**** BEGIN LOGGING AT Sat Oct 17 02:59:58 2009 Oct 17 06:42:13 Slyon: ping Oct 17 06:42:23 Slyon: what you did in setup.py does not work... Oct 17 06:42:29 install paths cannot start with a / Oct 17 06:44:20 <[Rui]> good moanin' Oct 17 06:45:18 [Rui]: morning ) Oct 17 06:45:47 <[Rui]> mrmoku: hear there's loots of goodies on your "oven" :) Oct 17 06:46:13 uhh.. yes :) Oct 17 07:00:41 TAsn: one thing though... we probably might have to keep the ophonekitd usage interface as is... Oct 17 07:01:16 as it's an actual use... and breaking stuff just for renamin is bad... we had that with our scenarii ;) Oct 17 07:16:47 hehehehehe Oct 17 07:17:19 my favourite Oct 17 07:17:23 joerg_42: I could finally identify my microphone problems. Oct 17 07:17:54 khiraly1: interesting. What was it? Oct 17 07:19:11 joerg_42: It seems, when I augment the sensitivity, it augment everything, especially the background noise. If I augment too much, it festers Oct 17 07:20:57 well that sounds like basics, yes Oct 17 07:21:05 And with combination of phones (samsung l700) and provider (vodafone hungary), when I augment too much the microphone, it start jag at the samsung l700 Oct 17 07:21:25 so he can only hear just some part of my voice Oct 17 07:21:37 and thats why he usually ends up recalling me from landline Oct 17 07:21:51 ooh, that's bad - of the 1700 ;-P Oct 17 07:22:22 joerg_42: no, its not a problem only with samsung l700. Just it is the most prone to fail (almost all the time) Oct 17 07:22:49 it happens on other mobiles (but not on landline), just they experience more rare Oct 17 07:23:25 landline people complain about the noisy environment (even when I sit in my own room, and they start to complain about bus(!!!) noise in the background) Oct 17 07:23:37 SHR: 03mok 07phoneuid * r36156f875b8b 10/ (55 files in 8 dirs): drop ophonekitd stuff Oct 17 07:23:41 SHR: 03mok 07phonefsod * reb4db1149c50 10/ (47 files in 7 dirs): drop the phoneuid part Oct 17 07:25:11 noisy environment is problem for everybody. So or I lower the microphone sensitivity, and I start shuttering in my phone all the time. Or I augment the microphone, but in that case the background noise are amplified too much too. Oct 17 07:25:53 in everycase It is impossible for me whispering in the phone Oct 17 07:26:01 ~tell khiraly1 about fr-audio Oct 17 07:26:09 So I cant say "Sorry Im on a meeting" Oct 17 07:27:41 joerg_42: yepp, you can redirect to it all the time, or do you want to say some specific thing to improve? Like background noise suppress ? Oct 17 07:28:46 please read about AT%Nxxxx in the wikipages referred above Oct 17 07:29:33 ~tell khiraly1 about fr-audio Oct 17 07:31:42 hmm Oct 17 07:31:47 no AT%N command Oct 17 07:31:52 http://wiki.openmoko.org/wiki/Neo_Freerunner_audio_subsystem Oct 17 07:31:55 http://wiki.openmoko.org/wiki/Neo1973_Audio_Subsystem Oct 17 07:32:02 but I remember lindi told me once Oct 17 07:32:30 j matchbox Oct 17 07:32:32 ups Oct 17 07:33:30 khiraly1: %N command is documented on the wiki Oct 17 07:34:30 PaulFertser: got an exact link? Im googling and also diving into irc logs, but cant find it Oct 17 07:35:28 khiraly1: http://wiki.openmoko.org/wiki/Neo_1973_and_Neo_FreeRunner_gsm_modem Oct 17 07:35:58 http://wiki.openmoko.org/wiki/Hardware:AT_Commands#Echo_Suppression_and_Noise_Reduction Oct 17 07:36:14 aham Oct 17 07:36:32 ok, I will try out with my friend's phone again, because it is the most easiest setup Oct 17 07:36:35 khiraly1: for extensive description look at http://git.freesmartphone.org/?p=framework.git;a=blob;f=framework/subsystems/ogsmd/modems/ti_calypso/channel.py;h=c86ceee43685acefcbed1e0bec5e1b382f7a51d3;hb=043cef6d2122bb62dcbae3211ed94c21d9c7d901 Oct 17 07:36:39 sorry, never realized it's actually missing link on audio page Oct 17 07:37:23 I will test it next week with my friend, because it is the worst setup ever. Could not believe to my own ear when I heared it yesterday Oct 17 07:38:12 too bad, samsung l700 cant record audio Oct 17 07:38:27 would be awesome to post it, to you guys Oct 17 07:39:48 oh dear there are payed software for l700 to call recording. What a stealing (49$) Oct 17 07:47:00 ok, I will play with it next week. mickeyl wrote me once how to issue these commands while talking (From commandline), I need to digg the irc logs a bit more Oct 17 07:48:16 khiraly1: I'd guess starting mickeyterm and simply entering the AT%N1234 would be enough Oct 17 07:48:35 to test during call Oct 17 07:48:51 ok, I will try it Oct 17 07:48:57 with fso-simplemixer too Oct 17 07:49:35 thanks for the tip DocScrutinizer, I will report once I completed my experience. Oct 17 07:49:51 Im really happy, that I found the most problematic phone counterpart Oct 17 07:49:55 khiraly1: I hope you didn't apply any nasty "patches" to framework (as suggested some 9months ago) Oct 17 07:50:05 DocScrutinizer: nope Oct 17 07:50:13 bare shr unstable like a month ago Oct 17 07:50:13 good Oct 17 07:50:59 just I need to figure out, how to record test call on the samsung side Oct 17 07:51:04 for permanent setting there's a line somewhere in the fso config files Oct 17 07:51:15 maybe with an another freerunner and audio recording software?;) Oct 17 07:51:39 because on the freerunner side, everything seems great Oct 17 07:51:40 ;) Oct 17 07:51:52 clear loud voice Oct 17 07:55:37 still I wonder why the samsung is so bad, And I doubt you can basically fix that with FR setup. FR *does not* care about far end, it begaves always the same way Oct 17 07:55:48 beHaves Oct 17 07:56:37 khiraly1: records won't help much I guess. What should we do with them? Oct 17 08:00:53 DocScrutinizer: others have the same problems, just not that worst, and occasionally Oct 17 08:01:05 also the background noise is a general problem for me Oct 17 08:09:05 khiraly1: bg noise should be killed by "noise gate" inside the calypso provided you do not gain it too much in the wm8753 Oct 17 08:14:41 but noisegate also can not reduce bg noise and same time amplify voice. It's just a "switch" cuting out audio that's below a certain threshold Oct 17 08:15:15 so usually most problems come from too *much* noisegate Oct 17 08:16:07 as it cuts off beginning/end of words where volume of the voice drops below threshold of nnoisegate Oct 17 08:17:41 DocScrutinizer-8: probably khiraly1 just amplificates mic input too much. Oct 17 08:23:16 PaulFertser: yep, to compensate noisegate cutting out low voice when he "whispers" to the phone Oct 17 08:24:10 DocScrutinizer-8: heh, then there's no way he can fulfil both purposes without dynamically manually adjusting controls :) Oct 17 08:24:31 probably when simply deactivating noisegate all together and then recalibrating the whole micpath will yield better results Oct 17 08:26:38 (manuall dynamically adjust) hmm, a PTT button might be a nice feature for a few people Oct 17 08:55:45 did something change in the buildhost config recently? I see no new images, and I can't get any updates on my SHR and it makes me wonder :) Oct 17 08:57:14 mrmoku|away, I don't think that many people use the ophonekitd interface, anyway, it's ok to break stuff, as we're in a super unstable branch, and renaming that as well will help preventing confusion. Oct 17 09:00:21 TAsn: (confusion) see line above from rebelThor Oct 17 09:00:21 mrmoku|away: i fixed the setup of shr-today, can you rebuild pls :) Oct 17 09:00:41 hehe :) Oct 17 09:01:15 TAsn: seems SHR "improves" beyond all usability Oct 17 09:01:23 rebelThor, don't worry, the pie is in the oven, we'll take it out when it's ready .:) Oct 17 09:02:04 DocScrutinizer-8, the change we did is too big for minor changes. Oct 17 09:02:10 *minor upgrades Oct 17 09:02:23 muhaha, the only pie that take weeks and months and nobody hopes it ever will get tasty Oct 17 09:02:24 so we had to do a freeze Oct 17 09:02:44 hey, I hope it'll be tasty. Oct 17 09:02:46 TAsn: that's cool, but I'd like to test the intermediary versions and report the bugs, not just the final version ;) but I guess the freeze was necesarry.. Oct 17 09:02:47 actually I know it will. Oct 17 09:03:00 rebelThor, you can, if you want. Oct 17 09:03:08 though it's BETA Oct 17 09:03:16 actually less than BETA :) Oct 17 09:03:29 http://build.shr-project.org/tests/mrmoku/unstable/images/om-gta02/ Oct 17 09:03:41 if you feel adventures Oct 17 09:03:51 TAsn: i sure do :D thanks! Oct 17 09:04:03 btw Oct 17 09:04:08 after you install it Oct 17 09:04:12 I suggest you'll do a Oct 17 09:04:23 opkg remove notifier ; opkg install shr-today ; opkg upgrade Oct 17 09:04:57 TAsn: ok, thanks, I'll do that. The trac still works as a bugtracker for this version, right ? Oct 17 09:05:08 yes. Oct 17 09:05:12 though please know this: Oct 17 09:05:23 people already posted some bugs about this version Oct 17 09:05:27 full image build :) Oct 17 09:05:30 so please don't forget to search Oct 17 09:05:37 and we know about missing features Oct 17 09:05:46 like names not resolving everywhere Oct 17 09:05:51 so there's no point reporting that Oct 17 09:05:54 as we are working on it. Oct 17 09:06:04 yep, I'll be kind when reporting anything :D Oct 17 09:06:27 TAsn: the master baker of the kingdom boke the best cake of the world ever to feed the people. Alas when the cake finally got finished and actually was the best ever seen, all of people had died from hunger meanwhile and nobody left to appreciate the fabulous cake :-S Oct 17 09:07:22 DocScrutinizer-8, that's why we also give people cookies sometimes Oct 17 09:07:26 and like what I did now Oct 17 09:07:30 with rebelThor Oct 17 09:07:37 we also give people to taste straight from the mold Oct 17 09:07:38 :) Oct 17 09:08:42 Slyon, Oct 17 09:08:44 File "/usr/bin/shr-today", line 137 Oct 17 09:08:44 elif (name == "RESUME" or name = "resume") and self.SMOOTH_SUSPEND == "True" and self.make_proxies(): Oct 17 09:08:44 ^ Oct 17 09:08:44 SyntaxError: invalid syntax Oct 17 09:08:49 noob. ;) Oct 17 09:09:02 TAsn: rebuild, it's already fixed ;) Oct 17 09:09:08 lol. Oct 17 09:09:18 i just got the same error :P Oct 17 09:09:33 that's ok, just fixed it here :) Oct 17 09:10:00 TAsn: okay, you are on tests/mrmoku aren't you? Oct 17 09:10:05 I am. Oct 17 09:10:09 though I don't have build rights. Oct 17 09:10:14 I built it for myself. Oct 17 09:10:34 oh right I can make it pop with suspend/resume. Oct 17 09:10:48 TAsn: no, just you image on the FR. could you test if the smooth_suspend thing works? Oct 17 09:11:00 what? Oct 17 09:11:04 what do you want me to do? Oct 17 09:11:37 just test suspend and resume with latest shr-today if display shuts off and back on correctly Oct 17 09:12:45 awesome Oct 17 09:12:48 but Oct 17 09:13:00 the logo looks weird Oct 17 09:13:05 :) Oct 17 09:13:09 but everything works great. :) Oct 17 09:13:12 looks really slick. Oct 17 09:13:26 okay :) Oct 17 09:13:34 and dimming the screen on suspend Oct 17 09:13:37 is an awesome idea Oct 17 09:13:38 :) Oct 17 09:13:51 maybe someone else should do artwork, which i can use afterwards :) Oct 17 09:14:12 Slyon, that's probably what will happen :) Oct 17 09:14:16 but that's fine. :) Oct 17 09:14:36 anyhow, to tell you the truth, we really want to incorporate shr-today into libphone-ui (our phone stack) Oct 17 09:15:30 yes, that'd be great. but therefore it has to be in C, hasn't it? Oct 17 09:16:20 kinda. Oct 17 09:16:31 though in the first step it doesn't have to Oct 17 09:17:30 we can use a module Oct 17 09:17:33 hmm, will see if i can dive into c a little bit, as it isnt much logic in the code.. just the handling of the dbus signals and the config Oct 17 09:17:33 that just runs apps Oct 17 09:17:43 Slyon, that'll be best Oct 17 09:17:45 just look in Oct 17 09:17:48 libphone-ui Oct 17 09:17:50 libphone-ui-shr Oct 17 09:17:54 (the second) Oct 17 09:18:01 there are elementary examples there Oct 17 09:18:04 and as for dbus Oct 17 09:18:08 you don't need to do anything Oct 17 09:18:19 as we'll listen to signals for you Oct 17 09:18:23 just make a Oct 17 09:18:33 shr_today_suspend Oct 17 09:18:42 actually Oct 17 09:18:44 we only need Oct 17 09:18:47 shr_today_show Oct 17 09:18:48 i just hope it uses less memory then the version now? Oct 17 09:18:48 and Oct 17 09:18:54 shr_today_hide Oct 17 09:19:02 zoff99, referring to what? Oct 17 09:19:22 Slyon, those are tho only two functions we need you to implement Oct 17 09:19:23 the version of shr-today i used, was using 15% memory all the time Oct 17 09:19:30 even when not visible Oct 17 09:19:36 thats why i didnt use it Oct 17 09:19:43 zoff99, visible or not visible makes no different Oct 17 09:19:47 and yes it'll use less. Oct 17 09:19:49 a lot less. Oct 17 09:19:58 ok hope so Oct 17 09:20:02 15% was too much :) Oct 17 09:20:04 TAsn: okay.. so i have to implement these functions in libphone-ui-shr, and what about the other dbus-signals shr-today listens to? Oct 17 09:20:16 Slyon, don't mind them Oct 17 09:20:25 as libphone-ui-shr doesn't listen to signals Oct 17 09:20:31 we'll add those to ophonekitd Oct 17 09:20:35 that'll call shr_today Oct 17 09:20:38 if i may suggest, try to use locale as much as possible Oct 17 09:20:47 especially for the DATE string Oct 17 09:20:56 Slyon, so from you we only need shr_today_show/hide/init/destroy ok? Oct 17 09:21:12 zoff99, Slyon is the chief architect. Oct 17 09:21:34 TAsn: okay, will see what i can do. but for now i'm off having some breakfast :) Oct 17 09:21:43 cheers :) Oct 17 09:22:06 zoff99: btw, if i rewrite in C i will use locale ;) Oct 17 09:22:29 cool Oct 17 09:22:49 those are some of the little things that bug me most, as a user Oct 17 09:22:57 not having dates in my locale Oct 17 09:25:30 zoff99, :) Oct 17 09:25:38 btw Slyon|away Oct 17 09:26:30 Slyon|away, we called shr_today idle_screen in libphone-ui-shr (actually not in shr yet, just in the underlying framework) so please conform to that name. Oct 17 09:33:07 Slyon, TAsn: rebuilding now Oct 17 09:35:54 mrmoku: please merge oe.dev for new binutils Oct 17 09:36:13 hi, is there any way i can manually download the packages used by opkg update and untar them into some folder? Oct 17 09:36:17 JaMa: ok, will do in a moment Oct 17 09:37:15 sorry, meaning download them myself instead of having opkg download them for me Oct 17 09:37:39 and then gunzip them to whichever location opkg checks for installing packages Oct 17 09:38:25 JaMa: done Oct 17 09:39:48 Slyon: huh? NOTE: /etc/X11/Xsession.d/slider.edj Oct 17 09:43:40 where does opkg update unzip the package archives to after it finishes downloading them? Oct 17 09:45:32 Slyon: huh? NOTE: /etc/X11/Xsession.d/slider.edj Oct 17 09:45:37 lol Oct 17 09:57:51 SHR: 03tom 07libphone-ui * r84596d04b2f8 10/src/ (phoneui.c phoneui.h.in): strated loading modules from the correct modules path Oct 17 09:58:01 SHR: 03tom 07libphone-ui * r6fc5de97ce85 10/src/ (phoneui-utility.h phoneui.c phoneui.h.in): fixed a couple of forgotten renames Oct 17 10:03:06 SHR: 03tom 07libphone-ui-shr * r22fe2bc7cfef 10/src/Makefile.am: it now installs to the correct location, /phoneui/modules Oct 17 10:03:16 mrmoku|away, got anything open in libphone-ui-shr or can I mass rename it? Oct 17 10:03:34 btw, as you can see Oct 17 10:03:39 TAsn: nothing open Oct 17 10:03:45 I changed libphone-ui and libphone-ui-shr Oct 17 10:03:52 so they'll install and load modules from the correct path. Oct 17 10:03:58 good Oct 17 10:04:02 mrmoku: (/etc/X11/Xsession.d/slider.edj) wtf did i do. fixed that Oct 17 10:04:07 :) Oct 17 10:04:20 Slyon|away: can rebuild then? Oct 17 10:04:26 yes Oct 17 10:04:29 ok, thanks Oct 17 10:04:41 hm.. why do we ship a pc file with efl2? Oct 17 10:04:44 I'm dropping that.- Oct 17 10:05:46 mrmoku: ah shit, stop there is a typo in the setup Oct 17 10:06:35 mrmoku: now its ok Oct 17 10:06:51 Slyon: well.. parsing bitbake files is slow enough :P Oct 17 10:06:57 ok^^ Oct 17 10:11:05 SHR: 03tom 07libphone-ui * r0ee999741c35 10/libphoneui.pc.in: fixed a couple of missing renames in libphoneui.pc.in Oct 17 10:16:15 Slyon, TAsn: now you can try if it works :) Oct 17 10:16:25 already built it locally :) Oct 17 10:16:47 mrmoku: i'll opkg install it :) Oct 17 10:23:18 phew. Oct 17 10:23:21 that was heavy :) Oct 17 10:23:24 SHR: 03tom 07libphone-ui-shr * re8f768036680 10/ (58 files in 6 dirs): did a mass libframeworkd-phonegui to phoneui rename and fixed a couple of stuff in the makefiles and such Oct 17 10:23:36 hehe :) Oct 17 10:23:38 58 files. Oct 17 10:23:38 :) Oct 17 10:23:46 I gotta say, I just love sed and bash. Oct 17 10:23:58 took me < 2 minutes. Oct 17 10:27:25 mrmoku, Oct 17 10:27:26 phoneui Oct 17 10:27:29 phoneuid Oct 17 10:27:32 I'm doing the split Oct 17 10:27:39 all the names in the data dir are correct, right? Oct 17 10:27:54 I can remove the ophonekitd stuff completely from the data dir Oct 17 10:27:55 right? Oct 17 10:28:00 btw, I hope you are not dirty ;) Oct 17 10:28:05 as I pretty much raped this Oct 17 10:28:12 Heinervdm: what do you use for checksums sorting? looks like script in ../dev/contrib/source-checker/oe-checksums-sorter.py sorts it just alphabetically and someone is sorting it based on version numbers too Oct 17 10:28:13 though actually, patching should just work since I did mv Oct 17 10:28:26 morning Oct 17 10:28:52 JaMa: i'm using that script, i just put the checksums at the and and then let it sort Oct 17 10:31:12 ok Oct 17 10:32:16 TAsn: NOOOO Oct 17 10:32:57 what? Oct 17 10:33:01 TAsn: please... DON'T TOUCH phonefsod ;) Oct 17 10:33:09 and I already did the splitting Oct 17 10:33:22 I meant renaming... Oct 17 10:33:29 using phoneui Oct 17 10:33:32 instead of phonegui Oct 17 10:33:34 and such. Oct 17 10:33:40 in libphone-ui though Oct 17 10:33:50 [12:28] TAsn | I can remove the ophonekitd stuff completely from the data dir Oct 17 10:33:53 scared me :P Oct 17 10:34:07 no Oct 17 10:34:14 I meant using that in phoneuid Oct 17 10:34:27 i.e changing every call from framewokrd-phonegui Oct 17 10:34:28 to phoneui Oct 17 10:34:46 mrmoku, please commit what you have Oct 17 10:34:53 so I'll be able to massively change it ;) Oct 17 10:34:57 (rename that is) Oct 17 10:35:01 TAsn: ok... phoneuid I'm clean Oct 17 10:35:07 it's phonefsod I'm working on right now Oct 17 10:35:11 ok. Oct 17 10:38:10 SHR: 03tom 07phoneuid * rd07aadd53cea 10/ (7 files in 2 dirs): using phoneui isntead of phonegui Oct 17 10:38:11 mrmoku, do we have recipes for all of the new software? Oct 17 10:39:35 bah Slyon shr-today looks beautiful, but then every time I unlock I get to see ugly niebiee :( not fair! :) Oct 17 10:39:52 TAsn: xD Oct 17 10:40:01 please disable unlocking or something :) Oct 17 10:40:06 I can live with only shr-today :) Oct 17 10:40:12 the theme is REALLY awesome. Oct 17 10:41:19 TAsn: doing the recipes soon now Oct 17 10:41:27 cool enough. Oct 17 10:41:30 TAsn: thanks :) Oct 17 10:42:01 I'm anxious to see if I broke anything with the renaming (and moving stuff and modules dir) or not. Oct 17 10:46:24 mrmoku, ping me when it's done. ;) Oct 17 10:46:28 Slyon, no, thank you. Oct 17 10:53:23 TAsn: have to eat first... as family is waiting :( Oct 17 10:53:35 and I want at least test if parsing works :P Oct 17 10:53:41 bbiab Oct 17 10:55:26 hehe ok ;) Oct 17 11:09:27 SHR: 03tom 07libphone-ui * r07847ec6a68a 10/src/phoneui-utility.c: fixed the broken include I accidentally broke with sed Oct 17 11:09:35 SHR: 03tom 07libphone-ui * r2f351abaf0e3 10/src/phoneui-utility.c: adedd a missing return Oct 17 11:19:26 TAsn: in the end... I did it without testing ;) Oct 17 11:19:36 pleas update and shout when it fails :P Oct 17 11:19:48 recipes? Oct 17 11:19:56 yup Oct 17 11:20:04 cool Oct 17 11:20:05 btw Oct 17 11:20:07 shr-devel Oct 17 11:20:17 please apply my patch. Oct 17 11:22:13 testing ui-shr Oct 17 11:25:51 SHR: 03tom 07libphone-ui * ra0e8ece3dae5 10/ (libphone-ui.pc.in libphoneui.pc.in): renamed the .pc.in file Oct 17 11:26:52 TAsn: just got back from lunch, trying the /tests/mrmoku/unstable image, but opgk remove notified complains about dependencies (task-shr-minimal-apps) and it's unable to find shr-today.. am I missing something, or do I need to modify the feeds to point to http://build.shr-project.org/tests/mrmoku/unstable/feed/ ? (now it seems it's pointing to http://build.shr-project.org/shr-unstable/) Oct 17 11:27:34 SHR: 03tom 07libphone-ui * r4131f9767d3c 10/Makefile.am: also renamed in makefile.am Oct 17 11:27:37 rebelThor, force remove. Oct 17 11:27:46 rebelThor, yeah Oct 17 11:27:52 you need to modify the feeds. Oct 17 11:27:59 cool, thanks! Oct 17 11:28:53 SHR: 03tom 07libphone-ui * r3a8b183d53a4 10/configure.ac: I feel like an ass, also renamed in configure.ac Oct 17 11:29:10 np. Oct 17 11:37:46 SHR: 03tom 07libphone-ui-shr * r3d7dfe86befc 10/src/Makefile.am: changed it to be a module and not a lib Oct 17 11:38:52 moin Oct 17 11:44:22 any way of retrieving the original elm- and phonegui-edj files from somewhere? Oct 17 11:45:21 the neo-theme fails to restore them on uninstall Oct 17 11:49:24 Tanuva: try here: http://git.shr-project.org/git/?p=shr-themes.git;a=tree Oct 17 11:50:37 Tanuva: or just -force-reinstall the original one Oct 17 11:51:10 mrmoku: do you know the package's name by heart? ;) Oct 17 11:51:44 by heart.. no. opkg list | grep theme is your friend :) Oct 17 11:52:06 oh. of course. dumb me. Oct 17 11:54:01 I always thought selection of theme was a matter of shr-settings, and original theme doesn't vanish on installing any fancy new theme o.O Oct 17 11:55:36 the neo theme somehow moves edj files round there, it's voodoo to me. and messes up font color (white instead of black) in apps using the original theme. Oct 17 11:57:13 SHR: 03tom 07libphone-ui * ra2a30c33077a 10/ (configure.ac libphone-ui.pc.in src/Makefile.am): fixed the .pc file library now works correctly Oct 17 11:58:54 SHR: 03tom 07libphone-ui-shr * r3d0ed2bde22f 10/po/ (POTFILES.in ar.po ca.po de.po es.po fi.po pl.po ru.po): renamed all the po files as well Oct 17 12:04:02 heyho Oct 17 12:05:17 JaMa: does one have to rebuild from scratch due to the new binutils? Oct 17 12:06:32 morphis: hey, still having fun with palm pre? Oct 17 12:06:54 mrmoku, hey Oct 17 12:06:56 I have a little issue Oct 17 12:06:59 mrmoku: jep, I am currently resizing my internal nand to get my further work starting Oct 17 12:07:02 everything works though not all the files ship Oct 17 12:07:05 mind looking into it? Oct 17 12:07:10 libphone-ui-shr Oct 17 12:07:12 * morphis was on holiday the last days Oct 17 12:07:22 NOTE: the following files were installed but not shipped in any package: Oct 17 12:07:22 NOTE: /usr/lib/phoneui/modules/shr.so Oct 17 12:07:23 .. Oct 17 12:07:23 .. Oct 17 12:07:23 .. Oct 17 12:07:26 * mrmoku hopes morphis had fun :) Oct 17 12:07:27 mrmoku: btw. you have anything done for the shr-palm-pre image? Oct 17 12:07:32 for sure :) Oct 17 12:07:35 morphis: somehow yes :) Oct 17 12:07:44 ok Oct 17 12:07:45 we started a new branch shr/merge based upon org.oe.dev Oct 17 12:07:54 and that includes stefans palm pre stuff Oct 17 12:08:13 ok, so I can simply switch my shr-unstable-image to the shr/merge branch? Oct 17 12:08:15 we're almost there to build from that branch instead of shr/import Oct 17 12:08:21 mrmoku, ? :) Oct 17 12:08:24 morphis: yup Oct 17 12:08:47 TAsn: just add /usr/lib/phoneui/modules to the package :) Oct 17 12:08:53 TAsn: moment Oct 17 12:08:56 where from? Oct 17 12:09:03 btw shr-devel commit my patch please. Oct 17 12:09:45 TAsn: done long ago Oct 17 12:09:52 ok, thanks :) Oct 17 12:09:56 anyhow, it's cool :) Oct 17 12:09:59 everything compiles Oct 17 12:10:10 and ui-shr is now a module. Oct 17 12:10:16 I'm now changing the default config. Oct 17 12:10:43 mrmoku: ok, than I will start to build me the first shr image for the pre Oct 17 12:11:00 TAsn: update and retry Oct 17 12:11:21 morphis: report when it explodes :P Oct 17 12:11:41 for sure Oct 17 12:12:44 TAsn: for pkgconfig... it is libphoneui, libphone-ui, phone-ui or phoneui? :P Oct 17 12:13:08 <[Rui]> should be consistent, if it isn't :) Oct 17 12:13:15 <[Rui]> having lunch. bbye! :) Oct 17 12:13:20 phone-ui Oct 17 12:13:25 ok Oct 17 12:13:28 i think Oct 17 12:13:34 check what I did in Oct 17 12:13:37 libphone-ui-shr Oct 17 12:13:37 :) Oct 17 12:13:42 (in the configure.ac) Oct 17 12:14:20 uhh... moment... I don't need that ;) Oct 17 12:14:20 SHR: 03tom 07libphone-ui * r894c4d0c1365 10/ (data/libphoneui.conf src/phoneui.c): now the config gets just the module name instead of the full library path Oct 17 12:14:43 yes you do Oct 17 12:14:45 in phoneuid Oct 17 12:14:58 not in phonefsod though :) Oct 17 12:15:00 but not in phonefsod where I'm working now ;) Oct 17 12:15:32 ok Oct 17 12:15:33 it's Oct 17 12:15:35 libphone-ui Oct 17 12:15:38 I'll add it to phoneui Oct 17 12:15:39 i Oct 17 12:15:41 phoneuid Oct 17 12:17:14 ok Oct 17 12:17:18 btw, mrmoku Oct 17 12:17:19 SHR: 03tom 07libphone-ui * r87be523d208d 10/src/phoneui.c: fixed a typo replaced a ; with a + Oct 17 12:17:21 in the new config file Oct 17 12:17:24 it's Oct 17 12:17:26 [dialer] Oct 17 12:17:28 module = shr Oct 17 12:17:29 :) Oct 17 12:17:35 instead of something ugly like it was. Oct 17 12:20:13 oh, oops rebuilt without waiting for your fix :) Oct 17 12:20:19 (for the inclusion) Oct 17 12:20:57 TAsn: nice (the config) Oct 17 12:21:05 I'm glad you like it. Oct 17 12:23:50 mrmoku, does phoneuid compile? Oct 17 12:23:51 does it work? Oct 17 12:24:04 * TAsn is trying. Oct 17 12:24:36 no the recipe looks broken. Oct 17 12:24:37 :| Oct 17 12:24:49 well, after you fix that i'll give it ago Oct 17 12:24:52 in the meanwhile Oct 17 12:24:55 food. Oct 17 12:39:50 hi PaulFertser I've updated my patch....there is an issue here how it looks: http://pastebin.com/m4818d564 Oct 17 12:40:41 is the AT commands sent without their result ok? Oct 17 12:41:53 basically I did kcalloc,copied the first buffer in the second one and printed it() Oct 17 12:42:15 (kcalloc takes cares of \0) Oct 17 12:42:37 but as I don't know the len of the second buffer called ptr I don't know how to pretty-print it Oct 17 12:42:53 mrmoku, took a look at phoneuid's recipe? Oct 17 12:43:03 because I wouldn't know the size of the kcalloc to do Oct 17 12:43:07 TAsn: not yet... trying to get phonefsod to compile Oct 17 12:43:18 hehe Oct 17 12:43:21 what are the issue? Oct 17 12:43:24 s Oct 17 12:43:38 renaming cleanup ;) Oct 17 12:43:42 PaulFertser, ah you seem not there Oct 17 12:43:44 sorry then Oct 17 12:43:48 mrmoku, oh ;) Oct 17 12:43:59 Gnutoo: strlen(str) is your friend Oct 17 12:44:18 DocScrutinizer-8, but if the buffer is not a string and that I'm in the kernel? Oct 17 12:44:30 DocScrutinizer-8, the string is not terminated by \0 Oct 17 12:44:48 I meant the second buffer Oct 17 12:45:14 DocScrutinizer-8, basically I do: Oct 17 12:45:33 second_buffer = kcalloc() Oct 17 12:45:51 but I don't know the size of the first buffer so I can't do kcalloc Oct 17 12:46:10 I'll post the code Oct 17 12:46:38 oh, if you don't know size of first buffer, and you don't have terminating \0 then you're out of luck Oct 17 12:46:45 mrmoku, my list of requests: phoneuid's recipe and correct inclusion of files in libphone-ui-shr :) Just that you won't forget. Oct 17 12:46:55 http://pastebin.com/m53c931ca Oct 17 12:47:08 I could do a custom len Oct 17 12:47:13 maybe Oct 17 12:47:28 maybe \n is the end of string Oct 17 12:47:50 TAsn: uhh... for phoneuid it's not the recipe... it's the missing files in data Oct 17 12:47:53 at least it is \r \n in the one I can print Oct 17 12:48:03 looks I lost them while splitting :( Oct 17 12:48:11 so the result AT command maybe has \r \n too? Oct 17 12:48:17 * mrmoku goes to look for his phone to ressurect them ;) Oct 17 12:48:32 mrmoku, hehe, ok:) here it doesn't compile anything Oct 17 12:48:39 as if the working dir of the recipe is wrong. Oct 17 12:49:00 the part is in smd_tty_notify Oct 17 12:49:15 anyhow, the list I gave you consists of what I couldn't solve and need in order to further test everything :) Oct 17 12:49:31 (the part where I have a problem) Oct 17 12:49:37 so maybe I do a for until \n Oct 17 12:49:40 and count Oct 17 12:49:47 and then calloc Oct 17 12:49:52 I'll try Oct 17 12:51:19 btw is there a kernel mode for emacs(because of space/tab thing) Oct 17 12:53:06 or I've got a bad idea Oct 17 12:53:15 ptr[len - 1 ] = Oct 17 12:53:22 '\n' Oct 17 12:54:10 Gnutoo: I find suspicious code in smd_tty_write() Oct 17 12:54:16 mrmoku: you probably should.. but I haven't preferred them now.. wanted to test it before pushing.. but now I want to flash full-image first and start using only shr/merge :) Oct 17 12:54:40 mmm Oct 17 12:54:51 mrmoku: if you rebuild shr-kms then there should be KMS in image name Oct 17 12:55:03 DocScrutinizer, which one? Oct 17 12:55:09 DocScrutinizer, the one I wrote? Oct 17 12:55:15 DocScrutinizer,do you want a diff Oct 17 12:56:22 TAsn: you did not update... already solved the ui-shr issue Oct 17 12:56:28 hmm difference between lite and full image is just 22MB Oct 17 12:56:39 mrmoku, I didn't. Oct 17 12:56:40 sec ;) Oct 17 12:56:43 Gnutoo: I think PaulFertser can help better Oct 17 12:56:45 maybe I'll do my second idea: print %s\n and replace the \n by \0 Oct 17 12:56:49 ok Oct 17 12:57:06 TAsn: and pull phoneuid Oct 17 12:57:08 SHR: 03mok 07phoneuid * r53c6b50b9cd1 10/data/ (6 files): add dbus configuration for phoneuid Oct 17 12:57:10 I added the missing files Oct 17 12:57:18 for me it builds now :) Oct 17 12:57:31 ah bad idea Oct 17 12:57:41 would need to copy the buffer anyway Oct 17 12:57:51 ah no Oct 17 12:57:59 if I re-put \n at the end... Oct 17 12:58:04 I'll do that Oct 17 12:58:15 Gnutoo: you still need the terminating \0 in last byte of buffer you hand over to print(%s), otherwise the print() doesn't know length of buffer and will print "infinitely" Oct 17 12:58:21 okie Oct 17 12:58:23 compiling Oct 17 12:58:27 mrmoku, thanks. Oct 17 12:58:31 btw, does phoneuid work? Oct 17 12:58:51 what do I need to do to test it? (what dbus signals) Oct 17 12:58:53 TAsn: installing it now Oct 17 12:59:04 I'm not even sure Oct 17 12:59:09 libphone-ui works properly :) Oct 17 12:59:11 TAsn: wait a second and I will commit phonefsod stuff Oct 17 12:59:26 mrmoku, what do we haev and what's missing? Oct 17 12:59:46 I talk faster than I think Oct 17 12:59:52 I would need the len anyway Oct 17 12:59:59 TAsn: hmm... shr-wizard is missing :P Oct 17 13:00:05 Gnutoo: I don't see the magic line "printbuf[i]=0x00, in line 150 Oct 17 13:00:10 mrmoku, I haven't wrote it yet Oct 17 13:00:15 but yeah, I'd appreciate a recipe ;) Oct 17 13:00:39 DocScrutinizer-8, indeed that's because of calloc Oct 17 13:01:10 ok, calloc might initialize the buffer with 0x00 Oct 17 13:01:35 NOTE: the following files were installed but not shipped in any package: Oct 17 13:01:37 mrmoku, ^ :( Oct 17 13:01:42 but you need to calloc len+1 anyway then Oct 17 13:01:42 SHR: 03mok 07phonefsod * rf64fa861b4af 10/po/Makefile.in.in: drop po subdir as we won't need that for the daemon ;) Oct 17 13:01:43 SHR: 03mok 07phonefsod * r686db398c8f5 10/ (27 files in 3 dirs): ophonekitd to phonefsod rename and main cleanups Oct 17 13:02:00 TAsn: which ones? Oct 17 13:02:07 same files Oct 17 13:02:08 ... Oct 17 13:02:11 all the modules Oct 17 13:02:13 indeed even the len - 1 Oct 17 13:02:17 hmm Oct 17 13:02:27 did you check for you? Oct 17 13:02:28 I'll check Oct 17 13:02:48 that's what I did Oct 17 13:02:49 printbuf = kcalloc( (len + 1) , Oct 17 13:03:20 oh, yes. sorry Oct 17 13:04:14 np Oct 17 13:04:20 so strlen(printbuf) should tell you actual charcount Oct 17 13:04:44 but also i will tell you Oct 17 13:04:55 and len will tell you Oct 17 13:05:01 mrmoku, Oct 17 13:05:04 you did Oct 17 13:05:05 TAsn: uhh.. yes I know Oct 17 13:05:05 FILES Oct 17 13:05:07 instead of Oct 17 13:05:08 * mrmoku stupid Oct 17 13:05:10 yep Oct 17 13:05:13 :( Oct 17 13:05:14 :) Oct 17 13:05:15 FILES_${PN} Oct 17 13:06:00 DocScrutinizer-8, indeed unless the string is not terminated by \0 Oct 17 13:06:11 NOTE: make DESTDIR=/home/tom/projects/openmoko/oe/shr-unstable/tmp/work/armv4t-angstrom-linux-gnueabi/phoneuid-0.0.0+gitr130+53c6b50b9cd1b89daf303bc53bfb6e41bc9a9a95-r0/image install Oct 17 13:06:11 make: *** No rule to make target `install'. Stop. Oct 17 13:06:11 FATAL: oe_runmake failed Oct 17 13:06:12 mrmoku, ^ Oct 17 13:07:31 Gnutoo: you told the stringbuffer printbuf gets the \0 via calloc, and I agreed it possibly does. So where's the problem now? Oct 17 13:07:50 DocScrutinizer-8, I'll explain: Oct 17 13:07:52 mrmoku, phoneuid still doesn't build Oct 17 13:08:02 but shr looks good :) Oct 17 13:08:03 there are 2 functions : the one that works is: Oct 17 13:08:25 smd_tty_write Oct 17 13:08:32 the one who has problem is: Oct 17 13:08:46 smd_tty_notify Oct 17 13:09:01 I'll fix tty_notify but that will be *very* ugly and dangerous Oct 17 13:09:23 Gnutoo: I don't see a single print() in tty_notify Oct 17 13:09:55 because I will copy the content of the ptr I get in a new buffer that will be calloced Oct 17 13:10:08 but in order to be calloced I need the len of the first buffer Oct 17 13:10:12 that is the ptr Oct 17 13:10:16 so I'll do : Oct 17 13:10:18 TAsn: ok, update again Oct 17 13:10:19 well, actually one with static text Oct 17 13:10:34 TAsn: I'm off for an hour or two now... brother came for a coffee :P Oct 17 13:10:37 bbl Oct 17 13:10:43 while (buf[i] != '\n'){i++} // very dangerous Oct 17 13:10:44 mrmoku|away, ciao Oct 17 13:10:45 have fun. Oct 17 13:12:30 mrmoku|away, same old same old. Oct 17 13:12:31 :| Oct 17 13:12:36 mrmoku|away, will have to wait for you. Oct 17 13:14:28 Gnutoo: anway, I have to leave now. Maybe Paul can help Oct 17 13:14:38 ok thanks a lot Oct 17 13:14:54 anyway my patch in smd_tty_notify don't work Oct 17 13:27:30 mrmoku|away: did you apply this patch? http://patchwork.dev.bearstech.com/patch/320/ its accepted in patchwork but haven't seen it in shr/merge :) Oct 17 13:37:33 mmm Oct 17 13:37:48 what non visible characters a response should have? Oct 17 13:37:53 for instance OK Oct 17 13:37:59 would be : Oct 17 13:38:28 'OK' or 'OK\n' or 'OK\r\n' etc.... Oct 17 13:48:08 mrmoku|away: I've resent that one as now it won't apply Oct 17 13:52:20 Slyon, I think the "incoming call" field should be bigger, though it still looks awesome! :) Oct 17 13:52:28 anyhow, I'm off Oct 17 13:52:29 ciao. Oct 17 13:52:32 bye Oct 17 13:52:55 Gnutoo: I don't think *kernel* should have *any* assumptions about that Oct 17 13:53:17 DocScrutinizer-8, yes but without the len...what could I do? Oct 17 13:53:31 I'll look if it finishes by \0 Oct 17 13:53:38 but if not I'm lost Oct 17 13:53:44 s/lost/out of luck Oct 17 13:53:53 if you don't have len, you're out of luck and did something wrong Oct 17 13:55:05 I didn't write the tty driver...but then I'm out of luck just the same Oct 17 13:55:47 as kernel itself is agnostic about actual content of the string and thus *needs* to know about len all times, otherwise the whole string is unusable - even for kernel Oct 17 13:56:39 so if you don't know len, you simply lost the info somewhere on your way Oct 17 13:56:49 but it has to be there Oct 17 13:59:27 still my comment I don't see any problematic print() or such in smd_tty_init() holds true Oct 17 13:59:56 and maybe that's just the wrong place to do whatever you want to do anyway Oct 17 14:01:58 I found some commented out print and kcalloc lines in notify (s/init/notify/ btw above). These lines don't look like they might compile Oct 17 14:02:38 e.g. len not defined I guess Oct 17 14:02:54 what seems to be related to your problem Oct 17 14:05:55 but as I never had a close look to the driver and its way to operate, as I didn't to your goals for that patch you try to do there, I really guess it's better you talk to PaulFertser about all that. As I got it he knows what all that is about... I don't Oct 17 14:12:55 I think I got the len Oct 17 14:13:23 avail = tty_prepare_flip_string(tty, &ptr, avail); seem to be the len Oct 17 14:13:27 according to: Oct 17 14:13:37 int smd_read(smd_channel_t *ch, void *data, int len); Oct 17 14:13:40 I'll try Oct 17 14:14:49 i have a sort of a newbie question over here: what happens if you don't extract the modules file and just flash the uImage and the main OS image? :D Oct 17 14:21:08 works now... Oct 17 14:21:20 DocScrutinizer, thanks a lot for your help Oct 17 14:21:55 it's somewhat verbose but it works Oct 17 14:22:01 thanks a lot Oct 17 14:26:57 howdy all! Oct 17 14:30:27 Gnutoo: yw (btw I suspected the avail thing) Oct 17 14:30:35 ok Oct 17 14:33:06 what's new? Oct 17 14:39:04 Gnutoo: btw don't expect it for granted one cycle of the for(;;) loop is always processing exactly one "line" of text - as from one to next. It might be arbitrary parts of lines, probably even spreading multiple lines Oct 17 14:40:19 ok Oct 17 14:43:24 Gnutoo: this won't matter as long as you simply copy text to printk() simply, incl. all and . But any additional header/trailers like timestamp, "AT>" or "\n\r" might not fall into place Oct 17 14:43:51 ok Oct 17 14:44:10 anyway it works... Oct 17 14:44:26 but is a little bit verbose Oct 17 14:44:34 but much better than before Oct 17 14:45:13 Gnutoo: usually. As calypso is introducing a pause after each line, and delivers chars fast enough so all get caught in a single loop cycle Oct 17 14:45:40 DocScrutinizer-8, ok (btw I use qualcomm_msm not caalypso) Oct 17 14:46:48 TAsn: hmm... for me it works... the only unpackaged thing I have is the debug file Oct 17 14:47:02 if calypso is changing (or qualcom) this timing behaviour, then you get your headers on the places where kernel caught up with speed of tty, which isn't necessarily exactly at linebreaks Oct 17 14:47:13 ok Oct 17 14:48:33 all that only if the reading function is operating the way I assume Oct 17 14:49:10 Immediately after resume is one place where assuming that linebreaks are useful will fail. Oct 17 14:50:27 depending on the mode in which the GSM is operating (MUX, or other settings), it is also possible for message content to be line-oriented, and further confuse the parser. As many have discovered, a GSM parser is difficult to get right. Oct 17 14:50:28 :( Oct 17 14:50:54 mwester: yep Oct 17 14:51:01 And then add unsolicited messages to the mix, and the nightmare is complete. Oct 17 14:51:17 this is a mere syslog style trace of tty driver though Oct 17 14:51:37 AIUI Oct 17 14:52:43 anyway, mwester you might want to take over, as I'm away for real now Oct 17 14:53:30 Later! Oct 17 15:02:17 TAsn: are you sure the modules get installed by the makefile? Oct 17 15:03:16 mwester: wasn't that ttyspy patch yours? maybe you should do more advertising for it (e.g. mention on wiki somewhere) Oct 17 15:16:43 freesmartphone.org: 03mickey 07specs * rb7dae334e876 10/ (3 files in 3 dirs): org.freesmartphone.GSM.SIM.GetAuthStatus(): allow 'unknown' as value Oct 17 15:18:13 hey mwester ! long time no see Oct 17 15:18:21 mickeyl: hi! Oct 17 15:18:26 hi khiraly1 Oct 17 15:18:33 how is palm pre hacking going? Oct 17 15:18:51 I learned, they are selling since okt 14 and the source is released since okt 16 Oct 17 15:19:02 okt -> oct Oct 17 15:19:03 pretty good, my time is not yet there, but the other guys are making great progress Oct 17 15:19:26 mickeyl: any public website/mailing list about the progress? Oct 17 15:19:30 yeah, thanks to a friendly "reminder", palm released the kernel sources pretty early this time Oct 17 15:19:46 khiraly1: yep, fso wiki -> palm pre challenge, webos-internal wiki, webos-internal mailing list Oct 17 15:19:47 yeah, Harald did a good job Oct 17 15:28:52 freesmartphone.org: 03mickey 07libfso-glib * r36763b09f5c8 10/ (.gitignore src/freesmartphone-gsm.vala src/fso-glib-1.0.vapi): remove generated file and regenerate Oct 17 15:29:32 freesmartphone.org: 03mickey 07vala-dbus-binding-tool * r54ffd64f7bc0 10/ (configure.ac src/Makefile.am src/vala-dbus-binding-tool.vala): do not link against vala, but rather gee upstream Oct 17 15:34:22 TAsn: ping Oct 17 15:37:35 any shr/import maintainers present? Oct 17 15:38:09 freesmartphone.org: 03mickey 07cornucopia * reb38865c6839 10/fsogsmd/ (7 files in 3 dirs): fsogsmd: org.freesmartphone.GSM.SIM.GetAuthStatus() done Oct 17 15:45:59 SHR: 03mok 07libphone-ui * ra046899b4013 10/ (configure.ac src/phoneui.h.in): expand LIBDIR and add a slash Oct 17 15:50:38 SHR: 03mok 07libphone-ui-shr * r4d53f4273726 10/src/view/views.h.in: views.h.in: fix path to edj files Oct 17 16:01:14 dcordes: whats up Oct 17 16:09:52 mrmoku, I am building shr-image for non moko device Oct 17 16:10:04 yup Oct 17 16:10:19 mrmoku, ran into some build errors. but I figured I used angstrom distro although you have shr Oct 17 16:10:28 I suppose it's necessary to use the shr distro? Oct 17 16:11:03 hmm, no idea... but you might want to use shr/merge... which is closer to org.oe.dev and where we're heading towards to Oct 17 16:11:11 I never tried angstrom Oct 17 16:12:21 TAsn: something went wrong :( Oct 17 16:12:35 mrmoku, ok. and where's the downside of shr/merge? Oct 17 16:13:01 dcordes: good question... hopefully none Oct 17 16:13:21 we might still have to sort out one or the other issue... though shr-lite-image seems to build Oct 17 16:13:49 dcordes: and you would have to rebuild from scratch :P Oct 17 16:14:50 not the first rm -rf tmp Oct 17 16:17:33 ohh... yeah... Oct 17 16:17:44 (today) Oct 17 16:17:47 rm -rf / tmp Oct 17 16:17:48 * mrmoku did rm -rf tmp at least twice a day the last week :P Oct 17 16:17:59 DocScrutinizer-8: do that on your box ;) Oct 17 16:18:09 :-P Oct 17 16:18:27 is there an shr sample conf? also do you hardcode anything arch specific? Oct 17 16:18:36 s/conf/local.conf/ Oct 17 16:18:36 dcordes meant: is there an shr sample local.conf? also do you hardcode anything arch specific? Oct 17 16:20:18 dcordes: well, we have our Makefile... local.conf gets setup from there Oct 17 16:20:53 but I don't want a moko device local.conf Oct 17 16:21:16 does Makefile setup anything else critical? Oct 17 16:21:20 machine get's specified in conf/auto.conf Oct 17 16:21:57 dcordes: nothing critical that I would be aware of Oct 17 16:24:29 ok, one outbound call ->framework borked Oct 17 16:24:53 I'm sooooooo tired of that crap Oct 17 16:25:20 so if I write conf/local.conf and conf/auto.conf in "shr Makefile style" I'm set ? Oct 17 16:25:22 SHR: 03mok 07phonefsod * r0790c8c61382 10/ (7 files in 2 dirs): some missing renames and cleanups Oct 17 16:25:45 Gnutoo, I totally forgot about the patches Oct 17 16:25:48 Gnutoo, sorry Oct 17 16:26:04 dcordes: yeah, you should be Oct 17 16:26:14 hmm.. we have site.conf too Oct 17 16:26:18 dcordes, I thought there was a problem and weren't accepted... Oct 17 16:26:38 dcordes, will you have the time to modify+push them? Oct 17 16:26:52 Gnutoo, yes after I got my shr build going Oct 17 16:26:58 ok thanks a lot Oct 17 16:27:19 Gnutoo, did you use "shr Makefile" to build for htcdream? Oct 17 16:27:37 no Oct 17 16:27:47 shr/import in oe Oct 17 16:28:34 Gnutoo, can you show me the conf files? Oct 17 16:29:03 Gnutoo, I also want to build "standard oe way" and I'm confused by the configurations Makefile writes Oct 17 16:29:14 http://shr.bearstech.com/git/?p=shr-makefile.git;a=blob;f=Makefile Oct 17 16:30:49 dcordes, ok you have to make some includes Oct 17 16:30:54 Gnutoo, can you pastebin the files needed? Oct 17 16:31:15 I'll do it...give me the time to find the local.conf Oct 17 16:31:32 ok thanks Oct 17 16:31:44 I'll look up your patches in the meantime ;) Oct 17 16:33:13 dcordes, http://pastebin.com/m19870cd9 Oct 17 16:37:19 Gnutoo, thanks Oct 17 16:37:44 Gnutoo, can you give me the full local.conf maybe? Oct 17 16:38:33 ok Oct 17 16:39:03 dcordes, remove the mplayer thing: http://pastebin.com/m7078183c Oct 17 16:43:02 Gnutoo, what do you setup in require conf/distro/include/shr-htcdream.conf ? Oct 17 16:43:38 ok I'll paste it too Oct 17 16:43:53 ok Oct 17 16:44:23 dcordes, oops: http://pastebin.com/m559f17b Oct 17 16:44:38 PREFERRED_PROVIDER_xf86-video-glamo = "xf86-video-fbdev" shouhd be removed Oct 17 16:47:16 waht's mesa? Oct 17 16:47:24 did X work for you with that configuration ? Oct 17 16:47:28 yes Oct 17 16:47:30 I think so Oct 17 16:47:41 basically I took the openmoko config and changed things Oct 17 16:47:48 I didn't look at it carefully Oct 17 16:47:59 mesa is software opengl imlpementation Oct 17 16:48:45 I understood that there is dri/drm that does hardware 3d and when something is not avaliable it uses mesa that does software Oct 17 16:48:47 3d Oct 17 16:50:00 I see Oct 17 16:50:24 ok let's roll Oct 17 16:50:27 ok Oct 17 16:52:04 freesmartphone.org: 03mickey 07cornucopia * rdad00705d7e2 10/fsogsmd/src/lib/ (atcall.vala atmediators.vala atunsolicited.vala): fsogsmd: hook to +CRING to detect incoming calls Oct 17 17:01:05 Gnutoo: so now you see only commands going to the modem but no replies? Oct 17 17:03:01 Gnutoo: show me the latest version of your patch and please give me a pointer to the driver sources. Oct 17 17:05:28 PaulFertser, no I've both now Oct 17 17:06:49 PaulFertser, I'm just waiting for the person who has adp/us phone to show up Oct 17 17:07:07 Gnutoo: nice Oct 17 17:07:52 Gnutoo, pushed Oct 17 17:08:55 ok Oct 17 17:09:42 dcordes, thanks Oct 17 17:10:04 any problems with SMD on dream? Oct 17 17:16:05 dcordes, no...it's only that some people with US/adp phones couldn't make work the reference ril so I made a patch for sniffing the proprietary ril Oct 17 17:16:52 so after that some people here like PaulFertser could interpret it and look what AT command to type to get it work Oct 17 17:17:16 ok no clue about android ril. Oct 17 17:18:18 dcordes, no problem...I bet there are some commands for setting the good frequencies like on some telit modems Oct 17 17:19:12 Heinervdm: shr-full-image from shr/merge booted ok too.. Oct 17 17:19:14 freesmartphone.org: 03mickey 07cornucopia * r30c2ed218a34 10/fsogsmd/src/ (4 files in 2 dirs): fsogsmd: callhandling: activate single call Oct 17 17:19:36 Heinervdm: that means no more commits to shr/import from me probably :) Oct 17 17:19:47 JaMa: i build one from scratch again, and it's 100 MB again... Oct 17 17:20:17 is your git branch clean? Oct 17 17:20:38 yes Oct 17 17:21:15 mine has about 10 commits but its all about Angstrom or Sharp Spitz... Oct 17 17:23:04 and did you try to compare package sizes? if its just more packages or some are much bigger than mine? Oct 17 17:23:36 It looks the same Oct 17 17:25:22 dcordes, did you forget to credit the original author here: http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=22a5ffec6f8ad3ced15aaac952163431a352bb5b ? Oct 17 17:25:36 JaMa: good thing to hear :) Oct 17 17:25:56 going to switch the Makefile to build unstable from shr/merge soon then Oct 17 17:26:44 Gnutoo, no Oct 17 17:26:51 now its time to test 2.6.31 :) Oct 17 17:27:04 ah so the web interface shows only shortlog then Oct 17 17:27:08 Gnutoo, martin is not him and I couldn't figure the original one. Oct 17 17:27:14 ok Oct 17 17:27:18 Gnutoo, so now there are three authors of the fb hack ;) Oct 17 17:27:31 :( Oct 17 17:27:41 lol Oct 17 17:27:48 anyone tried 2.6.31 or know what is missing from larsc branches to successfull boot and testing? Oct 17 17:28:08 you could have said that you didn't find the original author...but we can't change that once it has been commite Oct 17 17:28:09 d Oct 17 17:28:13 how far are ya of an working image? ^^ Oct 17 17:28:31 Gnutoo, it's more important to get rid of that crap and have it working correctly Oct 17 17:28:41 Hardy: me? Oct 17 17:28:45 Gnutoo, raster had a look at it once and shared some ideas. Oct 17 17:28:55 Gnutoo, he also owns an msm device Oct 17 17:29:01 i mean everyone ^^ who is working on SHR :D Oct 17 17:29:10 have to go now, ciao Oct 17 17:29:15 cya ^^ Oct 17 17:29:39 <[Rui]> hi Oct 17 17:29:42 Hardy: everyone who is working on SHR has sort of working image i guess :) Oct 17 17:30:01 Hardy: its just not so easy for use for public audience.. Oct 17 17:30:19 <[Rui]> a friend complains a lot about echo or "robotic" voice when I use the speakerset instead of the handset. Oct 17 17:30:25 i just want one which is stable Oct 17 17:30:28 <[Rui]> so I'm trying to figure out http://wiki.openmoko.org/wiki/Neo_1973_and_Neo_FreeRunner_gsm_modem#AT.25Nxxxx Oct 17 17:31:00 <[Rui]> a doubt I have is: do those AT%N commands replace each other (ie, only one is active) or they add-up upon each other? Oct 17 17:31:11 dcordes, ok I'll prepare and push to oe Oct 17 17:31:41 Hardy: it works OK, if you know what to tweak and can fix things when something get broken with latest update.. but now things will slow down a bit.. Oct 17 17:32:05 Gnutoo, yea it will be good to use htcraphael.conf as example Oct 17 17:32:17 JaMa: perhaps i can.. :D Oct 17 17:32:19 <[Rui]> also, how to make changes I make as permanent? Oct 17 17:32:19 dcordes, I used an htc-something as example Oct 17 17:32:47 Ah and someone should look at opim issue in shr/merge.. disabled opim is problably not an option in public image Oct 17 17:33:41 Hardy: ok, if you can fix things, then you probably can find mrmoku new images on buildhost and start testing with us Oct 17 17:33:53 okay Oct 17 17:34:01 mrmoku: ping Oct 17 17:34:13 morphis: exploded? :P Oct 17 17:34:21 mrmoku: jep Oct 17 17:34:33 mrmoku: got some error with opimd-utils Oct 17 17:34:49 Hardy: ah sorry, there isn't image in mrmoku test directory :/ Oct 17 17:35:08 morphis: pastebin? Oct 17 17:35:11 Hardy: you can try mine if you really want to test it and you have fallback distro Oct 17 17:35:31 2nd WSOD while booting KMS kernel :( Oct 17 17:35:40 yeah, if u want to package it Oct 17 17:36:13 mrmoku: jepp, wait a minute Oct 17 17:36:27 Hardy: http://jama.homelinux.org/org.openembedded.dev.merge/images/om-gta02/ Oct 17 17:36:30 I lost the message, but I am currently recompiling to get it back Oct 17 17:36:37 morphis: ok Oct 17 17:37:25 JaMa: ahh, there they are :D Oct 17 17:38:05 oww, i have to blow the dust away of my moko ;) Oct 17 17:38:50 Hardy: but you were warned.. its wip and I haven't called from it sofar :) Oct 17 17:39:14 perhaps somebody can wasting his time for port the wxpython to the opkg Oct 17 17:39:17 http://virtoreal.net/Blog/?show=65&vtrusr=Hardy Oct 17 17:39:20 i made this for it xD Oct 17 17:39:35 JaMa: okay... i just testing it :D Oct 17 17:39:40 to help a bit Oct 17 17:44:13 * JaMa leaving for dinner, gnite! Oct 17 17:44:37 cya Oct 17 17:55:14 mrmoku: Missing or unbuildable dependency chain was: ['shr-image', 'task-shr-minimal', 'opimd-utils-cli'] Oct 17 17:55:39 mrmoku: ERROR: '['/home/morphis/build/shr/shr-unstable/openembedded/recipes/tasks/task-shr-minimal.bb']' RDEPENDS/RRECOMMENDS or otherwise requires the runtime entity 'opimd-utils-cli' but it wasn't found in any PACKAGE or RPROVIDES variables Oct 17 17:55:43 NOTE: Runtime target 'opimd-utils-cli' is unbuildable, removing... Oct 17 17:55:46 Missing or unbuildable dependency chain was: ['opimd-utils-cli'] Oct 17 17:55:48 ERROR: Required build target 'shr-image' has no buildable providers. Oct 17 17:58:50 hmmm... I remember to have had that... but I thought it is resolved Oct 17 17:59:11 :) Oct 17 17:59:12 morphis: just open recipes/tasks/task-shr-minimal.bb and remove opimd-utils-cli Oct 17 17:59:18 ok Oct 17 18:04:44 <[Rui]> is AT%N0187 == ti_calypso_dsp_mode = aec+nr in frameworkd.conf ? Oct 17 18:15:27 JesusMcCloud1: ahh, you here too :D Oct 17 18:18:48 SHR: 03mok 07shr-specs * rd72ca163c02e 10/ (76 files in 9 dirs): rename, update and regenerate Oct 17 18:29:08 [Rui]: yes Oct 17 18:45:43 <[Rui]> DocScrutinizer: thanks Oct 17 18:45:51 yw Oct 17 18:46:13 <[Rui]> DocScrutinizer: strangely enough, when I sent that AT through mickeyterm, my colleague on the other side said the sound improved Oct 17 18:47:09 well, not absolutely sure about relation of symbolic andhex values Oct 17 18:47:48 also not sure about when, how often etc fso issues this cmd Oct 17 18:48:20 <[Rui]> DocScrutinizer: if it's not per call then it's lost Oct 17 18:50:27 DocScrutinizer: btw, did you buy the n900? Oct 17 18:53:05 [Rui]: have you seen my link to extensive documentation about %Nxxx command? Oct 17 18:53:29 [Rui]: it's effective immediately and btw FSO does emit it before every call. Oct 17 18:53:40 <[Rui]> http://wiki.openmoko.org/wiki/Hardware:AT_Commands#Echo_Suppression_and_Noise_Reduction or http://wiki.openmoko.org/wiki/Neo_1973_and_Neo_FreeRunner_gsm_modem#AT.25Nxxxx ? Oct 17 18:54:10 [Rui]: no Oct 17 18:54:25 <[Rui]> PaulFertser: I was on a call with a friend as I tried fiddling with the AT%N commands listed in those pages (the same, actually) and the hand and speaker states. Oct 17 18:54:30 [Rui]: http://git.freesmartphone.org/?p=framework.git;a=blob;f=framework/subsystems/ogsmd/modems/ti_calypso/channel.py Oct 17 18:54:33 <[Rui]> PaulFertser: then I guess I didn't Oct 17 18:55:23 <[Rui]> PaulFertser: ok, that's a nice map, but no more commands than those on the wiki pages I listed Oct 17 18:57:23 [Rui]: it's more descriptive and gives possible values for frameworkd config. Oct 17 18:57:54 <[Rui]> PaulFertser: indeed. Oct 17 18:58:09 any reason why glibc_2.9.bb do_package would fail with FATAL: kernel too old ? Oct 17 18:58:31 <[Rui]> dinner time Oct 17 18:59:30 dcordes: still using 2.4 ? Oct 17 18:59:50 mrmoku, I didn't even build a kernel package Oct 17 19:00:04 on the host I run 2.6.30 Oct 17 19:00:10 dcordes: yes, but which kernel _would_ it build? for the target... Oct 17 19:00:20 linux-msm Oct 17 19:00:25 which is? Oct 17 19:01:01 s/linux-msm/linux-msm7xxxx Oct 17 19:01:03 / Oct 17 19:01:17 2.6.27 Oct 17 19:01:17 hrhrrr O2-germany, on their palm pre tv advert, reduced duration they show the kbd to ~0.1sec. Oct 17 19:01:25 * DocScrutinizer-8 wonders why ;-) Oct 17 19:01:35 <[Rui]> idling Oct 17 19:02:00 dcordes: hmm... strange that definitely should be new enough Oct 17 19:02:13 I think 2.6.24 is the oldest kernel working... Oct 17 19:11:28 DocScrutinizer: lol Oct 17 19:11:43 palm pre physical keyboard is mostly a joke Oct 17 19:29:37 khiraly1: ack. and I don't like the c-ts as well Oct 17 19:30:01 capacitive touchscreen is awesome. ITs way more durable then resistive one Oct 17 19:30:09 actually pre is kinda... well kinda poor mans iphone Oct 17 19:31:00 khiraly1: (durable) yes, like a stone is more durable than a drumstick Oct 17 19:31:18 my freerunner has cracked screen because of my key (probably). So it has some color circles on the screen Oct 17 19:31:55 hi Weiss JaMa I have a question:there is that in oe in ./conf/machine/include/htc-msm7.inc : Oct 17 19:32:02 PREFERRED_PROVIDER_virtual/xserver = "xserver-kdrive" Oct 17 19:32:02 XSERVER = "xserver-kdrive-fbdev" Oct 17 19:32:03 to help that, a c-ts won't help much. Oct 17 19:32:06 so what about Xorg? Oct 17 19:32:10 SHR uses Xorg Oct 17 19:32:14 what should we do about that? Oct 17 19:32:48 SHR: 03mok 07phoneuid * r9eb80b9b8dd7 10/src/phoneuid.c: remove some obsolete code Oct 17 19:32:59 SHR: 03mok 07phoneuid * r55160f1ef0f1 10/ (12 files in 3 dirs): use dbus specs from shr-specs and rename Dialogs to Notification Oct 17 19:32:59 SHR: 03mok 07phoneuid * r988f40804da7 10/src/ (phoneuid-contacts.c phoneuid-contacts.h): rename Contacts.DisplayItem to Contacts.DisplayContact Oct 17 19:33:34 khiraly1: for a r-ts you can have more rigid protective glass *under* the ts foils than a c-ts ever can use for surface Oct 17 19:34:17 dunno, on my laptop I hace several color circles... Oct 17 19:34:36 like when you touch the screen with your finger really hard. Just it does not disappear... Oct 17 19:34:44 well, laptop is way larger Oct 17 19:37:31 and actually you don't want to use 3mm protective glass on r-ts just because with r-ts pinpoint resolution the paralaxis really does matter. For c-ts where you never see the point you touch the screen it's of minor importance if display plane and surface are 5mm apart from each otherr Oct 17 19:40:48 and I **WANT** to use my sylus and / or fingernails to pinpoint spot a pixel on screen. So c-ts never will find its way to my pocket Oct 17 19:41:41 I tried to take notes using the resistive touchscreen, but its clearly a no-go Oct 17 19:42:14 I tried to draw some simple diagrams (note taking during a course), but the 1bit depth is simply cant work for drawing Oct 17 19:43:13 drawing or handwriting recog on a c-ts? LOL Oct 17 19:43:52 nope, on resistive touchscreen Oct 17 19:44:32 as I cant draw on resistive touchscreen, I cant see any advantage using it (instead of capacitive touchscreen) Oct 17 19:44:47 well, I don't know any device that takes drawings and handwriting more paperlike than my N810 Oct 17 19:44:50 what I see, its way more fragile Oct 17 19:45:31 I have a wacom tablet. It is really nice to draw or handwrite with it Oct 17 19:45:43 it has 8bit depth resolution Oct 17 19:45:44 my beer glasses are way more fragile than a tin cup. Still I prefer them Oct 17 19:46:30 SHR: 03mok 07phoneuid * rbe82398c60f9 10/src/phoneuid-dbus-common.h: fix dbus path/names include file Oct 17 19:46:32 so what? which stylus??? Oct 17 19:47:26 BTW, do all of you know that "dead" seagate barracuda can in fact be "resurrected" fairly easy with just an UART-anything converter? Oct 17 19:47:38 or do you tell me you're writing and drawing with your finger on the wacom? Oct 17 19:49:02 PaulFertser: eh! Oct 17 19:50:21 khiraly1: btw a c-ts has not even exactly 1 accurate bit Oct 17 19:50:36 for pressure Oct 17 19:50:48 (21.47.38) DocScrutinizer-8: or do you tell me you're writing and drawing with your finger on the wacom? <-- using a pen Oct 17 19:51:01 I use a pen with freerunner's resistive touchscreen Oct 17 19:51:09 a special pen I guess Oct 17 19:51:09 and couldnt really draw *simple* diagrams Oct 17 19:51:16 or it is a r-ts then Oct 17 19:51:17 DocScrutinizer: a wacom pen, yes Oct 17 19:51:28 it has special tip Oct 17 19:51:55 with builtin responder electronic Oct 17 19:52:18 those ts are neither R nor C Oct 17 19:52:43 best descriptions would be inductive I guess Oct 17 19:53:50 you can't do that in transparent afaik, and you can't build a coil into your finger ;-) Oct 17 19:54:55 c-ts has a precision of roundabout the area of your fingerprint. Accept that, it won't change Oct 17 19:56:10 either you like that, and feel happy with restriction to specially adapted apps, or you want r-ts / any other technology Oct 17 19:59:09 :) Oct 17 19:59:32 I would really like more bit depth resolution screens Oct 17 19:59:36 not just on/off state Oct 17 19:59:47 wacom tablet was really nice in that sense Oct 17 19:59:54 yeah, it was something special Oct 17 20:00:02 I could change the pen's tip Oct 17 20:00:25 well, that's simpe. Just place a force sensor to the whole lcm mount points Oct 17 20:01:07 DocScrutinizer-8, wouldn't bending the display be bad ? Oct 17 20:01:18 you can even do a pretty good ts with three of those sensors, using simple vector arithmetics to calculate the poinnt of pressure Oct 17 20:01:34 I mean applying tensile strengh to it is a pretty bad thing for those crystals Oct 17 20:02:56 btw. my wacom pen acts as a transponder Oct 17 20:03:10 cz_jc: of course you don't press on the lcd. Instead test the force applied to a thick glass protective cover Oct 17 20:03:25 it has a tip with pressure sensor, it gets energized by RF radiation from the tablet and transmits back pressure Oct 17 20:03:36 cz_jc: (pen) every wacom stylus does Oct 17 20:03:40 DocScrutinizer-8, oh ! right then Oct 17 20:04:25 cz_jc: the "RF" actually is inductive coupling Oct 17 20:04:59 mrmoku, | configure: error: C preprocessor "arm-angstrom-linux-gnueabi-gcc -E" fails sanity check Oct 17 20:05:14 mrmoku, after a clean and rebuild of glibc Oct 17 20:05:23 brb Oct 17 20:07:42 cz_jc: if you have a table with 3 legs, and you place it on 3 gauges to test the weight, then you can tell mass and mass center position of any single object you put on the table Oct 17 20:08:32 DocScrutinizer-8, what if the surface properties interfere with the reading ? like if it bends or something Oct 17 20:08:56 I don't see such issue Oct 17 20:09:28 I'm very interested how pressure sensor pressure sensitive touchscreen would work out Oct 17 20:09:49 I mean you could make those things with glass and el cheapo piezo speakers Oct 17 20:09:50 if you put a mattras on the table, basically nothing changed Oct 17 20:10:20 yep. was planned for rasterphone ;-) Oct 17 20:10:27 SHR: 03mok 07phonefsod * r79a587978252 10/src/ (4 files): fix dbus paths Oct 17 20:10:28 SHR: 03mok 07phonefsod * ra00dccba7ab5 10/ (configure.ac src/Makefile.am): use official dbus specs from shr-specs to create the bindings Oct 17 20:10:36 exactly that setup: piezo speakers Oct 17 20:11:00 but what if the pressure evens out on the surface and you get the same reading on all sensors ? I'm not sure if it'll work Oct 17 20:11:05 additional benefit: nice moving-surface feedback Oct 17 20:11:45 DocScrutinizer-8, did someone try it? :) Oct 17 20:11:53 if the pressure is same on all three sensors, then obviously the pressure point has to be center of the triangle Oct 17 20:12:08 yep, back in the 80s Oct 17 20:12:16 on CRT Oct 17 20:12:41 or 70s? Oct 17 20:12:50 dunno, too long ago Oct 17 20:13:29 DocScrutinizer-8, I'm pretty sure thing like this would make hack a day :P Oct 17 20:13:36 and I think there were no piezo speakers back when ;-) Oct 17 20:14:03 I saw super old pressure sensor howto in amateur radio magazine I have here Oct 17 20:14:55 it was a cylindrical coil with metal piece on some hard rubber material inside Oct 17 20:15:19 yep, works as well Oct 17 20:16:35 DocScrutinizer, btw did you see LED based multitouch touchscreen ? Oct 17 20:16:55 DocScrutinizer, sorry that was for DocScrutinizer-8 Oct 17 20:17:06 uhh Oct 17 20:17:21 nope, not yet Oct 17 20:17:35 DocScrutinizer-8, basically you have a grid of IR leds under LCD that alternate quickly between radiating IR and acting as a photodiode in reverse mode sensing it back Oct 17 20:17:46 you have crappy resolution but multitouch Oct 17 20:17:49 but then, what's your def of led based? Oct 17 20:18:24 when finger goes over the leds, it reflects the ir back Oct 17 20:18:34 I c Oct 17 20:19:07 they tried it without the panel above it and got pretty good results, I think it could be made working with some smart software under a LCD display Oct 17 20:19:30 you could even use white leds and use them as backlight instead of IR leds maybe Oct 17 20:19:45 nope, as lcd isn't transparent, or at least the ir-led matrix isnt Oct 17 20:20:02 grid of IR leds <-- every door phone use this Oct 17 20:20:19 I didn't mean entire module, just the liquid crystal matrix without the backlight :) Oct 17 20:20:24 you could try this with active matrix (o)led screen Oct 17 20:20:42 oh yeah, oled could sense it back by itself actually thats a good idea Oct 17 20:20:46 http://www.arday.hu/images/pg007_1_03.jpg Oct 17 20:20:48 ^^ Oct 17 20:21:15 even with a crappy resolution, oled with touch capability with nothing added to the display whatsoever sounds very cool Oct 17 20:23:10 there were a lot of solutions similar to this for CRT. All have one big disadvantage: you need a bright screen (content) for this to work Oct 17 20:23:43 DocScrutinizer-8, that could be solved with additive to the LED matrix Oct 17 20:24:08 DocScrutinizer-8, if you made it emit strong IR radiation in addition to normal light, that would shine even with a lower voltage Oct 17 20:24:21 the technology called light-stylus had to cause a white flash of screen for each touch to detect cathode ray pass by Oct 17 20:24:41 I'm not sure about OLED radiation spectrum though, maybe it does this in some way already Oct 17 20:24:54 yep, with IR that's solvable problem Oct 17 20:26:16 but as my RC helicopter says, distinguishing IR radiation apparently sucks under direct sunlight Oct 17 20:27:04 u bet Oct 17 20:27:19 i better love the force gauge design Oct 17 20:27:35 yeah it sounds simple to read as opposed to the IR led method Oct 17 20:28:54 and it's tested. I already built prototypes for POC Oct 17 20:29:17 DocScrutinizer-8, really ? do you have any videos/anything? :) I'm pretty curious Oct 17 20:29:31 nope Oct 17 20:30:10 some time went by since then, and it was a hack of an hour, with an old bathroom persona gauge Oct 17 20:30:42 hmm I have an old LCD with blown low voltage transformer Oct 17 20:30:53 maybe I'll try it with it since its already apart once I repair it Oct 17 20:31:01 just have to extract 2 more piezos from some toys XD Oct 17 20:32:30 you'll probably find 4 force sensors (if it's a glass model). one each corner. they are serialized. if you short one, the reading goes down by the P-offset plus force of the particular corner Oct 17 20:35:23 the procedure is: short one sensor, calibrate to 0. unshort and reat out idle-reading. then apply force, short and read out offset, then unshort and calculate diff to former readings to get force applied to particular ccorner Oct 17 20:36:51 DocScrutinizer-8, yep, figured as much, thanks :) Oct 17 20:37:08 yeah I think I'll use glass, its already there over it Oct 17 20:52:59 well my transparent touchpannel had a really secure protective glass of 15mm thickness, and for accurate readout you had to press it with a force of several kilo. But was a 0-effort POC which worked quite encouraging Oct 17 20:54:39 DocScrutinizer-8, so what do you say? some hard sandwich material of different transparent plastic foils ? Oct 17 20:55:20 hmm? sorry don't get it. Oct 17 20:55:36 the bathhroom gauge was real glass Oct 17 20:55:47 DocScrutinizer-8, how thick ? Oct 17 20:56:00 you mean the 15mm glass one? Oct 17 20:56:00 15mm roundabout Oct 17 20:56:05 yep Oct 17 20:56:42 hmm thats bad.. Oct 17 20:57:25 physics say if you stack panels of different materials on top of each other, you get extremely high tensile strenght Oct 17 20:57:25 with four round silver feet each corner one that contained the force sensors, and a digital display uffer middle Oct 17 20:57:53 it works with wood + plastic, maybe it works with different plastics and maybe transparent glass Oct 17 20:58:28 point is to get thin and durable surface to touch, right ? Oct 17 20:59:03 well for mobile device 3+/- inch ts-displays I'd recommend polycarbonate glass ~1.5mm Oct 17 20:59:38 DocScrutinizer-8, how inaccurate the radout was when you just tapped the screen ? Oct 17 20:59:42 *readout Oct 17 20:59:53 with like slight touching force Oct 17 21:00:03 which screen? Oct 17 21:00:21 DocScrutinizer-8, the one with the 15mm thickness Oct 17 21:00:46 that was a transparent glass bath gauge. no screen at all Oct 17 21:01:13 yeah but how accurate was the triangulation? :) thats all that matters Oct 17 21:01:33 and the display/electronics it came with had an accuracy of 0.1kg Oct 17 21:01:57 with several kg it was quite accurate Oct 17 21:03:39 hmm glass panel I have handy here is 4.5mm thick.. well that sucks Oct 17 21:08:09 mrmoku, here? Oct 17 21:08:59 mrmoku, as I said, I'm not sure the module loading works Oct 17 21:09:04 but module installation works like a charm. Oct 17 21:11:41 sec Oct 17 21:11:44 I'll compile phoneuid Oct 17 21:14:53 actually, more than a sec Oct 17 21:14:55 cya tomorrow ;) Oct 17 21:15:08 unless you ping me soon and explain what's wrong. Oct 17 21:16:09 DocScrutinizer-8, do you think soldering on piezo element is dangerous ? Oct 17 21:16:24 I'm pretty fed up with things exploding in my face all the time already Oct 17 21:17:49 cz_jc: piezo is harmless :-D Oct 17 21:18:07 yeah I found one but without leads :/ Oct 17 21:18:09 allright XD Oct 17 21:18:26 but actually you don't solder it (at least I never seen a soldered one) Oct 17 21:19:28 probably thy break on heat, or soldering the piezo silver isn't very advicable Oct 17 21:19:40 damn it just doesn't stick too well X-) Oct 17 21:20:43 all piezo elements I've seen were contacted viia springs Oct 17 21:20:59 this one actually has little balls of lead on it already Oct 17 21:21:10 duh Oct 17 21:21:10 it seems like it was soldered to but it wasn't me.. Oct 17 21:21:13 still it doesn't stick Oct 17 21:23:01 damn that thing is scary Oct 17 21:23:23 I like had to head it for pretty long for the lead on it to even melt and its still stone cold all around it, it sure absorbed lot of energy Oct 17 21:23:28 *heat Oct 17 21:25:47 pheeew both leads are on and I'm not full of metal shrapnel or blind XD Oct 17 21:26:03 use goggles! Oct 17 21:26:13 I have glasses XD Oct 17 21:26:36 ok lets see how much voltage it does on touch Oct 17 21:26:38 aah, sometimes those are an advantage Oct 17 21:27:01 first test if it gives noise on 2V Oct 17 21:27:26 and probe for resistance >50MR Oct 17 21:27:51 just a quick test says it works Oct 17 21:27:57 cool Oct 17 21:28:00 when I leave it alone, it produces about 3-5 mV noise Oct 17 21:28:09 when I press it, it spikes at some 300mV Oct 17 21:28:34 so I'd suggest you immediately fixate the wires with two blobs of glue Oct 17 21:28:40 hehe thats uber easy to read actually Oct 17 21:29:29 haha I know what makes the noise now XD Oct 17 21:29:51 I have kinda bad habit that I'm shaking my leg weirdly when doing some experiment and its propagating through the table to the piezo XD Oct 17 21:29:54 your heart pulse? Oct 17 21:30:11 hehe Oct 17 21:30:38 when I go away from the table it goes to some 2mV and stays there, I think its because of the imprecise crappy multimeter Oct 17 21:30:50 and then when I even slightly touch it, it spikes Oct 17 21:31:35 cz_jc: the actual problem is you need a constant reading for a constant pressure Oct 17 21:31:57 I think it registers pressure change, doesn't it ? Oct 17 21:32:50 so you either need a meter with *very* high input-Z,or you integrate the voltage drop over a e.g. 1MR Oct 17 21:33:09 how much is *very*, do you consider 1 megaohm enough ? Oct 17 21:33:35 cz_jc: yep, it builds up some energy (*9)97>+l for a certain pressure change Oct 17 21:34:17 more like 1GR, 1 MR is dissipating the charge it created quite quickly Oct 17 21:34:42 DocScrutinizer, do you think sticking it to gate of some low power switching mosfet would work ? Oct 17 21:35:07 yep, probably Oct 17 21:35:24 ok, I'll try when I repair the panel then :) Oct 17 21:35:31 this was fun to play with actually XD Oct 17 21:35:46 hehe Oct 17 21:36:37 its so cool Oct 17 21:36:56 I can actually even on the stupid multimeter clearly tell diference between a slight tap, a press and a hard press against it Oct 17 21:37:27 but I guess it will have to be pretty precise for the touchscreen to work usably Oct 17 21:38:08 for a click it's simple. drags are a pain to get it right though Oct 17 21:38:48 DocScrutinizer-8, if you consider you have to press it and then depress it, you like have to remember how much charge went into it and expect how much will dissipate, right ? Oct 17 21:39:52 err, sounds somewhat right (as much as I get it) Oct 17 21:40:36 sounds like during a drag, you'd get depress on some sensors and more press on another Oct 17 21:40:48 yep Oct 17 21:41:42 that's the problem, as piezo doesn't show absolute pressure, rather it gives a signal proportional to pressure *change* Oct 17 21:42:14 so the first stage should translate that change to absolute pressure Oct 17 21:42:36 unless you probe with a gigaohms Z voltmeter, *and* the piezo itself has same high Z Oct 17 21:42:41 something similar is done in inertial reference systems in airplanes Oct 17 21:42:59 you get a change in position from various sensors and by that you guess where in the world you actually are Oct 17 21:43:01 (first stage) that would be aforementioned integration Oct 17 21:43:52 DocScrutinizer, I have an idea for that Oct 17 21:44:07 what if you put a coil above the pizeo Oct 17 21:44:29 use an opamp, with a 1kR and a 100uF on output? Oct 17 21:45:04 opamp with gigaohms input resistance? :P Oct 17 21:45:20 nah, megohms is enough then Oct 17 21:45:34 maybe even 100k Oct 17 21:45:35 for that ,even a stupid mosfet is enough, isn't it then ? Oct 17 21:45:42 oh right, its not linear Oct 17 21:45:43 yep Oct 17 21:46:00 no its no good Oct 17 21:46:08 and not symetric (for the negative voltage on de-pressing it) Oct 17 21:46:16 you need for the capacitor to discharge as much as it charges for the same pressure change on press/depress Oct 17 21:46:27 yes exactly Oct 17 21:46:57 it definitely needs to be symmetric to sense both polarities of changes Oct 17 21:47:58 I dunno what its called in english but opamp has a significant voltage range where you get neither negative nor positive output yet Oct 17 21:48:07 a simple opamp with bias on VDD/2 on +, and a gain of 1 for the 1M negative feedback, probably would work quite nice Oct 17 21:48:11 that will pose a significant problem when dragging Oct 17 21:48:24 the pressure will change slightly then I imagine Oct 17 21:49:33 DocScrutinizer-8, yeah that definitely sounds reasonable and a good place to start first experiment :) Oct 17 21:50:28 1k from +inp to VDD, 1k to GND. 1M from +inp to -inp. 1M from out to -inp. piezo to -inp and GND Oct 17 21:50:32 larsc, hi Oct 17 21:50:55 larsc, did you make xf86-input-tslib work for xorg Oct 17 21:51:00 and commit it? Oct 17 21:51:05 in oe.dev? Oct 17 21:51:17 Gnutoo, works for me on Gentoo Oct 17 21:51:29 cz_jc, wich xorg version? Oct 17 21:51:47 cz_jc, on oe it doesn't compile because of xorg abi ifdef that are not there Oct 17 21:51:56 s/abi/version Oct 17 21:52:10 cz_jc: 1k and series 47uF from out too GND Oct 17 21:52:49 Gnutoo, 1.6.2.901 (1.6.3 RC 1) Oct 17 21:52:59 connect a semi decent DVM (or scope) to 47uF Oct 17 21:53:08 from 26.7. 2009 Oct 17 21:53:21 (day/month/year) Oct 17 21:53:24 ah so not 1.7 Oct 17 21:53:27 cz_jc: then you should get a better signal Oct 17 21:53:27 nope Oct 17 21:54:25 shr uses 1.7 Oct 17 21:54:35 DocScrutinizer-8, yep, I really need to sleep though, gotta wake up early tommorow and go across the whole republic cheering up my ex again who is apparently on the edge of suicide Oct 17 21:54:38 I don't want to duplicate someone else's work Oct 17 21:54:45 so I ask if it was done Oct 17 21:54:49 and what's the state of it Oct 17 21:55:04 ouch Oct 17 21:56:02 cz_jc: good luck Oct 17 21:56:06 Gnutoo: it at least compiles Oct 17 21:56:16 ok have you pushed it? Oct 17 21:56:19 DocScrutinizer-8, thanks.. but I think it'll be allright.. again.. Oct 17 21:57:01 DocScrutinizer-8, and don't worry, I didn't cause it, her (now ex) bf did and I get to smooth out the damage :/ she takes it so seriously Oct 17 21:57:07 Gnutoo: no Oct 17 21:57:24 anyway I'm off.. goodbye and happy hacking, people :) Oct 17 21:57:35 larsc, can you send me patches? Oct 17 21:57:36 i can't test it. my xserver is complaining it can't compile an keyboard layout and won't start :/ Oct 17 21:57:37 git diff Oct 17 21:57:48 larsc, ok Oct 17 21:58:12 larsc, if It works: Oct 17 21:58:36 *where should I push/send for review it(oe.dev or shr-merge or shr-import etc...) Oct 17 21:58:43 *who should I credit Oct 17 22:02:44 Gnutoo: I have no idea Oct 17 22:03:29 larsc, you did the patches...so you have to be credited(or if you want to be anonymous I can credit an anonymous person) Oct 17 22:03:46 but I don't want to take credit for things I didn't done Oct 17 22:06:52 Gnutoo: yes, thats fine with me. I was referring to where it should go. Oct 17 22:07:02 ok Oct 17 22:07:09 I'll take care of that Oct 17 22:07:24 can I have your patch so I could test it? Oct 17 22:07:47 btw there are some people who commit stuff without testing on oe.dev Oct 17 22:09:43 larsc, a git diff,or only the patch for xf86-input-tslib would be fine Oct 17 22:09:52 any form of patch in fact Oct 17 22:10:43 larsc, if you're here tomorrow we could do the following: I test you commit...it would be easier Oct 17 22:12:29 Gnutoo: http://metafoo.de/tslib-xserver-1.7.patch Oct 17 22:12:37 larsc, thanks Oct 17 22:13:50 larsc, before you commit that put ifdefed for Xorg version such as: Oct 17 22:14:46 i know Oct 17 22:15:16 #if GET_ABI_MAJOR(ABI_XINPUT_VERSION) >= 7 Oct 17 22:15:16 ok Oct 17 22:15:37 but thanks a lot for the patch Oct 17 22:53:27 Gnutoo: I was able to test it now and it seems to work **** ENDING LOGGING AT Sun Oct 18 02:59:57 2009