**** BEGIN LOGGING AT Mon Sep 14 02:59:57 2009 Sep 14 06:34:07 mickeyl: fyi ofono people did look at 0710 implementations we have Sep 14 06:34:43 mickeyl: and they said that fso-abyss is basically unmodified "good" nokia code wrapped in a shit-vala useless dbus wrapper. Sep 14 06:37:42 wow Sep 14 06:40:19 Yes, they said no need for dbus here since nobody except the daemon itself should request a channel anyway, that's why they say they're going with a plain C library (and luckily for them nokia code is almost plain C-compatible) Sep 14 06:42:13 And btw, holtman still uses gta01 (and has recently ordered gta02). Sep 14 06:42:26 hehe Sep 14 06:51:06 mickeyl: he even used my instructions to "recover" half-flashed calypso in his gta01 ;) Sep 14 06:52:04 morning Sep 14 06:54:54 spaetz: morning Sep 14 06:57:16 mrmoku: got a working testing? Sep 14 06:57:43 Heinervdm: going to reflash now Sep 14 06:58:03 Heinervdm: the one before going to bed was fine after a small fso-abyss upgrade Sep 14 06:58:11 mrmoku: ok Sep 14 06:58:17 Heinervdm: dos1 rebuilt the image then... so it should be fine Sep 14 06:58:21 * mrmoku verifies that now Sep 14 07:00:25 lindi-_: you're just great. If only you were the head of OM QA department, we would have something decent long time ago. Sep 14 07:11:39 PaulFertser: heh. hopefully this bug exposes some more important bug, the backlight blinking is not itself very useful feature Sep 14 07:12:00 lindi-_: you give us too much work ;) Sep 14 07:12:51 lindi-_: but really, i hope i'll find time and force to look at it. Sep 14 07:14:49 * spaetz wonders why masquerading the FR connection through my Laptop stopped working Sep 14 07:14:59 I can't update. heck... Sep 14 07:16:04 mrmoku: are we continuing to work towards using dev.oe? Or should we just continue to patch teh SHR/import branch Sep 14 07:16:15 spaetz, iptables -t nat -L Sep 14 07:16:24 spaetz, definitely we have to work towards dev.oe! Sep 14 07:16:26 and is shr/import following dev.oe so we get all the improvments from there? Sep 14 07:16:31 cz_jc: thanks, will try Sep 14 07:16:42 spaetz, not yet... that's where we go to Sep 14 07:16:42 spaetz, noo, I mean send output :) Sep 14 07:16:52 spaetz, there should be some SNAT rule I think Sep 14 07:16:52 having a branch following dev.oe to get all the fancies Sep 14 07:16:55 lemme check my server Sep 14 07:17:03 (won't be named shr/import though ;) Sep 14 07:17:20 spaetz, SNAT all -- 192.168.0.0/16 !192.168.0.0/16 to:10.108.31.215 Sep 14 07:17:24 spaetz, for example Sep 14 07:17:46 cz_jc: I was using the masquerade script from a howto... IPTABLES -t nat -A POSTROUTING -o $EXTIF -j MASQUERADE Sep 14 07:17:54 it used to work, but stopped some time ago. Sep 14 07:18:02 will retry Sep 14 07:18:13 Guys, news about BFS: Sep 14 07:18:21 spaetz, maybe due to the rename from usb to eth? Sep 14 07:18:21 the whole script involved a bit more, I grabbed it from some masquerade howto Sep 14 07:18:34 CFS that is mainline has a knob that makes it behave as good as BFS: http://marc.info/?l=linux-kernel&m=125260525104922&w=3 Sep 14 07:18:35 PaulFertser: yes? Sep 14 07:18:54 PaulFertser: any noticable difference? Sep 14 07:19:00 PaulFertser, good news... no need to carry patches around ;) Sep 14 07:19:02 for the FR, I mean Sep 14 07:19:23 BTW, I get kernel panics when I use mokonnect to connect to ym wep based WLAN Sep 14 07:19:37 spaetz, hehe... DocScrutinizer2 has that too Sep 14 07:19:45 WEP seems to be deeply broken Sep 14 07:19:47 PaulFertser: are you interested in logs and how do I get them :-)? Sep 14 07:20:00 dammit, I don't want to use wpa :-) Sep 14 07:20:12 spaetz: yes, i think i am. Sep 14 07:20:35 spaetz: probably it'd be nice if you posted them to the OM trac, with category set to system software Sep 14 07:20:38 mrmoku: the weird thing is that using iwconfig it works just fine Sep 14 07:20:53 only when using mokonnect (which uses connman), I get the panic Sep 14 07:21:19 spaetz: can you trace the calls which connman uses? Sep 14 07:21:30 PaulFertser: no clue, I could give it a try. Sep 14 07:21:37 spaetz: i'm not sure i can easily reproduce your enviromnent on debian. Sep 14 07:21:57 spaetz: but it'd be nice to see what happens. Kernel backtrace would be nice anyway, please post it on trac :) Sep 14 07:22:03 PaulFertser: ok, I'll try to get some useful information and post to the list Sep 14 07:22:07 & trac Sep 14 07:23:48 spaetz: all tickets filed are posted to the kernel list automatically Sep 14 07:25:04 will need to find a way to get kernel backtraces though, never done that. But I'll google my way around Sep 14 07:25:16 spaetz: dmesg Sep 14 07:25:29 mrmoku: that would be a nice freature for shr-devel and shr trac too ;) (post filed tickets to ml) Sep 14 07:26:45 PaulFertser: but once I have a kernel panic, it's hard to call dmesg Sep 14 07:26:50 as everything is frozen :-) Sep 14 07:27:27 Heinervdm: it would be mostly spam. YOu can subscribe to all wiki and trac changes as an RSS feed (that is what I do) Sep 14 07:27:56 not real spam, butwhy force all changes on everyone if you have the RSS feed? Sep 14 07:28:11 Heinervdm, bug Ainulindale for it :P Sep 14 07:28:25 Heinervdm, hmm... for the tracer I might do it too Sep 14 07:28:27 * mrmoku looks Sep 14 07:28:37 spaetz: aha... Sep 14 07:28:42 spaetz, matching by interface is not a good idea.. Sep 14 07:28:42 spaetz: do you have a debug board? Sep 14 07:29:05 spaetz, also masquerade is mostly for dial up connections when IP can change Sep 14 07:29:10 spaetz: if not, is it too hard for you to selfcompile a kernel with lindi-_'s ramconsole? Sep 14 07:29:10 Heinervdm: For that "feature" I think it's annoying Sep 14 07:29:16 spaetz, snat is recommended over it whenever you can use it Sep 14 07:29:19 (we should really have it by default...) Sep 14 07:29:22 Ainulindale: ok :) Sep 14 07:29:23 I think it's only useless spam trac has a timeline feature Sep 14 07:29:43 But if you disagree just suggest the idea on the ml :-) Sep 14 07:30:09 PaulFertser: no, no debug board here Sep 14 07:30:20 cz_jc: thanks, I'll try to go the snat way then. Sep 14 07:30:29 Ainulindale: what i'm thinking is that a lot of tickets aren't seen, because nobody knows that there are new one's Sep 14 07:30:41 PaulFertser: compiling a kernel would be ok Sep 14 07:30:54 ramconsole, ok thanks for the hint Sep 14 07:31:11 btw. newest andy-tracking works for me really beautifully :) Sep 14 07:31:39 cz_jc: haha, it will break soon ;) Sep 14 07:31:52 cz_jc: transition to .31 won't be glitchless ;) Sep 14 07:32:07 PaulFertser, aah you're doing it ? :) Sep 14 07:32:17 cz_jc: not yet Sep 14 07:32:17 PaulFertser, then.. good luck.. I'll try testing :) Sep 14 07:32:24 cz_jc: i'm waiting for larsc directions Sep 14 07:32:45 PaulFertser, k ^^ Sep 14 07:33:45 PaulFertser, Ainulindale, and anyone with a gta01 in his hands: some one should test the new testing :) Sep 14 07:38:52 mrmoku: with the latest patch for the calypso? good. I think i'll try it today evening, when i get home from work. Sep 14 07:39:51 PaulFertser, which patch do you mean? Sep 14 07:40:52 PaulFertser, ogsmd: [TI CALYPSO]: on closing, drop out of muxing mode and power off the modem (AT@POFF) Sep 14 07:40:55 this one? Sep 14 07:41:02 mrmoku: yeah :D Sep 14 07:41:11 than it is in :) Sep 14 07:41:16 s/than/then/ Sep 14 07:41:17 mrmoku meant: then it is in :) Sep 14 07:42:22 Heinervdm, testing looks good so far Sep 14 07:43:21 mmh, do we have a new testing candidate? Sep 14 07:43:23 cool Sep 14 07:43:38 http://build.shr-project.org/tests/mrmoku/testing/images/om-gta02/shr-lite-glibc-ipk--20090913-om-gta02.rootfs.jffs2 Sep 14 07:44:10 mrmoku: oh, testing? I hoped you want me to test unstable Sep 14 07:44:17 PaulFertser, :P Sep 14 07:44:51 PaulFertser, I need someone with a GTA01 to give it a short test if it boots, registers and somehow works Sep 14 07:44:59 mrmoku: cool, will flash that Sep 14 07:45:30 and we do need a way to have people update translations without bothering the email list Sep 14 07:45:33 PaulFertser, and after that you will get something real unstable ;) Sep 14 07:45:40 mrmoku: hm, i think i can try that on my other SD. Sep 14 07:45:40 this is the most suboptimal way possible Sep 14 07:46:16 spaetz, you mean giving them git access? Sep 14 07:46:43 I have no problem with that... on the other hand with patchwork it is quite easy to handle those patches Sep 14 07:49:30 PaulFertser, would be very welcome :D Sep 14 07:51:19 mrmoku: i'm not sure it will register though because i have my battery connected through some wires but i'll try anyway. Sep 14 07:51:40 PaulFertser, ok, thanks Sep 14 07:59:27 mrmoku: not necessaritly git access. Sep 14 07:59:51 there's tools for that at launchpad.net and transifex.org(org?) where people can maintain translations Sep 14 08:00:22 and we only pull them once in a while. If there is a git tree maintained where we could pull from, all the better though :-) Sep 14 08:02:59 PaulFertser: I'm very interested if it will register Sep 14 08:03:33 joerg_42: ok. Unfortunately it has not only wires but also some old huge soviet switch of questionable quality. Sep 14 08:04:31 soviet switch never fails ;-) rock solid Sep 14 08:05:43 joerg_42: just sometimes requires fixing with a hammer (carmaggedon-style) Sep 14 08:07:34 mrmoku: can you please try this interactivity knob? Probably it worth setting as default for now: Sep 14 08:07:37 mkdir /debug; mount -t debugfs none /debug; echo NO_NEW_FAIR_SLEEPERS > /debug/sched_features Sep 14 08:10:09 could any of the linux system level cracks share a pointer to a site explaining foundation rationale and best practices of sysfs, please? I seen /dev drivers that simply interface to a /sys node - maybe do some sanity checking or muxing. And aiui sysfs is all owner:root and userland not supposed to have any access whatsoever Sep 14 08:11:16 PaulFertser: what's that magic? O.o Sep 14 08:12:52 DocScrutinizer-8: just disables some questionable for interactive systems feature of in-kernel CFS. Sep 14 08:13:32 wow - real voodoo Sep 14 08:14:09 DocScrutinizer-8: found by one of "end-users" in the middle of LKML discussion provoked by BFS announcement. Sep 14 08:14:40 what's B? bastard? Sep 14 08:14:41 PaulFertser: ok, will try Sep 14 08:14:51 DocScrutinizer-8: brainfucked scheduler Sep 14 08:14:58 hehe Sep 14 08:16:49 actually I thought about scheduler myself yesterday. Maybe tuning the timeslices might be worth it as well Sep 14 08:18:12 too much context switching will slow down the performance Sep 14 08:18:25 I can't see how to measure effect of that setting on FR anyway but i hope mrmoku can try. Sep 14 08:20:11 what's NO-NEW-FAIR-SLEEPERS doing? Sep 14 08:24:43 DocScrutinizer-8: if you think i'm that pro to know CFS internals, you're wrong ;) Sep 14 08:25:41 well, you know debug fs ;-P Sep 14 08:26:43 PaulFertser: do you know sysfs? Sep 14 08:26:58 DocScrutinizer-8: if i knew, i'd answered your question yesterday Sep 14 08:28:31 darn, I hoped it's a simple ignoring :-S Sep 14 08:32:04 PaulFertser: let's chat using our ignorance then :-) Sep 14 08:32:31 PaulFertser: you agree sysfs can't be chown'ed? Sep 14 08:32:48 or chmod'ed? Sep 14 08:34:57 PaulFertser: of which sane ways for a userland process could you think to access e.g. gmeters? (luckily we got FSO for this. I'm asking for other ways) Sep 14 08:41:01 huh? what? chanserv sick? Sep 14 08:55:44 DocScrutinizer-8: no, sysfs nodes can be chown and chmoded Sep 14 08:56:01 DocScrutinizer-8: gmeters are input devices, hence through input subsystem. Sep 14 08:56:37 ooooh Sep 14 09:08:02 PaulFertser: pong Sep 14 09:09:17 PaulFertser: shouldn't talk nonsense before first coffe kicks in Sep 14 09:10:00 rtp: hey Sep 14 09:10:13 rtp: so, are you the author of s3c udc driver? Sep 14 09:10:51 PaulFertser: according to the source code, yes :) Sep 14 09:11:18 rtp: sorry, didn't realise it. I assumed it was pretty much abandoned since nobody cared about that windows wsod thing ;) Sep 14 09:11:58 rtp: it would be very nice if you tried to solve g_file_storage problems then, i suspect they expose some bug/oddity in the udc driver. Let me pass you a link to om tracker. Sep 14 09:12:27 PaulFertser: the main problem is that I don't (and want) window$ so this doesn't help solving window$ related bugs /o\ Sep 14 09:12:56 rtp: bug is still bug, you know Sep 14 09:13:20 rtp: (g_file_storage) here it goes: http://docs.openmoko.org/trac/ticket/2278 Sep 14 09:13:28 PaulFertser: not necessary in the udc. there's a problem with 244x cores which sometimes are triggering an interrupt with no interrupt flags so this is treated as error /o\ Sep 14 09:13:52 PaulFertser: there's a hack for that in the om tree but I've heard it's not enough Sep 14 09:14:00 rtp: oh :( it'd be still nice to have a working workaround. Sep 14 09:14:40 PaulFertser: I'm waiting for my friend who made a better workaround to have time to send me it Sep 14 09:15:07 after that, if the bug is still there we'll have to investigate more. Sep 14 09:15:40 rtp: btw, i see you're involved in another driver -- s3c touchscreen. I have an observation that gta01 current is lowered by ~20mA by physically disconnecting LCM (with the display blanked and backlight off). We suspect it's TS drawing power. Do you probably have any ideas about it? Sep 14 09:17:11 the ts driver has no control over the touchscreen power Sep 14 09:18:06 did you try with ts driver unloaded ? it should at least power down the 24xx adc/ts device Sep 14 09:18:17 rtp: hm, i will, thanks for the tip Sep 14 09:18:52 rtp: there's some odd difference between gta01 and gta02 which i don't fully understand yet. gta01 uses some 4 additional pnp/npn transistors comparing to gta02. Sep 14 09:18:58 if with the ts driver unloaded the power consumption is lower, then it's really related to the ts /o\ Sep 14 09:19:51 rtp: (g_file_storage) it'd be nice if you added something to the trac ticket or should i do it for you? Sep 14 09:20:41 PaulFertser: it's strange that there are more transistors but I don't remember all diff between 2410 and 2442 cores so there may be good reasons (I hope) Sep 14 09:21:09 PaulFertser: I don't remember if I've an accound on OM track :( Sep 14 09:21:13 rtp: (TS) i guess there's no such a thing like "powering touchscreen". I can control SoC pins with userspace utility and just set all the relevant to input. Sep 14 09:21:35 rtp: (trac) np, i'll do it myself. Sep 14 09:22:38 what's happening if you put the pins as input ? power consumption drops ? Sep 14 09:23:57 rtp: haven't tried yet, sorry. Sep 14 09:24:03 rtp: only ts driver has control oper powering the readout voltage to X and Y plane Sep 14 09:24:24 s/ope/ove/ Sep 14 09:24:24 DocScrutinizer-8 meant: rtp: only ts driver has control over powering the readout voltage to X and Y plane Sep 14 09:24:56 rtp: basically that's main duty of ts driver Sep 14 09:26:23 rtp: (g_file_storage) interesting thing is that it works without issues with my old desktop and with all BIOSes i tried (i want to boot slax from my FR), but as soon as a kernel is loaded, it fails. Sep 14 09:27:09 rtp: ah, it's unclear: with my desktop it works like it should. With all other machines i tried only bios can access the emulated storage. Sep 14 09:27:13 rtp: 2442 has a dedicated set of function IO for this. 2410 needs to emulate this by explicitly powering switches to VDD and GND Sep 14 09:27:28 DocScrutinizer-8: iirc, the ts driver puts the ts in "waiting mode" which means that the X and Y plane stuff should not cosume Sep 14 09:28:39 rtp: waiting mode should connect one plane to GND *or* VDD and other plane senses pull-down/up in case of contact Sep 14 09:29:20 rtp: basically that means only *one* of the both switches for one plabe should be enabled Sep 14 09:29:59 the other plane floats on the ADC inputs internal pullup/down Sep 14 09:31:30 DocScrutinizer-8: I don't remember all the details, sorry. I Sep 14 09:31:31 if in waiting-mode the driver enables both instead of one switch, then obviously power is eaten across that "R" between VDD and GND Sep 14 09:33:20 rtp: PaulFertser: I suggest checking waiting-mode setup in gta01 ts driver Sep 14 09:33:44 DocScrutinizer-8: haha, it was already implied by previous messages ;) Sep 14 09:35:01 PaulFertser: I'm practicing to find more unambiguous wording Sep 14 09:46:21 DocScrutinizer-8: have you reported your crash bug somewhere? Sep 14 09:46:28 DocScrutinizer-8: is the same as http://docs.openmoko.org/trac/ticket//2309 maybe? Sep 14 09:46:47 lindi-: nope Sep 14 09:47:24 lindi-: I'm too lazy ;-) Sep 14 09:48:04 lindi-: ...and I know others will face that bug as well. So eventually there will be a ticket be opened for that anyway Sep 14 09:49:17 DocScrutinizer: i am not sure if I have hit the same bug or not Sep 14 09:49:26 DocScrutinizer: was yours somehow always related to wlan? Sep 14 09:49:29 lindi-: I spent much effort for talking people into having ramconsole by default. As long as there isn't I don't see any use in reporting kernel panics Sep 14 09:50:05 lindi-: yes, always related to GPRS disconnect afaict Sep 14 09:50:34 err, to connecting GPRS to WEP (or to my particular AP) Sep 14 09:50:55 I wish I knew more about kernel hacking to make ramconsole get included in mainline Sep 14 09:51:08 disconnect bug is an entirely different issue Sep 14 09:51:10 (I did send one mail to lkml about a related thing) Sep 14 09:51:31 DocScrutinizer: gprs or wlan? Sep 14 09:51:35 (I'm confused) Sep 14 09:51:50 duh. My brain still asleep. Sep 14 09:52:06 WLAN connect to WEP -> panic Sep 14 09:52:21 GPRS disconnect -> frameworkd freeze Sep 14 09:52:50 DocScrutinizer: thanks Sep 14 09:53:04 DocScrutinizer: then I have different bug, my crash is not wlan related Sep 14 09:53:15 not for these two 3-words-posts ;-) Sep 14 09:54:22 lindi-: though:: (guess!) GPRS disconnect during suspend -> black brick Sep 14 09:54:36 DocScrutinizer: uh Sep 14 09:54:41 DocScrutinizer-8: mtdoops already included in mainline and can be used if you run from SD and can spare some mtd partition. Sep 14 09:56:24 PaulFertser: as long as it needs "even me" to think how to activate it, I won't give it a try ;-) Sep 14 09:57:35 PaulFertser: how about you just include ramconsole patch together with a patch that sets mem=127M? Sep 14 09:57:47 PaulFertser: with a warning that the patch might be reworked in the future Sep 14 09:57:57 I gave my reasoning for this some months ago: We will never see a stable kernel if we dn't give a means to open meaningful tickets on oops/panic to the "normal user" Sep 14 09:58:12 lindi-: i guess we should finally do that, yes. Sep 14 09:58:22 lindi-: too bad we lack kernel maintainer Sep 14 09:59:07 PaulFertser: well you seem to have committed some patches to andy-tracking lately :) Sep 14 10:00:07 my suggestion been: make mem=127M (or better ramconsole=1M, or even 512K) the *default* that kernel starts with (if there are no real cmdline parameters). If you want to disable all that, then set RAMCONSOLE=0 Sep 14 10:00:32 lindi-: we're moving away from andy-tracking anyway Sep 14 10:00:46 yep Sep 14 10:05:54 kernel then does mem=mem - ramconsole (-> mem=127.5M, 0.5M buffer for ramconsole). Then the "writing" to this buffer gets an addititional "if (buffersize > 0) then..." stopping to trash memory of we gave RAMCONSOLE=0 Sep 14 10:05:59 tmzt, hi btw If I remember well the palm pre send a lot more and more personal infos than just df-h and partition layout Sep 14 10:07:19 Gnutoo: isn't that a nasty joke? :-S Sep 14 10:07:33 DocScrutinizer, found it...GPS coordinates Sep 14 10:07:36 Gnutoo: (of the palm devels) Sep 14 10:07:45 ouch!! Sep 14 10:07:50 I don't like it if it's a joke Sep 14 10:08:10 who developed this OS? homeland security?? Sep 14 10:08:16 from my slashdot feed: Sep 14 10:08:22 "AceJohnny writes "Joey Hess found that his Palm Pre was ratting on him. It turns out the Pre periodically uploads detailed information about the user to Palm, including the names of installed apps, application usage (and crashes), as well as GPS coordinates. This, of course, is without user consent or control. The only way he found to disable the uploads was to modify system files." Sep 14 10:09:04 all what I know is that they used openembedded and didn't gave back the recipes Sep 14 10:09:07 WTF?!?!! Sep 14 10:09:36 .oO(industry standard behaviour) Sep 14 10:10:14 yes indeed it's so common....everybody in proprietary software does it(spy its users) Sep 14 10:10:21 roh: *my* standard behaviour: write a virus to change exactly one URL ;-) Sep 14 10:10:29 everybody = microsoft,apple,real networks etc... Sep 14 10:10:56 Gnutoo i meant the stealing foss code for proprietary devices part Sep 14 10:11:08 maybe someone with enough oe skills + python or vala skills could change that and port SHR on the device Sep 14 10:11:11 s/palm.com/binladen.net/ XD Sep 14 10:11:12 the rest.. well.. i dont buy hw which requires use of binary tools anymore Sep 14 10:11:38 ah ok it depends...on the company Sep 14 10:11:42 no. Sep 14 10:11:49 depends on the behaviour. Sep 14 10:11:54 ok Sep 14 10:12:00 companies change. what they do is what matters Sep 14 10:12:16 roh: what about DSPs? Sep 14 10:12:48 if you got a compiler in foss code and have a way of uploading the compiled results to it without using binaries... fine Sep 14 10:12:49 for instance cisco+linksys volated the gpl a lot of time....the fsf even attacked them because of the lack of compliance.... Sep 14 10:12:52 same as cpus Sep 14 10:13:04 s/volated/violated Sep 14 10:13:25 most products which reach us in europe sadly aren't comppliant. Sep 14 10:13:33 ouch... Sep 14 10:14:00 but mostly not intended because of ill-will but about the fucked up process the industry and their sales structures run Sep 14 10:14:01 I know a non-compliant device...but would have bought it if it was compliant Sep 14 10:14:03 roh: i mean there's not much support for decent DSPs in free compilers. Sep 14 10:14:08 it was an internet radio Sep 14 10:14:33 PaulFertser there is also not much one can do with seriously anymore since we have simd instructions and multicore cpus Sep 14 10:15:46 roh: also afaik all fpga's require proprietary tools for tracing/layouting etc. Sep 14 10:16:42 u bet Sep 14 10:17:02 also they are expensive and difficult to code for (atleast at first) and thus not used much in consumer equipment (which is what we all use at home in the end) Sep 14 10:17:19 luckily it's not that easy to nasty in a dsp or fpga Sep 14 10:17:29 roh: but mad useful for manny hobby projects Sep 14 10:18:17 roh: that's true for fpga but not for dsp Sep 14 10:19:00 also one really needs a reason. those beasts are power-hungry Sep 14 10:19:37 Doc same true there for dsp Sep 14 10:20:53 DocScrutinizer-8: adding a dsp to some arm9 design is much more expensive and yields much less apps and 'bang' out of the buck than using an arm11 with simd and better caches with less engineering, and more productive coders since also debugging is much easier. Sep 14 10:20:55 roh: you wouldn't believe how many consumer devices have a DSP Sep 14 10:21:07 DocScrutinizer-8: dsp yes. e.g. my ac3-amp Sep 14 10:21:34 DocScrutinizer-8: but for mobile devices... not often. not neccessary. thats where asics save your ass and battery Sep 14 10:21:51 sure Sep 14 10:22:11 though each modem has a dsp nowadays Sep 14 10:22:26 once was... now also a asic again Sep 14 10:22:47 each dect phone has Sep 14 10:22:48 sometimes only tables for transformations get reloaded (some reciever/demods do that) Sep 14 10:23:09 well.. kids toys. (power wise) Sep 14 10:23:25 yep Sep 14 10:23:37 also mask-programmed or even asic-ified Sep 14 10:23:54 similar to calypso Sep 14 10:24:22 and I'm not afraid they will replace words in my speech they are processing Sep 14 10:24:58 if youre using dect you arent afraid of spublic speaking anyways #';) Sep 14 10:25:01 so I don't care neither DSP nor fpga code Sep 14 10:25:43 i wanted to play with cpld soon.. just for fun and learning vhdl or verilog (whichever i like more in the end) Sep 14 10:27:45 * DocScrutinizer-8 makes note to himself to check if wireshark has full control over maemos ip-stack Sep 14 10:28:30 hi guys...is possible to change sd card clock from booted system or i must do it another way? Sep 14 10:28:59 makyou can change on the fly Sep 14 10:29:09 Majkhii: ^ Sep 14 10:29:23 DocScrutinizer-8: how? Sep 14 10:30:08 there's some sysfs nodes in /sys/**glamo** Sep 14 10:32:47 module/glamo_mci/parameters/sd_max_clk? Sep 14 10:39:43 sounds about right Sep 14 10:48:54 mickeyl, ping Sep 14 10:50:39 mrmoku: moinmoin :-) Sep 14 10:51:39 DocScrutinizer-8, yo Sep 14 10:53:41 PaulFertser, difficult to judge... but my feeling is that application switching in illume is a bit more responsive with that thingie enabled Sep 14 10:54:04 PaulFertser, how to turn it off again to see the difference? Sep 14 10:57:22 hehehe Sep 14 11:03:17 mrmoku: there should be a sysfs file which lets you choose the scheduler Sep 14 11:04:03 Hi, I'm working on making libframework-phonegui-efl theme able easier.. any advice wrt my e-mail from yesterday? Sep 14 11:04:30 bumbl, though I did not change the scheduler... just set some option for it I think Sep 14 11:05:29 JaMa, mail to where? Sep 14 11:07:50 mrmoku: community list Subject: Re: nEo theme issues Sep 14 11:08:19 ahh, ok Sep 14 11:08:20 * mrmoku reads Sep 14 11:08:20 mrmoku: http://lists.openmoko.org/pipermail/community/2009-September/055159.html Sep 14 11:09:32 ie for libframework i just installed .edj files to shr subdirestory and in postinst created link default -> shr and all stuff from default is linked to libframework-phonegui-efl.. then with neo theme I can just change link default to neo and its done.. Sep 14 11:10:14 mrmoku: try ommitting the leading "NO_" Sep 14 11:10:15 JaMa: hmm I have no strong opinion about that... as long as it works :P Sep 14 11:10:22 bbiab Sep 14 11:11:17 mrmoku: then do a "cat /sys/kernel/debug/sched_features" Sep 14 11:11:26 mrmoku: ok, I'll send patch later Sep 14 11:11:40 JaMa: i suggest to use update-alternative Sep 14 11:12:11 DocScrutinizer: ls /sys/kernel/debug/ is empty on SHR Sep 14 11:13:49 spaetz: hmm, I c&p without scrutinizing it :-/ my fault. Probably that should be "cat /debug/sched_features" then Sep 14 11:13:58 Heinervdm: can it be used on directory links? Sep 14 11:14:24 Heinervdm: hmm seems so.. thanks.. Sep 14 11:14:34 JaMa: :) Sep 14 11:18:59 DocScrutinizer: ahh, had to mount mount -t debugfs None /sys/kernel/debug first Sep 14 11:19:24 cat /sys/kernel/debug/sched_features Sep 14 11:19:25 NEW_FAIR_SLEEPERS NORMALIZED_SLEEPER WAKEUP_PREEMPT START_DEBIT AFFINE_WAKEUPS CACHE_HOT_BUDDY SYNC_WAKEUPS NO_HRTICK NO_DOUBLE_TICK ASYM_GRAN LB_BIAS LB_WAKEUP_UPDATE ASYM_EFF_LOAD NO_WAKEUP_OVERLAP LAST_BUDDY Sep 14 11:19:37 spaetz: [2009-09-14 10:07:37] mkdir /debug; mount -t debugfs none /debug; echo NO_NEW_FAIR_SLEEPERS > /debug/sched_features Sep 14 11:19:53 yep, sorry missed that part :-) Sep 14 11:20:24 any way to benchmark changes? Sep 14 11:20:58 mrmoku: [2009-09-14 13:19:24] cat /sys/kernel/debug/sched_features... NEW_FAIR_SLEEPERS Sep 14 11:21:46 mrmoku: as expected this is the "leading NO_" pragma on that interface Sep 14 11:21:52 mmh, feels snappy. Although it's hard to measure it Sep 14 11:22:04 so I can't really tell whether it's better Sep 14 11:24:45 in some ml-post I read "as you're using x86, did you try NO_FAIR..." Sep 14 11:25:26 * DocScrutinizer-8 wonders what all that actually is supposed to do, and if it applies to FR at all Sep 14 11:25:48 DocScrutinizer-8, cat: can't open '/sys/kernel/debug/sched_features': No such file or directory Sep 14 11:26:03 mrmoku: read backlog Sep 14 11:26:10 hehe Sep 14 11:26:15 err backscroll Sep 14 11:26:41 mrmoku: I have that in my rc.local now: Sep 14 11:26:55 mount -t debugfs none /sys/kernel/debug Sep 14 11:26:55 echo NO_NEW_FAIR_SLEEPERS > /sys/kernel/debug/sched_features Sep 14 11:26:56 umount /sys/kernel/debug Sep 14 11:29:29 spaetz, do you notice a difference? Sep 14 11:29:51 ahh... backlog :P Sep 14 11:29:56 :) Sep 14 11:30:03 mrmoku: I hate those subjective measurements. Yes, scrollng my contact list feels snappy. But is it really faster or some placebo? I can hardly tell ... Sep 14 11:30:29 spaetz, lets test it today and make it default tomorrow :) Sep 14 11:30:44 at least the scheduler maintainer was for that change, so it can't be that bad :-) Sep 14 11:31:37 http://git.kernel.org/?p=linux/kernel/git/x86/linux-2.6-tip.git;a=commitdiff;h=3f2aa307c4d26b4ed6509d0a79e8254c9e07e921 Sep 14 11:31:57 Midori 0.1.10 hit OE Sep 14 11:32:47 as for improving performance, I think we should make x11-16 the default, that help definitely... :_) Sep 14 11:32:54 helps... Sep 14 11:33:21 * spaetz just ordered a 10" ebook reader. woot. 600€ down the drain... Sep 14 11:33:28 yes, segfaults significantly faster XD Sep 14 11:34:00 DocScrutinizer: bah. I haven't experienced segfaults yet. But might well be true Sep 14 11:34:25 Still, I take an occassional segfault over the slow-as-molasses-GUI any day Sep 14 11:39:49 mrmoku: DocScrutinizer-8: /me was away compiling some uber-shit driver written by morons, manually applying instructions from another compilation/installation script written by other morons (which included "tr -d '\r'", custom weird patch, etc) and then trying to use that driver with another bunch of badly written python and c apps. Sep 14 11:41:02 Ah, and also struggling with gedit, that useless app played a trick on me when i copy-pasted from it. While selecting i swapped to bytes by occasion (mouse selection, that's it), and it became tr -d ' \r' which removed all the spaces from sources. Sep 14 11:42:49 tztztz Sep 14 11:45:47 PaulFertser, sounds like fun ;) Sep 14 11:46:12 PaulFertser: any idea what that no-fair-sleepers-foo actually is doing to scheduler? Sep 14 11:54:03 DocScrutinizer-8: no, sorry. Ingo: "Btw., NO_NEW_FAIR_SLEEPERS is something that will turn the scheduler Sep 14 11:54:09 into a more classic fair scheduler (like BFS is too). Sep 14 11:59:08 PaulFertser: sorry. I repeat myself when half-asleep Sep 14 12:01:06 hmm, NO-NEW-FAIR-SLEEPERS seems to increase average power consumption of the system :-/ Sep 14 12:01:22 marginally Sep 14 12:05:48 or maybe not Sep 14 12:06:43 anyone knows a daap library written in C? Sep 14 12:07:29 hi Sep 14 12:07:37 Heinervdm: ping Sep 14 12:07:44 khiraly1: pong :) Sep 14 12:08:00 Heinervdm: how can I check which revision of elementary ships shr? Sep 14 12:08:47 khiraly1: for shr-unstable it will be bumped soon to HEAD Sep 14 12:10:02 khiraly1: shr-testing it's 41040 Sep 14 12:10:12 new shr-testing Sep 14 12:11:41 aham Sep 14 12:12:11 I want 42207 ;) Sep 14 12:12:22 http://trac.enlightenment.org/e/changeset?new=42207%40trunk%2FBINDINGS%2Fpython%2Fpython-elementary&old=41949%40trunk%2FBINDINGS%2Fpython%2Fpython-elementary Sep 14 12:12:26 ^^ these changes Sep 14 12:15:41 a lot of people are wating for newer EFL rev :) Sep 14 12:15:50 Heinervdm: how to force shr to update enlightenment? Sep 14 12:16:12 there is also an evas fixes too, what Im waiting for Sep 14 12:16:17 khiraly11: it will be updated after testing release Sep 14 12:16:28 and when is the testing release? Sep 14 12:16:39 hopefully today ;) Sep 14 12:16:41 .) Sep 14 12:17:41 Im waiting for 42115 too Sep 14 12:18:04 http://trac.enlightenment.org/e/changeset?new=42115%40trunk%2FBINDINGS%2Fpython%2Fpython-edje&old=41338%40trunk%2FBINDINGS%2Fpython%2Fpython-edje Sep 14 12:18:05 khiraly11: that's eve? Sep 14 12:18:18 nope, it adds evas box function to python Sep 14 12:18:25 with eve, hmm, wait a sec Sep 14 12:18:35 eve is up to date Sep 14 12:19:20 khiraly11: bumping efl is always a big pain, so this is done as seldomly as possible :-) Sep 14 12:20:38 spaetz: I made a small patch, and also forced e developers to fix an another bug for me Sep 14 12:20:45 all the two is pretty important. Sep 14 12:21:28 (first: you can obtain the evas object of an elementary widget, which is necessary if you want to mix evas with elementary. Needed by my gps app) Sep 14 12:22:01 second: i implement a suggestion box for the freerunner (it flows horizontally rather then vertically) Sep 14 12:22:25 nice Sep 14 12:22:51 so I *need* this new python evas and python-elementary bindings;) Sep 14 12:23:30 yep, that's the problem with living on the technology edge :-) Sep 14 12:24:19 Heinervdm: I started looking for browsers, and made some sketches for better ui. I studied nokia n97, apple iphone browser, palm os browser, blackberry, and the demo video of samsung jet phone Sep 14 12:26:27 Heinervdm: so Im collaborating with eve's developer (Gustavo) on this Sep 14 12:27:04 khiraly11: i like this approach to closely study what's already there Sep 14 12:27:10 khiraly11: nice, then it will get the best browser for FR, hopefully Sep 14 12:27:37 yepp. I keep in mind that freerunner lack of multitouch, which is key feature for iphone and palm os Sep 14 12:28:03 DocScrutinizer: I watched like 20 videos on youtube, and played with a friend's iphone Sep 14 12:29:05 DocScrutinizer: there are other browsers too, but they are mostly in the joke category Sep 14 12:29:06 like windows moob Sep 14 12:29:07 ;)) Sep 14 12:29:31 yep Sep 14 12:29:59 though opera really isn't that bad on some details Sep 14 12:30:56 yeah, I missed opera. I will look into it too. Sep 14 12:31:10 However I have a clean vision how it should look;) Sep 14 12:33:39 khiraly11: will gustavo enable fingerscrolling too? Sep 14 12:33:57 hope se Sep 14 12:33:58 so Sep 14 12:34:13 good Sep 14 12:34:33 mandatory, no? Sep 14 12:36:07 also some gesture to scroll down (/up) exactly one screen (- some overlap) is a must-have imho Sep 14 12:37:37 DocScrutinizer: dunno, especially the scrolling is seamless Sep 14 12:38:22 DocScrutinizer: btw, none of the most usable browsers have it (iphone, palm) Sep 14 12:38:35 check opera Sep 14 12:41:06 Gnu: I understand the gps data to be seperate from the rest and sent to qualcomm or your carrier, not palm Sep 14 12:41:21 there's a number of rather intuaitive ways to scroll in opera, depending on where scrolling gesture starts and where it stops Sep 14 12:42:17 as for the or recipes, that may be lack of a repo which they have not setup, just posted tarballs of patches Sep 14 12:42:22 and source Sep 14 12:42:37 but we can ask Sep 14 12:43:16 khiraly11: what I'm missing in opera is a gesture "home" and "end" - which might be to balistic scrolls in fast sequence Sep 14 12:44:01 s/be to/be two/ Sep 14 12:44:03 DocScrutinizer-8 meant: khiraly11: what I'm missing in opera is a gesture "home" and "end" - which might be two balistic scrolls in fast sequence Sep 14 12:46:33 khiraly11: it's all about position of "button-down", delay between button-down and start of drag, overall speed of drag, speed of drag immediately before button-up, and position of button-up Sep 14 12:48:22 khiraly11: e.g if speed of drag before button-up is high, then do balistic scroll, otherwise do proportianal scroll Sep 14 12:48:51 DocScrutinizer: for zoom please check samsung jet's demo Sep 14 12:49:00 anal scroll - nice XD Sep 14 12:49:28 what are they doing for zoom? Sep 14 12:50:03 press and hold, and when a small guidance appear slide up or down (for zoom in or out), and the guidance will change Sep 14 12:50:08 http://jet.samsungmobile.com/#/feature Sep 14 12:50:15 Feature->internet browsing Sep 14 12:50:24 (it requires flash btw) Sep 14 12:50:45 sounds good Sep 14 12:51:50 while opera requires double and triple taping Sep 14 12:52:46 though I somehow like the double-click zoom-in that's to be found on many browsers. Only for an additional means and keep some stepless zooming though Sep 14 12:53:04 hi all! Sep 14 12:53:35 khiraly11: iirc you do stepless zoom on opera rather similar way Sep 14 12:53:42 I find cool the new shr-today lock... but I think I should disable the old simple lock... Sep 14 12:54:06 where is the configuration so I keep only the new one? Sep 14 12:54:10 a zoom is never stepless imho Sep 14 12:54:21 a21_work: wrench-> screensaver Sep 14 12:54:47 you can disable the default lock through illume settings. Sep 14 12:56:15 DocScrutinizer: with double clicking there is a feature to fullscreen the actual paragraph of a webpage Sep 14 12:56:20 dunno in which browser I have seen it Sep 14 12:58:33 khiraly11: opera? Sep 14 12:58:44 nope. Sep 14 12:58:53 Im watching opera's features currently Sep 14 12:59:10 i think iphone has this feature Sep 14 12:59:16 * DocScrutinizer-8 starting fsckng winmoob to check opera Sep 14 12:59:29 as there is multitouch for zooming Sep 14 12:59:38 or palm os' browser Sep 14 13:00:20 khiraly11: if only FR had 3g or at least edge (and it was cheap here) or we had wifi spots everywhere... Sep 14 13:00:40 keycode-177? Sep 14 13:00:50 a21_work: just delete it Sep 14 13:00:57 in keybindings? Sep 14 13:01:02 ""glofiish - minds on the move"" I wonder where to... XD Sep 14 13:01:04 yepp, in keybindings Sep 14 13:01:46 PaulFertser: wifi spots, and some kind of usb 3g stick Sep 14 13:02:03 would be nice if we could have one with low power consumption Sep 14 13:02:31 not these huaweea shit Sep 14 13:03:31 ouch... again netsplit on irc :P Sep 14 13:03:45 are we here? Sep 14 13:04:36 khiraly11: can u read me? :S Sep 14 13:04:37 yepp Sep 14 13:04:42 nope Sep 14 13:04:43 fine... Sep 14 13:04:44 ;) Sep 14 13:04:51 khiraly11: no, operas zoom slider is via app-menu which pops up on hold-button-down Sep 14 13:05:11 so, I deleted in key bindings key177, and now it's ok, right? Sep 14 13:05:21 a21_work: try out Sep 14 13:05:30 there are two keybindings one for power button and one for aux Sep 14 13:05:35 I hope you deleted the aux one Sep 14 13:06:09 yes yes, so this is ok :D Sep 14 13:06:09 because your instruction was a quite different ;) Sep 14 13:06:34 yes, now tested and I get the aspected behaviour Sep 14 13:06:45 hehe, I pretty much got fsckd by trying to define a new "keycode" for keybindings Sep 14 13:08:09 there actually *is* a way out of that pitfall, but I can't recall what it was. Anyway quite fubar Sep 14 13:08:26 now we have shr-today... shouldn't it be without k177 by default? Sep 14 13:10:38 anyway basically whole illume-wrench isn't exactly adapted to FR :-/ Sep 14 13:10:50 quite annoying Sep 14 13:11:13 yeah, this keybinding screen does not fit into 480x640 Sep 14 13:11:32 all this *might* work on a 1280*1024 Sep 14 13:11:54 khiraly11: try defining a *new* key ;-D Sep 14 13:12:10 dont want to. I erased all the two Sep 14 13:12:11 lol... easy! set your FR @1280*1024 ;) Sep 14 13:12:24 hahahah Sep 14 13:12:24 for powerbutton I defined suspend via frameworkd Sep 14 13:12:26 khiraly11: "press key to learn, or ESC to abort" Sep 14 13:12:56 http://wiki.openmoko.org/wiki/User:Khiraly#Power_button Sep 14 13:13:00 ^^ powerbutton setup Sep 14 13:13:02 sth like that Sep 14 13:13:38 raster told me there's some utility to decompile E config, change anything you want and recompile it. Sep 14 13:14:31 ok, illume topbar [X] works for "ESCAPE" Sep 14 13:14:58 no need to mess with illume's keybindings Sep 14 13:15:18 decompile E config -> invent a new gui to manage settings in shr -> recompile :D Sep 14 13:16:02 a21_work: huh? what's wrong with shr-settings Sep 14 13:16:29 beyond all the things I pestered dos1 already ;-D Sep 14 13:17:14 a21_work: probably you're talking about illume settings? Sep 14 13:17:27 a21_work: nah, there's vim for changing any config Sep 14 13:17:38 illume settings->new page in shr-settings Sep 14 13:17:54 which for sure would gain a lot from simply reducing default font to 3pt Sep 14 13:19:19 we could get rid of the illume settings and unify all them in shr-settings Sep 14 13:19:40 all the stuff in a single place Sep 14 13:19:47 and _usable_ Sep 14 13:19:52 on our FR Sep 14 13:21:36 actually the random split into two places where to tweak settings feels rather odd. Especially since some of them seem redundant Sep 14 13:22:06 I agree Sep 14 13:22:43 eg: the new display engine setting Sep 14 13:23:20 if we can get every relevant illume configuration in shr-settings... we can remove also the wrench ;) Sep 14 13:23:49 not every, just what is really needed Sep 14 13:23:58 like 16 bit Sep 14 13:24:56 that's totally wrong Sep 14 13:25:04 and like keybindings ;) or accelerations and sesitivity of the display Sep 14 13:25:18 why? Sep 14 13:25:54 DocScrutinizer: wrong? Sep 14 13:26:52 from a usability and gui design POV this split and even more this duplication is not even pre-beta and not even rady for a proto showoff Sep 14 13:27:11 wasn't you? "actually the random split into two places..." Sep 14 13:27:28 will confuse hell out of users Sep 14 13:28:03 I'm saying to unify and use only shr-settings Sep 14 13:28:12 a21_work: I was replying to khiraly11 Sep 14 13:28:17 aah ok :P Sep 14 13:28:55 but I agree with khiraly11 about "not every, just what is really needed" Sep 14 13:29:05 DocScrutinizer: not really if you mark at the settings, that this can be also changed in wrench Sep 14 13:29:12 actually each function of illume-wrench duplicated in shr-settings is not improvement but a bug Sep 14 13:30:17 in shr settings could be a page named "illume settings on diet" Sep 14 13:30:30 and we would cherry pick some of the illume settings Sep 14 13:30:31 I agree Sep 14 13:30:40 and remove the wrench Sep 14 13:30:48 well, that sounds like it might be faintly bearable Sep 14 13:30:54 because it's not usable on FR Sep 14 13:31:03 (diet that is) Sep 14 13:31:33 no need to remove the wrench anymore then. Because it is not that big anyway Sep 14 13:31:49 if somebody want to use, go ahead. For us mere mortals, there are shr settings Sep 14 13:31:54 khiraly11: it is just confusing Sep 14 13:32:02 better fix wrench rather than remove it Sep 14 13:32:08 anyway I dont really care at the end, I set 16bit, and removed keybindings, and dont use illume settings anymore Sep 14 13:32:47 imho the wrench is a tool only for devs Sep 14 13:33:03 nonsense Sep 14 13:33:48 that's opinion of windoof users living all their live happily with default wallpaper Sep 14 13:33:56 ;) Sep 14 13:34:19 I have no problem with wrench as I dont use it at all;) Sep 14 13:34:54 I use it less than I'd like to, as it's fubar Sep 14 13:35:37 we can configure wallpaper from shr-settings ;) Sep 14 13:35:39 anyway setting launcher to doubleclick e.g. saved my nerves a lot Sep 14 13:35:56 and no, I do not use macroshit's software :P Sep 14 13:36:01 a21_work: another true windoof-user attitude Sep 14 13:36:37 ok, let configure everything by vim :P Sep 14 13:37:05 ""which icon to click to make this pic my wallpaper? can I please get a specialized tool for that?"" Sep 14 13:38:03 ""what??? a spelling error in default app? let's write a better one then"" Sep 14 13:38:48 sorry but... in shr-settings we already have "themes" Sep 14 13:38:53 ""... and it could even hide the menu entry of the default app then"" Sep 14 13:40:06 DocScrutinizer-8: Fixing illume-wrnech is jsut more difficult as that is code in the e repository Sep 14 13:40:10 a21_work: I bet illume-wrench had themes before shr was even thought of Sep 14 13:40:23 and we cannot just modify stuff to make it fit better to FR/SHR Sep 14 13:40:37 while we can play with shr-settings as much as we want Sep 14 13:41:02 spaetz: I agree Sep 14 13:41:09 so let's bin illume then, if we can't make it work with FR Sep 14 13:41:36 in the end DocScrutinizer convinced me;) Sep 14 13:41:43 * DocScrutinizer-8 shakes his head and wanders off Sep 14 13:41:55 DocScrutinizer-8: not the worst of all ideas, given that people already code replacements for a) the settings and b) the desktop drawing (aka launcher) Sep 14 13:42:20 there's not much of illume left then. It's just being used as a window manager and keyboard displayer, isn't it :-) ? Sep 14 13:44:39 last time I heard "we can't fix that" was when I reported a bug in windoof active-x. Never heard that in FOSS Sep 14 13:45:35 raster said multiple times, that illume's keyboard cant be transparent and above of other windows Sep 14 13:45:44 (although dont remember the exact phrasing) Sep 14 13:47:21 khiraly11: yep, I remember that too. That reminds me. I still want a keyboard that uses nearly the full screen (besides one input line), just like the N800 has. Sep 14 13:47:46 Use literkey Sep 14 13:48:11 we could toggle the keyboard with AUX button (be as it now, or fullscreen but transparent) Sep 14 13:48:22 DocScrutinizer: I do like illume keyboard, but could be better Sep 14 13:48:30 the problem, that nobody touched the code since 9 months Sep 14 13:48:40 and it is burried in illume code Sep 14 13:49:19 I like both literki and qwo Sep 14 13:49:38 DocScrutinizer-8: what the heck is literkey? Sep 14 13:49:50 Google only knows one page that doesn't help Sep 14 13:49:54 but I was never able to make qwo trasparent :S Sep 14 13:50:13 but it can be built trasparent :P Sep 14 13:50:56 literki is on opkg.org Sep 14 13:51:01 * spaetz checks Sep 14 13:51:09 and is cool Sep 14 13:51:38 you get also feedback on keypresses as vibro Sep 14 13:52:47 yeah, we should resurrect the illume's keyboard to its own project Sep 14 13:52:55 arggh... missed stefan_schmidt :( Sep 14 13:52:56 not bounded with illume Sep 14 13:53:02 mickeyl, ping :) Sep 14 13:53:05 you can easily change the layout or everything else trhrough configuration files under /etc Sep 14 13:53:26 mrmoku: hu, I'm just undercover. Hidden by an _ :P Sep 14 13:53:33 stefan_schmidt_, hehe Sep 14 13:53:41 stefan_schmidt_, one question about the oe.dev merge Sep 14 13:53:48 * stefan_schmidt_ blames 24h reconnect Sep 14 13:53:54 mrmoku: sure Sep 14 13:53:57 stefan_schmidt_, our recipes as single commits per recipe or one big one? Sep 14 13:54:06 (I mean only the SHR stuff) Sep 14 13:54:16 mrmoku: better one per commit. Sep 14 13:54:25 stefan_schmidt_, yeah... thought so Sep 14 13:54:32 Such big commits are hard to review or spot bugs in later Sep 14 13:54:41 mrmoku: I'm willing to help here Sep 14 13:55:03 stefan_schmidt_, and following the policy I can just commit our recipes... don't have to go through oe-devel Sep 14 13:55:05 mrmoku: Sadly I had no time to come up with a list of recipes yet Sep 14 13:55:25 mrmoku: Sure, for recipes that touch nothing else that is fine Sep 14 13:55:30 stefan_schmidt_, well... I did not come up with much either Sep 14 13:55:36 but I eliminated shr.bbclass :) Sep 14 13:55:45 mrmoku: Sometimes it's better to ask for review if you have things that are not that clear Sep 14 13:56:11 mrmoku: yup, seen it. And I gave moko-autorev.inc a nice death :) Sep 14 13:56:33 yay Sep 14 13:56:33 stefan_schmidt_, one such thing would be where to stuff something like SHR_MIRROR... Sep 14 13:56:34 mrmoku: All new shr recipes should be not a big problem Sep 14 13:56:41 mrmoku: no regression potential Sep 14 13:56:50 as long as it parses ;) Sep 14 13:57:03 mrmoku: well, build would be better. :) Sep 14 13:57:32 mrmoku: Problems will come with recipes that where changed both in .dev and shr. midori and such Sep 14 13:57:43 yup... that will be fun :) Sep 14 13:58:12 But that can be sorted out. First the shr recipes. That should shrink the stuff a lot Sep 14 13:58:24 mrmoku: hmm, what do you use SHR mirror for? Sep 14 13:58:31 clear case for diff-merge Sep 14 13:58:32 mrmoku: just a distro mirror? Sep 14 13:58:42 stefan_schmidt_, like the fso mirror thing... to point to our git Sep 14 13:59:01 DocScrutinizer: depends. Often it makes sense to go through such things by hand. But first the easier ones. :) Sep 14 13:59:09 mrmoku: hmm Sep 14 13:59:25 mrmoku: That depends on the shr distro. Sep 14 13:59:49 mrmoku: How is the distro structured atm? an include on angstrom or an own config? Sep 14 14:00:11 stefan_schmidt_, conf/bitbake.conf:FREESMARTPHONE_GIT = "git://git.freesmartphone.org" Sep 14 14:00:20 that is what I mean Sep 14 14:00:22 stefan_schmidt_: diff-merge is interactive Sep 14 14:00:56 DocScrutinizer-8: ok, need to check then Sep 14 14:01:00 mrmoku: aahhh Sep 14 14:01:10 mrmoku: That should not be a problem then imho Sep 14 14:01:28 stefan_schmidt_, I just add it there then (bitbake.conf) along with all the others Sep 14 14:01:28 mrmoku: That's the only thing shr recipes depend on now? Sep 14 14:01:35 can we switch to dev.oe and use shr/import as overlay, so that we can gradually phase out and remove stuff from shr/import without breaking compiling? Sep 14 14:01:38 mrmoku: yeah Sep 14 14:01:40 mmh... latest upgrade of shr-u and I cannot get gsm service up Sep 14 14:01:41 stefan_schmidt_, not even that... I hardcoded it for now Sep 14 14:01:58 do you know why? Sep 14 14:02:16 mrmoku: ok, but it should not be a problem. Same boat as fso here. Sep 14 14:02:24 ok Sep 14 14:02:50 spaetz, I don't think that makes sense... we still continue to build from shr/import until all stuff hit org.oe.dev Sep 14 14:02:55 spaetz: hmm, no idea if that is worth the effort. Should be possible thought. A lot coordination needed... Sep 14 14:03:05 can I paste few lines of log? Sep 14 14:03:07 then we will create a new branch based on org.oe.dev and build from that Sep 14 14:03:18 and make damn sure to keep close to it ;) Sep 14 14:03:23 stefan_schmidt_: check KDE Kompare e.g. Sep 14 14:03:48 how to keep close to it? rebasing on top of it every day? Sep 14 14:04:04 DocScrutinizer: ok, used meld so far. Sep 14 14:04:13 JaMa, nah.. make sure all stuff we commit to the new branch goes into org.oe.dev as well Sep 14 14:04:16 mrmoku: full ack Sep 14 14:04:20 and then pull from org.oe.dev from time to time Sep 14 14:04:29 (to get all the other nice fancies :) Sep 14 14:04:37 mrmoku: so we'll not know if stuff works out until we've brought everything over and make a big bang switch? Sep 14 14:04:38 mrmoku: Upstream first is a policy that works well for other projects here. Sep 14 14:04:50 Redhat has it for the kernel and OE/stable uses it as well. Sep 14 14:05:06 A bit hard at the beginning, but makes the life easier later. Sep 14 14:05:08 * spaetz is jsut afraid that we'll always play catchup with that many changes happening in shr/import Sep 14 14:05:26 mrmoku: But I leave that for you guys. It's your distro business now. :) Sep 14 14:05:48 spaetz, we just need some discipline to get stuff into org.oe.dev too until we switch Sep 14 14:06:02 mrmoku: ok, true it's not that many people with committ access Sep 14 14:06:06 so it should be possible Sep 14 14:06:17 yep :P Sep 14 14:06:19 spaetz: as mrmoku says plus you can build oe.dev image sin between for testing Sep 14 14:06:33 excuse me... any known issue in latest ophonekitd? Sep 14 14:06:41 spaetz: normally I hate flag days, but for this one it feels a lot easier. Sep 14 14:06:49 I'm not able to get gsm up Sep 14 14:07:16 stefan_schmidt_: visual source safe had a nice diff3-merge for exactly that. Dunno, seem to remember Kdiff3, but can't find it anymore Sep 14 14:07:18 we should just throw the switch immediately and compile using dev.oe as branch. Gives more incentives to fix things quickly and bring them over :) Sep 14 14:07:49 DocScrutinizer: Aren't kdiff3 and meld the same? fork? whatever! :) Sep 14 14:08:08 spaetz: well, I don't think your users would like that. :P Sep 14 14:08:16 stefan_schmidt_: and no, I don't recommend to use visual source safe if you want to keep your mental health ;-) Sep 14 14:08:28 DocScrutinizer: heh Sep 14 14:09:55 stefan_schmidt_, gitk org.openembedded.dev..shr/import is nice Sep 14 14:10:36 mrmoku: ah right, that graphical git thing I always ignore. Even shoragan points it out to me again and again. :) Sep 14 14:10:41 Need to try it. Sep 14 14:11:16 mrmoku: Will let you know before I touch anything so we can coordinate. Sep 14 14:11:37 hi Sep 14 14:11:38 mrmoku: btw, for oe.dev we need checksums. Not sure if shr is up to date in the regard Sep 14 14:11:41 hi dos1 Sep 14 14:11:47 mrmoku: how's testing image? Sep 14 14:12:09 dos1, looks good so far :) Sep 14 14:12:14 GPS works Sep 14 14:12:15 GSM works Sep 14 14:12:21 !!! :D Sep 14 14:12:28 just one thing... Dialer clickin on contacts segfaults Sep 14 14:12:29 mrmoku: we have a fso meeting on thursday and I hope to sneak in some hours for the .dev merge and pull in the others for help. Sep 14 14:12:59 stefan_schmidt_, ok good :) Sep 14 14:13:35 dos1, and the segfault was just once... Sep 14 14:13:37 now it works Sep 14 14:13:48 mrmoku: so we have testing image! :) Sep 14 14:13:48 brb Sep 14 14:14:03 dos1, well a candidate to test still ;) Sep 14 14:14:19 mrmoku: i built also om-gta01 and eten-m800 images, dunno how they behave Sep 14 14:19:15 mrmoku: have you tried subversion again? Sep 14 14:20:21 dos1: huh, testing image? !! Sep 14 14:21:37 rhkfin: candidate for -testing, yep Sep 14 14:21:42 (testing image) I suggest waiting for mickeyl fixing the freeze-on-gprs-disconnect issue Sep 14 14:22:04 DocScrutinizer-8: well, i think here gprs works ok Sep 14 14:22:08 awesome Sep 14 14:22:26 DocScrutinizer-8: i used it about 20 days ago, and it worked flawlessly Sep 14 14:23:01 dos1: yeah, 20 days ago I was happy with it as well Sep 14 14:23:05 stefan_schmidt_: checksums are quite ok in shr/import, I'm building images with checksum check.. Sep 14 14:23:16 waiting to test & enjoy testing image Sep 14 14:23:33 DocScrutinizer-8: hmm... i didn't see any change about gprs in fso these days ;/ Sep 14 14:24:04 * JaMa is waiting for experimental image :) Sep 14 14:24:16 JaMa: cool, good to know somebody does it. Seems we are doing well in preparations for getting things into dev. Sep 14 14:24:54 * JaMa already started building against org.oe.dev :) Sep 14 14:25:19 JaMa: cool Sep 14 14:25:21 but just angstrom instead of shr :) Sep 14 14:25:31 aww :( Sep 14 14:25:40 need to test that patch with KMS and xorg there... Sep 14 14:26:03 mrmoku: proper xorg after -testing is out? Sep 14 14:26:13 sure Sep 14 14:26:13 as it's not sheduled to be merged to shr/import :/ Sep 14 14:26:23 or is it? :) Sep 14 14:26:44 rhkfin, no, sorry not yet... because I don't know what to do about it Sep 14 14:27:00 my autotool foo is very limited :( Sep 14 14:27:28 dos1: do you think that's enough of an evidence there is not issue with gprs, though I see it all the time, and mickeyl admitted there are issues with pppd? Sep 14 14:29:04 dos1, I'm still waiting for Ainulindale to test it on gta01 ;) Sep 14 14:31:18 mrmoku: ok, thanks Sep 14 14:33:51 any bzr freak around here? Sep 14 14:34:18 * mrmoku wants to package apathy... and that is in a bazar repo :( Sep 14 14:34:19 you should try #ubuntu :-) Sep 14 14:34:47 spaetz, nah... they will kick me... I'm a convinced Fedora user ;) Sep 14 14:34:54 apathy? Is that sth like telepathy or empathy? Sep 14 14:35:07 spaetz, elementary sip thingie Sep 14 14:35:09 mrmoku: just that ubuntu's are the biggest bzr pushers Sep 14 14:35:17 http://www.openapathy.org Sep 14 14:35:21 cool, I am missing a sip thingie on the FR Sep 14 14:35:27 yep, me too Sep 14 14:35:57 the url for that is lp:apathy ... but OE doesn't like that ;) Sep 14 14:36:27 ou can find most of the apathy developers in the #apathy IRC channel on irc.freenode.net. Sep 14 14:36:37 ask them directly :) Sep 14 14:39:46 bzr branch lp:no-such-project Sep 14 14:39:47 bzr: ERROR: Not a branch: http://code.launchpad.net/no-such-project/ Sep 14 14:40:03 so lp: is really just a shorthand for http://code.launchpad.net/ Sep 14 14:43:18 advancedcaching bbfiles ready :) Sep 14 14:43:32 there's a project which uses telepathy with elm Sep 14 14:44:00 spaetz, hehe, thanks :) Sep 14 14:44:45 haven't tried if it actually works. but it does make sense Sep 14 14:45:00 btw. mrmoku gstreamer is missing an HTTP/MMS source in shr feed Sep 14 14:46:41 soltys: hey, did you already test new image on gta01? Sep 14 14:51:10 in shr-today, the battery image seems badly resized looking at those segments... Sep 14 14:54:44 dos1: nope I;ve got to charge gta01 battery first ;) Sep 14 15:02:51 spaetz, works :) Sep 14 15:32:37 baaah, shouldn't have messed with illume wrench... -> enlightenment segfault, withdelayed action Sep 14 15:32:44 DocScrutinizer-8: funny thing with gta01 charging :-/ Sep 14 15:33:26 PaulFertser: hm? Sep 14 15:33:33 DocScrutinizer-8: when i plug usb and force 0.4 Imax with "pmu 2b=3a" it charges with only ~45mA Sep 14 15:33:50 ooh Sep 14 15:33:58 DocScrutinizer-8: with pmu 2b=0a it discharges with ~130, with usb unplugged with ~150 Sep 14 15:34:09 echo NO_NEW_FAIR_SLEEPERS > /debug/sched_features Sep 14 15:34:19 someone tested it here? Sep 14 15:34:25 dos1: what to test exactly? Sep 14 15:34:27 yes Sep 14 15:34:43 lindi-: Sep 14 15:34:45 > > > mkdir /debug Sep 14 15:34:46 > > > mount -t debugfs none /debug Sep 14 15:34:48 > > > echo NO_NEW_FAIR_SLEEPERS > /debug/sched_features Sep 14 15:34:49 > > > Sep 14 15:34:50 dos1: no detectable diff Sep 14 15:35:01 DocScrutinizer-8: when 2b=3a and suspended to ram it charges with ~187mA Sep 14 15:35:01 lindi-: and some people says everything is then faster :P Sep 14 15:35:14 mrmoku: cool Sep 14 15:35:20 PaulFertser: that's strange Sep 14 15:35:30 DocScrutinizer-8: yes, very strange Sep 14 15:35:58 dos1: "everything"? Sep 14 15:36:11 dos1: I tested it yes, but no clue whether things actually change Sep 14 15:36:15 dos1: I need a specific test case Sep 14 15:36:38 dos1: btw s/\/debug/\/sys\/kernel\/debug/ Sep 14 15:36:45 lindi-: i can't provide you such case, as i don't know anything about that :P Sep 14 15:36:57 DocScrutinizer-8: one can mount debugfs wherever he likes Sep 14 15:37:09 PaulFertser: I know Sep 14 15:37:16 but as that change went upstream (disabling the fair sleeper thingie) it seems safe to also disable it or wait until we get it Sep 14 15:37:44 dos1: hard to test then Sep 14 15:37:58 lindi-: no diff found here Sep 14 15:38:07 I found a bug in zhone . I plan to look after it (it looks like a simple flag not set properly in a use case). Though is this is known issue : when calling the cancel or release then if using abyss the phone UI stay, then click on the four square button instead of quit, it gives the dialer UI but a DBusException: org.freesmartphone.GSM.Network.Unauthorized on each key pressed (and cannot dial) Sep 14 15:38:29 the workaround is to click on quit and then phone UI Sep 14 15:39:42 DocScrutinizer-8: after i resumed i got WSOD and when i blank display i get 74mA charging current with the same 2b=3a setting Sep 14 15:41:56 DocScrutinizer-8: but it charges great after power off, 450mA :) Sep 14 15:42:18 lol Sep 14 15:44:19 but in theory using NO_NEW_FAIR_SLEEPERS should increase performance on such limited hardware as freerunner, shouldn't it? Sep 14 15:45:20 (performance in switching tasks, i don't suppose other black magic from it ;)) Sep 14 15:45:36 Hm, looks like Qi on gta01 has different ideas about bad blocks, not compatible with the kernel :( Sep 14 15:56:55 dos1: one would think so, although it is hard to quantify. Interactivity is such a subjective measure Sep 14 15:57:18 mrmoku: the SHR testing image is using gsm0710muxd by default?! Sep 14 15:57:34 PaulFertser: it should use fso-abyss Sep 14 15:58:01 dos1: ti_calypso_muxer_type = gsm0710muxd Sep 14 15:58:13 wtf? Sep 14 15:58:18 PaulFertser, on gta01? Sep 14 15:58:23 dos1: how do i prove i downloaded the latest image? Sep 14 15:58:24 dos1: yes Sep 14 15:58:29 mrmoku: yes Sep 14 15:58:37 PaulFertser: ohhh Sep 14 15:58:47 then possible, dunno if i changed it also for gta01 Sep 14 15:58:47 dos1, fix it ;) Sep 14 15:58:56 mrmoku: man Sep 14 15:58:57 checking Sep 14 15:59:06 mrmoku: i downloaded from http://build.shr-project.org/shr-testing/images/om-gta01/ Sep 14 15:59:10 mrmoku: obviously it's not latest Sep 14 15:59:23 :) Sep 14 15:59:49 ti_calypso_muxer = fso-abyss Sep 14 15:59:57 PaulFertser: ok then :D Sep 14 16:00:17 dos1: it's not that ok but well, how am i supposed to test the image if it's not there? ;) Sep 14 16:00:45 PaulFertser: it's supposed to not be there ;) Sep 14 16:00:52 PaulFertser: it's there ;) in tests/mrmoku ;) Sep 14 16:00:58 PaulFertser: http://build.shr-project.org/tests/mrmoku/shr-testing/images/om-gta01/ Sep 14 16:00:58 or sth like that Sep 14 16:01:25 mrmoku: why haven't you warned me? :-/ Sep 14 16:01:27 mrmoku: ;) Sep 14 16:01:31 PaulFertser: sorry Sep 14 16:01:39 PaulFertser: http://build.shr-project.org/tests/mrmoku/testing/images/om-gta01/ Sep 14 16:01:39 dos1: np Sep 14 16:03:00 DocScrutinizer-8: this pcf50606 makes me feel stupid Sep 14 16:04:22 PaulFertser: check your hex values. some of those peripheral chips have rather starnge data format Sep 14 16:04:48 DocScrutinizer-8: well... Sep 14 16:04:50 also make sure no kernel things interfere Sep 14 16:05:01 DocScrutinizer-8: they do not Sep 14 16:05:17 well, no better advice here Sep 14 16:06:37 pcf506xx statemachine is fsckng complex Sep 14 16:09:20 indeed Sep 14 16:12:34 mrmoku: btw, you can refrain from publishing ubifs images, they're bogus anyway. Sep 14 16:13:45 PaulFertser: well, "bitbake shr-image" builds jffs2, tar.gz and ubifs at once, so i don't care when building :D Sep 14 16:13:53 dos1: i know Sep 14 16:13:58 dos1: but you give false hopes Sep 14 16:14:12 ;) Sep 14 16:14:17 mrmoku: so fix ubifs images or disable them ;) Sep 14 16:15:01 hi Sep 14 16:15:05 I even explained how to do that ;) Sep 14 16:15:40 do what? fix or disable ;-P Sep 14 16:16:34 * DocScrutinizer-8 -->doener Sep 14 16:25:06 dos1: btw, I could solve my elementary question of yesterday;) Sep 14 16:30:35 DocScrutinizer-8: it registered even with my DMM! Sep 14 16:33:22 Man, i want to use SHR as my daily phone :-/ Sep 14 16:36:21 PaulFertser: I do use freerunner as my daily use, but not really reliable. Especially if you manage hangup the exact same time as the other party. Everything hangs, I need to take off the battery... Sep 14 16:36:50 khiraly1: take off the battery? I can't really beleive it, sounds too weird and hard to explain. Sep 14 16:37:21 khiraly1: i also use my FR as a daily phone, just not SHR, it's debian with zhone for me. Sep 14 16:37:21 hmm, better: paroli hangs. frameworkd seems running, but in fact no matter how many time I relaunch frameworkd, it does not work Sep 14 16:37:33 (even if I launch shr telephony suite instead of paroli) Sep 14 16:37:46 khiraly1: that just can't be. Every time i properly restart frameworkd on gta02, it works again. Sep 14 16:38:00 for me, it doesnt. Sep 14 16:38:26 It occurs really rarely. Although it depends how many time you make/receive phonecalls Sep 14 16:38:40 I manage to crash this way 4 times the last week Sep 14 16:39:08 khiraly1: it just can't be. The modem is fully restarted, so is frameworkd. At least on gta02. Sep 14 16:39:40 what bugs me, is there is no anomaly in /var/log/frameworkd.log Sep 14 16:40:05 PaulFertser: next time, when it occurs, I will document how I relaunched, and what tried exactly Sep 14 16:40:34 but trust me, it is always the worst time, when it occurs. Like you are ordering the cinema tickets, and everybody calling you ... Sep 14 16:43:55 PaulFertser: if there is a dbus command where I could make sure that the gsm chip responds, I would highly appreciate it Sep 14 16:43:57 to test it next time Sep 14 16:44:28 khiraly1: just look at frameworkd.log Sep 14 16:44:37 there is nothing unusual Sep 14 16:44:51 as far as I remember Sep 14 16:45:16 btw, lindi wrote a watchdog, where it sends a command to the gsm chip, and test if he gets a respond or not Sep 14 16:50:56 whom can I send patch against shr-themes.git? :) Sep 14 16:51:20 JaMa: send it do shr-devel and someone will apply it Sep 14 16:52:22 mrmoku: preliminary result: seems like the latest mickeyl's patch didn't work. I can't say for sure since i tried only once and now i need to recharge the battery and the only way to do it on my gta01 currently is to turn it off fully. Sep 14 16:52:23 yes but i guess that patchwork runs against shr/import .. not against shr-themes.git Sep 14 16:52:40 JaMa: patchwork runs for everything Sep 14 16:53:32 khiraly1: it also checks that the response is sane and indicates that you are really logged in to the network Sep 14 16:54:05 lindi-: gimmie Sep 14 16:54:25 dos1: ok.. as you're the one updating shr-themes, do you want neo theme? if so where it should be? do you want single top directory neo and there illume, etk, elementary, gpe, icons, gtk or split it to already existing directories? Sep 14 16:54:32 khiraly1: the program? iki.fi/lindi/openmoko/gsm-watchdog but it's not really distribution quality yet Sep 14 16:55:32 JaMa: look how niebiee theme is placed in sht-themes.git and do the same ;) Sep 14 16:55:56 JaMa: and of course, i want neo theme in it ;) Sep 14 16:56:20 JaMa: but then make sure authors of that theme can access to shr-themes, so they'll be able to update it Sep 14 16:56:32 lindi-: need to rewrite in python, and integrate into paroli/ophonekitd Sep 14 16:56:46 or even make a small test app with elementary gui Sep 14 16:57:20 mrmoku: other than that i can confirm that testing on gta01 boots, registers to network, shows messages, incoming calls, etc. Can't really try sound because i've only bare board currently. Sep 14 16:57:50 mrmoku: but i'd very much like to see that bug fixed, so shr users on gta01 will be able to properly disable and enable gsm. Sep 14 16:58:31 dos1: Bernd Prünster just send e-mail to community list that he would be glad if i push themes to shr-theme.git.. but I'm not author or artist .. so I would import it and then he should ask for access when he has time to improve it Sep 14 16:59:13 khiraly1: i wrote it in perl for a reason :) Sep 14 16:59:23 for what reason? Sep 14 16:59:31 khiraly1: so that if there's a bug in python's dbus bindings the watchdog is not affected by that Sep 14 16:59:52 khiraly1: and yes we have had dbus issues Sep 14 17:00:15 khiraly1: i also think the watchdog should be a separate process for stability Sep 14 17:00:28 khiraly1: just like normal /usr/bin/watchdog is Sep 14 17:00:58 then we should be able to ask watchdog by dbus Sep 14 17:01:06 hmm Sep 14 17:01:27 khiraly1: well i don't trust dbus ::) Sep 14 17:01:56 anyway would be nice to have a button: "test now" Sep 14 17:01:58 so have a gui Sep 14 17:02:45 khiraly1: gui's are unreliable :) Sep 14 17:04:40 PaulFertser, ok, thanks for testing Sep 14 17:05:03 lindi-: so how could we notice the user? Sep 14 17:05:07 MUHAHA! Sep 14 17:05:14 I finished calculus! Sep 14 17:05:14 the gsm broke, we needed to restart it Sep 14 17:05:22 (well at least the first course out of three) Sep 14 17:05:31 (well at least if I got a good grade...) Sep 14 17:05:31 :) Sep 14 17:05:41 (damn, I'll have to study a lot more calculus) Sep 14 17:06:17 JaMa, so send me your pubkey then :) Sep 14 17:10:08 TAsn: both facts are cool imho Sep 14 17:10:40 btw, have I told you about the mess my knee pads caused? Sep 14 17:10:40 :( Sep 14 17:11:05 TAsn: nope Sep 14 17:11:11 bah, well Sep 14 17:11:17 they have a metal part Sep 14 17:11:31 mrmoku: patchwork is enough for me as long as clicking in patchwork doesn't bother you :) Sep 14 17:11:31 and I ride with my knees hugging the motorcycle Sep 14 17:12:06 JaMa, no problem Sep 14 17:12:10 vibrating metal on paint = havoc. Sep 14 17:12:11 TAsn: your new motorcycle? Sep 14 17:12:23 aye :( Sep 14 17:12:29 well not new, second hand, but yeah. Sep 14 17:12:37 TAsn: what model? Sep 14 17:12:37 had 0 scratches Sep 14 17:12:43 yamah xt125R 2007 Sep 14 17:12:57 good old xt :) Sep 14 17:13:06 :) Sep 14 17:13:10 so much fun Sep 14 17:13:16 TAsn: paint is something that can be fixed easier than knees, so you shouldn't worry too much. Sep 14 17:13:37 PaulFertser, or Sep 14 17:14:15 I should just stick small "anti scratch" stickers on the metal parts of the pads Sep 14 17:14:35 TAsn: so, how fast can it go? Are you comfortable on it? Sep 14 17:14:58 PaulFertser, It's very comfortable Sep 14 17:15:09 can reach 110 Sep 14 17:15:17 I only got to 105 with it. Sep 14 17:15:26 (can probably reach more than 110) Sep 14 17:15:31 all the roads here are limited to 90 anyway :) Sep 14 17:15:36 so I'm not missing *that* much Sep 14 17:15:52 khiraly1: it prints that to syslog Sep 14 17:16:25 khiraly1: you configure syslog to deliver certain messages to a helper that can then broadcast them over dbus to some notification applet if youwant Sep 14 17:18:43 TAsn: i'll probably ride Moto_Guzzi_California soon for some 5-10 km. Sep 14 17:19:50 never saw on in r/l though looks cool :)) Sep 14 17:20:15 TAsn: mad heavy :-/ Sep 14 17:20:53 yeah, saw a pic. Sep 14 17:21:01 today I saw a crazy harley Sep 14 17:21:13 man, the rear tire was wider than the length of my motorcycle :) Sep 14 17:21:22 crazy. Sep 14 17:21:37 Oh, those bikers are crazy indeed. Sep 14 17:22:16 :) Sep 14 17:22:24 well, I'm off. Sep 14 17:22:25 ciao. Sep 14 17:22:34 TAsn: btw, i don't like floorboards and clutch lever that one needs to press with his heel. Feels stupid. Sep 14 17:22:48 TAsn: cy and drive safe. Sep 14 17:23:18 thanks. :) Sep 14 17:23:26 ah bad edje_decc neo.edj generates wrong neo.edc file :/ leaving now.. Sep 14 17:27:01 (19.16.25) lindi-: khiraly1: you configure syslog to deliver certain messages to a helper that can then broadcast them over dbus to some notification applet if youwant <--seems hypercomplicated Sep 14 17:51:09 dos1, here? Sep 14 17:53:29 TAsn: no ;x Sep 14 17:53:33 TAsn: but you can ask ;) Sep 14 17:54:15 ok Sep 14 17:54:19 although I forgot what I wanted to ask Sep 14 17:54:27 hm.. Sep 14 17:54:29 oh yeah Sep 14 17:54:37 do you have many missed/incoming/outgoing calls in your phonelog? Sep 14 17:54:46 (many = >100 total) Sep 14 17:54:58 TAsn: i have about 700 Sep 14 17:55:21 cool Sep 14 17:55:35 mind checking something for me? Sep 14 17:55:37 :) Sep 14 17:56:23 TAsn: no ;x Sep 14 17:56:25 ;) Sep 14 17:56:27 TAsn: kidding, of course Sep 14 17:56:38 I'll take that as a yes. Sep 14 17:57:05 hm.. please check to see if incoming/outgoing/missed show correctly in the "all" page, because here they do not... Sep 14 17:57:19 it's opimd related Sep 14 17:57:26 either the query I'm doing or opimd itself (either way I need you to fix it...) Sep 14 17:57:29 :) Sep 14 17:57:39 ;] Sep 14 17:57:43 hmm Sep 14 17:57:57 i.e Sep 14 17:58:08 look at the top in incoming/missed/outgoing Sep 14 17:58:15 are they the first of their kind in the main list as well? Sep 14 18:00:01 TAsn: incoming and missed look ok Sep 14 18:00:07 TAsn: but i'm not sure about outgoing Sep 14 18:00:22 I'm not talking about the list Sep 14 18:00:29 I'm talking about the presence of those in the general list Sep 14 18:00:38 (oh, it seems like we are talking on the same thing, cool.) Sep 14 18:00:45 TAsn: i'm talking about presence of those in the general list too Sep 14 18:00:46 ;) Sep 14 18:03:58 I don't quite remember what misbehavior I noticed Sep 14 18:04:01 so something is wrong with outgoing calls Sep 14 18:04:02 but this could be it :) Sep 14 18:04:05 anhow, I'm off again. Sep 14 18:04:05 ciao. Sep 14 18:04:12 dos1, probably :) Sep 14 18:10:58 * [Rui] waves Sep 14 18:11:46 <[Rui]> it would probably be worth it to set NO_NEW_FAIR_SLEEPERS by default on the moko kernel Sep 14 18:12:00 <[Rui]> so far I feel an improvement Sep 14 18:12:03 [Rui]: i couldn't see any difference Sep 14 18:12:15 what does it do? Sep 14 18:12:26 <[Rui]> dos1: while the task list always lagged something, and sometimes a lot, now it's quite snappy Sep 14 18:12:36 it's fast enough without it, and it's still fast the same with it :P Sep 14 18:12:56 <[Rui]> selecting a running app is also blasingly fast, compared with previously Sep 14 18:12:56 i prefer NO_NEW_FAIR_SLIPPERS=y Sep 14 18:13:34 <[Rui]> dos1: what I don't know is if there any battery performance hit due to it Sep 14 18:13:43 <[Rui]> but in terms of responsiveness, it's a clear gain Sep 14 18:13:48 <[Rui]> at least I do notice it Sep 14 18:14:00 well, i didn't notice it at all;x Sep 14 18:14:36 <[Rui]> even the artifacts on screen rotation disappear faster Sep 14 18:15:21 placebo? ;) Sep 14 18:16:01 <[Rui]> dos1: I'll try doing some "time" Sep 14 18:16:20 <[Rui]> but to be fair I'll need to reboot it a few times Sep 14 18:16:56 <[Rui]> and ways to end the applications in time... Sep 14 18:18:20 <[Rui]> dos1: I don't think it's placebo effect, the task list lag was quite noticeable here. Sep 14 18:19:04 you guys always have some strange problems i don't have ;) Sep 14 18:19:47 <[Rui]> dos1: or that you've gotten so used to it you don't notice it Sep 14 18:20:18 [Rui]: but after NO_NEW_BLAHBLAH i didn't notice anything faster ;) Sep 14 18:21:37 <[Rui]> how do I remove this flag? Sep 14 18:21:44 <[Rui]> just echo NEW_BLABLAH ? Sep 14 18:21:52 <[Rui]> instead of NO_NEW_BlABLAH? Sep 14 18:22:25 is anyone using zhone ? and do not have issue with the dtmf view ... or what are you using (I stick to debian though I bet it is not that overkill to port a few of the shr application packaged ipk to it (at least done harshly) Sep 14 18:23:24 ciao Sep 14 18:24:32 because zhone needs a bit of work (preferences get you stuck on an empty screen if anything goes wrong with regards to preferences or usage) Sep 14 18:26:16 mrmoku: Sep 14 18:26:19 ERROR: Multiple .bb files are due to be built which each provide bluez-libs (/home/shr/shr-testing/openembedded/recipes/bluez/bluez-libs_3.33.bb /home/shr/shr-testing/openembedded/recipes/bluez/bluez4_4.39.bb). Sep 14 18:26:20 This usually means one provides something the other doesn't and should. Sep 14 18:27:19 and? This is long since like this... fortunately it selects the correct one :P Sep 14 18:28:14 dos1, btw... I built the feed Sep 14 18:28:34 mrmoku: are you sure? last time i checked it it was 3.33 Sep 14 18:28:36 <[Rui]> dos1: it may be a very small difference, but at least for me it seems to be the difference from a slow to a bit slow Sep 14 18:28:48 mrmoku: aaah, that's why it's building only metapackages now :D Sep 14 18:30:22 PaulFertser: race between local eisconnect and far end disconnect -> ogsmd stateengine locks up? Sep 14 18:30:52 DocScrutinizer-8: probably, but nevertheless frameworkd restart always cures that. Sep 14 18:30:56 PaulFertser: we had a freeze on far end disconnect a long time ago Sep 14 18:31:22 yep that's true. If you manage to kill it fist instance Sep 14 18:32:21 ogsmd / abyss seems to love to freeze frameworkd recently. See gprs disconnect Sep 14 18:33:11 mickeyl: ping Sep 14 18:35:40 dos1: can you build pisi on buildhost? Sep 14 18:35:42 mickeyl: could there be a deadlock in ogsmd (or dunno what) if a local call disconnect intercepts a far end disconnect - or vice versa? khiraly1 claims his framework occasionally locks up on local and FE ending call same moment Sep 14 18:35:59 Heinervdm: "bitbake pisi" started Sep 14 18:36:04 :P Sep 14 18:36:09 :) Sep 14 18:36:15 DocScrutinizer: it occured to me an hour ago Sep 14 18:36:29 just, I restarted the phone, because I was in urge Sep 14 18:37:07 khiraly1: hmm, so it seems it's no big diff then if it's 60min or 60h ago Sep 14 18:37:23 no, i dont think so Sep 14 18:37:40 I was in call with my mum, and I hanged up. (I think my mum hanged up at the same time) Sep 14 18:38:24 well, don't you own a second phone, so you could reliably reproduce? Sep 14 18:39:13 <[Rui]> DocScrutinizer: that would appear to be something I've run into, as well, but as I haven't been able to deterministically hit it, I haven't recorded any buk Sep 14 18:39:13 <[Rui]> bug Sep 14 18:39:14 DocScrutinizer: I tried, you really need chance to trigger it. Even if you hit the same time, maybe it does not register the same time at the freerunner side Sep 14 18:39:15 <[Rui]> and has invariable happened to be far from any computer I could ssh into it. Sep 14 18:39:33 btw: how would I manage to tweak the system so no logs ever get deleted on boot? Sep 14 18:39:45 DocScrutinizer: although Im wondering if this bug only happens on me, because nobody else whined about it yet Sep 14 18:40:43 <[Rui]> DocScrutinizer: that's the problem with these things... calls do cost! Sep 14 18:41:00 khiraly1: I seen freeze on outbound call when gprs was enabled. I think gprs disconnects on calls longer than maybe 3min, and the freeze was usual gprs-discon freeze Sep 14 18:41:20 which, according to dos1, never happens :-/ Sep 14 18:41:53 yeah, it is not cheap at all. It costs 0.2EUR/try Sep 14 18:43:26 <[Rui]> khiraly1: I'd be willing to spend some money if it was easier to make happen Sep 14 18:43:59 DocScrutinizer: but when it happens, what command should I issue to further debug? Sep 14 18:44:01 <[Rui]> one of the times I think it happened was when I had received a call from my wife, who would pay more than I to her, so I release the call. Sep 14 18:44:03 yes, compared to this el'cheapo hardware that's no reasonable relation :-P Sep 14 18:44:18 <[Rui]> I fire up dialer, call her back, but in the mean while she calls me back Sep 14 18:44:21 <[Rui]> bang! Sep 14 18:44:43 <[Rui]> sometimes it locks up indefinitly, sometimes it seems to get ok after a short while, sometimes it doesn't happen Sep 14 18:44:58 yeah, pretty random Sep 14 18:45:25 khiraly1: first of all: save all logs! Sep 14 18:45:48 which means? Sep 14 18:45:49 khiraly1: ps eax >>mylogfile.txt Sep 14 18:45:53 DocScrutinizer-8, by logging in /var/log ? maybe it only works there because I am running on sd Sep 14 18:45:56 /var/log/frameworkd.log? Sep 14 18:46:09 yes Sep 14 18:46:26 plus the other one same folder Sep 14 18:47:04 plus a dmesg >>mylogfile.txt Sep 14 18:47:29 there should be dmesg in /var/log ? Sep 14 18:47:33 and/or (I never know for sure) logread >>mylogfile.txt Sep 14 18:47:58 I mean it is already there (with previous boot logs too) Sep 14 18:48:19 prahal: not here afaik Sep 14 18:48:32 khiraly1: i think syslog is the right place to log about such failures Sep 14 18:48:55 lindi-: huh? Sep 14 18:49:06 oh sorry using debian (rsyslogd which does sysklogd and syslogd) Sep 14 18:51:08 dos1: wifi works too :D Sep 14 18:51:31 mrmobil: :)))) Sep 14 18:51:40 mrmobil: annouce this fscking testing and let's break unstable!!! Sep 14 18:51:47 dos1: you can sync ;) Sep 14 18:51:50 :D Sep 14 18:51:58 Noo Sep 14 18:52:19 Heinervdm: pisi built, syncing now Sep 14 18:52:22 dos1: prepare for spanking if testing has gprs bug Sep 14 18:52:52 DocScrutinizer-8: sure it will have it ;) Sep 14 18:53:14 DocScrutinizer-8: it was re gsm-watchdog failures Sep 14 18:53:30 but the fix will hit testing as soon as it exists Sep 14 18:53:32 lindi-: aah Sep 14 18:54:50 mrmobil: http://patchwork.dev.bearstech.com/project/SHR/list/ :P Sep 14 18:55:26 * mrmobil back upstairs Sep 14 18:55:32 ;D Sep 14 19:00:41 DocScrutinizer-8: have you seen my GSM registration report? Sep 14 19:00:52 yup Sep 14 19:01:21 makes me think you're not that far away from BTS, and your bat not that depleted Sep 14 19:01:40 dos1, what is with patchwork? ;) Sep 14 19:01:41 and you got a good DMM ;-D Sep 14 19:04:38 mrmoku: hehe ;) Sep 14 19:04:41 mrmoku: i'm building feed now Sep 14 19:04:51 DocScrutinizer-8: i'm not far, true. And my DMM is set to 10A shunt currently, so the drop is not that high it seems. Also funny thing is that my GSM ant is just lying under the board (touching the pad of course), obviously not the right way to ensure a good contact :) Sep 14 19:07:48 Nice evening today: several of my students joined a jabber conference and learning together how to use git in general and github in particular. I so much hope at least some of them will get some joy from my training. Sep 14 19:09:47 * PaulFertser wonders where did larsc go ... Sep 14 19:22:30 silly gprs allows irc connections but blocks http Sep 14 19:29:34 mrmoku: btw, i think you can feel free to remove wpa_supplicant from default gta01 image Sep 14 19:43:11 **sigh** reconfigured my AP to use WPA with what i thought was internal radius. Maemo refused to detect the AP were WPA and didn't connect. Then I cranked up my laptop wifi, tried to connect ->panic. I don't expect THAAT to work on FR now :-S Sep 14 19:43:25 PaulFertser, not needed? Sep 14 19:43:49 mrmoku: without integrated wifi i'd say yes. Sep 14 19:43:51 DocMobilizer, try anyway ;) Sep 14 19:43:58 PaulFertser, ok :) Sep 14 19:44:06 * mrmoku did not know that :P Sep 14 19:44:17 fsckng ZyXEL! Sep 14 19:45:01 * mrmoku remembers the time where he posessed an analog ZyXEL modem and was proud of it :P Sep 14 19:47:58 baaaahhh!! killed FR by pressing powerbutton during suspending/resuming (dunno), before even my desktop was completely up again Sep 14 19:48:45 actually 8sec powerdown worked Sep 14 19:48:56 so it's not been a kernel panic Sep 14 19:49:01 seems X borked Sep 14 19:51:58 not my day :-/ Sep 14 19:52:17 mrmoku: i also know how to scroll! :D Sep 14 19:52:17 dos1, what are you building? Sep 14 19:52:24 feed Sep 14 19:52:46 got apathy to build :) Sep 14 19:52:47 mrmoku: look at screen Sep 14 19:52:49 ;) Sep 14 19:52:55 mrmoku: that was building :D Sep 14 19:53:00 apathy? Sep 14 19:53:09 mrmoku: screen is yours, do what you want Sep 14 19:53:10 dos1, ok Sep 14 19:53:35 tmzt, http://www.openapathy.org/ Sep 14 19:54:18 mrmoku: and is it working? Sep 14 19:54:28 good, that's what I've been looking for Sep 14 19:54:38 I guess that means fso api is wrapped as telepathy? Sep 14 19:55:10 khiraly1, did not try it yet... it builds at least :P Sep 14 19:55:19 tmzt, no Sep 14 19:55:52 and... looking at that page SIP VoiP might not even work yet :( Sep 14 19:56:12 gsm? Sep 14 19:56:25 tmzt, google talk ;) Sep 14 19:56:29 and messaging Sep 14 19:56:50 oh, well I was looking for cell calls as well Sep 14 19:56:55 SIP, SMS and GSM to come Sep 14 19:56:55 in one ui Sep 14 19:56:59 right, ok Sep 14 19:57:18 I've been looking at maemo5 though Sep 14 19:57:40 we can write a telepathy backend for handling gsm *hides* Sep 14 20:01:39 dos1, did you have time to figure out the mdbus call do disable auto-sleep? Sep 14 20:01:53 this *REALLY* goes on my nerves! Sep 14 20:02:08 tomboy65: look at shr_display.py in shr-settings Sep 14 20:02:18 mmmkay Sep 14 20:02:36 mdbus -s org.shr.ophonekitd.Usage /org/shr/ophonekitd/Usage org.shr.ophonekitd.Usage.RequestResource Display Sep 14 20:02:41 tomboy65: try it (from memory) Sep 14 20:02:46 thx Sep 14 20:02:55 i know it's using non-standard way, but fiddling with SetResourcePolicy is even worse :P Sep 14 20:03:03 tomboy65: alternatively you can start some app with fsoraw Sep 14 20:03:35 <[Rui]> PaulFertser: did you notice anything with that flag set on Linux's CFS? Sep 14 20:04:12 can mdbus or anything else get a full introspection report from dbus services? Sep 14 20:04:26 I'm trying to get one for the dbus gsm service on the Pre Sep 14 20:04:52 tmzt: yup, it should Sep 14 20:05:04 just start "mdbus" or "mdbus -s" Sep 14 20:05:09 and then mdbus -s busname Sep 14 20:05:18 (or without -s) Sep 14 20:05:24 [Rui]: well, my desktop at work seemed to be a bit more responsive with distributed computing client running in the background and firefox sometimes lagged quite a bit, with that flag it seemed to be not that long... Who knows if it's placebo or not but Ingo disabled that flag in his tree already, i guess it says much. Sep 14 20:05:24 ok Sep 14 20:05:25 and then mdbus [-s] busname path Sep 14 20:05:29 hi, any1 knows how are build modules for resulting SHR image? I see they are just about 6MB in size Sep 14 20:05:36 is there an ipkg somewhere, static if possible? Sep 14 20:06:19 <[Rui]> PaulFertser: yeah, it says we can't rely on it :( Sep 14 20:06:54 <[Rui]> PaulFertser: I didn't test on my laptop as it seems fine, but on the freerunner some things felt snappier. Sep 14 20:07:07 [Rui]: no, it says that he changed the default to be with this flag disabled. Sep 14 20:07:34 [Rui]: means that he's convinced that current default behaviour is suboptimal (read broken). Sep 14 20:08:45 <[Rui]> PaulFertser: ah, so we'll benefit from it, great! :) Sep 14 20:08:45 <[Rui]> so it seems that fair sleepers weren't so fair after all ;) Sep 14 20:09:00 <[Rui]> PaulFertser: well, for the time being, could shr-u kernel be configured that way? :) Sep 14 20:10:17 [Rui]: why not just do it on boot with initscript? Sep 14 20:11:16 <[Rui]> PaulFertser: what, you mean a few more things slowing down the boot time? Sep 14 20:11:28 [Rui]: yes ;) Sep 14 20:11:32 <[Rui]> versus a small change in Linux configuration file? Sep 14 20:11:42 <[Rui]> which get's to be built anyway? Sep 14 20:12:01 shit Sep 14 20:12:07 Ainulindale, could you get me bzr on the buildhost? Sep 14 20:17:13 there was NO noticable diff with NO_MORE_FAIR_SLEEPER on FR though! Sep 14 20:17:38 DocScrutinizer, speedwise or powerconsumptionwise? Sep 14 20:18:09 PaulFertser: [Rui] I don't think testing scheduler beahviour on a multicore x86 is telling us much about results on ARM Sep 14 20:18:23 mrmoku: neither Sep 14 20:18:57 <[Rui]> DocScrutinizer: perhaps not, but that's not what I did, I tested *only* on the freerunner Sep 14 20:18:59 DocScrutinizer: there's a test utility developed to measure latency on any machine. Sep 14 20:20:04 PaulFertser: I give a shit about tests, as long as nobody can point to a site where expected operation basics of that tweak are explained Sep 14 20:20:45 <[Rui]> DocScrutinizer: intuitively I'd expect that it places some sleeps in certain tasks in order to be fairer Sep 14 20:20:57 <[Rui]> DocScrutinizer: and that loosing them may benefit some usecases Sep 14 20:21:01 PaulFertser: otherwise I could suggest to grind down a dried pea and place that on the CPU. And claim it will take less power then Sep 14 20:21:16 <[Rui]> DocScrutinizer: as Paul said, that setting will be the default real soon now Sep 14 20:21:29 <[Rui]> DocScrutinizer: so it seems there's some merit it in Sep 14 20:21:40 <[Rui]> s/it in/in it/ Sep 14 20:21:40 [Rui] meant: DocScrutinizer: so it seems there's some merit in it Sep 14 20:21:45 [Rui]: not on FR Sep 14 20:21:49 DocScrutinizer: i didn't claim anything in fact Sep 14 20:22:15 <[Rui]> bzzbot: nice bot Sep 14 20:22:32 actually as I didn't find ANY diff, I don't care at all Sep 14 20:23:24 but it's silly to test a mod you don't even understand which parameters it will tweak so you could look at the right corners Sep 14 20:24:06 <[Rui]> DocScrutinizer: it's a mod that upstream will set as being there by default Sep 14 20:24:29 <[Rui]> DocScrutinizer: I just suggest we do that in the meantime :) Sep 14 20:24:41 maybe this parameter has effect only on new started processes? maybe it is to be observed on heavy load only... who knows where to look at and how to tell random noise in test results from what we expected to see Sep 14 20:25:02 <[Rui]> DocScrutinizer: I just mount, echo, and notice Sep 14 20:25:13 ...nothing Sep 14 20:25:23 I noticed NOTHING Sep 14 20:25:24 Heinervdm: oh, i built pisi for testing ;x Sep 14 20:25:39 [Rui]: i also noticed nothing :P Sep 14 20:25:40 dos1|neo: :D Sep 14 20:26:10 Heinervdm: building for unstable now Sep 14 20:26:16 <[Rui]> dos1|neo: what can I say... Sep 14 20:26:44 <[Rui]> I just noticed another thing Sep 14 20:26:59 <[Rui]> sometimes on rotation, the illume launcher had a black strip in the bottom Sep 14 20:27:09 <[Rui]> I haven't been able to reproduce that ever since Sep 14 20:30:29 tbh I don't see how any scheduler policy can get us a lot of boost on a device that definitely has other bottlenecks than just taskswitching Sep 14 20:31:37 well, if our max-timeslice were something like 1ms then I'd suggest to make that 20 or even 50, to get down with load of context switching Sep 14 20:32:17 <[Rui]> DocScrutinizer: not a lot of boost, just a little bit more responsiveness which improves things a bit Sep 14 20:32:25 but I don't think it's much of a difference if a sleeping process is rescheduled with a somewhat shorter or longer latency Sep 14 20:33:16 [Rui]: wishfull thinking Sep 14 20:34:02 [Rui]: what do you do on xmas afternoon? :-P Sep 14 20:34:20 mrmoku: libtelepathy breaks QA checking Sep 14 20:34:38 mrmoku: so nothing can be built after it Sep 14 20:34:39 <[Rui]> DocScrutinizer: well... two years ago I bought a domain and made a website: imposto.ms ;) Sep 14 20:35:45 [Rui]: so maybe this year learn about CFS and come up with suggestions based on real reasons ;-) Sep 14 20:36:36 <[Rui]> DocScrutinizer: puah! :) Sep 14 20:36:41 [Rui]: the chances you will be happy on xmas night are higher then, than they were by hoping in random tweaks Sep 14 20:37:58 smells once more today like windoof spirit - - gimme that new cute registry optimizer, gimme gimme Sep 14 20:38:47 ooops, had side effects. Nah never mind, was a cute hack anyway Sep 14 20:38:51 <[Rui]> DocScrutinizer: http://lkml.org/lkml/2009/9/14/229 Sep 14 20:39:15 <[Rui]> 1.68 improvement for apache ab Sep 14 20:39:37 on **which machine**??? Sep 14 20:40:06 multicore? cachesize? Sep 14 20:40:13 swap size? Sep 14 20:40:29 number of concurent processes? Sep 14 20:40:57 and W*H*Y??? Sep 14 20:41:19 <[Rui]> "This test profile measures how many requests per second a given system can sustain when carrying out 500,000 requests with 100 requests being carried out concurrently." Sep 14 20:41:32 <[Rui]> at least one of your questions has an answer, the rest I don't know Sep 14 20:42:23 yep, number of concurrent processes is roughly 100 + X Sep 14 20:42:49 all with heavy load and never sleeping Sep 14 20:42:57 dos1|neo, ouch Sep 14 20:43:15 dos1|neo, and apathy dos not build due to missing bzr :( Sep 14 20:43:56 talking about sleeping... Sep 14 20:43:59 * mrmoku off to bed Sep 14 20:44:01 night all Sep 14 20:44:26 [Rui]: it's that simple: I tested it, found no difference, ergo won't care about it until somebody comes up with a good reason why I should Sep 14 20:45:29 or until I have to learn it borks something. Then you really get me started Sep 14 20:45:58 <[Rui]> DocScrutinizer: well, no matter, it'll pretty soon be the default unless some major problem comes up Sep 14 20:46:32 fine, so why bother about it at all? Sep 14 20:48:54 <[Rui]> DocScrutinizer: I don't see why not set it up, but I have no problem setting it for myself and hammered on by rc.local or something like it. Sep 14 20:48:54 <[Rui]> brb Sep 14 20:53:53 heh, this time for something nice ;-) WPA2 works with mokonnect Sep 14 20:55:18 so obviously it's something in the combination zyxel-p660 WEP FR mokonnect that causes the freeze/panic Sep 14 21:04:23 alas it breaks on suspend. But can be restarted by simply stopping mokonnect, powercycling WiFi in settings, and start over Sep 14 21:16:39 DocScrutinizer-8: hm, wpa_supplicant itself survives suspend properly, it reassociates with the AP. Sep 14 21:16:57 DocScrutinizer-8: most probably you just need to get DHCP lease again after resume. Sep 14 21:17:37 no idea. anyway route was reset and interface had no associated IP anymore Sep 14 21:18:49 DocScrutinizer-8: that's expected Sep 14 22:34:40 fsodeviced and fsonetworkd built Sep 14 22:34:46 now fsotimed (needed libgee bumping) Sep 14 22:38:54 ok, libgee fails ;p Sep 14 22:39:00 so no fsotimed yet Sep 14 22:43:14 PaulFertser: my modem seemingly rebooting out of nothing in WPA mode is not expected :-/ Sep 14 22:44:55 and even ethernet not recovering from that until I do a rcnetwork restart on my laptop is less expected yet Sep 14 22:47:00 freesmartphone.org: 03mickey 07vala-dbus-binding-tool * rf7dcc12e2e61 10/src/vala-dbus-binding-tool.vala: catch up with new 'async' keyword for methods. replaces 'yields' Sep 14 22:47:55 PaulFertser: (nokia wrt. shitty vala) did they really said it in this particular charming way? Sep 14 22:50:42 freesmartphone.org: 03mickey 07vala-dbus-binding-tool * r1c1b485b2b6b 10/src/vala-dbus-binding-tool.vala: fix position of async qualifier Sep 14 22:54:08 mickeyl: i built fsodeviced and fsonetworkd on shr/import Sep 14 22:54:32 mickeyl: fsotimed needs more work due to libgee 0.3.0 dependency Sep 14 22:55:00 dos1|neo: ok, sounds good. just use libgee git Sep 14 22:55:11 should have a recipe somewhere Sep 14 22:55:36 mickeyl: i tried cherry picking from oe.dev Sep 14 22:55:54 mickeyl: but it failed on compiling, dunno why, i'll check tomorrow Sep 14 22:56:36 ok, probably you need to use a more current vala as well Sep 14 22:56:55 so i'll leave it to mrmoku ;) Sep 14 22:56:59 we're in deep development right now, we can probably snap a vala version pretty soon once the async stuff is stable Sep 14 22:57:52 i already have 4 patches waiting for him about that cornucopia stuff ;) Sep 14 22:58:26 :) Sep 14 22:59:16 ok, when this build finishes i'll sync and go to sleep Sep 14 22:59:34 fsotimed will have to wait ;) Sep 14 23:02:01 if my msg about gps in settings made it to irc: cancel that Sep 14 23:02:23 was just framework not responding Sep 14 23:03:16 mickeyl: any plans about at least one from fsogpsd, fsopreferencesd, fsoeventsd or fsopimd? Sep 14 23:03:39 for sure, but Sep 14 23:03:45 mickeyl: fsopreferencesd->dconf, yup? but what with rest? Sep 14 23:03:59 fsopreferences will change rapidly (due to dconf) Sep 14 23:04:10 gypsy or has somethign replaced that? Sep 14 23:04:12 fsopimd we can think about when the interface got some more looks Sep 14 23:04:27 fsogpsd i hope i can pester alphaone to start writing soon Sep 14 23:04:36 fsoeventsd will have to be rethought completely Sep 14 23:04:37 as for pim, maybe tinymail and it's version of camel? Sep 14 23:04:46 tmzt: no, that'd be a step backwards Sep 14 23:05:01 we started fsopimd since there was nothing with a nice dbus API available Sep 14 23:05:03 err Sep 14 23:05:05 opimd even Sep 14 23:05:15 mickeyl: (opimd interface) ok, that's mainly my job ;) Sep 14 23:05:18 there is a camel with dbus used on maemo, that's the one I mean Sep 14 23:05:32 yeah Sep 14 23:05:47 the experts have scrutinized all that years ago Sep 14 23:05:58 and found it unsuitable, that's why GSoC 2008 had the opimd project Sep 14 23:06:11 mickeyl: and any news about (fs)olocationd? Sep 14 23:07:07 dos1|neo: I'm afraid not. it's pending more time/motivation from the other fso folks Sep 14 23:07:29 i definitely want that Sep 14 23:07:42 and i want onen to get on board for that, too Sep 14 23:08:00 yup, it was great on FSOSHRUDCON slides Sep 14 23:08:43 i was watching olocationd presentation, and i was impressed :) Sep 14 23:08:51 freesmartphone.org: 03mickey 07cornucopia * r9f97aa0006e9 10/libfsoframework/fsoframework/ (fsoframework-2.0.vapi subsystem.vala): libfsoframework: interfaces: add well-known paths for org.freesmartphone.GSM Sep 14 23:08:53 freesmartphone.org: 03mickey 07cornucopia * rb64b4de07cb6 10/fsogsmd/ (6 files in 3 dirs): Sep 14 23:08:53 freesmartphone.org: fsogsmd: first async dbus method implemented! Sep 14 23:08:54 freesmartphone.org: disable libfso-glib for now as it doesn't compile Sep 14 23:10:05 * mickeyl hints at the milestone in b64b4de07cb6 Sep 14 23:10:25 this was the hard bits, the rest is looong and booring work Sep 14 23:10:41 luckily i'm on vacation without computer soon ;) Sep 14 23:10:55 * mickeyl hopes someone else does the boring work Sep 14 23:10:56 mickeyl: ok, so keep doing great work, and maybe find some way to motivate other folks :) Sep 14 23:11:22 dos1|neo: (motivation) trying my best. meeting the other coreteam guys on thursday Sep 14 23:11:29 will kick them L) Sep 14 23:11:32 s/L/:/ Sep 14 23:11:32 mickeyl meant: will kick them :) Sep 14 23:11:36 mickeyl: hehehe, i hope the same about few parts of my opimd todo ;D Sep 14 23:15:54 hi Sep 14 23:16:08 are there some shr developers here? Sep 14 23:16:25 sre_: yup, but going to sleep Sep 14 23:16:53 mh Sep 14 23:17:00 sre_: most of SHR devs are european and there is ~1 AM now here in Poland ;) Sep 14 23:17:13 it's 1:16 here, too ;) Sep 14 23:17:20 elektranox@freerunner ~ % DISPLAY=:0.0 shr-dialer Sep 14 23:17:20 ** (process:5549): DEBUG: Trying to get the system bus Sep 14 23:17:20 ** (process:5549): DEBUG: Adding signals. Sep 14 23:17:20 ** WARNING **: No such file or directory Sep 14 23:17:20 aborting... Sep 14 23:17:38 perhaps a quick idea what may be wrong? Sep 14 23:17:45 (I'm just packaging shr apps for debian) Sep 14 23:17:46 sre_: hmm... try to strace it Sep 14 23:18:05 sre_: do you have libframeworkd-phonegui-efl? Sep 14 23:18:10 yep Sep 14 23:18:15 ophonekitd is running Sep 14 23:18:22 sre: cooool, someone who does debian stuff... Sep 14 23:18:23 hmm Sep 14 23:18:25 and if I call a window pops open :) Sep 14 23:18:30 try to strace Sep 14 23:18:39 sre: some packages are _so_ old... Sep 14 23:18:58 Wonka: yes, we're short on manpower :( Sep 14 23:19:00 open("/etc/phone-utils.conf", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) Sep 14 23:19:11 and sooo annoying Sep 14 23:19:19 sre: i'm short on time too... Sep 14 23:19:30 sre_: you need libphone-utils Sep 14 23:19:50 I have the library Sep 14 23:19:56 do I need the config file? Sep 14 23:19:58 hmm Sep 14 23:20:06 sre_: dunno, check Sep 14 23:20:12 maybe it needs it Sep 14 23:21:30 probably it needs it. those are 5 lines though Sep 14 23:21:42 cool :D Sep 14 23:21:46 it's working now Sep 14 23:22:23 sre_: nice :) Sep 14 23:23:41 ok, going to sleep Sep 14 23:24:17 night Sep 14 23:24:23 i have to wake up in ~5,5 hours ;x Sep 14 23:24:33 good night! Sep 14 23:33:00 how do you distribute /etc/phone-utils.conf in SHR? Sep 14 23:33:26 it contains "country_code = 49" for germany, did you make this default or is there some sort of selection? Sep 14 23:36:01 sre_, wait some time. this will be the default country code in the future Sep 14 23:37:23 sre_, opkg search /etc/phone-utils.conf Sep 14 23:37:23 libphone-utils0 - 0.0.2+gitr23+0a3a19e93237bf525a1bce6e56456dd53936e302-r2 - /etc/phone-utils.conf Sep 14 23:38:04 .oO( huh? i thought those ambitions were over since about 1947... ) Sep 14 23:38:47 :> Sep 14 23:41:18 git tells me you should spank Sep 14 23:43:14 Tom Hacohen Sep 14 23:48:48 sre_: it's the plan shr-settings will include some setup for phoneutils.conf afaik. Or something in first-boot setup Sep 14 23:49:19 mh ok Sep 14 23:49:26 I will do it the debian way then :) Sep 14 23:49:29 using debconf Sep 14 23:50:00 iirc the idea was to get some default from gsm, if available Sep 14 23:50:32 I think one can get a prefix from the sim card at least Sep 14 23:50:53 AC usually should be code for operator network you're on Sep 14 23:57:10 sre_: though I know there a "plans" (like contract) that make your cellphone like a landline in your hometown. So then area code should of course match that towns area code. The idea is if you dial the "basename" of any number (w/o any prefix) then the far end you are connected to shall be the same as when you call same number "fully qualified" Sep 14 23:58:18 :) Sep 14 23:59:19 I'll write that ugly statement to my wall, as an example how complicated I can talk about simple things ;-) Sep 15 00:02:00 btw. would be nice to have a proper error message when the config file is missing Sep 15 00:17:13 DocScrutinizer-8: in the us? Sep 15 00:17:36 I know there are a few places like that but it's not a contract feature that I know of Sep 15 00:17:49 only a few places even have seven digit dialing anymore Sep 15 00:17:53 for landlines Sep 15 00:18:54 tmzt: well, I know it was an option in Germany many years ago. No idea how they handle that today Sep 15 00:19:21 tmzt: length of number doesn't matter for this topic Sep 15 00:19:57 by seven digit dialing I mean prefix+lastfour Sep 15 00:19:59 3+4 Sep 15 00:20:06 which does not include the area code Sep 15 00:20:09 3+3+4 Sep 15 00:20:40 10 digit dialing is dialing with area code, so that setting shouldn't matter in that case Sep 15 00:21:05 this is without 1, which is our country code and the notation for long distance (what's that??:) ) Sep 15 00:22:09 what I'm saying is that in most of the US you have to dial the area code or you get the "your call could not be completed as dialed" message Sep 15 00:23:06 tmzt: all very interesting but irrelevant for this purpose Sep 15 00:23:07 if I understand right this is what libphone-utils is supposed to do Sep 15 00:23:20 what's the purpose? Sep 15 00:25:05 no, it's sole purpose is to expand basename-numbers or national-numbers to a fully qualified number for comparing purposes, as inbound numbers usually are fully qualified Sep 15 00:25:52 s/numbers/numbers from contacts/ Sep 15 00:25:54 DocScrutinizer-8 meant: no, it's sole purpose is to expand basename-numbers from contacts or national-numbers to a fully qualified number for comparing purposes, as inbound numbers usually are fully qualified Sep 15 00:25:59 but according to one of your last messages it's planed to be used for completing numbers, too Sep 15 00:26:44 e.g. I type "1234" and it will be completed to 1234 Sep 15 00:26:49 no, dialing a number from a contact never sees any mangling Sep 15 00:27:12 1234 Sep 15 00:27:18 that's completely wrong for us Sep 15 00:27:45 US Sep 15 00:40:32 tmzt: why is that completely wrong? please explain what's better Sep 15 00:42:47 I can't see any scenario where I would expect to dial four numbers and reach a specific phone, even in the same city Sep 15 00:43:12 why four ? Sep 15 00:43:16 even houses/businesses on the same street can have different exchanges Sep 15 00:43:23 1234 Sep 15 00:43:38 1234 Sep 15 00:43:52 exchange is three numbers Sep 15 00:44:16 telephone numbers here in germany are variable length, 1234 was just an example Sep 15 00:44:34 right Sep 15 00:44:40 that's what I'm saying, it's different here Sep 15 00:44:45 we don't have city code + number Sep 15 00:44:52 we have area code + exchange + number Sep 15 00:44:58 you always dial the exchange + number Sep 15 00:45:09 and in most places area code + exchange + number Sep 15 00:45:14 3 + 3 + 4 Sep 15 00:45:26 so what's the relevance of knowing the "exchange" then? Sep 15 00:45:34 number is always four digits Sep 15 00:45:43 because you always have to dial the exchange Sep 15 00:45:53 and again, we are not talking about any particular length of number Sep 15 00:46:32 right, but your autocomplete does not work here because we would never dial the last four, outside of a PBX Sep 15 00:46:53 huh? Sep 15 00:47:13 you can omit the last 4 numbers?? Sep 15 00:47:16 I don't know how to explain it better Sep 15 00:47:17 no Sep 15 00:47:25 but you can't omit the exchange either Sep 15 00:47:42 nobody even thinks about that Sep 15 00:48:05 basically we don't care about if there's any exchange or whatever Sep 15 00:48:06 thinks about what? Sep 15 00:48:14 ah, I guess I'm getting it Sep 15 00:48:26 or if your base phonenumber is 3 or 1w digits Sep 15 00:48:29 but we always break it down like nnn-nnnn when saying it Sep 15 00:49:12 and phoneutils removes the "-" for you ;-) Sep 15 00:50:14 then it checks if the leading part is a valid area or international code as specified in config Sep 15 00:51:21 if not, it expands the number from your contacts. in the end it becomes a fully qualified number in GSM format (leading +) Sep 15 00:55:12 ~botsnack Sep 15 00:55:12 DocScrutinizer: :) Sep 15 00:55:33 sorry, my modem reboots :-( Sep 15 00:55:56 last chars were "leading +)" Sep 15 00:56:56 there was no text between "leading +)" and your return Sep 15 00:58:22 then we check if that expanded number matches the number of the inbound call -> if yes we display this contact for inbound call Sep 15 00:58:26 on dialing a number from contacts we don't do this processing Sep 15 00:58:47 any particular reason? Sep 15 01:00:30 yes, we don't want to be smarter than the user. If you dial a 5 digit number, you get what you asked for, as with every usual cellphone Sep 15 01:01:36 is there a mean to git clone via git method (it requires authentication with the uri provided on the web view of the repo) Sep 15 01:01:58 I used http for a few ... though it does not support compression and such Sep 15 01:11:40 sre_: for instance I store some contacts with national number in 030-1234567 format as I don't want to dial them via intercontinental call - e.g. my pizza service ;-) Sep 15 01:13:06 also 112 mustn't be prefixed with international and area prefixes Sep 15 01:13:19 ah didn't thought of 112 Sep 15 01:13:45 probably there are further cases Sep 15 01:14:00 would be bad if one couldn't do emergency calls because the phone tries to be intelligent :D Sep 15 01:14:11 yes Sep 15 01:14:23 does modem provide you with emergency numbers? Sep 15 01:15:11 tmzt: afaik it does not Sep 15 01:15:12 I'm not sure about that. At least FSO doen't check for them atm afaik Sep 15 01:16:02 but I think there was a EU policy, that all EU countries should support 112 Sep 15 01:17:16 it's supported. And modem of course can dial emergency numbers. But it doesn't tell you what these numbers are Sep 15 01:18:24 though I think the modem itself knows about emergency numbers and treats them differently Sep 15 01:19:31 maybe you even can query modem about what are the e-# in current network, but FSO doesn't do this query afaik Sep 15 01:20:36 :) Sep 15 01:20:41 from wikipedia Sep 15 01:20:42 112 (one-one-two) is the emergency telephone number in the European Union member states, the Republic of Colombia (South America) and worldwide (on GSM mobile networks). Sep 15 01:20:55 in short: just use 112 :P Sep 15 01:23:53 afaik the sim card has a safed country code Sep 15 01:24:20 playya: yes Sep 15 01:26:04 playya: but you might have contacts stored with a different concept of which countrycode to omit Sep 15 01:26:59 DocScrutinizer, your provider is $$$ enough to not recognize you are in the country you used the international number for ? and bill for international call .... Sep 15 01:27:06 e.g. when I was in taipei I used one of OM's testsims as my own provider had no roaming agreement with taiwan Sep 15 01:28:54 if the compare/expand used that taiwan countrycode then none of my shortform contacts would match an inbound number Sep 15 01:36:57 so no way to get git shr via git scheme ? **** ENDING LOGGING AT Tue Sep 15 02:59:57 2009