**** BEGIN LOGGING AT Thu Jul 30 02:59:57 2009 Jul 30 03:02:07 uploaded to opkg -> http://www.opkg.org/package_258.html Jul 30 04:23:42 added mooedit to opkg.org too, another text editor with more features :D Jul 30 04:23:44 http://www.opkg.org/package_259.html Jul 30 05:34:55 raster: Hi :) Jul 30 05:35:18 raster: i've a question for you. Now about batget itself, the internal powel_supply method. Jul 30 05:36:17 raster: i exposed approximate "capacity" (in percents) in gta01 dumb battery driver. But batget doesn't use it properly as it looks at charge_now or power_now or voltage_now and does some extra calculations. Jul 30 05:37:11 raster: ok, i workarounded that by exposing fake charge_now equal to capacity and charge_full = 100. Works good on gta02 but on gta01 we have current_now readings and that'll probably mess everything up again, right? Jul 30 05:39:38 raster: what should i do? Gta02 doesn't have _any_ chance to obtain battery current, gta01 gives some very rough and noisy (+-100mA) idea of it. Jul 30 05:41:34 hi,just a question...I don't actually use the freerunner as a phone,but I heard that dbus doesn't loose calls anymore...so maybe with a good alsa state file I could use it as a phone? Jul 30 05:44:29 GNUtoo: i was using it as a phone since November. I can't say i lost many calls. Jul 30 05:44:41 ok thanks Jul 30 05:44:53 GNUtoo: do you have a buzzfix? If yes, why not to try phone functionality? Jul 30 05:45:06 no I've no buzzfix but no buzz either Jul 30 05:45:27 do you know a good alsa state file where I can put the phone near my ear and mouth ? Jul 30 05:45:33 because for now: Jul 30 05:45:36 GNUtoo: everybody has buzz. Sooner or later. Jul 30 05:45:37 in SHR-unstable Jul 30 05:45:42 ouch Jul 30 05:45:55 * Blu3 is suffering withdrawal during buzz fix period Jul 30 05:46:00 I can't have mine buzzfixed Jul 30 05:46:05 GNUtoo: why? Jul 30 05:46:28 I had exams or wasn't in the right place in the right moment Jul 30 05:46:38 GNUtoo: what about now? Jul 30 05:46:50 bearstech thing is over Jul 30 05:47:03 GNUtoo: do it yourself / with a help from your friends. Jul 30 05:47:05 I must buzz-fix it in italy and I fear the postal services(they stole things) Jul 30 05:47:12 ok maybe I'll do that later Jul 30 05:47:32 GNUtoo: don't you really have no electronics repair shop nearby? Jul 30 05:47:42 I don't know I'll look Jul 30 05:47:59 I know one but I don't know if they do phones Jul 30 05:48:19 mmm Jul 30 05:48:47 I can't buy an A7 so I will need to buzzfix it later Jul 30 05:49:03 GNUtoo: don't ask about phones! Ask if they could resolder one easily accessible 0402 resistor. Jul 30 05:49:13 ok Jul 30 05:49:40 GNUtoo: say you know exactly and for sure what and how to do. That you have photos and stuff. That the resistor is in the right place and that you'll disassemble the device yourself. Jul 30 05:49:45 I'll do it in september then because right now I've not the time to do it Jul 30 05:49:52 ok Jul 30 05:50:07 is it easy to disassemble? Jul 30 05:50:11 GNUtoo: it'd be best to have resistor and the needed capacitor too. Jul 30 05:50:17 ok Jul 30 05:50:36 GNUtoo: you can ask somebody (Joerg or probably Danial) to send it to you. Jul 30 05:50:45 GNUtoo: dissassembly is damn easy. Jul 30 05:50:50 ok I'll do that in december Jul 30 05:50:59 s/december/september Jul 30 05:51:16 by the way I hesitate about the fact of buying a debug board Jul 30 05:51:36 if the phone get unsuported at a point it would be easy for kernel debuging Jul 30 05:51:40 GNUtoo: if the shop feels friendly you can get #1024 fixed there too. Disassembling is a bit harder but you can do it yourself. The work is approximately the same, need to change one 0805 capacitor. Jul 30 05:51:51 ok Jul 30 05:52:14 GNUtoo: i don't think you really need a debug board. Jul 30 05:52:33 ok Jul 30 05:57:08 good morning Jul 30 05:57:20 PaulFertser: does one *have* to solder on a new capacitor? Jul 30 05:57:34 blindcoder: sure, that's the main point of the rework. Jul 30 05:57:38 PaulFertser: according to the second to last comment in #1024: All current distributions have a satisfying workaround, the real issue is still undergoing research wrt. GSM firmware. Jul 30 05:58:00 blindcoder: satisfying is subjectively, you know. Jul 30 05:58:15 PaulFertser: yeah, for some people a reboot might be satisfying ;) Jul 30 05:58:30 blindcoder: if you're happy with disabling deep sleep and some extra ~6mA in suspend, that's a satisfying workaround. Jul 30 05:58:45 friend of mine has a soldering station and necessary skills, guess I'll ask him then Jul 30 05:58:55 PaulFertser: does SHR have deep sleep enabled? Jul 30 05:59:29 blindcoder: i think it has "adaptive" mode by default. Means if frameworkd detects recamping it disables deep sleep for some time (an hour? i don't remember the details). Jul 30 05:59:50 PaulFertser: ah, okay. thanks :) Jul 30 06:00:05 blindcoder: be careful dismounting the can, better spend more time than break the holder. Jul 30 06:00:34 PaulFertser: of course. Jul 30 06:00:53 blindcoder: wifi module is simply glued on top of it. Just slowly pull it off. Jul 30 06:04:52 "glued on". nice... Jul 30 06:06:16 on another topic Jul 30 06:06:35 gpsdrive came with a script to download map files within a given area and LoD Jul 30 06:06:45 does tangoGPS have something similiar Jul 30 06:06:46 + Jul 30 06:15:55 blindcoder: tangogps has a gui method for that. Jul 30 06:17:05 it'd be nice if they both used the same map files, same methods of display, same methods of fetching Jul 30 06:23:32 PaulFertser: I only see "Auto download map files" but when I'm on vacation I won't have internet connection Jul 30 06:24:25 blindcoder: you go to the area you want to see, click to bring menu, then there's download tiles line. Jul 30 06:24:52 Blu3: it looks like navit is more promising, as it works with vector data directly. Jul 30 06:25:06 I've heard some people managed to make it useable on FR. Jul 30 06:28:43 PaulFertser: nice, thanks! Jul 30 06:36:16 good Jul 30 07:03:53 wtf... Stale NFS file handles? Jul 30 07:04:01 I haven't mounted anything on NFS... Jul 30 07:08:18 rm: cannot stat 'Maps/OSM/6/33/23.png': Stale NFS file handle Jul 30 07:08:22 wtf? Jul 30 07:19:31 alright, just a minor fs corruption Jul 30 07:21:17 blindcoder: yup, seen this here ;-) fsck will fix Jul 30 07:23:34 blindcoder: ext3? Jul 30 07:24:54 DocScrutinizer: ext2 I think Jul 30 07:25:06 jupp Jul 30 07:25:26 anyway, just did a backup of the partition with dd, jic :-) Jul 30 07:30:35 17k tiles... that'll take a while :) Jul 30 07:32:55 blindcoder: did you see the size-0 tiles problem? Jul 30 07:35:14 DocScrutinizer: not yet Jul 30 07:35:30 root@om-gta02 ~ $ find Maps/ -size 0 Jul 30 07:35:30 root@om-gta02 ~ $ Jul 30 07:35:37 maybe it's fixed now Jul 30 07:36:10 otherwise just delete them, as they never get downloaded again ;-) Jul 30 07:37:26 root@om-gta02 ~ $ find Maps/ -size 0 -exec "rm {} ; " Jul 30 07:37:47 when I re-run "Download Maps" will it just download the missing ones? Jul 30 07:37:58 ie: could I just retry? Jul 30 07:38:01 dunno, hope so Jul 30 07:38:39 for usual dl-on-demand (for current map) it does Jul 30 07:39:02 tangogps that is Jul 30 07:39:16 yeah Jul 30 07:39:26 but gprs connections abroad would be just a tad bit expensive :) Jul 30 07:40:40 hmm, interesting Jul 30 07:40:53 X doesn't react anymore, screen's flickering Jul 30 07:40:56 but ssh still works Jul 30 07:42:35 root@om-gta02 ~ $ ps aux | grep xinit Jul 30 07:42:35 Unknown HZ value! (93) Assume 100. Jul 30 07:42:37 :( Jul 30 07:43:14 ÜÜÜ Jul 30 07:43:23 err ??? Jul 30 07:43:42 yeah, my words exactly ;) Jul 30 07:44:41 aux led flashing... Jul 30 07:44:44 this looks bad :( Jul 30 07:44:45 suspend/resume, maybe the whole cpufreq shit fixes itself Jul 30 07:45:08 that's a kernel ooops :-/ Jul 30 07:45:12 yay! Jul 30 07:47:01 prolly suspend action during data transfer? BUG?? Jul 30 07:47:24 no, it doesn't suspend when usb plugged in Jul 30 07:47:32 but did I mention that I made a dd backup of my rootfs? Jul 30 07:47:36 the oops is a panic Jul 30 07:47:40 couldn't mount root filesystem Jul 30 07:47:42 *grr* Jul 30 07:48:55 I need another filesystem Jul 30 07:49:26 jupp, completely crashed rootfs Jul 30 07:50:08 fsck borks? Jul 30 07:50:48 *sigh* Jul 30 07:51:17 jup Jul 30 07:51:23 fsck basically gives up Jul 30 07:51:25 Hi ! Is it possible and how can I change the shr splash screen, I want to create one myself Jul 30 07:51:44 Think-Free: you can configure it in the E settings Jul 30 07:51:53 Think-Free: top-shelf, then the wrench in the upper left corner Jul 30 07:52:22 but I want to build one myself .. no way without enlightenment ? Jul 30 07:53:26 Think-Free: you mean the splash screen of the kernel or the X loading screen? Jul 30 07:53:45 oh sorry, the one of the kernel Jul 30 07:54:15 Think-Free: that's also possible to do, but I don't know how right now. IIRC it is compiled directly into the kernel as a .xpm file Jul 30 07:54:27 Think-Free: but don't take my word for it. it _is_ possible though Jul 30 07:55:20 I've found this folder : /home/christophe/NFS/usr/share/shr-splash/themes/shr-splash-theme-simple Jul 30 07:55:32 It contain a ppm image Jul 30 07:55:43 but changing it does nothing Jul 30 07:55:57 there is also a boot.fbz Jul 30 07:56:16 but I don't know if it's this file and how to build it :/ Jul 30 07:56:41 Think-Free: that's the X loading picture Jul 30 07:58:34 morning Jul 30 07:58:40 moin methril|work Jul 30 08:04:22 hmm Jul 30 08:04:34 with 400 MHz the freerunner might be just slow enough for some of my old dos games... Jul 30 08:05:12 blindcoder: i want a nes emulator Jul 30 08:05:14 :D Jul 30 08:05:39 i'll compile one eventually Jul 30 08:05:55 just wanted to say, there should be some of them Jul 30 08:06:07 fceu is nice Jul 30 08:06:18 used it some times Jul 30 08:06:45 PaulFertser, if you could explain me how to configure away correctly with irssi/bip... Jul 30 08:06:49 I certainly will do :P Jul 30 08:07:11 mrmoku: on irssi I just do /away gone Jul 30 08:07:11 :) Jul 30 08:07:39 blindcoder, I know Jul 30 08:07:47 but in combination with bip? Jul 30 08:07:55 I close my irssi Jul 30 08:08:08 and bip sets the nick and keeps the backlog Jul 30 08:08:54 bip is a bouncer? Jul 30 08:10:04 yep Jul 30 08:12:00 okay, I have no experience with those Jul 30 08:12:06 I just keep my irssi open in a screen Jul 30 08:12:41 blindcoder, yeah, but I want to connect with multiple clients and have just one connection to the outside Jul 30 08:12:48 this is what bip does for me Jul 30 08:13:39 I see Jul 30 08:23:02 good morning Jul 30 08:26:57 realizing the sheer amount of old original dos games lying around is always frightening me a little bit Jul 30 08:27:40 "IBM DOS-CD ROM" "© 1995" Jul 30 08:31:17 morning. Does someone know how the gsm pin dialog window is displayed on shr ? I'm wondering how hard it would be to modify this to handle bluetooth pairing Jul 30 08:33:37 mrmoku: it works in combination with bip as well. Jul 30 08:33:54 mrmoku: ah, you want to close irssi... Jul 30 08:33:59 mrmoku: i never do that :D Jul 30 08:54:53 mrmoku: what's expected semantics of "away" with bip and multiple clients?? Jul 30 08:57:55 mrmoku: imagine you have sth like "auto-away" and hop clients. eventually one ofe the idle clients will send an away while you're busy typing on another client o.O Jul 30 09:16:29 hehe, mallorca maps lod 9-15 downloaded, 76.5 mb :) Jul 30 09:20:46 hmm... is it possible to automatically restart the keyboard on screen rotation? Jul 30 09:22:51 * PaulFertser is still hoping that raster will clarify the situation with batget power_supply readings... Jul 30 09:27:54 hallo harlekin Jul 30 09:27:59 s/hallo/hello/ Jul 30 09:28:00 Hey, admiral0 Jul 30 09:28:01 admiral0 meant: hello harlekin Jul 30 09:29:02 s/^.*$/Cool./ Jul 30 10:02:44 harlekin: bzzbot substitute is very restricted. No fun Jul 30 10:03:24 DocScrutinizer, is there any good reason for that? Like security issues I'm not seeing? Jul 30 10:04:01 PaulFertser: (batget power_supply readings) duh? I thought things had been sorted out for good now... Jul 30 10:04:20 Is the feature helpful or is it just toying around / adding noise to the channel? Jul 30 10:04:41 harlekin: dunno, but yes, probably also security reasons. 's/x/xx/g' Jul 30 10:05:05 xxxxxxxxx Jul 30 10:05:06 DocScrutinizer: /lastlog raster 20 Jul 30 10:05:22 This wouldn't apply recursively on itself, would it? Jul 30 10:05:28 [2009-07-30 12:05:20] [Fehler] LASTLOG: Unbekannter Befehl. Jul 30 10:05:50 xxxxxxxxx Jul 30 10:05:59 s/x/xxxxxxxxxxxx/g Jul 30 10:05:59 DocScrutinizer meant: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Jul 30 10:06:09 That's annoying but not really risky. :) Jul 30 10:06:13 trunkates ;-) Jul 30 10:06:44 harlekin: check bzbot src, it's FOSS Jul 30 10:07:05 DocScrutinizer, I'll do. Thanks. :) Jul 30 10:09:30 DocScrutinizer: i exposed approximate "capacity" (in percents) in gta01 dumb battery Jul 30 10:09:34 driver. But batget doesn't use it properly as it looks at charge_now or power_now or voltage_now Jul 30 10:09:37 and does some extra calculations. Jul 30 10:09:43 DocScrutinizer: ok, i workarounded that by exposing fake charge_now equal to capacity Jul 30 10:09:46 and charge_full = 100. Works good on gta02 but on gta01 we have current_now readings and that'll Jul 30 10:09:49 probably mess everything up again, right? Jul 30 10:10:27 yup, maybe Jul 30 10:10:27 (that was my question) Jul 30 10:10:39 To someone who's responisble for the batget :D Jul 30 10:14:07 to me it seems power sysnodes are poorly defined/implemented all over the different platforms and E (aka raster ;-) is trying to handle it in a platform agnostic way. I doubt that will ever pan out Jul 30 10:14:38 Yeah.. Jul 30 10:15:31 maybe it woks great on raster's laptop, but may fail even on my different brand battery Jul 30 10:16:25 raster should inplement a bandaid config option Jul 30 10:17:33 I bet it works fine on all laptops, they don't use dumb batteries. Jul 30 10:18:49 don't mess with me ;-) I'll send you sysnodes od my siemens laptop. you'll be buffled Jul 30 10:19:17 DocScrutinizer: along with "uevent" sysnode contents please :D Jul 30 10:19:58 life link or what ? :-P Jul 30 10:20:04 DocScrutinizer: btw, i'm using today a dumb battery, works fine with my latest (not yet published) patches. I'm not sure many users would see any considerable difference. Jul 30 10:20:20 DocScrutinizer: just a snapshot will do. Jul 30 10:20:23 that's fair enough Jul 30 10:20:36 DocScrutinizer: in fact it's the only thing needed, it mentions all the nodes and their contents. Jul 30 10:20:46 ok, remind me later on please Jul 30 10:22:59 halley:/sys/devices/LNXSYSTM:00/device:00/PNP0C0A:00/power_supply/BAT0 # cat uevent Jul 30 10:23:01 POWER_SUPPLY_NAME=BAT0 Jul 30 10:23:02 POWER_SUPPLY_TYPE=Battery Jul 30 10:23:04 POWER_SUPPLY_STATUS=Discharging Jul 30 10:23:05 POWER_SUPPLY_PRESENT=1 Jul 30 10:23:07 POWER_SUPPLY_TECHNOLOGY=Li-ion Jul 30 10:23:08 POWER_SUPPLY_VOLTAGE_MIN_DESIGN=11100000 Jul 30 10:23:10 POWER_SUPPLY_VOLTAGE_NOW=12181000 Jul 30 10:23:11 POWER_SUPPLY_CURRENT_NOW=0 Jul 30 10:23:13 POWER_SUPPLY_CHARGE_FULL_DESIGN=5200000 Jul 30 10:23:14 POWER_SUPPLY_CHARGE_FULL=4914000 Jul 30 10:23:16 POWER_SUPPLY_CHARGE_NOW=4914000 Jul 30 10:23:18 POWER_SUPPLY_MODEL_NAME=e01241401B38B18500819 Jul 30 10:23:20 POWER_SUPPLY_MANUFACTURER=MAIN SMP SDI Jul 30 10:23:22 POWER_SUPPLY_SERIAL_NUMBER=0092 Jul 30 10:24:34 DocScrutinizer: hm, no capacity but i guess batget would manage that. Jul 30 10:25:36 well, KPowerSave Info get's rather confused most of the time Jul 30 10:26:20 s /'s/s/ Jul 30 10:26:48 POWER_SUPPLY_CURRENT_NOW=0 ? Jul 30 10:27:04 hehe, it's on mains of course Jul 30 10:27:34 but still it's POWER_SUPPLY_STATUS=Discharging Jul 30 10:30:21 for Ur convenience... pulling plug... Jul 30 10:31:07 halley:/sys/devices/LNXSYSTM:00/device:00/PNP0C0A:00/power_supply/BAT0 # cat uevent Jul 30 10:31:08 POWER_SUPPLY_NAME=BAT0 Jul 30 10:31:10 POWER_SUPPLY_TYPE=Battery Jul 30 10:31:11 POWER_SUPPLY_STATUS=Full Jul 30 10:31:13 POWER_SUPPLY_PRESENT=1 Jul 30 10:31:14 POWER_SUPPLY_TECHNOLOGY=Li-ion Jul 30 10:31:16 POWER_SUPPLY_VOLTAGE_MIN_DESIGN=11100000 Jul 30 10:31:17 POWER_SUPPLY_VOLTAGE_NOW=11802000 Jul 30 10:31:19 POWER_SUPPLY_CURRENT_NOW=2412000 Jul 30 10:31:20 POWER_SUPPLY_CHARGE_FULL_DESIGN=5200000 Jul 30 10:31:22 POWER_SUPPLY_CHARGE_FULL=4914000 Jul 30 10:31:23 POWER_SUPPLY_CHARGE_NOW=4846000 Jul 30 10:31:25 POWER_SUPPLY_MODEL_NAME=e01241401B38B18500819 Jul 30 10:31:26 POWER_SUPPLY_MANUFACTURER=MAIN SMP SDI Jul 30 10:31:28 POWER_SUPPLY_SERIAL_NUMBER=0092 Jul 30 10:32:33 notice ::: POWER_SUPPLY_STATUS=Full Jul 30 10:32:41 muahahaAAAAAhaaa Jul 30 10:33:19 ? Jul 30 10:33:29 could you explain to humans? Jul 30 10:33:35 XD Jul 30 10:34:10 admiral0: my Siemens Esprimo reports "bat=discharging" when hooked up to mains, but "bat=full" when run by bat Jul 30 10:34:14 admiral0: he's just crazy, ignore him ;) Jul 30 10:34:16 ;D Jul 30 10:34:38 lol Jul 30 10:35:46 QED - power sysfs is a mess all over the place, no matter which platform Jul 30 10:36:43 no way for a owever smart batget to get it right without "a little help" Jul 30 10:38:57 or let me put it this way: kernel has several bugs in sysfs power-node, so batget needs to catch these and work around. As long as we don't know what all the bugs might look like, we just can try to get it right for Neo-kernels and fix batget to work OK with these Jul 30 10:40:09 or, even another paraphrase: Laptop implementation isn't the reference for FR/1973 Jul 30 10:49:32 imho it's the kernel's job to expose power sources sanely Jul 30 10:49:50 if such soruces report bizarre info - i dont see it as a userspace job to go workaround every possible bizarre platform Jul 30 10:49:59 soon enough userspace looks like a Jul 30 10:50:11 if (platformx) {workaround x} Jul 30 10:50:22 else if (platformy) {workaround y} Jul 30 10:50:25 else if (platformz) {workaround z} Jul 30 10:50:28 else if (platforma) {workaround a} Jul 30 10:50:31 else if (platformb) {workaround b} Jul 30 10:50:32 ... Jul 30 10:50:52 when it is already the kernel that is doing per-platform/subsystem drivers anyway Jul 30 10:51:06 userspace sees an abstracted/generic set of info Jul 30 10:51:29 :) Jul 30 10:51:54 the freerunner kernel already does workarounds Jul 30 10:52:08 it exposes an apm device AND a /sys/class/power info blob Jul 30 10:52:25 it exposes apm to work around qtopia as its so old it doesnt support newer kernel api's Jul 30 10:52:29 for example Jul 30 10:53:03 there is a special set of #ifdefs and #defines for the gta02 - anything required to make power supply (batter etc.) look sane should be there. Jul 30 10:53:43 the day all arm devices and platforms are using a single generic arm kernel for all of them is the day i'll consider it a job for userspace to workaround :) Jul 30 11:00:55 raster: ack Jul 30 11:01:01 and hi! Jul 30 11:02:11 raster: see my last two statements. they basically were: get kernel right on Neo, get batget right to work with correct sysnodes Jul 30 11:03:08 so in the end Neo is the reference platform, and if it doesn't work on my Siemens Esprimo laptop, then that's kernels fault Jul 30 11:06:40 raster: my only implicit critics was "we have to make sure the implementation of batget as is now is conformant to a sane sysfs definition for power. We mustn't rely on a particular (raster's) laptop implementation for reference" Jul 30 11:09:22 and honestly I thought that has been the bottom line of our last discussion of the topic some few months ago... so nothing new. Jul 30 11:10:37 [12:04] PaulFertser: (batget power_supply readings) duh? I thought things had been sorted out for good now... Jul 30 11:18:50 raster: please reread my question :) Jul 30 11:19:30 raster: i already expose from the dumb battery all i have and all i can guess. That being Voltage now, Charging status and Capacity (%) estimate. Jul 30 11:19:56 raster: i guess it can work with HAL method but it certainly doesn't with Internal. Jul 30 11:20:22 raster: what else should i do from the kernel side to make you happy? ;) Jul 30 11:20:45 PaulFertser: that's the way ;-D Jul 30 11:24:48 PaulFertser: no idea. i havent turned on my gta02 in ages. Jul 30 11:25:00 raster: you don't need to. Jul 30 11:25:06 raster: please listen to me :) Jul 30 11:25:18 raster: i made the dumb battery driver work on gta02. Jul 30 11:25:29 then it should work Jul 30 11:25:30 raster: i expose the values i know and can guess. Jul 30 11:25:36 the bat n gta02 has a cc Jul 30 11:25:44 you can report current charge Jul 30 11:25:45 raster: i want to use a dumb battery on FR. Jul 30 11:25:52 (and design capacity is known) Jul 30 11:25:55 raster: because they're easily available and dirt cheap. Jul 30 11:26:02 oh the voltage thing Jul 30 11:26:12 raster: i expose voltage_now Jul 30 11:26:32 i have n o code to handle that Jul 30 11:26:45 as this now is the only device i know of to have that Jul 30 11:26:47 raster: the only difference between dumb battery on gta01 and gta02 is that gta01 can also provide some approximate noisy (+-100mA) current readings. Jul 30 11:27:40 raster: but i do expose "capacity", why do you think that's not enough? Jul 30 11:28:13 raster: several weeks ago we were talking about guessing percentages from voltage. Ok, done now. Jul 30 11:29:27 raster: do you really think that i should add a special sysfs node for user to supply design capacity of the battery he uses just to report it back in power_supply node later? Jul 30 11:30:08 i don't know Jul 30 11:30:18 a % is probably fine Jul 30 11:30:28 i dont have the code in front of me atm Jul 30 11:30:41 raster: i do :) Jul 30 11:32:47 doesnt help me if i dont have e's code Jul 30 11:33:03 right now - i cant comment. i have mundane stuff like procject schedules to do Jul 30 11:33:04 :) Jul 30 11:33:12 raster: i see :-/ Jul 30 11:33:14 and dinner to get out of the oven in 10 mins Jul 30 11:34:03 raster: but can you please take a look at http://trac.enlightenment.org/e/browser/trunk/e/src/modules/battery/batget.c sometime later and fix it to work with a driver that exposes only voltage_now, status and capacity please? Jul 30 11:34:32 raster: TIA :D Jul 30 11:42:46 hi Jul 30 11:42:52 litephone sources not available yet? Jul 30 11:43:35 raster: oh you can just forget about it and i'll stick a workaround to report percentage as CHARGE_NOW and 100 as CHARGE_FULL with a commit message blaming you for forcing to use kernel workarounds ;) Jul 30 11:44:58 DocScrutinizer: current doesn't affect percentage report, only time_left which i don't care about as we don't have reliable current readings anyway. Jul 30 11:45:20 DocScrutinizer: so after i reread the code i think i'm fully satisfied with a workaround. Jul 30 11:45:20 ack Jul 30 11:47:03 DocScrutinizer: unfortunately i haven't yet received a single comment on the stuff i changed to get gta02 working (neither from OM folks nor from upstream), so i'll push only the gta01 parts for now :( Jul 30 11:47:36 DocScrutinizer: i thought there's more interest in using dumb batteries with FR... Jul 30 11:48:11 PaulFertser: change sysnode name of charge_now to charge_now_percent. Create correct content of charge_now in uAh(?) Jul 30 11:48:50 DocScrutinizer: impossible since i don't know full capacity for random dumb battery. Jul 30 11:49:28 so fake it for unknown dumb bat Jul 30 11:49:55 const charge_max = 1200mA Jul 30 11:50:35 charge_now = charge_max * charge_now_percent / 100 Jul 30 11:51:08 DocScrutinizer: why 1200? Jul 30 11:51:23 err 1200000 Jul 30 11:51:35 uAh Jul 30 11:51:44 Do you think faking is better than not providing at all? Jul 30 11:52:24 I seem to remember that's the definition of the values: uA(h), uV Jul 30 11:52:35 and no % at all Jul 30 11:52:50 No, there's "capacity" node, for % Jul 30 11:53:08 ok, exactly Jul 30 11:53:37 this seems generic, as I don't see it here either (on my laptop) Jul 30 11:54:31 so batget shouldn't use capacity (which is a mere charge_max / charge_now anyway) Jul 30 12:00:54 PaulFertser: 1) define or find out about a sane sysfs power implementation. 2) fix OM kernel to support all madatory nodes *correctly*. 3) fix batget to operate on the mandatory nodes only Jul 30 12:01:32 PaulFertser: it seems capacity is NOT a mandatory node Jul 30 12:02:31 charge_now and charge_max are, it seems Jul 30 12:03:38 if I'm not wrong... ;) didn't check it here, just trying to recall what I read about power sysfs a few weeks ago Jul 30 12:04:16 PaulFertser: anyway start with 1), as it's easier than to start with 3) ;-> Jul 30 12:05:37 DocScrutinizer: no, there's spec about power_supply. Jul 30 12:05:50 DocScrutinizer: expose everything that hardware gives. And nothing more. Jul 30 12:06:17 bah, that's no spec, that's silly Jul 30 12:06:46 and especially fuzzy Jul 30 12:10:53 for dumb bat I don't see how a good guess for charge_max is worse than a non-existent (hw-wise, according to this "spec") capacity in percent Jul 30 12:12:40 if you want to kling to these specs verbatim and still have sth like batget working, you *need* to either config charge_max *or* batget heuristics Jul 30 12:12:56 I prefer the first one Jul 30 12:13:10 seems raster does as well ;-) Jul 30 12:17:08 so I suggest to hardcode a charge_max of 1.2Ah for dumb bat driver for first approach, and expose charge_max and charge_now to sysfs Jul 30 12:21:04 DocScrutinizer: i suggest 850 as it's what "nokia original" bl-5c seem to have. Jul 30 12:23:02 ok, fine with me as well Jul 30 12:23:53 though I'd refer to GTA01 bat and more recent BL-xC replacements Jul 30 12:24:11 make it a kernel cmdline parameter ;) Jul 30 12:27:06 PaulFertser: for sure this is a value that should err to the low side, to give user more standby than batget estimates. That's conservative for this kind of estimations. So I'm happy with 850mAh, even with 100 for completely unknown type of bat (e.g. external USB-supply etc) Jul 30 12:35:00 DocScrutinizer: on gta02 we won't have _any_ estimation of remaining runtime. And gta01 lacks any decent current readings as well. Jul 30 12:35:32 PaulFertser: btw my original BL-5C here in front of me doesn't state any max-capacity at all :-/ Jul 30 12:35:44 DocScrutinizer: same here Jul 30 12:35:49 PaulFertser: (runtime) ack Jul 30 12:35:53 DocScrutinizer: i just googled it. Jul 30 12:36:39 PaulFertser: (runtime) except those from bq27k Jul 30 12:37:34 PaulFertser: bq27k even provides an estimation for standby-current :-)) Jul 30 12:37:51 which seems rather good btw Jul 30 12:39:20 probably drops the short spikes for GSM-TX though Jul 30 12:40:59 DocScrutinizer: yeah, bq27k is good enough in this regard :) I'm still waiting for the upstream decision on my pcf50633. Jul 30 12:41:10 see ISLC, STTEL/H Jul 30 12:41:35 anyway, cya Jul 30 12:41:41 * DocScrutinizer off for meal Jul 30 12:41:44 CU Jul 30 12:41:49 Enjoy your meal :) Jul 30 12:53:27 anyone know what gpsd's "Satellite data no good (x of x)" message means? Jul 30 12:56:09 PaulFertser: ping Jul 30 12:56:20 jc: pong Jul 30 12:56:25 jc: i'm fast, heh? Jul 30 12:56:26 ;) Jul 30 12:56:47 PaulFertser: gismo has fixed the certificate issue couple days before. Jul 30 12:56:57 PaulFertser: should be ok now, right? Jul 30 12:57:35 jc: yeah :) Jul 30 12:57:52 PaulFertser: :) Jul 30 12:58:01 jc: thanks for caring Jul 30 12:58:34 PaulFertser: not much I can do, somehow this is an easy one for me. passing the message. Jul 30 12:58:54 PaulFertser: by the way, do you use gforge? Jul 30 12:59:16 jc: no Jul 30 13:00:30