**** BEGIN LOGGING AT Sun Apr 26 02:59:57 2009 Apr 26 03:16:36 Success! Apr 26 03:17:55 :) Apr 26 03:24:13 Did you a choice of two themes when you first booted the new image? Apr 26 03:24:47 yes, Illume and Illume-SHR Apr 26 03:24:59 Hmm.. Apr 26 03:25:05 I wonder why I do not... Apr 26 03:25:18 * mwester goes to start Yet Another Full Build From Scratch. Apr 26 03:25:30 (YAFBFS) Apr 26 03:26:42 quick question for you. Do you like the way SHR-Settings handles profiles? Does anything about it seem awkward to you? Apr 26 03:36:31 I haven't really messed about with profiles; my phone has never worked well enough for it to matter. Apr 26 03:36:39 ahh ok Apr 26 03:41:35 I seem to be missing libXv, odd Apr 26 03:50:33 apparently it's libxv1. Apr 26 08:28:48 has someone tested sqlite-contacts opimd backend? Apr 26 08:29:15 i've fixed custom fields handling Apr 26 08:30:53 noone? Apr 26 08:33:31 lames ;p Apr 26 08:33:36 ;) Apr 26 08:35:02 i ate them all Apr 26 08:36:23 raster: were them tasty? Apr 26 08:37:02 crunchy Apr 26 08:37:45 * dos1 needs to eat breakfast Apr 26 08:38:56 moin raster Apr 26 08:39:10 DocScrutinizer: panties! Apr 26 08:43:26 still online. amazing Apr 26 08:47:42 raster: try some of my 2440 chips. With guacamole ;D Apr 26 08:48:13 DocScrutinizer: why do you have any of those 2440's Apr 26 08:48:23 its like having 5-year old pizza in the fridge Apr 26 08:48:24 :) Apr 26 08:48:46 they stay crunchy ;) Apr 26 08:50:20 well, no kidding. could practice bga-soldering without letting smole out of really expensive new ones, for instance Apr 26 08:50:37 aaaah Apr 26 08:50:39 s/mole/moke/ Apr 26 08:50:44 u bga solder by hand? Apr 26 08:50:53 not yet Apr 26 08:50:59 wait - regualr bga.. not micro? Apr 26 08:51:15 whatever Apr 26 08:51:44 best i knew u couldnt do microbga by hand? Apr 26 08:51:45 I have a whole bag full of all kinds of surplus parts Apr 26 08:52:34 prolly not. ubga is hard even at industrial level Apr 26 08:54:44 yeah Apr 26 08:54:59 but anyway, all those chips at least make nice IO-extenders with lots of GPIO controllable via JTAG ;) Apr 26 08:55:10 hahaha Apr 26 08:55:32 DocScrutinizer: but how would the sound alert tell me which channel has activity? Apr 26 08:55:45 huh? Apr 26 08:56:25 lindi-: sorry, not up to 100% throttle yet. don't fet it Apr 26 08:56:49 lindi-: ahhhhh, xchat?!? :D Apr 26 08:56:54 DocScrutinizer: yep Apr 26 08:57:28 on N810 the channel is highlighted in blue on channel-list Apr 26 08:57:57 until you focus on it Apr 26 08:58:22 is there any elementary design howto? i want to make a window 480x220, not full screen, how can i do that? python-elementary Apr 26 08:58:51 DocScrutinizer: but the channel list does not fit to screen Apr 26 08:59:15 yeah, that's a FR problem for sure Apr 26 09:00:04 lindi-: i didn't test xchat on Neo yet Apr 26 09:00:04 m0nt0: what do u want to do wiuth the rest of the screen then? Apr 26 09:00:35 leave it trasparent, a small window to the bottom of the screen Apr 26 09:01:21 m0nt0: transparent is NOT same as smaller window Apr 26 09:01:21 raster, any docs with elementary? (please don't tell me to look at the source) Apr 26 09:01:28 when u add the window - make it of tyle DIALOG Apr 26 09:01:33 u'll get a titlebar tho Apr 26 09:01:43 the rest is up to the wm and its management policies Apr 26 09:01:53 raster, ok Apr 26 09:02:01 freesmartphone.org: 03seba.dos1 07framework * r2e58d83c395b 10/framework/subsystems/opimd/backend_manager.py: Apr 26 09:02:01 freesmartphone.org: Revert "opimd: don't fail when default backend isn't specified. Use some, which supports specified domain." Apr 26 09:02:01 freesmartphone.org: That isn't correct way to do this. Apr 26 09:02:01 freesmartphone.org: This reverts commit a100a6b73023abb4ed385cf0e3615c1f303f6cc4. Apr 26 09:02:03 freesmartphone.org: 03seba.dos1 07framework * r6d56c641ddd9 10/framework/subsystems/opimd/pimb_sqlite_contacts.py: opimd: fix custom fields handling in SQLite-Contacts backend Apr 26 09:02:08 http://trac.enlightenment.org/e/wiki/Elementary + src (test.c) + Elementary.h Apr 26 09:02:22 DocMobilizer, it was to give an example, i want a small window, not a big one with a tasparent part Apr 26 09:02:36 raster, thanks a lot! Apr 26 09:03:02 + read the src Apr 26 09:04:40 lol i was fearing that answer Apr 26 09:08:03 raster: fyi if somebody will ask you again why battery meter on FR with HAL doesn't work correctly, that's because HAL sees 3 batteries (real battery, apm emulation, usb is considered to be battery too due (according to hal sources)). One of the problems is due to the fact that "real" battery (/sys/class/power_supply) doesn't present all the nodes commonly found on laptops (like design_capacity etc). HAL irc channel doesn't seem to be very .. Apr 26 09:08:09 ... helpful but i'm going to write an e-mail asking for claricifactions on kernel sysfs requirements and about proper usage of hal by battery widgets etc. Apr 26 09:09:17 ok Apr 26 09:11:41 roh: ping Apr 26 09:13:34 need to go, thanks a lot raster Apr 26 09:13:35 PaulFertser: are you sure you don't want to png jc rather than roh (who's vacant like the other 80%) Apr 26 09:14:08 DocScrutinizer: i want to ask someone who has admin privileges on OM kernel ML. I guess only roh does. Apr 26 09:14:30 DocScrutinizer: Who is jc? Apr 26 09:14:55 nah, jc should have as well. he's the only "official" admin left over Apr 26 09:16:05 DocScrutinizer: as far as i know any sane mailing list should accept mails from unsubcribed members. I think last time i tried it with OM lists resulted in "your mail awaits moderation" spam sent back to me. Therefore OM lists are not appropriate for conversations with people outside OM community which i feel to be very wrong. Apr 26 09:16:25 DocScrutinizer: What's jc's irc nick? Apr 26 09:16:25 PaulFertser: actually most won't Apr 26 09:16:38 PaulFertser: we disabled public posting years ago Apr 26 09:16:42 most mailing lists did Apr 26 09:16:45 PaulFertser: that's correct and intentional Apr 26 09:16:55 and no sane person runs a mailing list with public posting allowed Apr 26 09:17:02 yup Apr 26 09:17:16 raster: how come all kernel devs can communicate on LAKML, ASOC, ide, usb etc lists without problems. I don't assume all they're subscribed to all mailing lists. Apr 26 09:17:16 it's a rather insane mailing list that allows is Apr 26 09:17:17 it Apr 26 09:17:18 :) Apr 26 09:17:32 raster: linux-usb allows it for sure - i tried it myself recently. Apr 26 09:17:39 PaulFertser: my bet is they are subscribed. i am subscribed to about 30 lists myself. Apr 26 09:17:53 well its an insane list Apr 26 09:17:55 an exception Apr 26 09:18:00 most people turned it off due to spam Apr 26 09:18:01 raster: are you really sure proper SPF verification is not enough? Apr 26 09:18:02 long ago Apr 26 09:18:24 i turned it off long before spf existed Apr 26 09:18:31 and no way am i turning it on Apr 26 09:18:47 raster: so i assume you don't have solid grounds to have it disabled now, do you? ;) Apr 26 09:18:58 absolutely Apr 26 09:19:21 as the mailing list admin - there are about 500+ spams per day awaitying moderation on sf.net's mailman mailing lists Apr 26 09:19:32 i no longer approve mails from non subscribers Apr 26 09:19:40 finding their mail in the mountain of spam is too hard Apr 26 09:19:51 raster: what's their SPF status? How can you prove that SPF is not enough? Apr 26 09:19:53 no susbscribe - no mail. simple. Apr 26 09:20:43 raster: i'm quite sure that's how mailing lists are intended to be used. Apr 26 09:20:47 and so sane as well Apr 26 09:20:54 s/that's/that's not/ Apr 26 09:21:21 intended by whom? copyright-holder? Apr 26 09:21:34 DocScrutinizer: intended by sane persons designed the mailing lists system. Apr 26 09:22:13 they supposed ML not being used for spam Apr 26 09:22:26 DocScrutinizer: imagine i want to have subscribers of one mailing list to see some discussion on the other? It's easy -- just Cc the list. But if you require everybody to be subscribed, that becames impossible. Apr 26 09:22:39 PaulFertser: sf.net offers no option to turn spf on/off for lists Apr 26 09:22:44 so either they dont implemente it Apr 26 09:22:49 or they do and it has zero effect Apr 26 09:22:53 DocScrutinizer: Somebody's gotta prove first that there's considerable amount of spam left after SPF verification. Apr 26 09:22:58 SHR: 03seba.dos1 07shr-themes * rc012c38d8be6 10/frameworkd/frameworkd-config-shr/ (5 files in 5 dirs): frameworkd-config-shr: add opimd configuration Apr 26 09:23:14 go ahead and volunteer Apr 26 09:23:22 i know i've tried a mountain of things rto get rid of spam Apr 26 09:23:37 greylisting only mildly reduces it Apr 26 09:23:53 raster: i don't trust sf.net a bit. The guys who are still not providing users with any DVCS are ... hm... Apr 26 09:25:48 on twinkle ML we even switched to moderated subscribe as spammers subscribed by script on zillion ML and spammed them Apr 26 09:26:16 DocScrutinizer: reading through my gmail spam folder is enough to prove? I'm seeing only softfail and neutral here, no single permitted yet. Apr 26 09:26:56 I don't care about any gmail spam folder Apr 26 09:27:29 DocScrutinizer: Why? I get a perfect load of spam there. And if every spam letter in there fails SPF check, then it proves my point that SPF generally is enough. Apr 26 09:28:14 you're mixing private mail account with ML Apr 26 09:28:35 ML is a high prio target for spammers Apr 26 09:28:48 DocScrutinizer: ok, how can i prove that ML won't be spammed with proper SPF verification then? Apr 26 09:29:18 dunno. the question is, who'd care if you do Apr 26 09:29:35 Fuck Apr 26 09:29:36 PaulFertser: gmail surely doesnt only use spf to check Apr 26 09:30:14 raster: yes, but i did. I just entered my "spam" folder. And checked headers of some 20 mails manually. Apr 26 09:30:23 every kind of spam filtering i know of lets spam through Apr 26 09:30:27 its just a q of how much Apr 26 09:30:31 Exactly Apr 26 09:30:42 on e's mailing lists we have had 0 spams Apr 26 09:31:07 (in the past year) Apr 26 09:31:13 i thnk we have had 1 or 2 over the many years Apr 26 09:31:29 and each bred a whol trhead of 20+ mails of people asking for it to be shut up/removed Apr 26 09:32:40 raster: that proves that your method is efficient, indeed. But it doesn't prove that strictly enforcing SPF policies is not enough to be acceptable. Apr 26 09:33:35 considering i've been through greylisting, spamassasin, bogofilter, and more, all of them have let throughh at least 5-10 spams per day Apr 26 09:34:54 raster: that's completely different measures. I don't say that SPF is an uber-method capable of solving all the spam problems with spam out there. But it'd be nice to have a real proof that it doesn't work. Apr 26 09:34:58 PaulFertser: the point is: it works the way it's now, so nobody will consider changing this proven scheme to enable sth you think is a desirable feature but generally is considered Xposting and deprecated anyway Apr 26 09:35:41 bingo Apr 26 09:35:43 DocScrutinizer: i don't see how Xposting is considered deprecated. And for some reason linux-usb mailing list works the way i describe... Apr 26 09:35:54 for everyone any other form is going to have more holes than we currently do Apr 26 09:36:36 and i am a developer - i am in no mood to be moderating out spams Apr 26 09:36:42 or continualyl chasing blacklists Apr 26 09:36:48 i have code wot write Apr 26 09:36:58 a susbscribe-to-post method is 0 work on my back Apr 26 09:37:04 and actually works Apr 26 09:37:30 we've had forum spam (web forum) with subscribe-bots that subscribe and post Apr 26 09:37:34 had wiki spam too Apr 26 09:37:37 susbcribe and edit pages Apr 26 09:37:59 i am in no mood to open up holes considering even if you try and close them - they still get through Apr 26 09:38:11 and i am sure that pretty much all mailing lists are the same Apr 26 09:38:16 you may have the odd one that isnt Apr 26 09:38:27 but the vast majority do require subscription Apr 26 09:41:42 LAKML: Recently, we have had to impose a restriction on the mailing lists. You must be subscribed to the mailing list in order to post messages to that mailing list. This is because of the UK Data Protection laws. Only subscribe to these lists if you accept the legal notice displayed on the relevant pages; by subscribing, you accept the terms laid out in the legal notice. Answers will be copied to you. Apr 26 09:41:58 raster: was it because of spam? No. Apr 26 09:43:55 raster: LKML: No, you don't have to be subscribed to the list to post to it. The address of the list is linux-kernel@vger.kernel.org. And you should indicate on your message that you wish to be personally CC'ed the answers/comments posted to the list in response to your posting. Apr 26 09:47:52 raster: i don't say you must do something with E MLs ;) Just that it's not that uncommon to allow unsubcribed postings. Apr 26 10:36:02 pff, APN is reset to "default" internet, even without boot and without disconnecting GPRS :-( Apr 26 10:39:03 this is: enter e.g "pinternet.interkom.de", connect, close settings-connectivity-gprs window, irc or surf, open settings-connectivity-gprs window, irc or surf, open connecticity-gprs again -> still connected, but APN is reset to "internet" Apr 26 10:39:44 errr, sorry Apr 26 10:41:46 this is: enter e.g "pinternet.interkom.de", connect, close settings-connectivity-gprs window, irc or surf, open connecticity-gprs again -> still connected, but APN is reset to "internet" Apr 26 10:44:26 DocScrutinizer: that's normal. shr-settings doesn't store apn, login and password ATM Apr 26 10:45:58 but Ididn't even boot. So it is absolute nonsense as I'm obviously still connected to pinternet.interkom.de while settings says "connected, APN:internet" Apr 26 10:47:14 dos1|away: how hard could it be to implement decent persistent storage and display of actual value here? Apr 26 10:48:23 s/here/in whole settings/ Apr 26 10:49:04 bzzbot: alive? Apr 26 10:49:29 bzzbot: stupid lazy bot Apr 26 10:49:52 ~uptime Apr 26 10:49:59 ~seen dos1 Apr 26 10:50:32 mwester: kick ass your lazy bot! Apr 26 10:51:44 bzzbot: die! Apr 26 10:54:40 dos1|away: i couldn't do this if APN:internet was a correct display Apr 26 10:55:36 so this isn't a flaw or missing feature, this is a BUG Apr 26 10:57:15 DocScrutinizer: What does the update to moko11 do? Apr 26 10:57:29 huh? Apr 26 10:57:36 it works great Apr 26 10:57:49 It seems that this supports reduced mic volume Apr 26 10:58:00 Just an observation here.. Apr 26 10:58:13 alphaone: outch. possible Apr 26 10:58:21 how so? Apr 26 10:59:30 there's a mic sensitivity setting in calypso flash somewhere. I asked Dieter about how to change this, and he told me it's stored in filesystem and isn't easy to change Apr 26 11:00:06 but basically it could happen if firmware would change the DSP register settings Apr 26 11:00:38 argh Apr 26 11:00:42 alphaone: Dieter said all settings have to be done with tITAT Apr 26 11:00:57 titat? Apr 26 11:01:34 TAT tool to adjust / calibrate calypso Apr 26 11:01:45 (was it TAT?) Apr 26 11:02:01 So where can I get that? Apr 26 11:02:09 anyway: WAAAAAAAHH! Apr 26 11:02:10 How do I use it? Apr 26 11:02:26 1) install win2000 :-) Apr 26 11:02:44 2) check internal wiki and fileserver Apr 26 11:03:23 fuck Apr 26 11:03:37 Okay, I'm reverting to moko8 here Apr 26 11:03:38 3) attach via hs to FR Apr 26 11:03:45 See if that helps Apr 26 11:03:58 If it does we need to seriously think about this Apr 26 11:04:06 yeah, please triple check if that's actually the cause of the issue Apr 26 11:04:23 will do Apr 26 11:04:33 I planned to do sth wrt improving those settings anyway Apr 26 11:04:46 great Apr 26 11:04:50 you'll get mail Apr 26 11:05:02 DocScrutinizer: Where can I find moko8? Apr 26 11:05:20 people.jr? Apr 26 11:06:23 I'll search Apr 26 11:13:09 alphaone: http://people.openmoko.org/joerg/calypso_moko_FW/gsm_ac_gp_fd_pu_em_cph_ds_vc_cal35_ri_36_amd8_ts0-gta02bv5-moko8.m0 Apr 26 11:13:29 or Apr 26 11:13:32 just found it on people.i.om.org Apr 26 11:13:33 thanks Apr 26 11:13:35 alphaone: http://people.openmoko.org/joerg/calypso_moko_FW/gsm_ac_gp_fd_pu_em_cph_ds_vc_cal35_ri_36_amd8_ts0-gta02bv6-moko8.m0 Apr 26 11:13:45 what's the difference between these two anyway? Apr 26 11:13:53 nothing ;-) Apr 26 11:13:58 And why don't we have two versions Apr 26 11:14:00 aha.. Apr 26 11:14:26 U got mail Apr 26 11:14:46 Well, that's not completely true Apr 26 11:14:53 they differ in two places Apr 26 11:15:22 version string Apr 26 11:15:42 okay Apr 26 11:15:50 So nothing ;-) Apr 26 11:16:12 it's kinda silly to create a new version hat differs in version string XD Apr 26 11:16:20 :-) Apr 26 11:16:21 true Apr 26 11:17:09 for MOKO10 we fixed this by unifying version string not to show hw revision of device anymore Apr 26 11:17:56 DocScrutinizer: but gta01 and gta02 versions still differs only in version string, doesn't it? ;) Apr 26 11:17:58 btw bv5 is an invalid pcb-version anyway Apr 26 11:18:07 yup Apr 26 11:18:14 afaik Apr 26 11:18:17 yep, it is gta02v5 Apr 26 11:18:30 hw doesn't differ, so fw shouldn't as well Apr 26 11:18:30 bvX was with gta01 Apr 26 11:18:36 yup Apr 26 11:19:04 actually its gta02A6 Apr 26 11:19:25 nevermind, doesn't care Apr 26 11:20:13 see TAT now? Apr 26 11:21:47 alphaone: you tried AT+AUL or whatever it's called? Apr 26 11:22:03 DocScrutinizer: Not yet, but that is an idea, yeah Apr 26 11:22:08 to load audio table0 (default)? Apr 26 11:22:14 Was it just AT+AUL=0? Apr 26 11:22:34 can't remember axactly, please check back in docs Apr 26 11:22:41 jup Apr 26 11:22:46 sth like that yeah Apr 26 11:22:50 WOW Apr 26 11:22:57 WOW? Apr 26 11:23:04 Mic sensitivity is WAY up there with moko8 Apr 26 11:23:10 WAAAAY up ther Apr 26 11:23:11 e Apr 26 11:23:14 FUCK!!! Apr 26 11:23:18 okay, reflashing moko11 Apr 26 11:23:31 triple checking Apr 26 11:24:13 DIIIIIIIEEETER!!!!!!! appear here! Apr 26 11:24:40 at least it's reversible :-) Apr 26 11:24:56 alphaone: ok, I got a mail ready to send. Tell me when to fire :-) Apr 26 11:25:19 short note, this device is an A5 with 0 Ohm R4303 Apr 26 11:25:24 orignal with moko8 Apr 26 11:25:45 didn't try moko11 before buzzfix, though Apr 26 11:26:14 And there were other A5s with 0 Ohm that worked just fine Apr 26 11:26:36 But maybe the calibration values are almost as erratic as buzz appears... Apr 26 11:28:57 DocScrutinizer: And FYI, it is 'revision': "GSM: gsm_ac_gp_fd_pu_em_cph_ds_vc_cal35_ri_36_amd8_ts0-Moko11" which is making trouble Apr 26 11:29:45 Now it's working Apr 26 11:30:03 DocScrutinizer: Apr 26 11:30:09 alphaone: 0R doesn't affect gsm-fw differences Apr 26 11:30:33 DocScrutinizer: Is there some autocalibration happening in moko8 perhaps? Apr 26 11:31:25 I always thought there is ALC for mic. Tho it never was stated explicitly anywhere in docs afaik Apr 26 11:32:34 alphaone: nevertheless - you confirmed now moko8 is performing better than moko11 wrt mic sensitivity? without any change whatsoever in rootfs? Apr 26 11:32:37 what was the reason for running everything as superuser, again? Apr 26 11:32:53 b0ef: laziness? Apr 26 11:33:05 DocScrutinizer: Okay, the deal was as follows: Rework buzzfix -> Call okay -> update Moko8 to Moko 11 -> Call volume quiet -> downgrade Moko8 -> call volume good (I believe better than the first try) -> update moko11 -> call volume the same (good) Apr 26 11:33:44 hmm, that's RATHER strange. Makes no story yet Apr 26 11:35:26 ALC might get upset by "hearing" some switch-on plop from mic(_bias), thus leveling down mic sensitivity completely Apr 26 11:36:04 hmm Apr 26 11:36:28 well, I'll test the other FR that I have here which exhibits a similar behaviour Apr 26 11:36:32 alphaone: try to preload gsmhandset.state before call is established Apr 26 11:37:03 DocScrutinizer: okay Apr 26 11:37:07 make sure the state isn't switched back and forth during call accept Apr 26 11:37:24 Or do make sure and check Apr 26 11:37:30 so we could keep eventual plops away from calypso - for test Apr 26 11:37:56 hmm, my gta02 does not boot if I have debug board v2 connected to it Apr 26 11:38:01 anything extra I should check? Apr 26 11:38:19 lindi-: power up debubo? Apr 26 11:38:25 DocScrutinizer: ah, that is needed? Apr 26 11:38:28 doublecheck FPC? Apr 26 11:38:38 lindi-: dunno, wild guess Apr 26 11:38:41 DocScrutinizer: the instructions say that it should power up before I power up debug board Apr 26 11:38:55 at http://wiki.openmoko.org/wiki/Debug_Board_v2 Apr 26 11:39:05 I never cared about this Apr 26 11:39:12 well :) Apr 26 11:39:43 lindi-: in the end I want to power-down / power-up FR with debubo connected ;-) Apr 26 11:40:09 yeah it boots when debug board has power Apr 26 11:40:35 sure, reset not driven low Apr 26 11:40:54 makes no sense other way round Apr 26 11:41:17 well me away now for some sunday activity :-> Apr 26 11:41:47 alphaone: seems I have to kill my pre-prepared mail ;-) Apr 26 11:42:30 alphaone: well maybe I can edit and recycle it later Apr 26 11:42:54 alphaone: you got mail? Apr 26 11:43:10 the one, yeah Apr 26 11:43:22 see attachment.zip Apr 26 11:44:06 audio_tables? Apr 26 11:45:58 nah, the TAT screenshots Apr 26 11:46:18 alphaone: I'm concerned about: " Apr 26 11:46:19 das Kommando zum Laden von Audio Tables aus dem Filesystem gibt Apr 26 11:46:21 es ja schon (funktioniert aber nicht wirklich) Apr 26 11:46:22 " Apr 26 11:46:32 hmm, yeah Apr 26 11:47:28 DocScrutinizer: Can I query these parameters via AT commands? Apr 26 11:47:32 yup audio_tables.zip Apr 26 11:47:51 nope, that's the problem Dieter elaborates Apr 26 11:48:05 you only can check/modify with TAT Apr 26 11:49:38 DocScrutinizer: Okay, I got a FR here that has really low volume Apr 26 11:49:41 5min, then bbl Apr 26 11:49:43 with moko11 Apr 26 11:50:01 okay, I'll keep this channel posted :-) Apr 26 11:50:23 please don't mess with volume vs sensitivity. Be clear on that in your wording Apr 26 11:50:50 sorry, okay Apr 26 11:51:17 at%AUL=0 Apr 26 11:51:17 EXT: I Apr 26 11:51:17 ADC 0 = 1035, ADC 1 = 0, ADC 2 = d, ADC 3 = 721, ADC 4 = 2a4, ADC 5 = 259, ADC 6 = 3ec, ADC 7 = 0, ADC 8 = a5cd Apr 26 11:51:17 OK Apr 26 11:51:18 also clearly state FR/local vs Far End Apr 26 11:51:21 Has no effect Apr 26 11:51:37 but a cool output ;-) Apr 26 11:51:59 try =1 Apr 26 11:52:07 It's the far end that has low volume which leads to the assumption that mic sensitivity is reduced Apr 26 11:52:20 correct :-))) Apr 26 11:52:45 at%AUL=1 Apr 26 11:52:46 EXT: I Apr 26 11:52:46 ADC 0 = 1035, ADC 1 = 0, ADC 2 = d, ADC 3 = be8, ADC 4 = 74a, ADC 5 = 6dc, ADC 6 = 5b0, ADC 7 = 0, ADC 8 = aca3 Apr 26 11:52:56 the ADC values are always different Apr 26 11:53:04 (current values?) Apr 26 11:53:13 I guess. ADC??? Apr 26 11:53:34 smells like A/D-converter current values Apr 26 11:53:40 yeah Apr 26 11:53:54 tho I don't get the meaning with AUL Apr 26 11:54:22 doesn't make any sense to me Apr 26 11:54:30 seems Dieter is right, it's broken Apr 26 11:55:13 at%AUL already gives this output Apr 26 11:55:26 anything I append to that doesn't seem to have an effect Apr 26 11:55:43 at%AULFOO Apr 26 11:55:43 at%AULFOO Apr 26 11:55:43 EXT: I Apr 26 11:55:43 ADC 0 = 102e, ADC 1 = 0, ADC 2 = d, ADC 3 = 624, ADC 4 = c6a, ADC 5 = ac9, ADC 6 = a11, ADC 7 = 0, ADC 8 = 4be4 Apr 26 11:55:43 OK Apr 26 11:55:46 U checked %AUL is the "load user audio table" command? Apr 26 11:56:33 I'm pretty sure Apr 26 11:56:34 but no Apr 26 11:57:03 have fun with ~20Meg of pdf docs ;-) Apr 26 11:58:35 DocScrutinizer: Seems like it's AT%A Apr 26 11:58:35 EXT: I Apr 26 11:58:35 ADC 0 = e19, ADC 1 = 0, ADC 2 = d, ADC 3 = 6ea, ADC 4 = a25, ADC 5 = 9fc, ADC 6 = 96d, ADC 7 = 0, ADC 8 = 6 Apr 26 11:58:43 As seen in http://wiki.openmoko.org/wiki/Neo_1973_and_Neo_FreeRunner_gsm_modem#AT.25VER Apr 26 11:59:22 which pdf are you reading? Apr 26 11:59:25 AT@AUL Apr 26 12:00:20 AT@AUL=? Apr 26 12:00:20 AT@AUL=? Apr 26 12:00:20 @AUL: (0) Apr 26 12:00:20 OK Apr 26 12:00:20 AT@AUL="0" Apr 26 12:00:20 AT@AUL="0" Apr 26 12:00:22 ERROR Apr 26 12:00:24 [edit] AT@AUL Apr 26 12:00:24 AT@AUL=0 Apr 26 12:00:25 audio table load. Apr 26 12:00:26 AT@AUL=0 Apr 26 12:00:28 ERROR Apr 26 12:00:30 So that's not working Apr 26 12:00:36 :-) Apr 26 12:01:25 okay, so now that that's not working Apr 26 12:01:26 probably because we got no audio_user_table file at all to load Apr 26 12:01:37 probably Apr 26 12:01:43 ...back to TAT Apr 26 12:01:44 is that some NDA'd pdf? ;) Apr 26 12:02:05 TI docs, yes Apr 26 12:02:18 Now recall with gsmhandset already enabled Apr 26 12:02:32 DocScrutinizer: Do you think I need to turn off the modem for that? Apr 26 12:02:40 nope Apr 26 12:02:44 okay Apr 26 12:02:47 let's see Apr 26 12:02:49 since i have audio quality issues too, and i am about to report a bug when i get enough data Apr 26 12:03:34 well /me away for 5min. will have an occasional look to irc Apr 26 12:04:28 'Error: jtag.c:1338 jtag_examine_chain(): JTAG communication failure, check connection, JTAG interface, target power etc.' Apr 26 12:05:16 hmm, both devices dropped from usb bus Apr 26 12:07:46 alphaone: please check control63 in gsmhandset for "mic 2" !! Apr 26 12:08:16 DocScrutinizer: is totally bypassing automatic level control a good thing? Apr 26 12:08:43 yes, it's mandatory for echo cancellation Apr 26 12:09:22 DocScrutinizer: but then you have no automatic level control? Apr 26 12:09:31 DocScrutinizer: That's the one labeled "Mic2" in alsamixer? Apr 26 12:09:35 It's at full power Apr 26 12:09:36 DocScrutinizer: which is not good since it won't adjust to the current environment? Apr 26 12:09:36 also wolfson alc isn't configured at all, so result might be random, possibly leading to the issue U see Apr 26 12:10:33 lindi-: calypso is supposed to do its own alc Apr 26 12:10:58 DocScrutinizer: I'm pretty convinced the mic sensitivity is reproducably low. Apr 26 12:11:01 alphaone: control63 is a muxer Apr 26 12:11:41 name 'Mic Sidetone Mux' Apr 26 12:11:41 value 'Right PGA' Apr 26 12:12:08 WRONG Apr 26 12:12:26 &alue "mic 2"!!! Apr 26 12:12:38 If I select Mic2 it's better Apr 26 12:13:10 rright pga gain undefined Apr 26 12:13:13 thats why Apr 26 12:13:28 But it's still not the way it was Apr 26 12:14:48 okay, will reboot, retest, reflash moko8 and retest again Apr 26 12:14:56 anyway, please use correct statefile to rule out effects from there Apr 26 12:15:14 Will do Apr 26 12:15:28 Will also need to correct all the other statefiles... Apr 26 12:15:37 FSO is currently doing it wrong Apr 26 12:15:51 DocScrutinizer: i'm sure many distros are using "wrong" statefile since they haven't heard of anything better Apr 26 12:16:12 http://docs.openmoko.org/trac/attachment/ticket/2121/gsmhandset.state.new Apr 26 12:16:29 don't use http://docs.openmoko.org/trac/attachment/ticket/2121/gsmhandset-a7.state Apr 26 12:17:11 DocScrutinizer: but it is labeled as being only for A7? Apr 26 12:17:22 forget about that shit Apr 26 12:18:00 a correct statefile should (and will) work on all gta02 Apr 26 12:19:52 you might want to readjust volume and sensitivity to your own taste. But that's not like "this is an A7-only file" Apr 26 12:19:56 * pavo_baxter loves raster Apr 26 12:20:37 pavo_baxter: too late, but you may get invitation to our marriage ;-) Apr 26 12:20:41 DocScrutinizer: i'm just trying to argue why distros are not picking this :) Apr 26 12:21:24 cause of BIIG confusion all over the place wrt statefiles Apr 26 12:21:44 and echo Apr 26 12:21:59 and operation principles of wolfson at large Apr 26 12:22:04 audio at large Apr 26 12:23:52 and Harald made a mistake when defining audiopathes back for gta01, as he didn't consider bad effects of external ALC to calypso AEC, sp path was via left(!) PGA originally (iirc) Apr 26 12:24:31 yup Apr 26 12:24:59 see http://wiki.openmoko.org/wiki/Neo_1973_audio_subsystem#using_phone-internal_microphone_and_speaker Apr 26 12:25:14 okay, so rebooted and mic sensitivity / far end call volume was better with the new settings, but still annoyingly silent Apr 26 12:25:19 evetually it changed to "right PGA" and nobody noticed Apr 26 12:26:07 alphaone: U checked http://wiki.openmoko.org/wiki/Neo_Freerunner_audio_subsystem#Alsamixer_channel_controls ? Apr 26 12:27:12 compare http://people.openmoko.org/joerg/ALSA/doc/WM8753_control_diag_gsmhandset_mic_std.png Apr 26 12:27:18 DocScrutinizer: Will do now, but this still doesn't explain where the differences are if only updating the ti firmware Apr 26 12:27:30 yup Apr 26 12:27:58 :-) okay, will do Apr 26 12:28:25 so I suggest flashing back AND forth from Moko11 to MOK11 and check if it's a random effect or you can relate it to particular moko-version Apr 26 12:28:49 okay, will do that as well Apr 26 12:29:06 If it is I should be able to (un)trigger it at random intervals Apr 26 12:29:19 * DocScrutinizer away now definitely (though I'll take FR with pidgin with me ;) Apr 26 12:30:08 alphaone: is there some test recording I should play through FR in various environments? Apr 26 12:30:20 nope Apr 26 12:30:24 lindi-: Not yet, no Apr 26 12:30:45 have fun DocScrutinizer Apr 26 12:31:00 since my goal is to play it through fr and a nokia phone at the same time in same environment Apr 26 12:31:05 turns out ar6k is fscked up in latest andy tracking (with and without preemption), fails on iwconfig Apr 26 12:31:14 something regarding radio switch Apr 26 12:31:16 jc_: works here Apr 26 12:31:33 lindi-, commit pls ? Apr 26 12:31:33 jc_: b4136a36f31a65d0 Apr 26 12:31:39 lindi-, thanks :) Apr 26 12:31:44 rfkill? Apr 26 12:31:46 jc_: packaking defconfig Apr 26 12:31:51 packaging even Apr 26 12:31:54 lindi-, do you have config handy by chance ? Apr 26 12:31:59 jc_: i did not modify it Apr 26 12:32:01 lindi-, ah... ok :) Apr 26 12:32:10 what do you mean by radio switch? Apr 26 12:32:24 some in kernel rf switch subsystem Apr 26 12:32:36 it oopsed somewhere in it Apr 26 12:32:58 traceback shows numerous ar6k functions Apr 26 12:33:17 does anyone have ar6k datasheet by any chance? (with register descriptions and all that) Apr 26 12:33:22 oh, ok Apr 26 12:33:24 lindi-: 100cm away from device, shout some sentence to someone next room same loudness Apr 26 12:33:32 best testsound Apr 26 12:35:36 DocScrutinizer: but it's not reproducible :) Apr 26 12:36:24 hood enough Apr 26 12:37:51 we're not interested in sub-db diffs. its subjective anyway Apr 26 12:39:08 yuocan tell how loud speaker was shouting by listening to the sample Apr 26 12:41:21 so except for distance to dut its even better than synthesized audio Apr 26 12:42:58 yuo can tell distance as well from reverb but not that accurate Apr 26 12:43:55 bbl Apr 26 12:44:41 hey was there a sysctl change recently for wifi in kernel ? Apr 26 12:44:45 looks like that might be it Apr 26 12:47:41 hhm, seems I like to take N810 and klofisch with me anyway. big screen, hw-kbd, and HSDPA are tempting ;-) Apr 26 12:48:38 freesmartphone.org: 03mickey 07vala-dbus-binding-tool * rd85b7927c3e9 10/.gitignore: add .gitignore Apr 26 12:51:37 usability-wise there's no sweetspot between size of device and size of screen+kbd. N810 is close to this tho Apr 26 13:01:36 lindi-, allright I'm gonna try your commit with slightly altered configuration, wish me luck :) Apr 26 13:01:47 lindi-, did you notice any problems with it ? Apr 26 13:07:47 want to raise a bug on the shr trac, but when I "register" Apr 26 13:08:01 it wants a user/pass to register - what gives? Apr 26 13:23:17 lindi-, hey, wifi works in your commit, looks like something got borked after that or my config was wrong Apr 26 13:24:14 jc_: i'd go for broken config Apr 26 13:24:23 lindi-, yeah, probably Apr 26 13:24:24 jc_: it has WSOD Apr 26 13:24:26 lindi-, thanks :) Apr 26 13:24:34 lindi-, WSOD? whats that ? Apr 26 13:26:09 jc_: white screen of death Apr 26 13:26:23 lindi-, when does it happen ? Apr 26 13:26:36 on resume Apr 26 13:29:32 should debug board v2 work with freerunner using openocd from debian lenny? Apr 26 13:30:03 lindi-, I'll try Apr 26 13:30:29 lindi-, works perfectly fine here Apr 26 13:30:51 lindi-, does it happen every time after resume for ya ? Apr 26 13:30:59 jc_: it is non-deterministic Apr 26 13:31:05 jc_: see the bug report Apr 26 13:31:32 lindi-, aah.. I'll let you guys know if my system ever crashes Apr 26 13:32:25 DocScrutinizer: Okay, the one phone where reflashing moko8, then flashing moko11 worked has a REALLY high sensitivity. Apr 26 13:33:06 control.63 was set to Right PGA and when I set it to Mic2 it was almost unbearably high. Apr 26 13:34:37 DocScrutinizer: and regarding flashing moko11 ontop of moko11 - it doesn't work Apr 26 13:34:48 fluid or calypso will refuse to do anything Apr 26 13:34:51 lindi-, is it this one? http://docs.openmoko.org/trac/ticket/1841 Apr 26 13:35:46 <\marco> hi to all Apr 26 13:36:43 lindi-, hey I actually saw a patch to your exact problem recently that restores all glamo register on resume Apr 26 13:36:45 <\marco> last upgrade seems to has broken my e17 :D Apr 26 13:36:52 lindi-, (somewhere in git log) Apr 26 13:37:16 jc_: nope, much more recent bug Apr 26 13:37:28 <\marco> ( I'm on shr-unstable ) Apr 26 13:37:42 jc_: #2274 Kernel regression: white screen of death reappeared with 2.6.29 Apr 26 13:38:09 2.6.29 has been here for quite a while Apr 26 13:51:48 hello guys, i updated the zenity package at http://www.opkg.org/package_181.html with a finger-scroll feature Apr 26 13:52:13 anyone would like to test it? Apr 26 13:55:01 roby: have you sent the patch upstream? Apr 26 13:55:09 i have the bb recipe and the patches as well, but i'll wait some test before to push it up Apr 26 13:55:16 @lindi not yet Apr 26 13:55:51 * mwester wishes to add the bb file to oe Apr 26 13:57:57 i should send them to Angus Ainslie, correct? Apr 26 14:00:49 Unless they are completely specific to Openmoko, you should send them to the OE mailing list. Apr 26 14:01:55 * lindi- has gdb working with debug board. bugs beware! Apr 26 14:02:36 @mwester ok, i'll subscribe to it Apr 26 14:13:44 Is there any currently accepted way to store PIM data? Apr 26 14:14:17 thrashold: opimd Apr 26 14:14:26 (contacts) Apr 26 14:14:35 'Info: arm7_9_common.c:291 arm7_9_add_breakpoint(): sw breakpoint requested, but software breakpoints not enabled' Apr 26 14:14:38 hmm Apr 26 14:15:00 PaulFertser: bip dead :-/ Apr 26 14:15:53 lindi-: gdb - coool Apr 26 14:16:17 dos11: opimd is part of FSO Framework, right? And it doesn't have calendar support yet? Apr 26 14:16:27 \marco: yes. :( Blame raster. Apr 26 14:16:32 And if I use it, my data would technically be stored system-wide? Apr 26 14:16:48 thrashold: yep. Apr 26 14:17:15 thrashold: if you want to write calendar support, you are welcomed ;) Apr 26 14:17:47 ;) Apr 26 14:19:09 'monitor arm7_9 force_hw_bkpts enable' seems to workaround this Apr 26 14:24:35 lindi-, why software breakpoints ? Apr 26 14:24:46 lindi-, the soc supports debug via jtag quite nicely Apr 26 14:24:50 jc_: 'next' in gdb seemed to like those Apr 26 14:24:51 dunno what the debugboard uses Apr 26 14:25:08 but you can basically singlestep, trap, read memory, reprogram, whatever, make the cpu your bitch over that thing :) Apr 26 14:25:24 jc_: yep except i have to say 'monitor arm7_9 force_hw_bkpts enable' Apr 26 14:25:34 lindi-, ah.. allright then :) Apr 26 14:25:45 lindi-, I think it has a safety mechanism too (the hw debug unit) Apr 26 14:26:04 I think it has to be enabled somehow via some register Apr 26 14:28:01 dos1: what's new? 8) Apr 26 14:30:01 lindi-, looks like watchdog must be off when debugging Apr 26 14:30:07 lindi-, your config has it on Apr 26 14:30:22 alphaone: was offline :-/ Apr 26 14:30:35 Q-Master: SQLite backend tested and fixed :) Apr 26 14:30:45 alphaone: (BIP crashed) Apr 26 14:30:55 lindi-, does stepping and such work? :) Apr 26 14:30:57 bip? Apr 26 14:31:08 IRC-proxy Apr 26 14:31:08 dos1: had you made it default backend? Apr 26 14:31:17 Q-Master: in SHR: yes Apr 26 14:31:17 im here! Apr 26 14:31:18 ;) Apr 26 14:31:27 good Apr 26 14:31:29 DocScrutinizer: Okay, the one phone where reflashing moko8, then flashing moko11 worked has a REALLY high sensitivity. Apr 26 14:31:30 * bzzbot has quit ("Leaving") Apr 26 14:31:30 control.63 was set to Right PGA and when I set it to Mic2 it was almost unbearably high. Apr 26 14:31:30 * bzzbot (n=mwester@209-242-7-139.rev.dls.net) has joined #openmoko-cdevel Apr 26 14:31:30 DocScrutinizer: and regarding flashing moko11 ontop of moko11 - it doesn't work Apr 26 14:31:32 fluid or calypso will refuse to do anything Apr 26 14:31:36 jc_: after that change it does Apr 26 14:31:37 bzzbot: well again? Apr 26 14:31:44 jc_: hmm, i have hardware watchdog running Apr 26 14:31:46 ~hi Apr 26 14:32:09 lindi-, looks like it just stalls watchdog reset signal when cpu halted due to debug but it gets asserted anyway Apr 26 14:32:43 dos1: is opimd functional in SHR? or still in development, with everything still on the SIM for now? Apr 26 14:32:46 alphaone: reflashing same version can be --forced afaik Apr 26 14:33:44 Toaster`: there's no import from vcard script, coz dos1 refuses to make all the opimd names to correspond the vcard3.0 spec. 8) Apr 26 14:33:52 alphaone: still I don't know actual conclusion wrt mic-sensitivity issue Apr 26 14:34:28 Toaster`: contacts can be stored on SIM, CSV or SQLite backend. it misses only editing and deleting contacts Apr 26 14:34:29 alphaone: basic sensitivity setting is controlled by #48 Apr 26 14:34:37 Toaster`: messages can be stored only on SIM ATM. Apr 26 14:34:52 dos1: that sux too Apr 26 14:34:58 ok, cool. Thanks Apr 26 14:35:17 alphaone: push #48 up all the way and adjust #5 as needed Apr 26 14:35:44 bzzbot: stinky robot! Apr 26 14:36:19 ~seen bzzbot Apr 26 14:36:19 i haven't seen 'bzzbot', DocScrutinizer2 Apr 26 14:36:37 I believe THAT ;-D Apr 26 14:36:55 ~uptime Apr 26 14:37:12 okay, this is getting on my nerves... Apr 26 14:37:21 what? Apr 26 14:37:24 I need to build a proper image with proper statefiles Apr 26 14:37:33 lol Apr 26 14:37:39 s/ol/ul/ Apr 26 14:37:39 dos1 meant: lul Apr 26 14:37:40 for a clear reference Apr 26 14:38:03 first prerequisite for decent debugging anyway Apr 26 14:38:47 Q-Master: i will write SQLite-Messages backend, but first I have to understand how Messages domain works ;) Apr 26 14:39:05 dos1: great Apr 26 14:40:36 DocScrutinizer2: Are you sure wrt. #48 all the way up? Apr 26 14:40:47 It seems that introduces quite some moise Apr 26 14:40:50 noise even Apr 26 14:41:24 Q-Master: the problem is - i don't know how to handle incoming message, and store it to some backend instead to SIM Apr 26 14:41:38 alphaone: this should be compensated by setting #5 accordingly Apr 26 14:42:02 dos1: as I remember there was a dbus signal that message had arrived Apr 26 14:42:05 if you get clipping though you need to reduce #48 Apr 26 14:42:15 okay.. Apr 26 14:42:18 Q-Master: i know. SIM backend works on that signal. Apr 26 14:42:40 alphaone: clipping means audio is distorted even when setting #5 to very low levels Apr 26 14:42:42 dos1: so fetch the sms and put it to sqlite db removing from sim Apr 26 14:42:45 Q-Master: but how to store it by my backend, instead of SIM, as ogsmd do? Apr 26 14:43:12 DocScrutinizer2: No, doesn't seem to be the case Apr 26 14:43:14 Q-Master: that's not good way. imagine situation, when you have few backends Apr 26 14:43:36 alphaone: what doesn't seem to be tha case? Apr 26 14:43:43 clipping Apr 26 14:43:44 SQLite, MySQL, CVS... Apr 26 14:43:46 dos1: so? fetch. process by all backends needed. remove Apr 26 14:43:54 transactions roxx Apr 26 14:43:57 8) Apr 26 14:44:28 alphaone: so you *should* get _less_ noise when going up with #48 and going down with #5 same ammount Apr 26 14:44:31 Q-Master: oh... incomind message stored by default backend? Apr 26 14:44:33 dos1: better selecting 1 backend working in config. Apr 26 14:44:53 alphaone: that's the theory ;-) Apr 26 14:44:55 dos1: yep. that will be good desision Apr 26 14:45:27 DocScrutinizer2: Yeah, well it's hard to verify with the signal going through AEC/NR Apr 26 14:45:39 Q-Master: yep. backend in config = default backend Apr 26 14:45:39 ;) Apr 26 14:45:40 sure Apr 26 14:46:05 dos1: yep. that's wise. Apr 26 14:46:21 so - fetch, put, removw Apr 26 14:46:22 e Apr 26 14:47:12 dos1: my SE has an option of storing some smses to sim-card Apr 26 14:47:58 alphaone: another more precise SOP to set #48 is to adjust PGA to 0dB gain and then record from mic with arecord -vvv. Then adjust #48 for -12db .. -6dB on alsarecord while shouting loud close to mic Apr 26 14:48:56 it's difficult as there's quite some leveling in PGA as well as digitally prior to ADC Apr 26 14:49:10 hmm Apr 26 14:49:38 alphaone: the basic idea is to adjust #48 for -6dB in analog path of wolfson at maximum audio level Apr 26 14:50:00 then adjust #5 to accomodate calypso Apr 26 14:51:34 since you're talking about sound, guys Apr 26 14:52:22 do you think connecting a preamp with high input resistance to FR headphone out would get a reasonable sound out of it ? Apr 26 14:52:26 Well, my primary concern at the moment is not sending people buzz fixed but unusable Freerunner... Apr 26 14:52:37 jc_: sure Apr 26 14:53:05 alphaone: nobody complained so far Apr 26 14:53:15 hmm Apr 26 14:53:19 I did :-) Apr 26 14:53:20 the reworked devices we tested in TPE were ok Apr 26 14:53:31 DocMobilizer, so basically, just a small little box with an opamp, input and phone out would be ok ? :) Apr 26 14:53:42 does buzz fixing imply updating the calypso fw to moko11? Apr 26 14:54:01 tilman: not mandatory Apr 26 14:54:04 tilman: no Apr 26 14:54:09 its a shame though music sucks out of the box :/ Apr 26 14:54:17 alphaone: but you were planing on doing both? Apr 26 14:54:19 But updating the fw is a good idea in any case Apr 26 14:54:28 tilman: Until now I thought so, yeah Apr 26 14:54:33 * tilman looks for the changelog for moko11 Apr 26 14:54:41 jc_: sorry I couldn't move "the powers that be" to fix this Apr 26 14:54:45 tilman: Fixes some problem with lost calls Apr 26 14:55:21 ah Apr 26 14:55:24 anyway, I have to join family live again Apr 26 14:55:29 cya l8er Apr 26 14:55:35 bye Apr 26 14:55:43 DocScrutinizer, bye, thanks for info :) Apr 26 14:56:53 jc_: is there some way to have gdb break on kernel oops other than 'break __do_kernel_fault'? Apr 26 14:57:20 lindi-, I think so, but it would be literally on kernel oops Apr 26 14:57:37 lindi-, just put a breakpoint into the function that outputs oops :D Apr 26 14:58:44 (I'd grep the sources for the printf message you see, find function it is in, look for it in disassembly and put a breakpoint there) Apr 26 14:59:13 note I have *no* experience debugging linux whatsoever, it may have much better mechanisms for this Apr 26 14:59:49 jc_: the only problem with that is that you need to hit the oops twice :) Apr 26 15:00:02 if the think you're debugging does some ARM fault interrupt, thats easy then, just put breakpoints on all fault interrupt vectors Apr 26 15:00:27 lindi-, then 'just' pass the first when it occurs :) Apr 26 15:00:39 if it doesn't mind the little delay Apr 26 15:01:48 lindi-, what are you debugging ? Apr 26 15:02:03 jc_: thought http://docs.openmoko.org/trac/ticket//2240 would be something easy to start with Apr 26 15:02:43 lindi-, do you understand the subsystem in question ? Apr 26 15:02:49 USB gadgets Apr 26 15:03:22 jc_: somewhat Apr 26 15:03:42 i've struggled with gadgetfs bugs in the past (when i did not have debug board) Apr 26 15:04:49 lindi-, it probably uses the socs own usb host/device stuff, I'd also read the sections in datasheet Apr 26 15:05:06 jc_: yep read those Apr 26 15:05:12 some trace would be nice :) Apr 26 15:05:29 * lindi- hits 'submit changes' button -- now you have backtrace Apr 26 15:06:46 aaahj Apr 26 15:06:50 thats nice Apr 26 15:06:54 thas misalignment Apr 26 15:07:33 lindi-, when does it happen ? Apr 26 15:08:41 I see thumb bit disabled, don't some routines in kernel use thumb (especially fast memory stuff) ? Apr 26 15:08:53 nope Apr 26 15:09:41 shame it only gives virtual address Apr 26 15:09:53 otherwise we could grep disassembly of kernel and find where it occured Apr 26 15:10:48 jc_: on ifup usb0 Apr 26 15:10:59 lindi-, after module reload ? Apr 26 15:11:03 jc_: yep Apr 26 15:11:25 jc_: modprobe does not cause it Apr 26 15:11:35 dev_get_stats does Apr 26 15:11:41 noo thats LR Apr 26 15:11:45 thats where it came from Apr 26 15:11:59 it branched somewhere bad, didn't it ? Apr 26 15:12:16 oh yeah Apr 26 15:12:18 [ 1494.355000] Unable to handle kernel paging request at virtual address 2e780330 Apr 26 15:12:26 PC is at 0x2e780330 Apr 26 15:12:30 jc_: looks like it Apr 26 15:12:40 * lindi- reads openocd docs.. Apr 26 15:12:43 in LR usually is address it branched from if its a call Apr 26 15:12:57 +/- 4 or so because of prefetch Apr 26 15:13:03 I'm bad for those counts Apr 26 15:13:22 dev_get_stats branched somewhere it shouldn't have Apr 26 15:13:41 maybe it called function it thought was supposed to be there but turns out it isn Apr 26 15:13:42 t Apr 26 15:13:45 so .. bad pointer ? Apr 26 15:13:45 jc_: can prefetch cause fault? Apr 26 15:14:21 lindi-, I don't think so Apr 26 15:14:42 it prefetches next address, if it finds crap, it puts crap in cache and tries to execute it if you don't branch before that Apr 26 15:15:00 The kernel uses indirect function calls throughout. Apr 26 15:15:12 yeah, looks like indirect call somewhere bad Apr 26 15:15:22 possibly even direct Apr 26 15:15:35 the code might have thought something was supposed to be there at compile time Apr 26 15:15:39 Whenever you have a jump to a location that is outside the kernel space, it's probably because something did not fill in, or overwrote a structure associated with the device. Apr 26 15:16:00 is 2e780330 out of kernel space ? Apr 26 15:16:09 Not likely compile time because it worked at one point; it fails when you re-load the module. Apr 26 15:16:12 I'm not sure since it is a virtual address, whatevr could be there Apr 26 15:16:15 Yes, that's outside kernel space Apr 26 15:16:33 IIRC, virtual is mapped one to one for the kernel. Apr 26 15:16:46 so virtual = physical ? Apr 26 15:17:02 in that case, lemme check where it is Apr 26 15:17:44 0x2e780330... hmm... Apr 26 15:18:09 thats definitely a bad physical address Apr 26 15:18:14 thats before first SDRAM chip Apr 26 15:18:19 in a unmapped region Apr 26 15:18:22 = bad pointer Apr 26 15:18:27 something got overwritten Apr 26 15:18:33 clearly Apr 26 15:19:27 would be nice to see source for dev_get_stats Apr 26 15:19:40 I don't have resources for this here right now, unfortunately Apr 26 15:19:52 just git clone git://git.openmoko.org/git/kernel.git neolinux= Apr 26 15:20:00 takes a long time :D Apr 26 15:20:28 aah disregard that, I actually have it on my phone still Apr 26 15:21:18 Basically, to make it easier -- references to c0xxxxxx are in kernel space, generally. the kernel will not call outside of that address range. Apr 26 15:21:28 mwester, when disassembling kernel image, do I get adresses the same as they end up in physical space after uncompressing/preparing/whatever ? Apr 26 15:21:30 (that's on the gta0x) Apr 26 15:21:41 jc_: yep Apr 26 15:21:54 basically, yes. Look at the System.map file -- there is a one-to-one mapping. Apr 26 15:21:59 ok Apr 26 15:22:14 jc_: the function has just if (ops->ndo_get_stats) return ops->ndo_get_stats(dev); else return &dev->stats; Apr 26 15:22:18 lindi-, could you provide your map? just to make sure I'm in the right function Apr 26 15:22:19 At compile time, the physical and virtual addr for the kernel must be known. Apr 26 15:22:33 ops->ndo_get_stats(dev); Apr 26 15:22:39 thats a clear indirect branch Apr 26 15:22:46 ops structure could be incorrect Apr 26 15:23:00 I'd add a check there Apr 26 15:23:09 if its outta range, don't jump there and return some null stats Apr 26 15:23:12 what about that ? Apr 26 15:23:12 jc_: http://iki.fi/lindi/tmp/{vmlinux,System.map} Apr 26 15:23:15 Yep, the ops structure is bogus. Apr 26 15:23:31 does it have some checksum or something ? Apr 26 15:23:41 any way you could tell its bad Apr 26 15:23:52 So the question is, who is responsible for filling that in --- and since it works once, who *was* responsible for clearing it out or marking it invalid when the module was rmmod'd? Apr 26 15:23:59 checking if address is in ram doesn't seem so right to me, too device specific Apr 26 15:24:14 jc_: no, there are no checksums ; that would slow the kernel terribly. Apr 26 15:24:28 mwester, is this done only on insert/removal ? Apr 26 15:24:51 there are no checks at all, just that it is non-zero. And even that test is not for safety, that test is usually to handle the case where a driver does not support a function. Apr 26 15:25:07 There kernel implicitly trusts that most structures are correct. Apr 26 15:25:15 its easy to crash kernel, I see :) Apr 26 15:25:25 only root can do that anyway Apr 26 15:25:25 yes, a bad driver can easily damage the kernel. Apr 26 15:25:31 That's why only root can load modules. Apr 26 15:25:49 yeah so now, I'd look at who fills that structure Apr 26 15:25:51 That's also a valid technical objection to binary kernel modules... Apr 26 15:25:55 soo lets look at traceback Apr 26 15:26:22 rtnl_fill_ifinfo Apr 26 15:28:40 lindi-: you hooked me with that git cmdline ;-) Apr 26 15:28:46 looks like if you disabled netlink support, you could avoid it, couldn't you ? Apr 26 15:29:01 git is really nice (if I need to c&p a cmdline) ;D Apr 26 15:30:07 woohooo, 3% done already Apr 26 15:30:41 * DocScrutinizer2 axiously doing a `df -h` Apr 26 15:31:00 lindi-, I don't like this problem, it looks like it roots in some bogus call over a netlink socket Apr 26 15:31:20 the traceback isn't complete enough to tell the source unfortunately Apr 26 15:31:22 jc_: perhaps; but often problems like this travel in packs -- there may be other uninit/misinit structs as well. Apr 26 15:31:46 couldn't netlink be debugged ? Apr 26 15:32:01 you mean, like provide us with who sends what over netlink right before the crash Apr 26 15:32:26 ifdown usb0 calls dev_get_stats. netdev_ops = 0xc02cb8c4, Apr 26 15:32:32 I think the better approach would be to add printfs to the logic involved in modprobing and rmmoding g_ether. Apr 26 15:32:39 Something has gone wrong there Apr 26 15:33:07 dev->netdev_ops->ndo_get_stats = 0xc01b0dc0 , Apr 26 15:33:22 maybe its some function that notices iface has gone up and sends a new status, it might not have to do with modprobe/whatever Apr 26 15:33:28 we can't tell for sure Apr 26 15:34:03 hmm, i forgot to disable the watchdog Apr 26 15:34:06 but yeah, looking at that is a good start :) Apr 26 15:34:14 lindi-, watchdog doesnt corrupt memory Apr 26 15:34:18 it'd reboot your system Apr 26 15:34:19 but it reboots this :P Apr 26 15:34:24 yeah :D Apr 26 15:34:29 lol Apr 26 15:34:34 memory corrupting watchdog Apr 26 15:34:46 woof, I don't like that pointer in your mobile phone Apr 26 15:37:37 anyway, I'm gonna have to get shower now, lindi- lemme know if you find out who did that netlink call Apr 26 15:42:01 dos1: Do you mind if I take a stab at closing the gprs TODO in SHR-Settings? The one regarding persisting login data? Apr 26 15:43:18 Toaster`: go ahead!!!! :-)) Apr 26 15:44:01 Ok, but I don't have GPRS on my phone, anyone out there willing to test things in a little bit, maybe tomorrow? Apr 26 15:44:13 I'll be testing GPRS soon Apr 26 15:44:18 once my system is working Apr 26 15:44:38 now I'm focused at some ar6k bugs Apr 26 15:49:53 goddamit how can my kernel with preemption have 10ms ping latency when just untarring stuff Apr 26 15:50:00 some fscked up bus master or something ? Apr 26 15:50:06 anyway, off to shower already :) Apr 26 15:56:25 jc_: spinlock? Apr 26 16:05:48 DocScrutinizer2, no idea :) I wonder if its hard to profile the kernel with the debug board Apr 26 16:07:24 Is there an easy way to set my own OE buildhost as a opkg repo? Apr 26 16:08:07 I've build abiword and I want to give it a try, but don't want to install all the dependencies manually :P Apr 26 16:08:14 Sharwin_F, i think it should be matter of adding your repo to list of opkg repos config, refer to wiki, I'm sure it was somewhere there Apr 26 16:09:05 jc_, but, I'm suppoused to do something on the directory I have all the packages? ie, setup and http server Apr 26 16:09:28 Sharwin_F, of course, your phone needs to get into that dir somehow Apr 26 16:09:49 I took it as if you already made a repo Apr 26 16:10:14 jc_, nop, that's the part I don' know how to do :P Apr 26 16:10:28 jc_, is there some info in the wiki about it too? Apr 26 16:11:54 Sharwin_F, doesnt look like it, just info on creating packages Apr 26 16:11:57 I'd look at man opkg Apr 26 16:12:36 jc_, ok Apr 26 16:13:05 I did a fast look on google some days ago and didn't find so much info Apr 26 16:13:47 look in the contrib dir in oe Apr 26 16:14:15 you should find a script that will create a package index and copy the ipks to a dir; then setup your web server to serve that dir. Apr 26 16:14:21 then add it to the feeds. Apr 26 16:14:52 ok, thanks :) Apr 26 16:16:56 mwester, I'll wait till there's a new image to flash on build-shr-project.org though, because I'll probably finish destroying the image I have on my fr now xD Apr 26 16:57:56 With the framework dbus, if a reply_handler is set, the returns from the method are passed to that function? Apr 26 16:59:53 yes, that's how async dbus works. Apr 26 17:00:04 use error_handler as well Apr 26 17:00:14 ok, I wasn't sure. Thanks Apr 26 17:01:30 the gsm chip audio input and output is connected to the wolfson codec right, in gta02? Apr 26 17:01:41 I mean, the gsm chip does not drive the speakers itself?;) Apr 26 17:02:00 correct Apr 26 17:02:26 thanks Apr 26 17:03:04 so, is it gsmd that supports alsa directly? Apr 26 17:05:22 or you route it directly giving instructions to the wolfson codec? Apr 26 17:05:29 the latter Apr 26 17:05:54 well, this prevents alot of juicy things Apr 26 17:06:11 why not let the gsmd support SIP? Apr 26 17:07:42 b0ef: how do you let it support SIP? Apr 26 17:07:47 freesmartphone.org: 03mickey 07vala-terminal * r5d14c542cfa3 10/ (.gitignore autogen.sh configure.ac): use standard autogen.sh; check for vala version Apr 26 17:10:26 lindi-: well, maybe best is to just write a module for a SIP softswitch, such as freeswitch, to be able to get the streams from the gsm chip Apr 26 17:13:28 b0ef: i'm afraid that's not going to work very well :/ Apr 26 17:13:39 freesmartphone.org: 03mickey 07dbus-hlid * r0844f3eda298 10/ (.gitignore Makefile.am configure.ac src/obj.vala): remove warnings, check vala version Apr 26 17:14:24 b0ef: a good intermediary goal would be to let me use my usb headset with gsm Apr 26 17:15:02 lindi-: well, that is a audio stream routing problem, solved with JACK and pulseaudio Apr 26 17:16:59 b0ef: please try if you can make it even work Apr 26 17:17:36 freesmartphone.org: 03mickey 07libgsm0710 * r3bb80ba6cc9f 10/README: readme++ Apr 26 17:18:25 b0ef: 1) capture: arecord -D hw:0,0 -t raw -r 48000 -f S16_LE -c 2 - | pv > /dev/null Apr 26 17:19:08 b0ef: 2) playback: pv /dev/urandom | aplay -D hw:0,0 -t raw -r 8000 -f S16_LE -c 2 Apr 26 17:19:18 b0ef: can you test if you can run these both at the same time on your freerunner? Apr 26 17:21:01 lindi-: not sure I understand; you mean there is a hardware problem? Apr 26 17:21:32 b0ef: just try the above please Apr 26 17:21:39 lindi-: right Apr 26 17:21:54 b0ef: my freerunner is under JTAG right now, can't test Apr 26 17:23:03 b0ef: i think it lagged a lot last time i tried Apr 26 17:23:53 MarcOChapeau: how does JACK on freerunner go?;) Apr 26 17:25:22 b0ef: try avoiding pulseaudio at all means Apr 26 17:25:51 we can do all these fancy things on other platforms Apr 26 17:25:56 on the FR you can forget about it Apr 26 17:26:01 as it's sincerely lacking horsepower Apr 26 17:26:08 yup Apr 26 17:26:10 DocScrutinizer: no, I don't agree here, pulseaudio ran extremely fine after some tuning on the gta02 Apr 26 17:26:21 and PA is broken by design anyway Apr 26 17:26:44 last time i tried PA used 30% CPU for mixing zeros (read: in idle mode) Apr 26 17:27:00 mickeyl: well, then you didn't tune it Apr 26 17:27:03 PA is a realtime solution and designed without care about CPU hogging Apr 26 17:27:08 probably Apr 26 17:27:17 did you rewrite the mixer in arm asm? Apr 26 17:27:20 mickeyl: it also did that with me, but after I did something, which I can look up, it ran perfectly Apr 26 17:27:30 no;) Apr 26 17:27:51 i sincerely doubt that, but please prove me wrong by disclosing your magic tuning Apr 26 17:27:56 b0ef: it goes, but without optimization Apr 26 17:28:05 b0ef: still it goes :) Apr 26 17:28:18 I simply don't see the benefits of PA, other than changing audiodevices while stream open Apr 26 17:28:25 mickeyl: right, I'll look it up Apr 26 17:28:39 well, that's exactly a need Apr 26 17:28:50 we connect a bt headset and we want the audio to go there Apr 26 17:29:11 MarcOChapeau: bad in any way? Apr 26 17:29:13 ;) Apr 26 17:30:11 * MarcOChapeau wonders if jack could be a solution for the FR Apr 26 17:30:27 after all audio is critical on a phone Apr 26 17:30:51 jack + a rt kernel and some well written audio apps might work Apr 26 17:30:51 it sure is the final solution to audio together with pulseaudio Apr 26 17:30:58 defintely;) Apr 26 17:31:01 b0ef: well this is no problem, except if you want to do that (BT connection) on the fly while playback of sound. Sounds silly to me Apr 26 17:31:42 wtf?? rt-kernel? lol Apr 26 17:31:52 DocScrutinizer: yeah, but most of the times, a bug is found precisely when a user did something silly... Apr 26 17:31:55 :p Apr 26 17:32:07 b0ef: well, BT headset works already Apr 26 17:32:10 DocScrutinizer: what about the rt kernel ? Apr 26 17:32:10 well, you connect a pair of USB ghetto blasters to the freerunner, you want the audio to play there Apr 26 17:32:28 so what? Apr 26 17:32:29 and you still can use your bt headset to make a call while the ghetto blaster plays Apr 26 17:32:32 no problem Apr 26 17:33:09 (as long as you use supported bt headset) Apr 26 17:34:05 or you connect 2 bt headsets Apr 26 17:34:07 freesmartphone.org: 03mickey 07libeflvala * r4b7440b08b96 10/vapi/ecore.vapi: ecore.vapi: bind new function "EAPI double ecore_timer_interval_get(Ecore_Timer *timer);" Apr 26 17:34:11 one for your girl and one for you Apr 26 17:34:30 b0ef: then you fail :) Apr 26 17:34:33 b0ef: well, then don't get a girl... Apr 26 17:34:35 just if you want to switch from mp3-playback via builtin speaker to playback via BT headset, the mediaplayer has to open another audio device. That's sane anyway Apr 26 17:34:48 well, we want the final solution Apr 26 17:34:58 something that can scale and work for every possible scenario Apr 26 17:35:05 that's jack and pulseaudio Apr 26 17:35:13 pff Apr 26 17:35:16 MarcOChapeau: rt kernel doesn't make much sense for an ordinary freerunner user. Though low-latency-can-preempt-anywhere kernel would be good for multimedia purposes, it's not strictly real-time kernel anyway. Apr 26 17:35:19 go ahead Apr 26 17:35:32 * DocScrutinizer away for more interesting stuff Apr 26 17:36:22 PaulFertser: I'm pretty much aware of that. although, rt-kernel pretty much makes sense to me as far as the phone audio stream is concerned Apr 26 17:36:47 b0ef: please try my two commands Apr 26 17:37:00 lindi-: yes, I will Apr 26 17:37:16 b0ef: since i tried this but I think it was laggy, crashed kernel Apr 26 17:37:19 MarcOChapeau: do you know any non-experimental multimedia programs utilizing real-time capabilities of RTAI/Xenomai/etc? Apr 26 17:37:24 , but as MarcOChapeau ran jack, this should not be a problem Apr 26 17:38:45 lindi-: are you sure concurrent capture and playback with different rates is/should be supported? Apr 26 17:38:58 different rates? Apr 26 17:38:58 PaulFertser: actually I don't use RTAI/Xenomai Apr 26 17:39:06 we don't do different rates.. Apr 26 17:39:32 PaulFertser: you want your browsing of your contact list to interrupt your audio? Apr 26 17:40:02 PaulFertser: in fact it's the first time I hear of those two project... given the fact that I've been among the linux audio community for a while now, I'd say people that work with audio don't use these Apr 26 17:40:08 PaulFertser: i think they only work with different rates Apr 26 17:40:16 b0ef: Have you heard of those ? Apr 26 17:40:25 b0ef: RTAI/Xenomai Apr 26 17:40:27 PaulFertser: since my usb headset only does 48000 Apr 26 17:40:37 MarcOChapeau: yeah, I've heard of those, but that's some completely different beasts Apr 26 17:40:39 PaulFertser: and 44100 Apr 26 17:40:44 MarcOChapeau: so, what RT-kernels do audio people use? Apr 26 17:40:56 PaulFertser: patch from ingo molnar Apr 26 17:41:11 wow... I feel weird just saying it Apr 26 17:41:14 MarcOChapeau: it's not RT iirc, it's just heavy preemption everywhere. Apr 26 17:41:46 PaulFertser: well, it's not guaranteed, but pretty dependable Apr 26 17:41:57 lindi-: "-D hw:0,0 -t raw -r 8000" and "-D hw:0,0 -t raw -r 48000" ... I might be misunderstanding something... Apr 26 17:42:01 PaulFertser: ok, I know it's not RT... that's just it's nick name. although, it does work for heavy duty digital audio workstations Apr 26 17:43:06 realtime preemption is right for any audio application, so what the problem is?;) Apr 26 17:43:08 MarcOChapeau: is patching still necessary with modern kernels? Apr 26 17:43:23 no, just for extremely low latencies Apr 26 17:43:28 PaulFertser: for the hard style preemption, yeah Apr 26 17:43:36 just make sure full realtime preemption is enabled in the kernel Apr 26 17:43:38 MarcOChapeau: got a link? Apr 26 17:43:45 sure, hold on Apr 26 17:43:52 http://www.kernel.org/pub/linux/kernel/projects/rt/ Apr 26 17:43:54 MarcOChapeau: ;) Apr 26 17:44:05 http://www.kernel.org/pub/linux/kernel/projects/rt/ Apr 26 17:44:07 damn Apr 26 17:44:10 :-p Apr 26 17:44:11 8) Apr 26 17:44:42 b0ef: I'm sure you cheated, you must have a bookmark :) Apr 26 17:45:04 nah, just typed mingo realtime into google;) Apr 26 17:45:14 been doing it for years..;) Apr 26 17:45:17 b0ef: MarcOChapeau: no description, just raw patch? ;) Apr 26 17:45:32 I'll find it don't worry. Apr 26 17:45:36 PaulFertser: what ? you need a readme ? :D Apr 26 17:45:57 PaulFertser: actually there's a wiki somewhere on earth about this patch Apr 26 17:45:57 is there some way to get module relocation info via gdb? I can't access /sys since the kernel has crashed Apr 26 17:46:12 (I can't get symbols to loaded modules without that) Apr 26 17:46:19 PaulFertser: http://rt.wiki.kernel.org/index.php/RT_PREEMPT_HOWTO Apr 26 17:46:29 PaulFertser: shame on you. Which Russian ever read docs? 8) Apr 26 17:46:54 Q-Master: uhm, smart ones? Apr 26 17:47:01 lol Apr 26 17:47:05 8)) Apr 26 17:48:47 MarcOChapeau: nice link, thanks a lot! Apr 26 17:49:21 np Apr 26 17:51:57 PaulFertser: jackaudio.org might also be interesting;) Apr 26 17:52:48 b0ef: Thanks for this link too :) Apr 26 17:54:28 actually, that would be trac.jackaudio.org for the current docs Apr 26 17:54:47 we need to advertise this link on jack's front page... Apr 26 17:55:13 lindi-: you're using different rates in those commands Apr 26 17:55:20 Anyone out there who uses SHR-unstable and gprs will to be a guniea pig for a little bit? I lack gprs and would like to test my changes to SHR-Settings. Apr 26 17:55:34 b0ef: yep Apr 26 17:55:43 b0ef: can you make it work with 8000 only? Apr 26 17:56:14 lindi-: haven't tried it yet, but thinking about why you want to do that Apr 26 17:58:02 b0ef: well, my usb headset only does 44100 or 48000 Apr 26 17:58:50 hmm, yeah, and the GSM IO is probably 8k Apr 26 17:58:55 lindi-: but you specify hw:0,0 in your command Apr 26 17:59:23 b0ef: GSM modem is not connected by digital sound interfaces. Apr 26 18:00:05 PaulFertser: ?, the gsm modem outputs analog audio to the wolfson audio chip, doesn't it? Apr 26 18:00:25 b0ef: analog, yes. So you can capture it with any rate wolfson and SoC support. Apr 26 18:01:09 PaulFertser: right, but you don't want to sample at 48k there maybe Apr 26 18:02:05 b0ef: unless you want to output something at 48k through I2S interface of SoC, then you won't have any options, rates must match. Apr 26 18:03:12 PaulFertser: right, so we just fire up jack at 48k, but who does the sample rate conversion? Apr 26 18:03:22 ah, right Apr 26 18:03:38 so the gsm chip just samples this analog audio stream again on its input Apr 26 18:03:46 b0ef: hehe :) Apr 26 18:04:20 right, so then we just open the chip at 48k Apr 26 18:04:45 s/open/start jack on/ Apr 26 18:04:47 b0ef meant: right, so then we just start jack on the chip at 48k Apr 26 18:05:05 If there's enough computational/whatever power to do that, why not... Apr 26 18:05:32 MarcOChapeau: well, you're sort of the team leader on this one;) Apr 26 18:07:25 b0ef: but if you want to use a bluetooth headset via PCM interface of the bluetooth chip, then hw:0,1 must be configured with 8000 rate. And i'm not sure it can be correctly done while hw:0,0 is still open (in case we're lucky hw:0,0 will just stop transfering bytes while we're in bluetooth-compatible DAI mode). Apr 26 18:09:27 PaulFertser: not sure what you mean Apr 26 18:10:05 b0ef: check Neo1973_Audio_Subsystem wiki page to see about routing and DAIs. Apr 26 18:10:20 I might need to do that too Apr 26 18:10:22 ... Apr 26 18:10:37 http://wiki.openmoko.org/wiki/Neo_1973_audio_subsystem Apr 26 18:11:55 OMG, that's gonna hurt... Apr 26 18:13:21 hold on, this is specific to the neo1973 ? Apr 26 18:14:09 With DAI 0 mode it should be possible to simultaneously use different rates on SoC DAI and BT DAI but i can't guarantee it'll work without any problems... Somebody needs to try all that fancy things to be sure. Apr 26 18:14:18 and anyways, playing with jack with the way BT needs to be configured is going to be painful Apr 26 18:14:42 running jack on 2 different interfaces and making them talk is hell Apr 26 18:15:01 maybe one day with jack 2, but we're not there yet Apr 26 18:15:03 MarcOChapeau: no, the codec is the same on both neo1973 and FR, but channel names might differ slightly. The diagram with red channel names is for freerunner. Apr 26 18:15:28 hmm, so 0,1 must be 8k Apr 26 18:15:44 b0ef: do you see where this hurts ? Apr 26 18:16:10 jack + multiple sound card... hehe, it's been a while since I saw a use case Apr 26 18:17:26 well, do we even know if this is possible Apr 26 18:18:10 the wolfson codec must do resampling then Apr 26 18:18:15 BTW, i personally tried to switch bluetooth chip to route SCO audio through USB interface and succeeded. Apr 26 18:20:10 PaulFertser: not sure what you mean; you were able to get audio into the bluetooth chip without using the audio input on the bt chip? Apr 26 18:20:39 wow, I'm confused now;) Apr 26 18:20:53 I need to read a little on the specific chips;) Apr 26 18:20:56 b0ef: Yep Apr 26 18:21:12 well, if that's the case, then it's no problem;) Apr 26 18:21:33 b0ef: It's always doable with A2DP, but not that easy with SCO and only SCO supports recording from mic. Apr 26 18:21:46 I guess this way we'll get more latency though. Apr 26 18:22:24 recording directly from mic is not ideal, though Apr 26 18:22:47 sure, it's more latency, but hopefully not too much Apr 26 18:23:26 b0ef: and it'll allow recording of GSM calls made with the headset. Apr 26 18:24:32 right, I think every GSM/VoIP call should go through a softswitch, like freeswitch, so recording and arbitrary routing is easy Apr 26 18:25:25 you might want to transfer a GSM call to your desktop via SIP, f.ex Apr 26 18:25:33 'arm7_9_common.c:297 arm7_9_add_breakpoint(): no watchpoint unit available for hardware breakpoint' Apr 26 18:25:38 hmm, i run out of hardware breakpoints Apr 26 18:26:04 b0ef: switching SCO to USB is somewhat permanent, to temporarily switch back to dedicated audio interface one has to use some special command to route the next SCO through DAI. Nobody tried to use it yet, it seems. Apr 26 18:31:12 b0ef: this setting is stored inside BT chip eeprom. Apr 26 18:34:00 mwester: any tips on how to get more breakpoints with gdb and debug board? Apr 26 18:35:09 Is there anyone out there with GPRS who can help me test my changes? Apr 26 18:36:26 lindi-: do software breakpoints work for you? When the code is in RAM i think they should be available. Haven't tried that though, just a wild guess :-/ Apr 26 18:36:34 PaulFertser: how should I use those? Apr 26 18:36:51 PaulFertser: boot loader is going to overwrite those on boot at least Apr 26 18:37:06 i'd like to know how many hardware breakpoints this supports Apr 26 18:38:35 lindi-: this http://www.shakthimaan.com/installs/freerunner-debug-setup.html has "arm7_9 sw_bkpts enable" openocd command... Apr 26 18:39:21 freesmartphone.org: 03mickey 07libgsm0710mux * r08261d574acb 10/ (Makefile.am gsm0710mux-1.0.deps): add gsm0710mux-1.0.deps Apr 26 18:44:09 Toaster`: do you still need tester for GPRS changes? Apr 26 18:45:58 Someone know how the RAM is attached to the GTA02? Apr 26 18:46:28 iirc it's always one nibble internal, one nibble external Apr 26 18:46:55 dos1|away: yes, everything looks good on my end, I just can't verify I haven't broken anything. Apr 26 18:47:01 mwester: booting image from scratch now Apr 26 18:47:46 Toaster`: so i can be a guinea pig ;) Apr 26 18:48:03 i have packet for GPRS, so i can test it :) Apr 26 18:48:17 dos1: perfect. Shall I send you a diff or just commit the changes? Apr 26 18:48:46 Toaster`: as you want. what did you change? storing values? how did you do that? Apr 26 18:50:23 mwester-laptop: what problems did you encounter with your image? Apr 26 18:50:32 what should I check? Apr 26 18:51:17 dos1: I did several changes (as ususal). I cleaned up the createView code and added a couple classes to simplify the gui code. Persistence is using /etc/freesmartphone/persist/gprs.pickle, similar to the ogpsd data. I also got rid of the timer; I replaced it with a dbus signal call which seems to work well. Apr 26 18:51:39 Toaster`: commit :) Apr 26 18:51:48 looks good Apr 26 18:52:05 SHR: 03frazier.cameron 07shr-settings * r558629247939 10/shr_settings_modules/shr_gprs.py: [GPRS] Addition of persistent apn, login and password info. Data is stored in /etc/freesmartphone/persist/gprs.pickle. Not fully tested as I lack GPRS. Apr 26 18:52:33 The persistence is only triggered on a successful connection. Apr 26 18:52:40 mrmoku: only get a single theme available at first boot. Apr 26 18:52:51 mwester-laptop: I had both themes Apr 26 18:53:03 looks all good Apr 26 18:53:11 I committed another patch to the efl bb stuff -- it was not packaging some of the ecore libs. Not sure how that could have been missed. Apr 26 18:53:23 (unless nobody actually builds from scratch anymore) Apr 26 18:53:41 I built from scratch last night... No problems I've seen so far Apr 26 18:53:52 Right. I added the patch. Apr 26 18:53:56 mwester-laptop: guess you fixed it :P Apr 26 18:54:01 No, I hacked it. Apr 26 18:55:01 mwester-laptop: btw... glibc did not build without limiting locales Apr 26 18:55:09 does it build for anybody without limiting locales? Apr 26 18:55:11 I suspect the right fix is to fix the bb recipe in OE, but after this last go-round in OE with the E recicpes, I'm not touching those recipes! Apr 26 18:55:20 mrmoku: I build no locales at all. Apr 26 18:55:38 mrmoku: I only built with the one locale. Apr 26 18:55:52 Toaster`: dbus is borked... Apr 26 18:55:58 Toaster`: worked at first run Apr 26 18:56:12 Toaster`: i changed value, and now "dbus is borked" Apr 26 18:56:21 the "dbus is borked" error is a catchall Apr 26 18:56:29 Toaster`: debuging now Apr 26 18:57:34 dos1: Probably something foolish on my part Apr 26 18:58:22 hmm, i probably need some magic to break at module initialization code Apr 26 18:58:58 freesmartphone.org: 03mickey 07libgsm0710mux * rca6ff8b9e0e0 10/ (Makefile.am gsm0710mux-1.0.pc.in): fix header location and add necessary require entry in .pc Apr 26 18:59:29 freesmartphone.org: 03mickey 07fso-abyss * r53edb9c9af54 10/ (Makefile.am configure.ac src/main.vala): catch up with vala 0.7.x Apr 26 18:59:40 Toaster`: IOError: [Errno 9] Bad file descriptor Apr 26 18:59:55 File "/usr/lib/python2.6/site-packages/shr_settings_modules/shr_gprs.py", line 149, in loadConnectionData Apr 26 18:59:55 apn, login, password = pickle.load(pickleFile) Apr 26 19:00:30 really... ok, I'm proably abusing pickle inappropriately Apr 26 19:02:07 let me see what I can do Apr 26 19:04:37 ok, I know... Apr 26 19:05:37 dos, I'm opening the file wrong. Foolish copy/pste error Apr 26 19:05:47 s/dos/dos1/ Apr 26 19:05:48 Toaster` meant: dos1, I'm opening the file wrong. Foolish copy/pste error Apr 26 19:06:12 Toaster`: ok, then commit fix. BTW. can you change "Internet" to "internet" in default value? Apr 26 19:06:23 sure Apr 26 19:06:25 i think it's the most popular default value for GPRS APNs Apr 26 19:06:38 ok. thanks for your work :) Apr 26 19:07:22 freesmartphone.org: 03mickey 07fso-abyss * re24a6abb8069 10/ (.gitignore Makefile.am): do not let vala compile, just generate C files Apr 26 19:08:20 freesmartphone.org: 03mickey 07libgsm0710mux * re7538c53ef3e 10/Makefile.am: do not let vala compile; just generate C files Apr 26 19:09:27 <\marco> is there any news about the e17 break? Apr 26 19:09:55 <\marco> are -_- Apr 26 19:10:50 For all interested in RT preemption and such in current kernels, here's a very recent interview: http://lwn.net/Articles/319544/ Apr 26 19:13:57 Toaster`: when do you want to commit? Apr 26 19:14:35 I'm writing the comment now Apr 26 19:14:47 ok :) Apr 26 19:14:51 <\marco> mwester-laptop: is e17 still broken? Apr 26 19:15:50 SHR: 03frazier.cameron 07shr-settings * r3b1146a76741 10/shr_settings_modules/shr_gprs.py: [GPRS] Fixed file I/O error in previous commit Apr 26 19:16:29 \marco: i put a patch in place, but due to the sweeping impact, it requires someone (mrmoku?) who can clean out the tmp dir on the bearstech bulder, and force a full rebuild. Apr 26 19:17:59 Toaster`: it doesn't work... it fallbacks to "internet" every time Apr 26 19:18:11 what? Apr 26 19:18:26 Hrmmm... Apr 26 19:18:42 Toaster`: it is "internet" after loading, even when i changed it before Apr 26 19:18:58 <\marco> mwester-laptop: good.. thanks for the patch :) .. I'll wait for the full rebuild :) Apr 26 19:20:03 dos1: are you making the gprs connection? Apr 26 19:20:21 oh Apr 26 19:20:24 ok Apr 26 19:20:38 it only sets the data on a successful connection Apr 26 19:20:47 Ainulindale: just to let you know... doing a rebuild from scratch now Apr 26 19:20:49 does frameworkd mess with the clock? because my fr keeps rapidly losing time, despite running ntp often Apr 26 19:20:56 and it has the right time on boot Apr 26 19:21:13 it loses several minutes an hour Apr 26 19:21:19 Ainulindale: for unstable that is Apr 26 19:22:38 Toaster`: it works! thanks! :) Apr 26 19:22:46 :) Any time Apr 26 19:24:15 freesmartphone.org: 03mickey 07cornucopia * r4535efb1caed 10/libfsotransport/ (5 files in 3 dirs): fsotransport: catch up with vala 0.7.x Apr 26 19:24:17 freesmartphone.org: 03mickey 07cornucopia * r2bee14d692eb 10/libfsoframework/ (4 files in 2 dirs): fsoframework: catch up with vala 0.7.x Apr 26 19:24:18 freesmartphone.org: 03mickey 07cornucopia * r8c19b6464481 10/libfsoframework/ (6 files in 3 dirs): fsoframework: catch up with vala 0.7.x; fix header path Apr 26 19:24:19 freesmartphone.org: 03mickey 07cornucopia * r0917aead424f 10/fsodeviced/ (7 files in 6 dirs): fsodevice: catch up with vala 0.7.x Apr 26 19:24:20 freesmartphone.org: 03mickey 07cornucopia * re1ba5ce4be30 10/fsogsmd/ (9 files in 7 dirs): fsogsm: catch up with vala 0.7.x Apr 26 19:24:34 shoragan: ping Apr 26 19:24:57 hey onen Apr 26 19:25:05 alphaone: hallo! Apr 26 19:28:17 raster: ping Apr 26 19:28:24 raster: please look at http://trac.shr-project.org/trac/ticket/416 ;) Apr 26 19:28:47 DocScrutinizer: ping Apr 26 19:28:57 DocScrutinizer: got my statefile? Apr 26 19:31:51 Ok, I'm going for a walk. Cheers all. Apr 26 19:32:29 alphaone: how are things going with M5.5? pretty busy, I guess... Apr 26 19:34:36 Ainulindale: when we want to bump e in shr-unstable? Apr 26 19:35:23 Ainulindale: actual revision isn't so stable... segfaults really often when playing with top shelf Apr 26 19:35:34 actual = that one we have now Apr 26 19:35:52 dos1: just started a rebuild from scratch... would have been the correct occasion ;) Apr 26 19:37:36 freesmartphone.org: 03mickey 07libeflvala * r111247b229c1 10/ (14 files in 8 dirs): catch up with Vala 0.7.x Apr 26 19:39:11 onen: Well, things are starting to look up Apr 26 19:40:02 alphaone: with cellhunter db, yuck ;-) Apr 26 19:40:15 heh Apr 26 19:40:51 alphaone: you stated that monitoring interface gives sometimes incorrect values, with M5.5 Apr 26 19:41:17 alphaone: is it sth like lac and/or cid with a value of 0000 ? Apr 26 19:41:18 freesmartphone.org: 03mickey 07fso-term * re0d92222e3e5 10/Makefile.am: catch up with vala 0.7.x Apr 26 19:42:18 onen: it's with longer characters Apr 26 19:43:41 alphaone: do you think this is safe for us (obm) to work with M5.5? (monitoring is being used for neighbours, lac/cid/rxlev/c1/c2) Or better to wait for stable release? Apr 26 19:46:07 mwester, do you know if I can create the repo directory in the tmp directory used by bitbake in order to save space? (tmp/deploy) Apr 26 19:46:50 Sharwin_F: repo? i'm symlinking to tmp/deploy on my buildhost Apr 26 19:47:04 dos1, ok, thanks :D Apr 26 19:47:58 dos1, I'm trying to make the repo in my computer, I've to use oe/openembedded/contrib/deploy/nfs/geany oe-nfs-deploy-image, isn't it? Apr 26 19:48:08 hmm Apr 26 19:48:12 * dos1 is checking Apr 26 19:48:12 uops, without geany Apr 26 19:48:13 :P Apr 26 19:48:50 Sharwin_F: feed -> tmp/deploy/glibc/ipk Apr 26 19:48:55 did someone successfully compiled libfso-glib ? I get an error from the vala compiler: "--gir: unknown option" Apr 26 19:48:55 i have that symlink Apr 26 19:50:03 dos1, but the script I've to run is that one? Apr 26 19:50:14 i don't know Apr 26 19:50:17 to create the indexes and that kinf of stuff Apr 26 19:50:18 :P Apr 26 19:50:22 i'm using make index Apr 26 19:50:26 from SHR Makefile Apr 26 19:50:58 dos1, would that add into the index packages which aren't on the shr repo? Apr 26 19:51:12 yep Apr 26 19:51:21 so I might try that Apr 26 19:51:56 this one? http://shr.bearstech.com/Makefile Apr 26 19:55:57 Sharwin_F: yep Apr 26 19:56:28 Sharwin_F: make image; make distro; make index Apr 26 19:56:34 to get a fully fledged SHR repo Apr 26 19:56:51 (without fat image) Apr 26 19:57:07 mrmoku, I have a repo with abiword, atm I only need to make a repo to try it out Apr 26 19:57:15 mrmoku, can I make the index without building SHR? Apr 26 19:58:10 Sharwin_F: yep... just do make index... or '. setup-env; bitbake package-index' Apr 26 19:58:15 which is the same Apr 26 19:58:27 onen, you should try MS5.5 now Apr 26 19:58:54 the problem with monitor interface is understood and only occours very rarely Apr 26 19:59:17 mrmoku, so doing "setup-env; bitbake package-index" and pointing my http server to tmp/deploy/ipk would be suficient to make it work? Apr 26 19:59:44 onen, i'll be back tomorrow, see you then (and i still intend to write to nick) Apr 26 19:59:49 Sharwin_F: yep Apr 26 20:00:16 Sharwin_F: well... tmp/deploy/glibc/ipk me thinks Apr 26 20:00:18 mrmoku, what's setup-env for? I don't have it Apr 26 20:00:22 mrmoku, yes, sorry Apr 26 20:00:33 Sharwin_F: setup the environment Apr 26 20:00:37 you don't have it? Apr 26 20:00:47 it should be inside shr-unstable (or shr-testing) Apr 26 20:01:09 Sharwin_F: ahh... those make commands were *not* for the SHR makefile Apr 26 20:01:12 shoragan: ok, I try it *now* ;-) Apr 26 20:01:13 cd shr-unstable first Apr 26 20:01:32 mrmoku, I don't have shr-unstable build, so no shr-unstable folder :P Apr 26 20:02:24 hi, once again I'm fighting libtool-native. In do_compile I get: Apr 26 20:02:31 configure: error: source directory already configured; run "make distclean" there first Apr 26 20:02:45 I tried bitbake libtool-native -c distclean Apr 26 20:02:48 Sharwin_F: what do you have then? Apr 26 20:03:02 mrmoku, just builded abiword with bitbake Apr 26 20:03:11 and I want to setup a repo with the packages to try install it Apr 26 20:03:20 bitbake in where? Apr 26 20:03:26 shoragan: good. I was thinking you are pretty busy with M5.5. And wanted to know if you knew when you hope to have time to keep going with the discussion. There is no hurry, of course, it was just to have an idea. Apr 26 20:03:45 mrmoku, what do you mean exactly? Apr 26 20:04:06 shoragan: see you tomorrow then Apr 26 20:04:26 <\marco> the packages.gz files are missing due to the rebuild from scratch? Apr 26 20:04:36 Sharwin_F: bitbake abiword... in what folder did you call it? Apr 26 20:04:56 mrmoku, I've set the TMPDIR to /media/sdb2/tmp Apr 26 20:05:19 so nomatter where do I call it AFAIK Apr 26 20:05:35 Sharwin_F: hmmm... ok. Where is your OE tree? Apr 26 20:05:48 I've the ipks on /media/sdb2/tmp/deploy/glibc/ipk/ or similar Apr 26 20:06:06 /home/user/dev/oe/ Apr 26 20:06:30 (or maybe /home/user/dev/oe/openembedded/ ? :P) Apr 26 20:06:58 Sharwin_F: well... if bitbake abiword worked... bitbake package-index should work too Apr 26 20:07:25 mrmoku, yes, package-index worked Apr 26 20:07:52 so now I can use it as a repo when setting up an http server in the directory where ipks are? Apr 26 20:07:59 Sharwin_F: ok... fine then :-) just link it Apr 26 20:07:59 yep Apr 26 20:08:14 ok Apr 26 20:08:41 mrmoku, are you going to build a new shr-unstable soon? I'll surely broke the one I have on my neo xD Apr 26 20:09:26 <\marco> Sharwin_F: me too :( Apr 26 20:09:39 Sharwin_F: already building :-) Apr 26 20:09:56 btw, is there any lightweight http server? or do I' have to use apache? Apr 26 20:10:00 any recomendations? hehe Apr 26 20:10:15 I'm only using it for that on this computer Apr 26 20:10:16 Running task 465 of 6999 Apr 26 20:10:18 <\marco> mhm.. there was a lightweight webserver.. mhm Apr 26 20:10:18 I'm trying to find a simple HelloWorld python edje tutorial. Any tips? Is it so that all enlightenment GUI must be compiled outside the python app itself? Apr 26 20:10:19 will take some time :P Apr 26 20:10:23 lighthttpd Apr 26 20:10:27 cherokee Apr 26 20:10:31 busybox Apr 26 20:10:41 ^^Sharwin_F Apr 26 20:10:59 thanks :D Apr 26 20:12:02 * Sharwin_F giving lighttpd a try Apr 26 20:14:26 mwester-laptop: hey, do you have some gdb + debug board experience? I'd like to debug a kernel module init() routine. I found lsmod gdb macro and used 'add-symbol-file add-symbol-file ./lindi/drivers/usb/gadget/g_ether.ko 0xbf119000' at /* Start the module */ in module.c. I then single stepped.. Apr 26 20:15:01 mwester-laptop: and found myself at 0xbf12a1a4. disassembly confirmed that this was indeed the real init() of g_ether but gdb did not show me line number info -- any idea what's going on? Apr 26 20:15:20 <\marco> Sharwin_F: http://en.wikipedia.org/wiki/Tiny_web_servers Apr 26 20:15:31 <\marco> for a best comparison :) Apr 26 20:15:39 lindi-: I confess that I think "old-style"; I've never used the debug board except for very very early init (and suspend/resume) issues. For all else, I sprinkle the code with kprintfs. Apr 26 20:15:49 mwester-laptop: ah Apr 26 20:15:54 Sorry :( Apr 26 20:16:12 I know gdb would be far faster, but that's not a skill I've mastered, I'm afraid. Apr 26 21:02:18 guys, if there's anyone alive here using the latest SHR testing image, can you tell me if you are able to see the correct icons in Vagalume ? Apr 26 21:06:20 odd... my FR refuses to suspend Apr 26 21:07:03 I force it to and after about 30s it does. It then promptly resumes again, with no calls/messages Apr 26 21:10:50 Toaster`|cooking: you enabled GPRS? Apr 26 21:11:26 Toaster`|cooking: /var/log/framework.log tells you why it thinks it woke up Apr 26 21:28:58 DocMobilizer: No, I don't have gprs, and I've rebooted since I noticed the issue Apr 26 21:32:29 <\marco> is there a way to reinstall all the world? Apr 26 21:33:23 <\marco> an opkg install -force-reinstall $( opkg list_installed) ? Apr 26 21:34:24 <\marco> or there's a way more appropriate..? Apr 26 21:34:58 <\marco> opkg install -force-reinstall $( opkg list_installed | cut -f 1 -d '-') Apr 26 21:35:00 <\marco> :) Apr 26 21:40:33 \marco: dfu-util -a rootfs -R -D /path/to/shr-image.jffs2 Apr 26 21:41:14 <\marco> mwester: eheh maybe it's a a better idea :) Apr 26 22:10:17 ok, now i really need help, i've tried a lot of things but i'm unable to make a small elementary window appear at the bottom of the screen, python elementary on shr unstable, is it possible to make it? if not is there anything in the e17 libs that can help me? Apr 26 22:17:58 budfive: /var/log/framework.log doesn't tell me anything about why it resumed Apr 26 22:18:56 Toaster`|cooking: resume_reason file? Apr 26 22:19:31 lindi-: umm... where would that be? Apr 26 22:19:51 Toaster`|cooking: find /sys -name resume_reason? Apr 26 22:20:14 Toaster`|cooking: i don't try to remember sysfs paths by heart, they are way too long Apr 26 22:20:38 cat `find /sys -name resume_reason` solved that Apr 26 22:21:44 Is this the reason? "* EINT01_GSM" Apr 26 22:23:06 Toaster`|cooking: yes Apr 26 22:23:14 Toaster`|cooking: your GSM chip wanted to say something to you Apr 26 22:23:30 Toaster`|cooking: check what was read from ttySAC* Apr 26 22:24:22 lindi-: I'm sorry, but is that in the framework.log file? Apr 26 22:24:37 Toaster`|cooking: depends on your logging configuration Apr 26 22:24:50 Toaster`|cooking: log_level =DEBUG will put it there Apr 26 22:24:55 Standard SHR-unstable Apr 26 22:26:28 Changing frameworkd.conf and restarting frameworkd Apr 26 22:43:22 lindi-: even with `log_level = DEBUG` I see nothing regarding ttySAC*. Apr 26 22:44:28 Toaster`|cooking: ttySAC is not mentioned in the logs Apr 26 22:44:42 Toaster`|cooking: but you should get DEBUG prints Apr 26 22:45:47 I have the DEBUG prints, I'm just having a hard time figuring out what is going on. Apr 26 22:52:25 Toaster`|cooking: well, what's the first read after resume? Apr 26 22:55:54 lindi-: http://pastebin.ubuntu.com/158868/ Apr 26 22:56:17 That's a snip of the relevant area. Apr 26 23:05:10 Toaster`|cooking: odd, looks like it is sending to gsm chip and not vice versa Apr 26 23:05:16 Toaster`|cooking: please strace gsm0710muxd Apr 26 23:06:03 Toaster`|cooking: something like sudo strace -t -o gsm0710muxd.strace -e '!gettimeofday,poll' -s4096 -p `pidof gsm0710muxd` Apr 26 23:06:21 Toaster`|cooking: that'll log with timestamps everything it talks with gsm Apr 26 23:07:01 And grab that data across the suspend cycle? Apr 26 23:07:08 Toaster`|cooking: yep Apr 26 23:07:10 ok Apr 26 23:07:21 Toaster`|cooking: wait for it to resume Apr 26 23:07:31 Toaster`|cooking: how do you suspend btw? Apr 26 23:07:50 I have to use the Illume menu, as it never seems to timeout Apr 26 23:07:59 even then it take 30s or more Apr 26 23:10:40 lindi-: is using the Illume suspend fine for this logging? Apr 26 23:11:17 Toaster`|cooking: i guess Apr 26 23:11:22 Toaster`|cooking: it just calls frameworkd Apr 26 23:11:26 ok Apr 26 23:15:45 And now it's staying suspended :-/ Apr 26 23:16:00 :) Apr 26 23:24:53 Toaster`|cooking: are you sure you haven't enabled GPRS? Apr 26 23:25:41 DocScrutinizer: ok, foolish question, how can I check outside of SHR-Settings? I've never had to toy with it before Apr 26 23:25:43 I have same resume_after_30s issue as long as logged in to GPRS and actively receiving data Apr 26 23:25:57 I don't have a gprs account Apr 26 23:27:39 not having gprs account probably only means communication with AP doesn't result in registration and internet access Apr 26 23:28:05 so it may just keep hitting the AP tryign to connect? Apr 26 23:28:14 Toaster`|cooking: regarding check, I'm not sure Apr 26 23:28:39 but probably there's some ppp-daemon involved. Apr 26 23:29:12 also in frameworkd.log there should be indication if gprs ppp is active Apr 26 23:29:42 What would that activity look like? Would it be obvious? Apr 26 23:29:55 Toaster`|cooking: (keep hittin) that's what I'd expect, yes Apr 26 23:34:22 hrm, well, `tail -f frameworkd.log` isn't showing any ogsmd reports anymore Apr 26 23:38:30 Toaster`|cooking, take a look at /tmp/fso-monitor.log Apr 26 23:39:59 you should see CipherStatus changes Apr 26 23:45:09 playya: /tmp/fso-monitor.log is blank Apr 26 23:45:22 hmm. not good Apr 26 23:45:32 shr-testing? Apr 26 23:45:45 Maybe I should just install a fresh image Apr 26 23:48:24 Toaster`|cooking: sorry don't have any more log lines related to GPRS (due to boot). seem to remember it's been ppp you find in this line Apr 26 23:50:38 DocScrutinizer: `grep -i ppp frameworkd.log` gives me nothing... I think I'm just going to reflash to a clean image Apr 26 23:50:49 playya: shr-unstable Apr 26 23:52:07 It only takes a couple hours to build a new image, and I still have a backup of my last working image Apr 26 23:52:52 Toaster`|cooking: hm, why not use shr-testing image? Apr 26 23:53:42 DocScrutinizer: I find stability boring? :) Apr 26 23:54:34 I'm looking for some Elementary API documentation. Is there any? Apr 26 23:54:34 Since I'm working on some SHR stuff, I find it's easier to just run unstable. Apr 26 23:54:40 No one calls me anyways Apr 27 00:07:49 Toaster`|cooking: should that be Toaster`|burning by now? :p Apr 27 00:08:16 mwester: Bread is rising :) Apr 27 00:08:29 The nspy kernel on my web site (with modules) may shed some light on the issue, if you really really want to dig deep. Apr 27 00:09:34 mwester: It probably has to do with something I did but know not what. And I wouldn't know how to interpet the results anyways. I'm not to good at programing outside of a tad of python Apr 27 00:39:11 m0nt0: you can't choose where to put the window Apr 27 00:39:24 raster, :( Apr 27 00:39:35 by making it a dialog illume's policy will be to center it and make it minimum height for the window - that's it Apr 27 00:39:52 you have no choice on positioning or sizing of the window - wm has its policies and enforces it Apr 27 00:40:42 ok, may be i've done something wrong with the layout because also the dialog was fullscreen, no problem Apr 27 00:41:52 just last question, can i make a "hole" in the window using something similar to the trasparent icon example? Apr 27 00:42:10 yes Apr 27 00:42:15 (ehis will be a workaround) Apr 27 00:42:34 but thats just silly/wrong Apr 27 00:42:42 so not going to help you there Apr 27 00:42:45 i know :( Apr 27 00:42:54 no problem, i won't do it Apr 27 00:42:57 you are trying to do somehting the wm has not any policy for for layout Apr 27 00:43:04 you need to live with it Apr 27 00:43:38 trying to work around the wm's policy will just result in bugs, pain, bizarre behavior and in the end, if the wmchanges policy - it will all fall apart Apr 27 00:43:48 if you are tying to mimic the palm pre Apr 27 00:44:40 no no i just want my 2 buttons sitting on the bottom, but there's no problem living with the illume policy Apr 27 00:44:46 may be i've done something like the beginning of a "today" apps for shr Apr 27 00:45:16 what are you trying to do int he end anyway? Apr 27 00:45:24 why do you want just a window at the bottom? Apr 27 00:45:32 what is wrong with everything in your big window? Apr 27 00:45:48 you don't already have a window? Apr 27 00:46:01 i like the idea to see the message "1 missed call" at the bottom and still see the rest of the desktop Apr 27 00:46:10 just my idea Apr 27 00:46:49 and i'm learning elementary and e* libs with python so just wanted to see if it was possible Apr 27 00:47:33 aaah ok Apr 27 00:47:42 so you are tying to mimic palm-pre Apr 27 00:47:50 ie you want little notification windows at the bottom Apr 27 00:47:53 no :) Apr 27 00:48:02 yes you do Apr 27 00:48:07 gtk stak was doing that Apr 27 00:48:12 *stack Apr 27 00:49:15 raster, here is what i've done http://scap.linuxtogo.org/files/8db82546d2432b1457f33ed8b7d7723e.png Apr 27 00:49:28 http://www.rsacorp.com/images/pre.JPG Apr 27 00:49:35 on the left screenshot Apr 27 00:49:42 the 4 "bars" at the bottom Apr 27 00:49:49 a calendar meeting item Apr 27 00:49:51 a missed call Apr 27 00:50:00 message and a music player controller Apr 27 00:50:21 nice Apr 27 00:50:38 there is a whole ui policy on it Apr 27 00:50:50 (ie "wm" tho it doesnt run x and thus there is no wm) Apr 27 00:50:59 an ap can create a "notification window" Apr 27 00:51:10 and the "wm" will stack them at the bottom - most recent at the top Apr 27 00:51:22 and adjust the "active app" window accordingly Apr 27 00:51:27 THAT is what you really want Apr 27 00:51:35 yeah Apr 27 00:51:42 that's it Apr 27 00:51:44 a mechanism where any process can pop up a "notification" window/bar Apr 27 00:51:56 is it bad? Apr 27 00:52:06 and that can be clicke don to dismiss it or see the full info or simply do nothing (ie the music controller) Apr 27 00:52:11 no its not bad Apr 27 00:52:17 its how it should be done Apr 27 00:52:30 ok ok Apr 27 00:52:35 but illume has no such ppolicy in it Apr 27 00:52:40 i'd like to add it Apr 27 00:52:46 but i simply am not doing illume stuff atm Apr 27 00:52:54 i'm working on elementary/evas/edje Apr 27 00:53:17 don't worry i'll wait for that :D Apr 27 00:53:51 yup Apr 27 00:53:57 thats what you should wait for Apr 27 00:54:11 if i added that Apr 27 00:54:36 then your hacks would likely become buggy or not work or simply become totally inconsistent with the rest of the ui Apr 27 00:55:59 yes, it was an image on scap which was made me think how that could be done, it was posted some months ago maybe a mockup so i was just trying to get that behaviour Apr 27 00:56:25 don't want to make hack or workaround, that's bad, i know Apr 27 00:58:19 :) Apr 27 00:58:30 i have plans to work on illume and adda tonne of this Apr 27 00:58:38 but its all secondary to e17's release Apr 27 00:58:43 and evas/edje/elementary work atm Apr 27 01:00:33 you just made me happy Apr 27 01:01:53 the whole launcher in elementary needs actualy work Apr 27 01:02:14 it was a quick 2hr hack on an afternoon when i knew the "Slider" mode launcher than sean wanted was just going to sucka/piss people off Apr 27 01:02:20 so i quickly threw it together Apr 27 01:04:47 yeah i remember the "slider" idea Apr 27 01:05:06 not so bad, but more practical the "icon" view Apr 27 01:06:58 yes Apr 27 01:07:01 more practical Apr 27 01:11:25 well need to get to bed, 3 AM here, good night, see you tomorrow :D **** ENDING LOGGING AT Mon Apr 27 02:59:56 2009