dont know... Sep 21 15:00:09 C7, hmm,..ok thanks mate Sep 21 15:00:22 jpcass_: so we still have to wait to get gsmd 100% working Sep 21 15:00:44 Sup3rkiddo: oh, and youll have to create a folder playlists in your roots homedirectory Sep 21 15:01:26 vciaglia: search the irc logs and mailing lists, i remember some temporary work-around that involved something like /etc/init.d/gsmd restart in matchbox.session Sep 21 15:01:34 vciaglia: patching the kernel fixes things so the device doesn't lock up; gsmd still doesn't like to start correctly very often and requires some "persuasion" Sep 21 15:01:55 Sup3rkiddo: and perhaps even create a playlist by hand (http://www.xspf.org/quickstart/) Sep 21 15:02:50 thanks mwester Sep 21 15:04:00 woohoo..i just played some thing through ogg123 Sep 21 15:05:03 speakers arent too bad are they? Sep 21 15:05:17 try setting amplifier 3d sound in alsamixer... Sep 21 15:05:37 and... how can i use repositories and add new applications, like minimo? Sep 21 15:05:52 is there a wiki page? Sep 21 15:06:44 oh, here: http://wiki.openmoko.org/wiki/Minimo Sep 21 15:09:02 vciaglia: read this (http://lists.openmoko.org/pipermail/device-owners/2007-September/000449.html) to get gsmd working Sep 21 15:12:17 vciaglia: did you see openmoko-browser2? Sep 21 15:12:40 works pretty well Sep 21 15:12:46 C7: how can i use it? Sep 21 15:12:52 and where can i find it? Sep 21 15:13:00 ipkg install openmoko-browser2 Sep 21 15:13:04 ;) Sep 21 15:13:50 is there anyone working on an app to record a phone call as it takes place? Sep 21 15:13:55 i suppose exists "ipkg search package" ? Sep 21 15:14:02 i'm really new to openmoko world :) Sep 21 15:15:25 vciaglia: you can do ipkg list | grep yoursearchterm Sep 21 15:15:36 thanks C7 Sep 21 15:15:41 but im sure theres a nicer way to do Sep 21 15:17:35 -sh: ipkg: not found ... umh... sounds strange! Sep 21 15:18:14 ?? Sep 21 15:18:40 yep, i don't know why Sep 21 15:19:52 which image did you flash? Sep 21 15:20:50 * Sup3rkiddo is gonna try dualbooting qtopia and OM Sep 21 15:21:17 the latest one, C7 Sep 21 15:21:58 vciaglia: from here? http://buildhost.openmoko.org/OM2007.2/tmp/deploy/glibc/images/fic-gta01 Sep 21 15:22:01 but 5 minutes ago i used "ipkg" to upgrade the system. Sep 21 15:22:02 2007.2? Sep 21 15:22:26 ah ok so there was ipkg ;) Sep 21 15:22:26 yes C7 Sep 21 15:22:35 heh, there was Sep 21 15:23:10 im sure there still is Sep 21 15:23:35 root@192:~$ find / -name ipkg Sep 21 15:23:35 /etc/ipkg Sep 21 15:23:35 /var/lib/ipkg Sep 21 15:23:35 /usr/lib/ipkg Sep 21 15:23:35 /usr/share/ipkg Sep 21 15:23:37 maybe your $PATH is wrong Sep 21 15:24:53 i'd like to test Qtopia Phone Edition too Sep 21 15:25:29 * C7 would like to have his neo back :( Sep 21 15:25:41 i'm going to test it too Sep 21 15:25:46 downloading the latest sources Sep 21 15:25:50 C7, there...there...what happened? Sep 21 15:26:01 C7: who has your neo? Sep 21 15:26:08 some ass buggler :/ Sep 21 15:26:31 thief, even Sep 21 15:26:43 qtopia-opensource-sr 11% |******** | 11668 KB 00:05:50 ETAwget: Write Error: No space left on device Sep 21 15:26:45 d'oh! Sep 21 15:26:51 ass buggler? wtf is that! Sep 21 15:27:07 burgler I think Sep 21 15:27:11 snatcher, filcher, pilferer Sep 21 15:27:15 ya, burgler is the correct term Sep 21 15:27:26 ass bandit in UK Sep 21 15:27:26 yours wa funnier ;) Sep 21 15:27:29 highwayman Sep 21 15:27:29 heh Sep 21 15:27:36 :) Sep 21 15:27:38 burglar would be the word... or thief Sep 21 15:27:49 but this insult is very not polite to one portion of population Sep 21 15:28:04 well... is there a positive insult at all? ;) Sep 21 15:28:08 s/positive/polite/ Sep 21 15:28:09 mickeyl meant: well... is there a polite insult at all? ;) Sep 21 15:28:09 ass bandit isn't a thief Sep 21 15:28:15 hahaha Sep 21 15:28:19 * mickeyl chuckles Sep 21 15:28:20 mickeyl: did you shut down my neo before you put it in the backpack? Sep 21 15:28:28 ass burglar is a man who likes men :-) Sep 21 15:28:29 C7: no Sep 21 15:28:37 C7: by now it's probably out of battery though Sep 21 15:28:51 so he would have a deep discharged battery now ;) Sep 21 15:28:52 that's autoshutdown :) Sep 21 15:28:56 tomorrow my 6gb microSD will arrive :) Sep 21 15:29:08 :o Sep 21 15:29:08 C7: but he has a spare battery *cough* Sep 21 15:29:09 ssh c7.neo gllin :-D Sep 21 15:29:26 and will think its broken Sep 21 15:29:35 mickeyl: :) Sep 21 15:29:45 XorA: yeah Sep 21 15:31:08 i still have some hope that someone just took the wrong backpack Sep 21 15:31:11 .. Sep 21 15:31:19 hope dies at last ;) Sep 21 15:31:45 C7: hope that those batteries explode when he puts them in his nokia Sep 21 15:31:55 :) Sep 21 15:32:17 mickeyl: how can I make bitbake unpack a file into a directory name different from the tar filename? Sep 21 15:33:21 C7: actually monitor ebay.ch Neo's are rare enough one turning up would be unusual Sep 21 15:33:34 shame they carn't C4 ;) Sep 21 15:33:56 XorA: thats what im doing Sep 21 15:34:34 XorA: and ill have to ask my provider for my IMEI number Sep 21 15:35:33 just in case someone gets it to work again and puts in his sim ;) Sep 21 15:41:05 jpcass: you can't. if you need that, you need to do something like Sep 21 15:41:14 do_configure_prepend() { Sep 21 15:41:17 mv stuff around Sep 21 15:41:18 } Sep 21 15:41:35 or better... pester upstream to use correct filenames ;) Sep 21 15:42:17 CM: ping Sep 21 15:42:23 mickeyl: thanks - just looking at ckermit_211 as an example which overrides do_unpack() because tar filename != bb packagename Sep 21 15:42:31 right Sep 21 15:42:51 if it's just one tarball in SRC_URI, often it's enough to adjust S Sep 21 15:42:54 ScaredyCat: :) Sep 21 15:43:02 S = "${WORKDIR}/" Sep 21 15:43:07 mickeyl: in my case I am making recipe for "fuse" (zx spectrum emulator) but fuse already exists in packages (user space filesystem) Sep 21 15:43:17 understood Sep 21 15:43:20 in that case adjust S Sep 21 15:43:21 CM: just uploading a new qtopia image, fixed dialer Sep 21 15:43:29 so slow to unpack a 100 MB .tar.gz archive :) Sep 21 15:43:36 mickeyl: ok, I'll try that in first instance, thanks Sep 21 15:43:38 ScaredyCat: Ah, thanks. I'll give it a try right away Sep 21 15:43:39 np Sep 21 15:43:40 and that other bug on the homescreen Sep 21 15:44:21 ScaredyCat: What was that about? Sep 21 15:44:54 the preview src does't work Sep 21 15:44:57 hm, openembedded want's to download a file from a https-server - but the nocheck-argument for wget seems not to be set. how could I fix this? Sep 21 15:45:04 I found some snapshots... Sep 21 15:45:12 those work better... Sep 21 15:45:17 had to fiddle with bb files Sep 21 15:45:41 ok Sep 21 15:45:45 ScaredyCat: you mean the greenphone sources? Sep 21 15:46:00 that reminds me Sep 21 15:46:01 qtopia-opsensource..blah Sep 21 15:46:10 * mickeyl going to blog about his view on all that Sep 21 15:46:18 so little time before vacation *sigh* Sep 21 15:46:29 mickeyl: How long vacation? Sep 21 15:46:37 * ScaredyCat wonders if mickeyl is going to bitch Sep 21 15:46:38 ScaredyCat: eheh, i'm curious to test it Sep 21 15:46:41 CM: two weeks Portugal, Algarve Sep 21 15:46:48 ScaredyCat: you should know me better by now Sep 21 15:46:51 :D Sep 21 15:46:54 mickeyl: Ah, nice. Don't bring any phone! :P Sep 21 15:46:57 roflmao... Sep 21 15:47:16 CM: now way. bringing electronic devices is forbidden by law^wwife Sep 21 15:47:22 Hehe Sep 21 15:47:27 s/now/no/ Sep 21 15:47:27 mickeyl meant: CM: no way. bringing electronic devices is forbidden by law^wwife Sep 21 15:47:45 mickeyl: take heart that people want to dual boot.. ie qtopia gives them a phone now, but they really want Openmoko.. Sep 21 15:47:58 mickeyl: i recently broke that law ;-( Sep 21 15:48:07 ScaredyCat: i know and i understand Sep 21 15:48:09 jpcass: heh Sep 21 15:48:16 bad husband, you Sep 21 15:48:19 jpcass_: http://store.teptronics.com/hcn1.html?gclid=CP3TuYza1I4CFVB1OAodB2-MDA Sep 21 15:48:22 (or boyfriend) Sep 21 15:51:28 mickeyl: overriding S ( S = "zxfuse-${PV}" ) for fuse-0.8.0.1.tar.gz doesnt work, it still unpacks to fuse-0.8.0.1 ... Sep 21 15:51:28 CM: uploaded Sep 21 15:51:51 jpcass: overriding S doesn't change the unpacking directory, it changes the directory your build starts in Sep 21 15:51:55 ScaredyCat: downloaded Sep 21 15:52:01 jpcass_: can't you mv that dir Sep 21 15:52:14 you need to say Sep 21 15:52:15 mickeyl: aha...ok Sep 21 15:52:29 S = "${WORKDIR}/fuse-${PV}" Sep 21 15:52:37 then it'll work Sep 21 15:53:05 and won't step on fuse's toes since WORKDIR is depending on ${PN} anyways Sep 21 15:53:33 have to go for some hours Sep 21 15:53:34 bbiab Sep 21 15:53:38 yes: I will get work/arm.../zxfuse-0.8.0.1/fuse-0/8/0/1 ... Sep 21 15:53:40 thanks Sep 21 15:54:12 looks like its working... Sep 21 15:54:50 and...when will we get the stable release of OpenMoko? Sep 21 15:58:10 ALBUQUERQUE,NM, US 09/21/2007 8:53 A.M. OUT FOR DELIVERY Sep 21 15:58:14 oh my god Sep 21 15:58:24 come on UPS guys Sep 21 15:58:28 i'm still home Sep 21 15:58:29 :) Sep 21 15:58:42 vciaglia: you do't have to build it to test it :) Sep 21 15:58:58 * ScaredyCat phones UPS Sep 21 15:59:11 * ScaredyCat asks for alternate delivery address Sep 21 16:15:12 (script) wiki RecentChanges: [[Neo1973_Hardware]] [[Wishlist:Camera]] [[Community_nominees_for_GTA02_P0]] Sep 21 16:18:07 i still can't find "ipkg" :/ Sep 21 16:18:23 ScaredyCat: you can always track down the truck and ask him for it. :) Sep 21 16:18:34 lol... Sep 21 16:18:44 ScaredyCat: I've done that for work... alot Sep 21 16:18:49 he'll probably throw it over the fence ;) Sep 21 16:19:51 * mwester remembers chasing the Fedex truck down the street, carrying a package to be sent. Sep 21 16:24:39 moin Sep 21 16:26:41 ScaredyCat: I still can't get qutopia to turn off the screen. Sep 21 16:26:51 with my new image? Sep 21 16:26:55 ScaredyCat: It dims, but never turns off. Plugged in or not Sep 21 16:27:00 Yes, new image Sep 21 16:27:01 wow, nothing new since sep 17 on builthost? Sep 21 16:27:08 But calling works :) Sep 21 16:27:11 lemmie test Sep 21 16:27:36 need to build opemoko images too now :) Sep 21 16:27:37 I have Dim light 20s, Display off 40s, and Suspend Off for both Sep 21 16:27:39 did someone get the dual boot to work om/gt? Sep 21 16:27:50 hhf423: Don't know if anyone have tried yet Sep 21 16:28:17 CM: ghiottone did try, but that was like two or three days ago I think Sep 21 16:28:37 Ah, ok. Haven't heard anything about it though Sep 21 16:28:55 CM: then he probably did not get it to run Sep 21 16:29:35 any significant changes in OM in the last two days? Sep 21 16:30:02 s/significant// Sep 21 16:30:09 And still no.. ;) Sep 21 16:30:21 The AI interface is finally in, with the do-what-I mean gesture interface. Sep 21 16:30:39 You just squiggle on the screen while thinking of something, and it does it. Sep 21 16:31:07 did you click 'add to profile' CM ? Sep 21 16:31:08 harhar Sep 21 16:31:14 On the downside, it does occasionally take over the world, and kill everyone. Sep 21 16:31:33 ScaredyCat: Ah, didn't see that one Sep 21 16:31:38 I see that zecke and mickey are working on the browser Sep 21 16:32:01 whats libmokooui responsible for? Sep 21 16:32:45 ScaredyCat: Heh, I seem to have a really long timeout on the usbnet. Sep 21 16:33:05 I like that happening too. Unplug, go out do some GPS traces. Sep 21 16:33:12 Plugin, keep typing on ssh Sep 21 16:33:24 Exactly Sep 21 16:33:50 no news on GPS I guess? Sep 21 16:33:55 Nope. Sep 21 16:33:56 Now it doesn't even dim.. Sep 21 16:37:35 whwre are the qt snapshots? Sep 21 16:39:55 do youwant a build Sep 21 16:40:42 in the forums there is sort of an indication that there is now more recent stuff like 18.09? Sep 21 16:40:59 ScaredyCat: you have something more recent? Sep 21 16:41:57 I have a build from today in my buildhost Sep 21 16:42:10 oh? Sep 21 16:43:05 this? http://buildhost.automated.it/OM2007.2/OpenMoko-qtopia-image-glibc-ipk-P1-September-Snapshot-20070921-fic-gta01.rootfs.jffs2 Sep 21 16:43:24 yes Sep 21 16:43:34 use the latest kernel from there too Sep 21 16:43:46 cool, I'll try that later. any significant changes since three days ago? Sep 21 16:44:10 not really - I've put some minor fixes in there Sep 21 16:44:48 mainly just to be sure it was easy enough to rebuild Sep 21 16:45:23 there's still some way to go, but it is actually usable for a few hours, 5.5 on mine, wihtout calls. Sep 21 16:48:47 5.5 for battery life? Sep 21 16:48:53 yah Sep 21 16:49:24 I'm still wondering how I'm going to be able to use the gta01 neo as a gps in my car Sep 21 16:50:01 It will have to suck 500mW, not the default 100 Sep 21 16:50:11 Will it do that from a "dumb" usb charger? Sep 21 16:51:39 CM: is there not the possibility to tell the neo to use the fast or autofast charging from unix environment? Sep 21 16:51:39 dunno, you mean from the cigarette lighter socket Sep 21 16:51:40 ? Sep 21 16:52:15 CM: because if you set it to fast or autofast in uboot, it charges from a dumb charger. I did that when I was on vacation. Sep 21 16:52:38 CM: I was just not sure how to do that in linux, so I left it in uboot over night. Sep 21 16:53:50 has anyone tried more recent openmoko builds than 20070916? Sep 21 16:54:20 hum give it any inofficial information about gta02 and the release date? Sep 21 16:54:22 folken: see "pahse 1 software testing" in wiki Sep 21 16:55:16 the openmoko.com website has still the information for october Sep 21 16:55:17 s/pahse/phase Sep 21 16:55:42 hhf423: that is exaclty why i asked, since it hasen't been updated. :) Sep 21 16:55:48 i think this question is the most asked ;) Sep 21 16:56:28 folken: if you look onto buildhost, there is only an image from the 17.09. Sep 21 16:57:04 folken: and if you look into commitlog, you can see the answer why not testing anything later: why bother Sep 21 16:57:38 maybe on the weekend there will be a new image with the browser2 to test Sep 21 16:58:07 folken: if you want to try out stuff, have a look at the qtopia image, thats nice Sep 21 16:58:28 hhf423: where can i get informationen about the openmoko-browser2? Sep 21 16:58:38 hhf423: allready running it. But still has too many bugs for day to day operation. Sep 21 16:59:54 PBeck: from the changelogs Sep 21 17:00:30 hhf423: had openmoko-browser2 a project page - anywhere? Sep 21 17:00:43 PBeck: not that I am aware of Sep 21 17:01:22 where can i find any changelogs and pictures? Sep 21 17:02:30 PBeck: pictures: http://scap.linuxtogo.org/ Sep 21 17:02:57 changes: http://lists.openmoko.org/pipermail/commitlog/2007-September/date.html Sep 21 17:03:53 ScaredyCat: do you have any kernel patches in your build? Sep 21 17:03:58 hhf423: ah thanks Sep 21 17:08:03 is there a known good u-boot version? i'm running the latest from svn now, but there is a (reported) bug that makes it not boot the kernel sometimes Sep 21 17:14:28 mwester: it's the same kernel as the standard openmoko build Sep 21 17:16:14 jas4711: this one: http://buildhost.openmoko.org/snapshots/2007.08/images/ Sep 21 17:17:54 ScaredyCat: Ok - you have other updates in your tree that aren't (yet) appearing on buildhost, though? Sep 21 17:18:41 well, yes I guess so.. I build most days... Sep 21 17:18:43 I'm wondering if we need to provide an updated image with some key out-of-tree patches, for the benefit of those who flash the old, stale images and run into the same old problems... Sep 21 17:18:47 todayes are just uploading Sep 21 17:20:10 http://scap.linuxtogo.org/files/43091acec6b18755afc111cb45857f5c.png !!!! Sep 21 17:21:47 * mwester shakes his head sadly - he wants his GPS to work, too. Where the &%(@*&$ is that updated library from broadcom???!!!!!!! Sep 21 17:22:17 :( Sep 21 17:22:49 I remain convinced that selecting an alternate chipset for the GPS *before* Broadcom delivered on the updated library was really, really, naive business move on the part of somebody. Sep 21 17:23:04 it's reportedly not been decided Sep 21 17:23:18 there are reportedly two candidate models, one with each chipset Sep 21 17:23:31 I know *exactly* what priority my employer would place a customer that selected a competitor for the next generation. Sep 21 17:23:59 mwester: broadcom isnt exactly known for their linux friendlyness Sep 21 17:24:06 (did broadcom buy GL?) Sep 21 17:24:18 yes Sep 21 17:24:30 I think the "not yet decided" is back-pedaling. It's pretty clear that the OM core team wants a more opensource-friendly solution, and (as buz_ points out) that ain't Broadcom! Sep 21 17:24:38 i'd rather prefer it not to have broadcom in the phone Sep 21 17:24:42 exactly Sep 21 17:25:00 broadcom may well be the last of the truly opensource unfriendly chip makers Sep 21 17:25:06 The reverse engineering effort proceeds. Sep 21 17:25:09 Slowly though. Sep 21 17:25:21 RE might do jack with a new chip Sep 21 17:25:27 Especially now that there has been doubt raised as to its eventual functionality Sep 21 17:25:47 thats what i mean Sep 21 17:26:13 The annoying thing is that an opensource driver with the HH chip might be very flexible, allowing possibilities that are not likely to be in a chipset that gives a more cooked position. Sep 21 17:26:16 how much cpu does gllin eat? Sep 21 17:26:22 10-20% Sep 21 17:26:28 That's a whole other problem. What incentive is there to reverse-engineer a solution for such a limited user base. I fear that the P1 owners may end up with a non-functional GPS (to go along with the non-functional AT&T sims) Sep 21 17:26:32 SpeedEvil: count on broadcom fighting that tooth and nail Sep 21 17:26:50 buz_: they can't really fight it tooth and nail. Sep 21 17:27:00 sure, they will change hammerhead ever so slightly Sep 21 17:27:15 Umm. No. Sep 21 17:27:27 look at their wifi chipsets Sep 21 17:27:30 Hardware revisions are hideously expensive. Sep 21 17:27:43 they recook wifi chips for no apparent reasons Sep 21 17:27:45 wouldn't they just flash it Sep 21 17:27:51 i dont think it has flash Sep 21 17:27:58 A new chip mask costs millions. Sep 21 17:28:03 seems to be mostly a RF receiver Sep 21 17:28:39 if anyone wants the openmoko-browser2 - it;s in my repo Sep 21 17:28:51 i want a gta02 ;) Sep 21 17:29:05 I'll put one in my repo too ;) Sep 21 17:33:16 http://worsethanfailure.com/Articles/All-For-19-Seconds.aspx Sep 21 17:33:43 tho after seeing gta01 i was tempted to order one ;) Sep 21 17:34:55 hm, how do I turn of the vibration (something was killed while it was on and now it is vibrating all the time :P) Sep 21 17:35:04 OMFG Sep 21 17:35:04 that's why we have seconds since the epoch ;) Sep 21 17:35:17 find / -name "*vibra*" Sep 21 17:35:27 in that directory is a 'brightness' file Sep 21 17:35:31 echo 0 > that file Sep 21 17:36:35 him. Sep 21 17:36:37 oops Sep 21 17:39:00 hehe, brightness didn't work but the delay_on worked :P Sep 21 17:43:34 hm, they do different things both turning it on :P oh well :) Sep 21 18:07:35 mmp, how're things? Sep 21 18:09:04 mccarthy_: hello, as usual... Sep 21 18:09:22 during GSoC, h5000 support was improved quite a lot... Sep 21 18:09:35 I was pleasantly surprised to see that X now works "out of the box" with angstrom images. Nice work Sep 21 18:10:08 mccarthy_: it's work of angstrom developers; they bumped kernel version in h5000 angstrom images:) Sep 21 18:10:14 having trouble with usbnet (which I hate) ... but otherwise it looks nice Sep 21 18:10:32 is wifi closer to being easy to get working? Sep 21 18:11:14 err, not that much, I still need to figure out how to reasonably implement power management... Sep 21 18:11:43 mmp, where do you hide your nice kernel's these days? I looked at hh.org/~mmp/h5000 with no joy Sep 21 18:12:24 err, I put kernel image there only occassionaly:) Sep 21 18:12:46 (e.g. when someone asks for it:) Sep 21 18:15:10 mccarthy_: but you reminded me what I wanted to do:) Sep 21 18:15:14 (script) wiki RecentChanges: [[Community_nominees_for_GTA02_P0]] [[Mokopedia]] [[Wish_List_-_Hardware]] [[DECT]] [[User:Pgrodt]] [[Talk:Developer_preview]] Sep 21 18:15:30 * mmp patches his kernel tree with at76c503 driver Sep 21 18:18:59 ..a small patch for man, a giant wlan for my phone? ;-) Sep 21 18:20:04 Lynet: only for PDA :-/ Sep 21 18:22:09 hi all ---- a small announce: in whislist of software, there's someone who wrote of a program to upload quickly images from OpenMoko to Flickr. Someone has already created this kind of utility for iPhone, and has released it under GPL. http://mobilepushr.jottit.com/ My question: there is any possibility to port this application on OpenMoko?? Sep 21 18:22:23 roh - Could you reproduce my bug on your hw? Sep 21 18:23:38 The application's name's Mobile Push, and it's written in C Sep 21 18:23:49 Fradeve87 sure. but how would you load the images onto your neo? Sep 21 18:24:23 kristian-m: bluetooth? Sep 21 18:24:47 kristian-m: sometimes friends send you images from them cells tho your, using bluetooth Sep 21 18:25:20 Fradeve87 i see. do you have the url of that tool Sep 21 18:25:29 is in my precedent post Sep 21 18:25:36 http://mobilepushr.jottit.com/ Sep 21 18:25:58 I've seen the source code and it appears to be only C Sep 21 18:26:12 git://mg8.org/MobilePushr Sep 21 18:27:08 Fradeve87 hmm thats cocoa... Sep 21 18:27:32 Fradeve87 maybe it's redone in some hours in python Sep 21 18:28:07 mmm obviously it's only an idea, but I think that this tool could be useful Sep 21 18:29:06 Fradeve87 i guess the fist step for your usecase would be to build a bluetooth gui Sep 21 18:30:05 lol ---- it's not a bad idea :D Which is the situation of BT communications nowaday in OpenMoko? Sep 21 18:30:41 Fradeve87 you'd have to use the commandline and the bluez tools Sep 21 18:33:07 Fradeve87: there's a screenshot app already... Sep 21 18:33:13 sure -- but I currently do not have a Neo (I will buy GTA02); I've posted only to know the real possibility of a porting of Mobile Pushr Sep 21 18:33:31 so you could modify it to ul to flickr Sep 21 18:35:26 ScaredyCat: I don't know if there is a screenshot; however, I will insert the link to Mobile Pushr in OM Wiki Software Wishlist, in the part related with Flickr. Anytime, who wants to applicate on this, has a starting point ;) Sep 21 18:36:13 hey mokoers, check this out: http://uk.gizmodo.com/2007/09/21/tiny_phonebased_projector_in_a.html Sep 21 18:36:25 the existing app uploads to http://scap.linuxtogo.org/ Sep 21 18:37:28 webjames: resolution? Sep 21 18:38:02 ScaredyCat: not sure, but the idea would be cool, i'll watch the vid and report back.. Sep 21 18:38:29 webjames: i have seen this prototype on IFA, Berlin, this year Sep 21 18:38:56 it's <5 seconds showing it, and 15 seconds of nothing (interesting) Sep 21 18:39:02 did it look good.. Sep 21 18:39:05 ? Sep 21 18:39:12 not for daily use :) Sep 21 18:40:29 ScaredyCat: it's native res: 800 by 600 Sep 21 18:42:18 okay another idea: http://www.engadget.com/2007/09/21/seoul-semiconductors-420-lumen-led-puts-the-sun-to-shame/ Sep 21 18:47:18 mmp, any nice kernels for me, then? Sep 21 19:03:48 hi all Sep 21 19:05:06 hey Sep 21 19:10:50 hi Ghiottone Sep 21 19:11:28 hey kristian-m Sep 21 19:11:44 I've just released and tested the 0.7 version of SettingsGUI Sep 21 19:12:20 0.7-r1 should work out of the box. I have applied just a little patch to setup.py and it worked Sep 21 19:12:42 now I have to go out. just logged in to tell that. Sep 21 19:12:43 Ghiottone nice - I'll put that to the wikipage - I think manual installation is to complex Sep 21 19:12:53 Ghiottone thanks :-) cu Sep 21 19:12:59 :) Sep 21 19:13:01 bye Sep 21 19:36:15 Does anyone know where to look in the build files to see what compiler options are used for compiling .c files? Sep 21 19:36:28 openmoko: 03thomas * r3028 10/trunk/src/target/OM-2007.2/libraries/libmokoui2/ (ChangeLog libmokoui/moko-stock.c): Sep 21 19:36:28 openmoko: * libmokoui/moko-stock.c: (_moko_stock_add_icon): Start documentation Sep 21 19:36:28 openmoko: for moko-stock Sep 21 19:36:40 casten: grep -r FLAGS ? Sep 21 19:36:45 thank you very much Sep 21 19:37:13 (that was a guess, btw, if the ? didn't make it clear) ;) Sep 21 19:38:28 openmoko: 03thomas * r3029 10/trunk/src/target/OM-2007.2/libraries/moko-gtk-engine/ (ChangeLog data/gtkrc src/moko-draw.c): Sep 21 19:38:28 openmoko: * data/gtkrc: Add some gradient and border options to styles Sep 21 19:38:28 openmoko: * src/moko-draw.c: (moko_draw_box): Implement gradient option Sep 21 19:43:17 CM: http://www.meritline.com/car-charger-usb-chargers.html Sep 21 19:43:30 Output: 5Vdc, 500mA Sep 21 19:46:57 ScaredyCat: is this a passiv usb plug? Sep 21 19:47:25 no idea Sep 21 19:47:43 http://www.gomadic.com/unusbcarchar.html Sep 21 19:47:50 1 amp version Sep 21 19:52:46 i really would like to use a standard headphone with a adapter like this: http://www.handynow.de/handy-zubehoer/product_info.php?products_id=5325 - but on top of the 5euro you pay 8 for shipping :/ Sep 21 19:54:25 why can't you just plug a normal headset without it/ Sep 21 19:54:27 ? Sep 21 19:54:57 ScaredyCat: the normal headset are really bad sound quality. i think the neo is capable for more Sep 21 19:55:19 bluetooth ? Sep 21 19:55:35 as an alternative Sep 21 19:55:37 and a good headphone uses at least 3.5mm Sep 21 19:55:55 for what? i need a conn to a /good/ headphone Sep 21 19:56:17 you ccan get some good bt headphones Sep 21 19:56:23 the headset that comes with the neo is okay but not really good for music Sep 21 19:56:53 what does /good/ mean? i searched very lonng for the one i have now Sep 21 19:57:44 maybe he means a bt-to-3.5mm adapter :D Sep 21 19:57:52 no idea if that exists Sep 21 19:58:38 anrp: good idea ;) don't know if bluetooth is a good idea to transport sound... Sep 21 19:58:51 the neo ones look like a cheap Steinhauser knock off Sep 21 19:59:09 fic says they put a godd sound system in Sep 21 19:59:42 it's mono, in v2 Sep 21 19:59:57 it needs to be twice as good as it currently is then Sep 21 20:00:30 cjb: no, the speaker is, not the sound system Sep 21 20:00:34 cjb: mono - even via a headset? Sep 21 20:00:36 anrp: ? Sep 21 20:00:42 oh, ok Sep 21 20:00:44 if it's mono Sep 21 20:00:58 no, it's the same sound hardware as i understand that Sep 21 20:01:16 just one less speaker for space reasons ? Sep 21 20:01:17 only the speaker was replaced with wifi (whoever needs that one... :/ ) Sep 21 20:01:27 ScaredyCat: yes, as i understand Sep 21 20:04:21 i would like to know how long the battery lasts with wifi turned on - in fact you are free with network but need a cable for power... :D Sep 21 20:06:11 30seconds Sep 21 20:07:34 mickeyl: hey Sep 21 20:07:48 hi zecke ! Sep 21 20:09:51 mickeyl: Some of the images in the artwork dir are misnamed - mind if I fix that or would you rather I file a bug report again? Sep 21 20:10:30 abraxa_: go ahead. Sep 21 20:10:33 thanks Sep 21 20:10:39 Thanks :) Sep 21 20:11:22 mickeyl: btw - gstreamer still stutters on UI interaction on gta01 even with my patch Sep 21 20:12:13 mickeyl: top says the player eats ~40% CPU (incl. gstreamer), PulseAudio ~35% and openmoko-today 8~15% (wtf?) Sep 21 20:12:16 abraxa_: that's not surprising. we're going to a lot of indirections and just have 266MHz Sep 21 20:12:29 s/to/through/ Sep 21 20:12:30 mickeyl meant: abraxa_: that's not surprising. we're going through a lot of indirections and just have 266MHz Sep 21 20:12:51 i think further data can only be generated with exmap Sep 21 20:13:00 mickeyl: did you blog ? Sep 21 20:13:06 ScaredyCat: no, i was out eating :) Sep 21 20:13:13 ScaredyCat: doing it within the next 24h Sep 21 20:13:28 plain goes on Sunday, 14:00 Sep 21 20:13:31 eating is probably more important :) Sep 21 20:13:32 plane, even Sep 21 20:13:35 yes! Sep 21 20:13:41 and with a full stomach, everything looks bright ;) Sep 21 20:13:46 hehehe Sep 21 20:13:56 better than plaid Sep 21 20:14:02 yea Sep 21 20:14:11 anyone know where I can find the CFLAGS, etc. for the build? I tried grepping, but I'm not having any luck. Sep 21 20:14:18 sure Sep 21 20:14:23 conf/bitbake.conf Sep 21 20:14:24 * ScaredyCat feeds mickeyl cookies at regular ntervals to keep things looking bight Sep 21 20:14:28 hehe Sep 21 20:15:07 anything package-wise in particular you want me to include in the September snapshot? Sep 21 20:15:15 (script) wiki RecentChanges: [[SettingsGUI]] [[User:ChrisLord]] [[User:Javiroman]] Sep 21 20:15:26 i think we have 5 to 7 MB left until we need to stop stuffing software into it Sep 21 20:15:34 :o Sep 21 20:15:52 how much space does it leave? Sep 21 20:16:07 mickeyl: I'll keep trying to improve things and commit to SVN when I think I found something that works - is there anything I should/need to do to get my patch into OE? (if OE is even the right place in the first place) Sep 21 20:16:57 abraxa_: OE is the right place for everything that upstream will take a while (or will never) accept Sep 21 20:16:57 how about ooo? Sep 21 20:17:04 open office?????? Sep 21 20:17:05 .. Sep 21 20:17:06 * anrp runs away Sep 21 20:17:17 ScaredyCat: we need to leave ~10 MB free for jffs2 Sep 21 20:17:23 ScaredyCat: right now we are ~43 Sep 21 20:17:27 I'd rather stop @ 50 Sep 21 20:17:48 user data can go on the SD Sep 21 20:18:14 mickeyl: Okay, so when I commit I'll also send you a proper patch for OE - sounds good? Sep 21 20:18:18 it would almost make more sense to go with two microsd sockets, one internal and one external, instead of futzing around with nand flash for data storage Sep 21 20:18:23 abraxa_: sounds good. Sep 21 20:18:26 get a 4mb flash for kernel and use sd for the rest... Sep 21 20:18:53 yeah, no kidding. i thought about that as well Sep 21 20:18:57 yeah... I think it's probably a good idea to either use some flash (separate mtdblock?) for users settings so that they don;t get over writeen all the time or make it clear that user stuff should be on the sd card Sep 21 20:19:00 is there a way to see the current somehow, the power that is currently drawn from battery? Sep 21 20:19:12 Hmm, I don't see anything that looks like compiler flags in bitbake.conf Sep 21 20:19:17 microsd has large economies of scale Sep 21 20:19:31 i can pay $10, and get 8mb worth of flash chips in low quantity Sep 21 20:19:39 or, i can pay $10, and get 1GB of flash Sep 21 20:19:44 ... and it's physically smaller! Sep 21 20:20:06 casten: # Build flags and options. Sep 21 20:20:07 ... Sep 21 20:20:14 there is all you need Sep 21 20:20:21 + optimization flags in tune-.conf Sep 21 20:20:30 the rest comes from the actual Makefiles Sep 21 20:20:35 (of course) Sep 21 20:20:54 be sure to add -fomit-random-instructions Sep 21 20:21:12 heh Sep 21 20:21:16 * mickeyl just read that as -fvomit Sep 21 20:21:25 vomit? Sep 21 20:21:27 it makes code smaller Sep 21 20:21:48 yeah, -fomit-frame-pointer is often misread Sep 21 20:22:12 -O6 -fuck-upstream-maintainer... Sep 21 20:22:18 hehe Sep 21 20:25:43 ok, I was looking in bitbake/conf not OE/conf, that's why I didin't see anything useful Sep 21 20:26:36 i don't know if my question came through: any idea how to determine the current power that is used by the neo? Sep 21 20:27:37 no way without a scope Sep 21 20:27:43 well, measuring equipment Sep 21 20:28:13 PM is going to change radically in gta02, perhaps there's a change Sep 21 20:28:14 chance Sep 21 20:28:15 hm, that's sad for testing... Sep 21 20:28:29 what means PM? Sep 21 20:28:53 how is the plan for that - i read alot about pm with suspend-resume... i don't like that idea Sep 21 20:29:01 power manegment Sep 21 20:29:18 oh ... yes ;) Sep 21 20:29:38 mickeyl: give it any status updates to gta02? Sep 21 20:29:44 on* Sep 21 20:30:42 PBeck: didn't it just get a red nose and comes through the chminey? Sep 21 20:30:47 emdete: well, you definitely need suspend-resume otherwise you won't get standbye times >= 4 hours Sep 21 20:31:01 there's just no way without a much more complicated smartphone design Sep 21 20:31:09 mickeyl: how about freq change of the proc? Sep 21 20:31:22 emdete: yes, but that won't change a lot. the CPU is not the critical factor here Sep 21 20:31:35 emdete: ? Sep 21 20:31:46 (at least that's what i've been told... not being a hardware guy) Sep 21 20:31:47 so steady-on is no way with neo? Sep 21 20:31:51 PBeck: christmas? Sep 21 20:32:05 emdete: i think that are inoffcial information? Sep 21 20:32:06 emdete: can't for sure, but i highly doubt it from what the HW engineers told me Sep 21 20:32:31 PBeck: can't say more. my gta02v2 was (as always) intercepted by customs and will be here tomorrow Sep 21 20:32:38 which is of no use as I'm going on vacation on Sunday :D Sep 21 20:32:41 so what 'more complex design' may that be? Sep 21 20:33:01 emdete: more intelligent PMU Sep 21 20:33:04 mickeyl: you got one? he! :D Sep 21 20:33:33 mickeyl: ah ok. you aren't in taiwan? Sep 21 20:33:41 mickeyl: means what? what can that do more than switch parts of when they aren not used? Sep 21 20:33:54 PBeck: right. i'm in frankfurt at the moment Sep 21 20:34:12 mickeyl: gruesse nach frankfurt ;) Sep 21 20:34:22 mickeyl: grüße aus stuttgart :) Sep 21 20:34:37 emdete: don't ask me, i'm really out of my depth here. that'd be a question for laforge Sep 21 20:34:41 PBeck: danke und zurück :) Sep 21 20:34:51 ;) Sep 21 20:34:56 Does anyone know if it is possible to create compatible partially linked libraries using gcc without installing the whole OE build system? Sep 21 20:35:09 Sorry to ask so many annoying questions. Sep 21 20:35:18 what is a partially linked library? Sep 21 20:35:29 I mean an ELF I guess. Sep 21 20:35:32 we have a toolchain available, if you mean that Sep 21 20:35:46 it's just not polished very much, but it works Sep 21 20:35:47 mickeyl: i hope v2 had not a hell of a lot bugs :/ Sep 21 20:35:56 PBeck: no comparison to gta01 Sep 21 20:36:03 PBeck: layout was done inhouse Sep 21 20:36:09 pretty much everything worked right in v1 Sep 21 20:36:41 mickeyl: v2 external? Sep 21 20:36:49 ? Sep 21 20:37:03 * mickeyl prefers full sentences Sep 21 20:37:09 lack of context ;) Sep 21 20:37:34 so an ARM7TDMI .elf file made with another tool won't be able to be linked with an app written for this platform? Sep 21 20:37:35 hum my english ^^ which layout was done inhouse v1 or v2? Sep 21 20:37:43 how much of raphic accelleration is done for x11? Sep 21 20:37:49 * hhf423 looks at builthost, finds nothing new Sep 21 20:37:57 casten: oh, if it's gnueabi, it might work Sep 21 20:38:25 PBeck: ah. all the GTA02v* layout was done inhouse. all the GTA01[B]v* was done outsourced Sep 21 20:38:50 mickeyl: that mean a higher quality? Sep 21 20:38:57 PBeck: yes. Sep 21 20:39:00 mickeyl: any news to the gllin-eabi driver? Sep 21 20:39:03 ok super :) Sep 21 20:39:06 If it doesn't make any standard library type calls, then it's just a question of FP stuff, right? Sep 21 20:40:12 are arm7, arm9, arm11 in any way compatible? Sep 21 20:40:15 mickeyl: need we a version 3 for gta02 to fix all bugs? Sep 21 20:40:37 arm7 should run on arm9, arm 9 should run on arm11 I think Sep 21 20:40:56 * hhf423 plunders the scaredycat repository Sep 21 20:41:14 PBeck: yes. it's not only bugs btw., it's also adding some more features. hw design is incremental as well, i.e. v1 didn't include wifi nor bt ;) Sep 21 20:41:20 my neo is arrived! Sep 21 20:41:25 yay! Sep 21 20:41:26 casten: strange thing was that greenphone binaries did not work on neo. perhaps this eabi stuff? Sep 21 20:41:27 now the problem is: how do i charge it? Sep 21 20:41:33 what is the alternative to eabi? Sep 21 20:41:34 They mostly added instructions and FP support I think, ARM11. Sep 21 20:41:36 * hhf423 congratulates bedboi Sep 21 20:41:41 thnx Sep 21 20:41:49 fgau: no. i can't make it to work. i have requested more support from Broadcom :/ Sep 21 20:41:57 mickeyl: is december realistic for a relase? developer? Mass?? Sep 21 20:41:57 ~lart closed source Sep 21 20:41:58 * apt cats /dev/urandom into closed source's ear Sep 21 20:42:12 bedboi: hook it up to a linux system Sep 21 20:42:18 ok Sep 21 20:42:34 i connected it through a usb cable Sep 21 20:42:36 PBeck: if you ask me, i think we can sell the hardware in december. Note that .tw is some miles away and I'm not always superinformed regarding the hardware status, so take it with lots of salt ;) Sep 21 20:42:49 bedboi: before that, remove the battery for about 20mins to make sure its totally dead Sep 21 20:42:50 mickeyl: thanks, i hope for more Broadcom support Sep 21 20:42:56 fgau: yep, me too Sep 21 20:43:03 bedboi: yes, with usb Sep 21 20:43:03 ~bon appetit casten Sep 21 20:43:06 salt is bad for you Sep 21 20:43:54 mickeyl: which release you think? Mass-Release or only for developer? Sep 21 20:44:09 p1 02 ? Sep 21 20:44:13 bedboi: if you remove battery for 20mins, then charge on a linux system for 30mins it might have enough power so that you can turn it on Sep 21 20:44:23 PBeck: well, we will always want to build a run of 100 or so Sep 21 20:44:28 before thinking about real MP Sep 21 20:44:34 anything else is too risky Sep 21 20:44:35 hhf423: voodoo :) Sep 21 20:44:42 mickeyl: yes you are right. Sep 21 20:44:46 damn i lost my nokia phone :( Sep 21 20:46:12 mickeyl: btw => for the new agenda the openmoko-team should perhaps add a few moth to the calculated date and publish this ;) Sep 21 20:46:14 bedboi: yes, a little bit. the problem is that when you try to turn it on when it is not enough charged, it will "half" turn on or so and then not charge good any more, thats why you need to make sure that its totally dead (20 mins w/o battery) when you need to charge it from scratch Sep 21 20:46:45 bedboi: as long as the phone does turn on, it will always charge good when hooked up to a linux pc Sep 21 20:46:53 mickeyl: so had the openmoko team no stress ^^ Sep 21 20:47:01 PBeck: well... there's always this fight between marketing and engineers Sep 21 20:47:14 OM/FIC is no different in that regard Sep 21 20:47:17 mickeyl: i think you should hear on the engineers ;) Sep 21 20:47:24 heh Sep 21 20:47:25 tell me :) Sep 21 20:47:27 i'm the engineer Sep 21 20:47:34 hhf423: i rember worse things with the h5455 Sep 21 20:47:45 :( Sep 21 20:47:45 mickeyl: thats tricky ;) Sep 21 20:47:58 it is dead now, my poor h5455 :(, my first linux pda Sep 21 20:48:01 if it was me, we didn't even sell gta01 at all Sep 21 20:48:15 mickeyl: marketing wins ever over the engineers in you company ;) Sep 21 20:48:17 *your Sep 21 20:48:17 PBeck: adding months does not make sense, the development process will always take the available time. it is the job of a good PM to make sure that it is not much more than that :-) Sep 21 20:48:17 mickeyl: don't say that. Sep 21 20:48:24 mickeyl: i just got one! Sep 21 20:48:27 wtf Sep 21 20:48:27 well Sep 21 20:48:31 it's not that it's bad hardware Sep 21 20:48:31 :) Sep 21 20:48:35 eehhe Sep 21 20:48:36 it's that the software is not there Sep 21 20:48:43 mickeyl: selling gta01 was a good idea.... in my view Sep 21 20:48:45 and I would have prefered to sell hardware with complete software Sep 21 20:48:52 but luckily this is my personal opinion Sep 21 20:48:52 software is always feasible. Sep 21 20:49:13 hhf423: i make sens, because any news site and co promote the release and then will it be delayed Sep 21 20:49:14 i know. and you know Sep 21 20:49:17 *it Sep 21 20:49:18 but lots of people don't know Sep 21 20:49:24 they think that finished hardware must mean finished software Sep 21 20:49:32 and then are utterly disappointed Sep 21 20:49:43 like my idea was to not ship gta01 with a case Sep 21 20:49:48 to make it clear that it's a developer board Sep 21 20:49:48 :D Sep 21 20:49:57 but marketing didn't want that Sep 21 20:49:58 heh Sep 21 20:50:16 but then you would not even be able to take it around once it is finished Sep 21 20:50:30 right. it was only half earnest anyway ;) Sep 21 20:50:49 mickeyl: i am completly fine with that. i can show the neo to people an tell the idea. the big thing with it is oe - the build system. i start to /love/ it... Sep 21 20:50:50 mickeyl: one point for the marketing ;) October release => two points for marketing ;) Sep 21 20:51:08 mickeyl: I don't know about others, but I not not disappointed about the state of the software 2 months ago, I am disappointed about the slow progress in key areas, like gsmd and gps Sep 21 20:51:32 mickeyl: but you will win on the end ;) Sep 21 20:51:33 yep, who cares about all the other stuff. Sep 21 20:51:49 hhf423: it's important to at least being able to place a call. Sep 21 20:52:17 actually, a rough power management is also needed i guess Sep 21 20:52:35 * hhf423 just flashed the latest openmoko image from ScaredyCat and the top menu does not even bother to show up Sep 21 20:52:39 hhf423: agreed. sorry about that Sep 21 20:52:41 it will change Sep 21 20:52:44 promised Sep 21 20:52:56 no worries Sep 21 20:52:59 there's a lot of resources in the pipeline Sep 21 20:53:05 cool Sep 21 20:53:14 don't overload the pipeline :) Sep 21 20:53:20 hehe Sep 21 20:53:23 no, not that much Sep 21 20:53:28 so are the wories different - i would like to see a working graph-accel... Sep 21 20:53:33 that's even worse (/me is congestion control guy) Sep 21 20:53:49 emdete: who cares about that Sep 21 20:54:03 no, don't overload. "adding people to a late project makes the project more late" (from "the mythical man month", a book that I really recommend) Sep 21 20:54:05 iphone is for That Shit (TM) Sep 21 20:54:31 pipeline? You sure it's not a series of tubes ? Sep 21 20:54:37 hahaha Sep 21 20:54:42 bedboi: see Sep 21 20:55:18 bedboi: the people that see the phone here are Sep 21 20:55:20 emdete: imho, the most important stuff is having the phone to be able to place calls, it is a phone Sep 21 20:55:32 emdete: it's very questionable. Sep 21 20:55:40 emdete: in my case, i hate too bloated stuff. Sep 21 20:55:45 iphone is the case. Sep 21 20:56:16 voice + SMS + 10 hour standby and it's good enough Sep 21 20:56:32 bedboi: its not only a phone, its the phone to end all phones, the last phone you'll ever need. also the last PDA and the last computer. except of course a Linux PC to charge it. hrhrhr Sep 21 20:56:33 mickeyl: I vote for kexec-tools in the image, along with the kexec atags kernel patch. Also, I'd love to see one of the solutions in bug 788 in the kernel, or a fusion of both patches in the kernel... Sep 21 20:56:36 right now I'm more concerned about potential hardware problems like the gsm rf noise in the audio Sep 21 20:56:37 10 hour standby? woot? :P Sep 21 20:56:41 a phone should be sober enough Sep 21 20:56:47 bedboi: a stable working gsm and gprs are important also - shure Sep 21 20:56:48 these are show-stoppers Sep 21 20:56:49 Writchie: oh, i hear that too... any suggestions there? Sep 21 20:56:55 bedboi: but i dont care wlan at all Sep 21 20:56:56 Writchie: it's actually pretty bad sometimes Sep 21 20:56:57 one phone to end them all..? Sep 21 20:57:32 one phone to free them all! Sep 21 20:57:37 ehehe Sep 21 20:57:53 mickeyl: are the alsa .state files supposed to be in /etc or /etc/alsa, becuase current oe places them in /etc Sep 21 20:58:26 they moved to /etc/ a while ago Sep 21 20:58:26 ljp: this is taste... i have no strong opinion on that Sep 21 20:58:35 ljp: any guidance from major distros or FHS? Sep 21 20:58:40 * hhf423 wants the context sensitive todo list. wnats, wants, wants Sep 21 20:58:59 Hi mickeyl! :-) Sep 21 20:59:02 well, personally I think they look better in /etc/alsa instead of thrown in /etc Sep 21 20:59:05 greetings elrond Sep 21 20:59:07 btw ljp i sorted my snapshots and modded the .bb files... all works well now Sep 21 20:59:20 * ScaredyCat agrees with ljp... Sep 21 20:59:27 keep them out of the way Sep 21 20:59:53 everything in its place, a place for everything Sep 21 21:00:03 ok, off to bed for today, busy day tomorrow. ttyal! Sep 21 21:00:30 * ljp notices qtopia-neo iamge has 390 downloads, while greenphone image has a pathetic 26 Sep 21 21:00:38 :) Sep 21 21:00:40 good for TT Sep 21 21:00:43 mickeyl - I have strange behaviour with today2. I still wonder, if I imagine it: I switch to the "+"-tab (for starting something), then I choose "Games" from the dropdown list. I click somewhere. And it starts the App, that was at that position _before_ switching to Games (dialer in many cases) Sep 21 21:00:44 and for FIC Sep 21 21:01:04 well, fic is better at hardware than yuhuatel Sep 21 21:01:05 mickeyl - I feel quite "Heh, what's wrong with me, can apps go that wrong?". Sep 21 21:01:11 ljp: so how many phones do you have? Sep 21 21:01:17 ljp: the preview doesn;t work properly Sep 21 21:01:40 ScaredyCat: yes, I know. someone broke things while I was gone Sep 21 21:01:44 Elrond: not unlikely. i observed something like that myself. not recently though Sep 21 21:02:09 i rather suspect the input focus going wrong Sep 21 21:02:17 mickeyl - It happens for roh's image from people.om.o/roh/ Sep 21 21:02:54 mickeyl - Puhh, good to hear, it's not me. :) Sep 21 21:06:29 mickeyl - And the other thing, I noticed: When killing apps from the today-taskmanager, the whole today app goes white and doesn't react any more. (pressing the power button doesn't help either. I was in the bus, so I couldn't connect usb and test, if the kernel at least was alive.) Sep 21 21:06:30 oh my, tux Sep 21 21:06:33 * bedboi hugs tux Sep 21 21:06:48 Elrond: nasty. that calls for a bugreport Sep 21 21:07:11 * ScaredyCat 's wife makes tuxes for ScaredyCat Sep 21 21:07:21 lucky one ;) Sep 21 21:07:43 she refused to make the 5 foot one I asked for :( Sep 21 21:07:54 ScaredyCat - ;o) Sep 21 21:08:08 mickeyl - I have to find a way to reproduce it. ;-) Sep 21 21:08:39 Elrond: right Sep 21 21:08:52 mickeyl - Oh, and some good news: Some of the games are the first games on a mobile, that I can play without glases. :-) Sep 21 21:09:21 Elrond: cool! that's really good news Sep 21 21:09:44 i think accessibility is one area where we can do a lot to outperform closed source Sep 21 21:09:50 eventually... Sep 21 21:10:27 Yes, I think so too. :-) Sep 21 21:10:35 * CM too :) Sep 21 21:10:46 Took a while to read the backlog.. :D Sep 21 21:11:11 ScaredyCat: Will the Neo negotiate correctly with that charger? Sep 21 21:11:12 mickeyl - Okay, the white today thing does not lock up the kernel. :-) Sep 21 21:11:20 good Sep 21 21:11:29 i think for accessibility the gui must get much simpler Sep 21 21:11:33 root 1083 1082 5 03:39 ? 00:00:07 [openmoko-today] Sep 21 21:11:38 Okay, so today crashed. ;o) Sep 21 21:11:46 dunno... but you can add the uboot option to 'force' it can't you ? Sep 21 21:11:52 mickeyl: A question. Is there any way to let the neo suck 500mW default when plugged in on usb, for use with car charger? Sep 21 21:12:03 ScaredyCat: Ah, just wondered about that :-) Sep 21 21:12:56 root 1084 1082 2 03:39 ? 00:00:04 [openmoko-dialer] Sep 21 21:13:01 root 1087 1082 4 03:39 ? 00:00:05 [matchbox-panel-] Sep 21 21:13:08 Lot's of crashed stuff. ;o) Sep 21 21:13:35 dialer and panel are gsmd related Sep 21 21:14:30 ScaredyCat: the qtopia snapshots will work better :) Sep 21 21:14:30 CM: no. there's a lenghty thread about that on the mailing list. bottom line: we _never_ ever are allowed to do that. you could patch it in if you want to, but on your own risk Sep 21 21:15:04 Ok Sep 21 21:15:22 But then there will never be any way to charge the neo from a simple charger? Sep 21 21:15:33 I mean enough to have it powered on continiously? Sep 21 21:15:36 mickeyl - Yes. I just removed openmoko-panel-gsm from the default panel plugins. Let's see, if my panel works better then. ;-) Sep 21 21:16:06 ljp: read up... I found them, modified the .bb files to use them Sep 21 21:16:23 although a minor hack for the srcdata Sep 21 21:16:29 srcdate Sep 21 21:16:39 CM: I believe there is a trick with a pulldown resistor on pin 5. I would at least be possible to create a USB cable with such a resistor, that could then be used with any USB charger Sep 21 21:17:03 thseiler: Oh, that's good news. :) Sep 21 21:17:17 Huh, what's the charger issue? Sep 21 21:17:37 just don't use it on a port that can't handle 500ma Sep 21 21:17:57 But how does that negotiation work? Sep 21 21:18:02 thseiler: apparently not on the neo Sep 21 21:18:12 Will it automatically go up to 500 on a "dumb" charger? Sep 21 21:18:13 No hardware to read this Sep 21 21:18:15 I've no idea, ask SpeedEvil ;) Sep 21 21:18:24 SpeedEvil: ok, i probably mix up EZX with neo, sorry Sep 21 21:20:41 so, is today's openmoko build usable? Sep 21 21:22:17 mickeyl - Of course, when I have gdb connected to today, I can't crash it. ;o) Sep 21 21:23:27 Elrond: A watched pot never crashes... or something like that Sep 21 21:23:59 SpeedEvol Sep 21 21:24:30 SpeedEvil: Sorry, I must have mixed up something, i ment this bug : http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=685 Sep 21 21:24:56 where can i find the latest somewhat-usable image? Sep 21 21:26:06 bedboi - people.om.o/roh/ Sep 21 21:26:40 Elrond: heh. Heisenbug Sep 21 21:30:40 gotcha. Sep 21 21:30:42 on the charger issue: how about a small plug with usb-a-male == small chip that checks power and does the pin 5 resistor thing == usb-a-female Sep 21 21:30:51 Program exited with code 01. Sep 21 21:30:53 :-( Sep 21 21:31:04 zash`: it can be worked round in software Sep 21 21:31:38 zash`: you first turn on and listen for a host in suspend connected to the device. Sep 21 21:31:47 if it is, you send the appropriate signals to wake it Sep 21 21:32:48 hi all Sep 21 21:33:01 if it's not, and there is no USB bus active there, then you turn on the charger after a couple of seconds delay. Sep 21 21:33:23 Or whatever the specified minimum for USB-suspend saying 'no' is. Sep 21 21:34:05 morning Sep 21 21:34:11 I'd like to ask how to remove package/application for real - because I'm running out of space and ipkg remove appname removes it from it's repository but no free space increases Sep 21 21:34:15 perhaps. Sep 21 21:34:15 any ideas? Sep 21 21:34:43 read up on jffs2. Sep 21 21:34:50 zdanek: sometimes jffs2 takes a while to free space collect Sep 21 21:34:52 it's some complicated allocation procedure Sep 21 21:35:46 * zash` reads of dumb-usb-charger-thingy: "OUTPUT: DC5V/1000mA" Sep 21 21:36:02 zdanek - jffs2 will reacquire the free space, if it needs some for real. Sep 21 21:36:12 The garbage collector runs in the background, turning dirty blocks into free blocks. It does this by copying valid nodes to a new block and skipping obsolete ones. That done, it erases the dirty block and tags it with a special marker designating it as a free block (to prevent confusion if power is lost during an erase operation). Sep 21 21:36:21 http://en.wikipedia.org/wiki/JFFS2 Sep 21 21:36:27 zash`: yes, you basically have to find out that you're on a dumb charger byu it being a source of 5V that is not a functioning USB host Sep 21 21:36:54 can I call "garbage collector" manually? Sep 21 21:37:02 At least with current hardware zash`. Sep 21 21:37:12 I've rebooted phone and nothing changes Sep 21 21:37:14 :D Sep 21 21:37:42 SpeedEvil: preferebly that would be in the pmu Sep 21 21:37:42 SpeedEvil: but you can't do any of that when the battery is dead Sep 21 21:38:05 Writchie: no. But in principle, you can do it in early-uboot Sep 21 21:38:17 zdanek - Just try to use space. :-) Sep 21 21:38:18 zash`: not for gTA01, or 02 at least Sep 21 21:38:32 elrond: :D Sep 21 21:38:57 is there any command like df that calculates real free space? Sep 21 21:40:43 no df or du? Sep 21 21:41:44 except that to boot - you need to press buttons. Even the most primitive use case is plug a dead phone into a charger and have it charge overnight and we cannot meet that at present Sep 21 21:42:32 however solution to this bug may at least make an overnight charge possible http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=762 Sep 21 21:42:43 Writchie: RTC interrupt in principle Sep 21 21:42:56 not from dead battery Sep 21 21:43:08 someone posted an interesting thought on the wiki... add DECT to the phone too ... Sep 21 21:43:18 no - the PMU can be told to charge at 100mA when power comes back as far as I understand. Sep 21 21:43:25 if luser can be trained to not let the battery completely die then many things are possible Sep 21 21:43:27 zash: that's what confuses me - df shows disk usage befor garbage collecting Sep 21 21:43:28 then you can wakeup on RTC Sep 21 21:43:50 DECT is insane Sep 21 21:43:52 SpeedEvil: if you died you can't wake up Sep 21 21:44:01 zdanek: try du -sh / Sep 21 21:44:02 * mickeyl afraid of DECT Sep 21 21:44:11 i don't think there is any autoboot Sep 21 21:44:13 mickeyl: ? why ? Sep 21 21:44:15 DECT-usb-plugin? Sep 21 21:44:27 ScaredyCat: it has a bit of a bad reputation, emission wise. Sep 21 21:44:39 besides, we have enough problems with 4 receivers, 3 senders in the Neo Sep 21 21:45:01 getting RF to play along is very challenging Sep 21 21:45:01 ahh you mean at board level... Sep 21 21:45:04 right Sep 21 21:45:32 RF is such a complex beast. Sep 21 21:45:35 shame... would be nice to have one device for everything Sep 21 21:46:00 ScaredyCat: yes. it'd be awesome. Even Laf0rge would like to have that Sep 21 21:46:01 At work we "only" work at about 100MHz - 310MHz, but it's still scarry. Sep 21 21:46:23 ScaredyCat - I don't want one device for everything. :-) Sep 21 21:46:29 At least not todays. Sep 21 21:46:37 zash: it's cosmicaly sloow :) Sep 21 21:46:43 heheh mickeyl... Sep 21 21:46:48 and counts with SD card Sep 21 21:46:57 Elrond: then you buy the separates edition Sep 21 21:47:00 ;) Sep 21 21:47:05 and with mounted volatile memory Sep 21 21:47:08 zdanek: du -sh /path/to/sd/mount Sep 21 21:47:13 then Sep 21 21:47:19 and yes its slow Sep 21 21:47:39 cus it checks all filesizes and adds them together Sep 21 21:47:46 i know Sep 21 21:48:09 zash - du -x -sh / Sep 21 21:48:13 -x := one fs. Sep 21 21:48:25 there's a new ver of SettingsGUI available to install from Ghiottone's repo Sep 21 21:48:31 I recommend it Sep 21 21:48:54 elrond: works great Sep 21 21:49:00 and fast Sep 21 21:49:12 but what is the size of Flash memo? Sep 21 21:49:23 I get 119MB of data Sep 21 21:49:39 isn't it 64M at Flash? Sep 21 21:50:13 yes, but jffs2 is compressed Sep 21 21:50:26 okay Sep 21 21:50:32 this is why jffs2 is so slow Sep 21 21:50:36 :) Sep 21 21:50:53 might think about a cramfs for the static stuff once we are near a release Sep 21 21:51:00 damn, I'm having some problems with packages lately Sep 21 21:51:00 Configuring openmoko-icon-theme-standard2 Sep 21 21:51:01 Failed to write hash table Sep 21 21:51:01 postinst script returned status 1 Sep 21 21:51:01 ERROR: openmoko-icon-theme-standard2.postinst returned 1 Sep 21 21:51:15 everytime I do some upgrade Sep 21 21:51:21 eeks. that looks like your fs is full Sep 21 21:51:30 bingo! Sep 21 21:51:30 mickeyl, or mlock() for major proc/libs? Sep 21 21:51:43 so that's why I came here :] Sep 21 21:51:58 mjr: interesting. Sep 21 21:52:14 zdanek: try removing some files you don't need, then reboot, then ipkg again Sep 21 21:52:17 df says 1.1M free Sep 21 21:52:20 eeks Sep 21 21:52:22 that's more than full Sep 21 21:52:38 I removed Midori and no change Sep 21 21:52:38 allow jffs2 ~15% of the total flash for housekeeping Sep 21 21:52:44 otherwise you run into trouble Sep 21 21:52:47 I mean, I wondered earlier if a tmpfs for selected things could possibly be a good idea, but mlock might just be simpler... Sep 21 21:52:51 I'm into Sep 21 21:52:55 mickeyl: pm Sep 21 21:52:58 ipkg remove all-that-porn Sep 21 21:53:02 ehehe Sep 21 21:53:12 zdanek: like i said, remove some files (not ipkg remove, real 'rm'), reboot, then ipkg again Sep 21 21:53:16 ljp: ok? Sep 21 21:53:39 but nothing to remove because all I have are app installed via ipkg :( Sep 21 21:54:04 erm Sep 21 21:54:21 the only way to get out of this situation is to remove some files Sep 21 21:54:24 I think I will reflash FS and set ipkg to put files on SD Sep 21 21:54:26 that's all i can say Sep 21 21:54:53 Hello, everyone Sep 21 21:54:56 hi pavelm Sep 21 21:54:59 hi pavel Sep 21 21:55:01 * mickeyl receives a call from wifey Sep 21 21:55:04 bbiab Sep 21 21:55:34 zdanek - rm /usr/share/openmoko/sounds/startup_openmoko.wav Sep 21 21:55:39 hi pavelm Sep 21 21:56:03 Hi zecke/mickeyl! Sep 21 21:56:04 (that vies 1.7M) Sep 21 21:56:16 Pretty full of developers here :-). Sep 21 21:56:16 elrond: I will because I always remove it from pulse/session :D Sep 21 21:56:40 Hi zdanek, nice to see someone from eastern europe here ;-) Sep 21 21:57:15 Q: is there some reasonably recent image available? I could flash scaredycat's image, but I guess something newer is probably available...? Sep 21 21:57:20 pavelm: yeah! the same thought :D Sep 21 21:57:58 ! Sep 21 21:58:17 pavelm - ipkg upgrade? ;o) Sep 21 21:58:17 * ScaredyCat points at the date and time on his images Sep 21 21:58:51 scaredycat: sorry ;-). can I have url again? I have one of your images here, but not url any more :-( Sep 21 21:58:53 zdanek - /usr/share/openmoko/sounds has more files to kill. ;o) Sep 21 21:59:06 elrond: i can see Sep 21 21:59:06 http://buildhost.automated.it/OM2007.2/ Sep 21 21:59:20 especially notify_doorbell.wav Sep 21 21:59:26 doubled with .mp3 Sep 21 21:59:52 elrond: It contains qtopia just now, I do not think I want ipkg update. Sep 21 22:00:46 pavelm - Ahh, okay. Sep 21 22:00:50 scaredycat: I see you have qtopia images... did you experiment with qtopia and X in same image? Sep 21 22:01:02 actually you can ipkg update Sep 21 22:01:07 even with qtopia Sep 21 22:01:33 ...even with qtopia from trolltech? Sep 21 22:01:41 pavelm: no, not yet, I was getting a good qtopia build first... Sep 21 22:01:48 yes... Sep 21 22:01:58 it's an oe build Sep 21 22:02:14 so you get ipk's too Sep 21 22:02:58 I need rest otherwise I will crush my nose falling on the keyboard :) Sep 21 22:03:11 scaredycat: OpenMoko-openmoko-devel-image-glibc-ipk-P1-September-Snapshot-20070921-fic-gta01.rootfs.jffs2 is reasonable/current image to use...? Sep 21 22:03:17 goodnight guys Sep 21 22:03:37 goodnight zdanek. At least you are not going full day on horseback tommorow. Sep 21 22:03:39 http://buildhost.automated.it/OM2007.2/OpenMoko-openmoko-devel-image-glibc-ipk-P1-September-Snapshot-20070921-fic-gta01.rootfs.jffs2 Sep 21 22:03:45 is from today Sep 21 22:03:58 17:33 CET Sep 21 22:04:02 yeah Sep 21 22:04:09 cu Sep 21 22:04:16 cu Sep 21 22:04:31 cu -l /dev/ttySAC0 Sep 21 22:04:44 (sorry - bad pun) Sep 21 22:04:47 scaredycat: I see the timestamp, but there's openmoko-devel-image-fic-gta01.jffs2 ... which looks reasonable, too. Sep 21 22:05:01 it's the same file Sep 21 22:05:05 (symlink) Sep 21 22:05:19 scaredycat: Aha, so _that_ is the detail I did not see :-) Sep 21 22:05:33 :) Sep 21 22:05:44 * pavelm would like GPS for tommorow's trip. Sep 21 22:05:57 buy a tomtom :P Sep 21 22:06:11 s/tomtom/garmin/ ;o) Sep 21 22:06:21 Where are the alsa state files gone? Sep 21 22:06:30 /etc/ Sep 21 22:06:34 scaredycat: I already have fortuna handheld GPS. Sep 21 22:06:40 scaredycat; and pygps already beats _that_. Sep 21 22:06:59 ScaredyCat - Oh, directly under /etc/. Who invented that? Sep 21 22:07:02 does it work well? (pygps?) Sep 21 22:07:08 ...plus I keep claiming that openmoko is terrible phone but quite good GPS. Sep 21 22:07:24 Elrond: dunno, they moved from /etc/alsa to /etc Sep 21 22:07:40 scaredycat: reasonably well. shows position, shows sattelites, shows map with arrow "here you are". Sep 21 22:07:47 cool Sep 21 22:07:52 ...yep, it works quite well. Sep 21 22:08:08 is that with *ahem* a chroot environment ? Sep 21 22:08:32 yes, I need chroot environment for gllin :-( Sep 21 22:10:10 is there some development on the gllin side, btw? Like "at least statically linked gllin"? Sep 21 22:10:35 pavelm: hehe, don't ask for statically linked glibc binaries Sep 21 22:11:04 zecke: still better than chroot with all the neccessary libraries, right? Sep 21 22:11:28 pavelm: no, you can't do statically linked glibc binaries. Sep 21 22:11:41 pavelm: they will still do dlopen on the nss stuff... Sep 21 22:12:05 unless you disable the nss calls Sep 21 22:12:23 zecke: ouch :-(. does the gllin actually use the nss stuff...? I guess so, for aGPS support... :-( Sep 21 22:12:25 we'd settle for the soruce then zecke ;) Sep 21 22:12:25 Clint: at that point you can use uclibc as well... Sep 21 22:12:44 * Clint nods. Sep 21 22:12:51 scaredycat: actually documentation would work, too ;-) Sep 21 22:13:30 Playing WAVE 'ringtone_classy.wav': Yuck, that one is loud! Sep 21 22:14:51 Or maybe we can all go beg at Broadcom's corporate HQ to see if there's some way, any way, perhaps they can free up someone to rebuild gllin. How long can that take - 20 minutes on a P-III? Sep 21 22:15:09 gllin is Broadcom? Sep 21 22:15:14 They are now. Sep 21 22:15:16 (script) wiki RecentChanges: [[Wish_List]] [[Mokopedia]] Sep 21 22:15:23 huh. Sep 21 22:15:30 That explains a lot, doesn't it? Sep 21 22:15:38 hehe Sep 21 22:16:03 scaredycat: I had some talk with people at #qtopia, and qtopia should be able to run in X window, using qvfb or something. Sep 21 22:16:04 I can see it: 1 developer, who could compile it and email it out in 20 minutes on one side of the room. Sep 21 22:16:16 ...that should make integration with rest of openmoko possible. Sep 21 22:16:35 On the other: Broadcom's fleet of 10,000 lawyers confering among themselves to see how they get out of the contract. Sep 21 22:16:42 pavelm: ahh ok, i did see the -no-qvfb option in the build... Sep 21 22:16:48 I'll try one without it Sep 21 22:17:04 scaredycat: Yes, that would be useful. Sep 21 22:17:30 mwester: They don't have to get out of the contract. Send us the sources and we'll quietly go away ;-). Sep 21 22:18:33 That's exactly what Broadcom is trying NOT to do -- they have a record of obfuscation, legaleze bigger than the source code, all designed to make it hard for opensource. I'm pretty sure they really don't want to even release a binary driver. Sep 21 22:18:54 I cannot believe that the issue is technical in nature any longer. Sep 21 22:19:04 pavelm: About adding a program to the menu, add .desktop files to /opt/Qtopia/apps (might be /home/Updates/apps in the 4.3 image) Sep 21 22:19:34 mwester - jurisdical-technical. ;o) Sep 21 22:21:36 mwester: I know. I just wonder what they gain from this :-( Sep 21 22:22:23 have a nice weekend folks Sep 21 22:22:52 I've worked with large corporations in negotiations for a number of years -- it's not so much a matter of gain, it's just the corporate culture. "Preventionism" is the term I heard used once to describe it. Sep 21 22:22:58 CM: I have no /home/Updates, and no /opt/Qtopia/apps :-( Sep 21 22:23:23 pavelm: Oh.. Hmm.. Just picked that up in #qtopia while you were offline Sep 21 22:23:24 mwester: Unfortunately I work for one such big dinosaur, too, these days :-( Sep 21 22:23:39 CM: thanks ;-) Sep 21 22:24:23 pavelm: Anyway, ljp said that adding the .desktop file in the right folder would work, so I guess you can find where it is :) Sep 21 22:24:28 CM: anyway, that would allow me to run scripts, but not gtk+ stuff, so I guess I'll try merging qtopia into X images, first. Sep 21 22:24:38 Ok, cool. :) Sep 21 22:24:56 CM: yep, unfortunately I do not know what structure .desktop file should have, so that will have to wait. Sep 21 22:26:07 I think it's a standard format, just look at any .desktop file on your desktop :) Sep 21 22:26:19 * CM hands pavelm_afc a k and takes away the c ;) Sep 21 22:29:04 * pavelm_afc looks confused. He thinks about taking his console (but not keyboard) into the shower, and decides better not to. Sep 21 22:44:07 * CM is away from his keyboard :D Sep 21 22:46:13 poor keyboard, all alone :( Sep 21 22:48:17 poor CM's keyboard can talk to my console ;-) Sep 21 22:49:49 scaredycat: so I flashed your image... it seems gsmd still does not work :-(. Sep 21 22:50:08 pavelm_afc: Using the latest kernel? Sep 21 22:50:38 ScaredyCat: Are there any patches applied? Sep 21 22:51:09 CM: using http://buildhost.automated.it/OM2007.2/uImage-2.6.22.5-moko11+svnr2937-r2-fic-gta01.bin Sep 21 22:51:13 I've only gotten gsmd working once, and that was the build on september third. Since then it's never worked Sep 21 22:51:23 I'm using that one too, but with qtopia right now Sep 21 22:51:37 ScaredyCat's latest build version Sep 21 22:51:51 And calls and sms work just fine Sep 21 22:54:24 cm: I've seen gsmd working randomly -- like on some boots it worked, in most it did not; and it can be usually brought back to liefe by starting it manually. Sep 21 22:54:42 rwhitby talked about some timing issues Sep 21 22:54:51 But I don't know what they were Sep 21 22:56:24 * CM is off to read some more Tad Williams Sep 21 22:56:28 night Sep 21 22:56:29 CM: my timing stuff was actually just the registration state machine, which was fixed. Sep 21 22:56:36 good night. Sep 21 23:00:05 hm, funny, I see changes in gsmd in OE instead of in OM as I would expect... Sep 21 23:04:19 And emotion is still broken in OE, preventing me from compiling the feed. Sep 21 23:13:13 * cesarb is starting to get the feeling the OE people don't care about amd64 Sep 21 23:13:33 2928, 2992... Sep 21 23:14:35 noone cares about amd64 :-). Sep 21 23:14:52 ...unless you have >2G of ram, it only slows you down. Sep 21 23:15:11 cesarb: I set up a debian i386 chroot under my gentoo amd64 system for openembedded :) Sep 21 23:15:21 pavelm_afc: even with the DOUBLING of the number of registers? Sep 21 23:15:58 thomasg: hah. I'm more persistent than you... I insisted on using /bin/dash as /bin/sh when everyone was saying it shouldn't be done (and, in fact, am still using it) Sep 21 23:16:32 cesarb: Mostly yes. Because it also _doubles_ pointer size. Sep 21 23:16:32 brrr posix Sep 21 23:16:40 :) Sep 21 23:16:43 cesarb: I had no problem with building the openmoko-devel-image under my amd64 system. but as soon I was trying something more special, I ran into trouble Sep 21 23:16:48 ...and pointer size actually matters, as you have to transfer that out of main memory. Sep 21 23:17:17 thomasg: only gentoo people seem to be having trouble building things under amd64 (that is, until bug 2928 started breaking things) Sep 21 23:17:52 pavelm_afc: do you have any benchmarks that show running in pure-64 mode actually slows things down? Sep 21 23:18:21 (I don't have any benchmarks at all) Sep 21 23:19:15 Until I see hard benchmarks, with the slowest grade of DDR-2 memory (which is what I have), I'm more inclined to believe all these things about 64-bit mode slowing you down even with the extra registers are nothing more than FUD Sep 21 23:19:43 cesarb: I saw such benchmarks some time ago. Jan Hubicka was working hard to make most benchmarks as fast as i386, but it was impossible for pointer-heavy applications. Sep 21 23:19:51 cesarb: Get SPEC comparisons somewhere. Sep 21 23:20:37 uhuh, just copying /opt from qtopia into scaredycat's images does most of the trick, good. Sep 21 23:20:41 qpe start, and qpe comes up. Sep 21 23:22:18 so... what's with gsmd these days? I had things working a couple of weeks ago, but now it seems to be having problems again... do I need to put back in the gsmd stop/gsmd start in the matchbox session? Sep 21 23:22:55 Yes. Sep 21 23:23:10 The problem is exactly that: there's nothing with gsmd these days! :p Sep 21 23:24:09 mwester: I'd noticed there were changes to libgsmd...and theres been changes in the Dialer. I now get "Dialer: Error setting antenna power when it boots Sep 21 23:24:45 * mwester things everyone is waiting about for a fix, because it *MUST* be coming soon, surely! Therefore nobody is investing a lot of effort in fixing gsmd, if it's "just a few more days" before the real fix comes out... Sep 21 23:24:48 bye bye everyone. Seems like qpe integration was easier then expected... but I have to disconnect network for testing. Sep 21 23:25:01 s/things/thinks Sep 21 23:26:19 I looked at gsmd for while last night. It really can't be that big a problem to resolve, but it would take a lot of effort to understand the code, and there's been nothing but total and absolute silence from the OM core team on the future of gsmd. Sep 21 23:26:36 huh Sep 21 23:27:38 mwester: I tried as hard as I could in bug 766, and failed Sep 21 23:28:07 mwester: with my patch on that bug it works fine, IF it gets an answer from the modem. The problem is that sometimes the modem doesn't answer! Sep 21 23:28:42 mwester: and I exausthed all my guesses as to the cause Sep 21 23:29:23 (btw, this is why I sort of "went silent" for the last week... sorta got too annoyed at my failing at finding the root cause, and gave the whole of openmoko a break) Sep 21 23:32:55 yep, putting the stop/start back into session "cured" it... was able to place a call. Sep 21 23:33:35 jseghers_: I'm refusing to use such a non-fix Sep 21 23:33:43 cesarb: the *IF* part is where I got stumped as well. Sep 21 23:34:36 mwester: yeah, it's pretty annoying, even more since we cannot look at the other side! (the FIC guys perhaps can...) Sep 21 23:34:44 cesarb: I cannot reproduce the problem with cu, or any other tool -- looking that the kernel counters, the serial driver reports data the moment the modem powers on and the CTS line is enabled. Sep 21 23:35:23 mwester: did you try with the two debug patches I added to bug 766 (not the main one which fixes the bug I found, but the other two ones which only print stuff), or some equivalent? Sep 21 23:35:24 cesarb: I can understand that... I can't take such a viewpoint since I need call processing to test the stuff I'm working with... so I just need it to work...whatever workarounds I have to use Sep 21 23:35:50 mwester: it can be really enlightening, since it reports what gsmd is *actually* sending to the modem, not just what it has queued to send Sep 21 23:35:56 I came to the conclusion that writing a mock gsmd from scratch to duplicate the initial handshake might be the next logical step in debugging, but that's where I began to wonder if the effort would be time well spent or not. :( Sep 21 23:35:59 So how to play "Pattern" without a right mouse button? ;o) Sep 21 23:36:10 openmoko: 03zecke * r3031 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (9 files in 2 dirs): (log message trimmed) Sep 21 23:36:10 openmoko: 2007-09-21 Holger 'Siddhartha' Freyther Sep 21 23:36:10 openmoko: Make the "GO"-Page, "Open-Pages"-Page and the "Current"-Page Sep 21 23:36:10 openmoko: share a GtkListStore. This avoids to recreate it in Open-Page on every Sep 21 23:36:10 openmoko: activation, will keep the GtkTreeSelection and will help us to Sep 21 23:36:14 openmoko: bind that to JavaScript. Sep 21 23:36:16 openmoko: * src/Makefile.am: Sep 21 23:36:24 openmoko: 03zecke * r3030 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (ChangeLog src/browser-data.h): Sep 21 23:36:27 openmoko: 2007-09-21 Holger 'Siddhartha' Freyther Sep 21 23:36:29 openmoko: Place the asterisk right. Sep 21 23:36:33 openmoko: * src/browser-data.h: Sep 21 23:36:46 cesarb: I knew you were working on it, but I didnt' think to check a bug report for patches. I'll take a look, and poke around a bit more. Sep 21 23:38:18 mwester: please do, especially with the debug patch which prints what is actually sent. without that one, you are debugging in the blind :P (and I do think it should be added upstream, it's that useful) Sep 21 23:38:45 (there is no need for a patch reporting what was actually *received*, since upstream gsmd already does that) Sep 21 23:40:17 Good grief - that's good information in that bug report. It also confirms that there are some issues in the code to be worked out yet... Sep 21 23:40:46 mwester: yeah, I went a bit too offtopic in that bug report, and dumped everything I found out in it Sep 21 23:41:28 mwester: the bug itself is already fixed (the main patch attached to that bug fixes it)... but I couldn't let it go... Sep 21 23:42:22 I started from the other end, actually. I presumed that we might have a bug in the serial driver, and started investigating there. I didn't record the debugging details, however. Sep 21 23:43:10 Basically what I found is that if you power up the modem, nothing will happen until you enable CRTSCTS - at that point the modem will dump the accumulated data (the greeting usually). Sep 21 23:44:26 If you toggle the modem and mess with the serial port enough before the CRTSCTS line is set, it still works -- but the driver gets strange data - most often a buffer overrun, but sometimes a framing error from the UART, and in some cases break characters. Sep 21 23:45:03 None-the-less, it all stabilized quickly, and everything I tried resulted in a response of "OK" to the "AT" command. Sep 21 23:47:28 mwester: hm... but on my gsmd tests, sometimes ATZ doesn't get an answer... sometimes _ATE0V1_ doesn't get an answer, even if ATZ just got an answer... and it only happens on the first power up, if you kill gsmd and restart it (by hand, not init script), it works Sep 21 23:47:59 mwester: and I failed to find any hypothesis to explain that behaviour Sep 21 23:49:26 One thing I noticed a couple of weeks ago (with start/stop in session file) was that if I booted the neo from cold (battery out) the radio would NOT come up. If I then rebooted the phone, it would come online. Sep 21 23:51:50 jseghers_: that was my main testcase (no start/stop in session file, though) Sep 21 23:52:05 jseghers_: pull the battery, wait a bit, put battery in (no SIM), and boot Sep 21 23:52:17 jseghers_: it gets reproducibly stuck (the modem doesn't answer) Sep 21 23:53:16 jseghers_: specifically, the modem didn't answer to the initial ATZ (perhaps sometimes it got to the ATE0V1 before not answering, don't remember), which got everything stuck Sep 21 23:53:22 how to activate gsm? Sep 21 23:53:25 on 2007.2? Sep 21 23:53:47 jseghers_: then, I kill -9 gsmd and start it by hand (no init script, no modem power fiddling)... and it works, showing the modem wasn't hung after all! Sep 21 23:53:50 wait... Sep 21 23:53:58 * cesarb just thought of something Sep 21 23:54:02 when I kill -9 gsmd Sep 21 23:54:08 it closes the UART fd Sep 21 23:54:15 what does the kernel do when that happens? Sep 21 23:54:25 I'd suppose it _drops flow control_! Sep 21 23:55:21 perhaps if I made gsmd also drop flow control (RTS or CTS, always forget which is in and which is out) and then release it again... Sep 21 23:55:42 * cesarb will look at it after finding out what's broken with raster's evas on amd64 Sep 21 23:56:00 i'm for the blackberry/moto usb dumb charging technique...but neo can't detect? Sep 21 23:56:19 can G Sep 21 23:56:26 can GTA02? Sep 21 23:56:51 cesarb: ugh! Sep 21 23:56:57 my place of biz does NOT allow personal devices to be connected to work machines Sep 21 23:56:59 jaebird: no, and harald (IIRC) already said detecting whether D+ is shorted to D- on the USB cable is too complex Sep 21 23:57:26 "too complex" Sep 21 23:57:30 jaebird: (I had a bug report on it on openmoko) Sep 21 23:57:57 jaebird: well, actually there's more to that in the specification... it says _how_ you are supposed to detect that short, and perhaps that's what is too complex to do Sep 21 23:58:04 then we need to get a usb hub chip and stick it in a "dumb" charger to make it a "smart" charger Sep 21 23:58:16 jaebird: you cannot just randomly pull the line up Sep 21 23:58:49 this issue just seems so silly...especially with china "requiring" usb chargeble devices now Sep 21 23:58:50 jaebird: one thing you can do is to simply add a u-boot menu entry to enable autofast, and also add a menu entry to the main GTK UI to do the same Sep 21 23:59:02 jaebird: plug the dumb charger, enable autofast, done Sep 21 23:59:07 cesarb: that is good, wiki? Sep 21 23:59:35 jaebird: a very smart implementation could even disable autofast after it stops charging (charger removed), so you would not forget it enabled by accident... Sep 21 23:59:43 u-boot already...but need it for when powered on and booted Sep 21 23:59:54 cesarb: That's a good thought. You actually have no control over the CTS line with the serial driver if its in flowcontrol mode, so you need to turn off and on CRTSCTS mode. Sep 22 00:00:21 There are actually no modem control signals on the serial port, just CTS and RTS -- the others are just dummied up by the driver. Sep 22 00:00:25 jaebird: I don't recall where the uboot trick is on the wiki (speedevil can give you a premade menu recipe for it), but for the linux kernel you just have to fiddle with the mode in sysfs (you want fast_cccv) Sep 22 00:01:02 mwester: well, actually, if it's what I'm thinking, you just have to close the fd and reopen it (perhaps with a sleep() in the middle) Sep 22 00:01:45 i think the "just works" philosophy should be applied to this Sep 22 00:01:49 mwester: I'd have to dive in linux's scary tty layer to check if it does clear the flow control when closing, however Sep 22 00:02:03 If you close the fd, the tty settings will revert to what they previously were - (and it does clear the flow control on close) Sep 22 00:02:12 especially when the "big" vendors have already solved the problem Sep 22 00:02:23 !ombug 685 Sep 22 00:02:57 jaebird: that's the bug, you will find on it a link to the relevant specification (which is very recent... gta01 was already started when it was released) Sep 22 00:03:12 right...but what about gta02? Sep 22 00:03:20 jaebird: I know from the "internets" that at least my motorola charger does short D+ to D- Sep 22 00:03:40 jaebird: too late for gta02 too ;-) (according to harald) Sep 22 00:03:45 nice Sep 22 00:03:45 Another (easier and broader) test would be to disable the entire tty setup stuff in gsmd, and rely on an "stty" command to do it. At least for debugging. That would ensure that gsmd can open/close the fd without resetting any of the tty settings. Sep 22 00:03:57 jaebird: it will be. It needs work done first. Sep 22 00:04:00 the PMU should be taking care of it, imo Sep 22 00:04:10 Unfortunately the PMU is dumb Sep 22 00:04:14 it can't talk USB Sep 22 00:04:48 and to do it right, we need to talk over the USB bus to see if we are connected to something that cannot supply 500mA, and if not, don't enable that mode Sep 22 00:04:53 but their should be some hardware hackery that would allow it to detect the short or resistance, et al Sep 22 00:05:00 There is not. Sep 22 00:05:09 There will not be in GTA02 Sep 22 00:05:22 It is possible to do it in software. Sep 22 00:05:32 It's just that the software has not been written. Sep 22 00:05:39 right Sep 22 00:05:41 s/it/most of it/ Sep 22 00:05:41 SpeedEvil meant: It's just that the software has not been wrmost of itten. Sep 22 00:06:30 mwester: well, the hypothesis is that _just opening and closing_ the fd is enough to flip the relevant line if crtscts is enabled Sep 22 00:07:31 SpeedEvil: detecting a short between D+ and D- does not require actually talking over the bus... just a softare-controlled pullup plus a comparer on the other end, if I'm not spouting nonsense... Sep 22 00:07:48 s/softare/software/ Sep 22 00:07:49 cesarb meant: SpeedEvil: detecting a short between D+ and D- does not require actually talking over the bus... just a software-controlled pullup plus a comparer on the other end, if I'm not spouting nonsense... Sep 22 00:08:31 SpeedEvil: on the other hand, it's impossible to do in software, since the CPU does not allow using the lines as GPIOs Sep 22 00:09:05 and the PMU is dumb :) Sep 22 00:09:12 It's impossible to do directly in software. Sep 22 00:09:28 no lines at all..eh? Sep 22 00:09:41 You can eliminate 'we are connected to a USB bus in suspend' and 'we are connected to a USB bus that cannot suppply 100mA' Sep 22 00:09:46 in software. Sep 22 00:09:52 You can eliminate 'we are connected to a USB bus in suspend' and 'we are connected to a USB bus that cannot suppply 500mA' Sep 22 00:10:30 in early-uboot - and mandate that we keep 5% power - or whatever the minimal startup charge is Sep 22 00:10:55 SpeedEvil: we can detect we are connected to a suspended USB bus? cool Sep 22 00:11:23 cesarb: I assume so - I haven't dug into it, I just know there is a way for a device to signal a USB bus to wake Sep 22 00:11:59 The other cases are 'we are connected to a host with a crashed USB bus' and 'We are connected to a charger' - the cases that have 5V on the USB pin that is. Sep 22 00:12:05 SpeedEvil: ok, waking up a sleeping computer is a bit of a bad behaviour, but is much better than the alternative Sep 22 00:12:28 Suspended busses are not guaranteed to provide any power Sep 22 00:12:36 any significant power Sep 22 00:12:44 SpeedEvil: can we detect we are connected to a windows box without the drivers for the neo installed, through a small bus-powered hub in the back of a monitor? Sep 22 00:12:49 so if we want to charge, we need to wake them. Sep 22 00:13:11 SpeedEvil: (or if we are connected to the same windows box, but directly so we can draw the whole 500mA?) Sep 22 00:13:13 gadget-mass-storage shoulf fix that Sep 22 00:13:23 SpeedEvil: ah, great idea! Sep 22 00:13:29 SpeedEvil: or perhaps HID Sep 22 00:13:41 From what I recall of the power managment chip, there is not a 'draw x00mA' mode. Sep 22 00:13:43 SpeedEvil: there's no way Win would ignore a poor little HID device :D Sep 22 00:13:43 does one know of a "smart" charger that isn't a pc? Sep 22 00:13:57 SpeedEvil: HID would be easier to implement in the Bootloader, me thinks... Sep 22 00:14:09 SpeedEvil: there's fast_cccv, which is 500mA, and the qualification/precharge, which is less than 100mA, right? Sep 22 00:14:23 Just 'charge, drawing up to 500mA from the USB bus', and '100mA' mode Sep 22 00:14:31 SpeedEvil: I think right now, the procharge is more like 43ma Sep 22 00:14:40 it is Sep 22 00:14:51 thseiler: you are speaking to the guy who measured it ;-) Sep 22 00:14:59 oups Sep 22 00:15:08 SpeedEvil has some sort of measuring contraption (sp?) Sep 22 00:15:09 '100mA' is what it was referred to Sep 22 00:15:18 in some docs. Sep 22 00:15:26 some copper strip thingy Sep 22 00:15:36 (script) wiki RecentChanges: [[User:Avanc]] [[User_talk:Avanc]] [[Geocaching]] [[GPS]] Sep 22 00:15:52 i suppose it is still a bit to early for folks to "vote" on features besides hacking it themselves :) Sep 22 00:16:15 SpeedEvil: Yep, I was looking trough the docs of the PMU, and i think all that is needed to fix this is to programm MBCC3 . CURRAT to 10 insteat of 01. Sep 22 00:16:25 in the bootloader Sep 22 00:16:30 upon every boot. Sep 22 00:18:25 cf page 69 of the datasheet... Sep 22 00:21:08 thseiler: what does that do? Sep 22 00:21:47 Its setting the current ration between fast charging mode and trickle charge Sep 22 00:22:08 Ah - so that boosts trickle charge to 100mA? Sep 22 00:22:21 Is that allowed by the _battery_ datasheet? Sep 22 00:22:31 Right now its set to 01 which means a factor of 0.10 * I_Fast according to the datasheet Sep 22 00:22:32 Umm - what? Sep 22 00:22:59 ah Sep 22 00:23:11 yes, you can safely charge the battery at 100mA Sep 22 00:23:57 SpeedEvil: a value of 10 means 0.20 * I_Fast. Sep 22 00:24:07 0.20 * 500mA = 100 mA Sep 22 00:24:18 Sounds plausible. Sep 22 00:24:27 Sorry to do what so many people must be doing at the moment - but I've popped in to find out if anyone has any more information on when gta02 will be available for people wanting to develop applications on it. I'm trying to decide whether to just cut my losses and get an a780 or a1200 instead and us OM on one of them... Sep 22 00:24:36 SpeedEvil: So I wonder if we programm that register via Bootloader upon every boot, this might fix the 43 ma tricklecharge issue... Sep 22 00:24:53 Probably. Sep 22 00:25:10 Anybody here with bootloader hacking expirience ? Sep 22 00:25:58 * jaebird runs away Sep 22 00:26:18 should have gotten the dev board Sep 22 00:30:13 Ok, I have a devboard, but I have never used it before. Sep 22 00:31:01 How do I flash or run gdb on the bootloader with a debug board ? Sep 22 00:37:33 ok, found all the infos about the debug board in the wiki... Sep 22 00:38:01 thanks to all you folks, that keep the wiki up to date Sep 22 01:41:24 When I get a "Working copy '.' locked" when running 'make update', where exactly is '.' so I can run 'svn cleanup'? Sep 22 01:47:47 tRubl, most probably the openmoko dir. try svn cleanup openmoko Sep 22 02:02:31 Thx Ghio, it's not complaining anymore, but apparently it's not resolving the lock issue.' Sep 22 02:07:31 tRubl, Sep 22 02:08:36 tRubl, try to pastebin the error lines Sep 22 02:13:37 http://www.pastebin.ca/706600 Sep 22 02:15:34 what did svn cleanup say? Sep 22 02:15:42 verbose mode if its available with that option Sep 22 02:15:45 the line before the error was about bitbake dir Sep 22 02:15:55 try svn cleanup bitbake Sep 22 02:21:48 cleanup doesn't support verbose...I will try bitbake now Sep 22 02:23:41 yeah, that was it, thank you Sep 22 02:27:24 np Sep 22 02:27:28 nite all now Sep 22 02:41:40 hm, from my reading of the code, it seems the evas/edje crashes are in libpng's mmx assembly code, or perhaps in gcc's interpretation of it (it seems to be ignoring a clobber of %ebp) Sep 22 02:41:57 (which means they are not raster's fault ;-) ) Sep 22 02:46:22 A simple fix seems to be to upgrade from libpng-1.2.19 to libpng-1.2.20, which utterly removed all the assembly code :D Sep 22 02:46:32 * cesarb prepares his local overlay for a test **** ENDING LOGGING AT Sat Sep 22 02:59:56 2007