**** BEGIN LOGGING AT Wed Oct 15 02:59:57 2008 Oct 15 04:52:39 good morning Ainulindale Oct 15 07:42:13 Ainulindale, please don't use tabs in python files Oct 15 07:43:03 shoragan: I know, I know, that's vim fault ! Oct 15 07:43:19 just use :set expandtab Oct 15 07:47:29 * OdyX` always uses tab for indentation in any programming language Oct 15 07:50:26 OdyX`, but when changeing existing code one should continue to use the existing style Oct 15 07:50:34 true Oct 15 07:50:35 ;) Oct 15 07:51:06 even if the style sucks ;) Oct 15 07:52:10 G'Morning Oct 15 07:58:52 OdyX: I guess you've read this already... ;) http://www.jwz.org/doc/tabs-vs-spaces.html Oct 15 08:23:11 CM_work: anyway ;) "religious point #1" => I don't agree ;O) Oct 15 08:25:05 :) Oct 15 09:03:42 Hi there. Oct 15 09:03:54 AinuFallback: hail Oct 15 09:07:03 rwhitby ? Oct 15 09:49:09 AinuFallback: yes? Oct 15 09:50:49 rwhitby: WEird error on the buildhost, conflict merge on mtn update Oct 15 09:50:57 Do you know where that could come from ? Oct 15 09:51:02 (shr buildhost by the way) Oct 15 09:51:10 git update sorry Oct 15 09:51:14 http://shr.bearstech.com/autobuild-tail.log Oct 15 09:51:36 AinuFallback: that means that someone editing the files on the buildhost and now the git pull has a conflict Oct 15 09:51:41 never edit files on an autobuilder Oct 15 09:51:56 I suspected that, but is there any way to know which file ? :-) Oct 15 09:52:01 git reset --hard HEAD and git clean -f -x will fix it Oct 15 09:52:07 Ok thanks Oct 15 09:52:10 (that will blow away any edits) Oct 15 09:52:19 That was my guess I was just checking :-) Oct 15 09:52:43 I probably need a makefile target for that Oct 15 09:53:03 And do you happen to know how to say in a bb recipe that it should regenerate the package from local sources no matter what ? Oct 15 09:53:12 Or to md5sum the files to see if there's any differences ? Oct 15 09:53:13 nope Oct 15 09:53:40 Looking the bb doc right now about that but can't really find information :-/ Oct 15 10:32:30 freesmartphone.org: 03charlie 07framework * rd8068e77c75f 10/tests/ (README opreferencesd.py sim.py test.py tests.conf): Oct 15 10:32:30 freesmartphone.org: tests: Added 'request' decorator to the tests Oct 15 10:32:30 freesmartphone.org: This can be used to specify that a test can only be run under specific conditions (like if the sim card is present.) Oct 15 10:32:30 freesmartphone.org: The current test condition can be specify in the file 'tests.conf' Oct 15 10:32:33 freesmartphone.org: I also created a README file for the tests and started a sim test suite. Oct 15 10:50:10 mm.. latest testing updates seemed to do some impressive bricking Oct 15 11:27:02 freesmartphone.org: 03mickey 07framework * r0be912f962a4 10/framework/subsystems/ousaged/generic.py: ousaged: readd parameter checking which seems to got lost on the way (tsss...) Oct 15 11:38:58 hi tehre Oct 15 11:49:09 http://onlinedev.blogspot.com/2008/10/weekly-engineering-news-412008.html <- good good good good ! Oct 15 11:50:07 :) Oct 15 11:50:13 *nod* Oct 15 12:10:57 mickeyl: do you happen to know how to be able to regenerate a bitbake package each time with local sources ? Oct 15 12:11:02 Without explicitely cleaning it ? Oct 15 12:11:11 can anyone tell me how i can get the call ringing tone after i use lgsm_voice_out_init ? I am able to make a call successfully but only problem is, i cannot hear the ringing tone. Oct 15 12:11:57 trupheenix: heh, still using libgsm :) Oct 15 12:12:04 trupheenix: you need to adjust the alsa scenario for gsmhandset Oct 15 12:12:16 mickeyl: yea i want to learn that first. Oct 15 12:12:29 *shrug* Oct 15 12:12:43 mickeyl: i'm not fond of py ;) it hides everything Oct 15 12:12:54 hmm Oct 15 12:12:56 trupheenix: why don't you use libframeworkd-glib then ? Oct 15 12:13:01 that goes for dbus as well? Oct 15 12:13:55 AinuFallback: you could (manually) erase the stamps for the bitbake tasks. we should have an option like --force in bitbake though. will talk to the bb guys soon Oct 15 12:13:55 AinuFallback: I went through the freesmartphone code. I find libgsmd lot simpler and cleaner for the time being. Oct 15 12:14:38 trupheenix: I thought quite the contrary Oct 15 12:14:45 And in fact experienced that libgmsd was a pain in the ass :-/ Oct 15 12:14:59 mickeyl: in fact this is related to the buildhost Oct 15 12:15:09 mickeyl: as right now it won't compile latest frameworkd changes Oct 15 12:15:16 Nor ophonekitd/lfg/lf-phonegui changes Oct 15 12:15:43 AinuFallback: unstable should use AUTOREV, so it should work. rwhitby ? Oct 15 12:15:53 AinuFallback: i will start on freesmartphone once i get a hang of libgsmd. Oct 15 12:16:20 mickeyl: by the ay Oct 15 12:16:22 +w Oct 15 12:16:28 SHR will be out before 2009 :-p Oct 15 12:16:41 mickeyl: yes, unstable should be using autorev Oct 15 12:16:53 AinuFallback: hmmm Oct 15 12:17:00 AinuFallback: hopefully :) Oct 15 12:17:10 mickeyl: that's not hope but fact Oct 15 12:17:16 It's working more than enough right now Oct 15 12:17:18 awesome. Oct 15 12:17:20 Only problem is the absence of PIM Oct 15 12:17:22 AinuFallback: ok i will chek out the SHR repos. i would like to see SHR on debian Oct 15 12:17:36 PIM is going to land in FSO soon Oct 15 12:17:41 mickeyl: and I had some issues with suspend/resume which resulted in a discussion with alphaone Oct 15 12:17:49 and some git revert of my dumb patch :-) Oct 15 12:17:54 AinuFallback: yeah, i've seen some commits Oct 15 12:18:06 if it lead to fixing the right way i'm happy Oct 15 12:18:10 It did Oct 15 12:18:13 godo Oct 15 12:18:15 good, even Oct 15 12:18:15 Thanks to alphaone Oct 15 12:18:36 rwhitby: regarding mickeyl's question Oct 15 12:18:42 I tried to build a shr-srcrev.inc Oct 15 12:18:46 But didn't manage to include it properly Oct 15 12:19:13 one thing you must know is that to make it detect newer sources you need a full reparse Oct 15 12:19:16 i.e. touch conf/local.conf Oct 15 12:19:31 That's what rwhitby did in his makefile if I recall correctly Oct 15 12:19:52 yep Oct 15 12:20:22 rwhitby: could you point me to some resources about this include thing ? couldn't find proper documentation Oct 15 12:20:29 Or help me even :-) Oct 15 12:21:38 AinuFallback: dunno, I've never done overlays myself Oct 15 12:22:06 mickeyl: by the I still have problems with the idle notifier Oct 15 12:22:33 THis morning I woke up and my phone was at brightness max Oct 15 12:22:41 (on USB) Oct 15 12:22:50 strange. do you have logs? Oct 15 12:22:56 for the .inc Oct 15 12:23:17 we can add shr-autorev.inc into OE if you want Oct 15 12:23:21 Unfortunately no Oct 15 12:23:33 and then you can 'require conf/distro/include/shr-autorev.inc' if you want Oct 15 12:23:34 Because right now ogpsd is filling my logs Oct 15 12:23:43 as there I still have the unfixed frameworkd version Oct 15 12:23:49 mickeyl: this sounds good to me Oct 15 12:23:59 I also would like you to include the libframeworkd-glib.bb file Oct 15 12:24:05 sure thing Oct 15 12:24:09 just send me that stuff Oct 15 12:24:53 http://shr.bearstech.com/shr-sources/ Oct 15 12:24:57 ah, even better Oct 15 12:25:09 http://shr.bearstech.com/shr-sources/openembedded/conf/shr-autorev.inc <= Oct 15 12:25:33 Feel free to modify it if necessary Oct 15 12:26:15 ok. i will push the first batch and eventually if you don't mind maintaining that in OE i'd like to give you OE access as well Oct 15 12:26:31 No problem with me Oct 15 12:26:44 In any case I won't modify anything in OE, just maintain that Oct 15 12:27:02 cool, lets do it that way then Oct 15 12:27:08 means more exposure for SHR Oct 15 12:27:11 and less work for us all Oct 15 12:27:13 :) Oct 15 12:27:25 Exactly Oct 15 12:28:34 * rwhitby thinks mickeyl should be the community manager, since he does the things that I recommend a community manager should do ... Oct 15 12:29:00 ... bringing the developers into the fold ... Oct 15 12:29:05 mickeyl: did you think about writing documentation on how to get into frameworkd development ? Oct 15 12:29:09 Ah I see rwhitby was just thinking the same Oct 15 12:30:52 rwhitby: yeah, I probably would even enjoy that position, but I can't do it in addition to all other stuff. Oct 15 12:31:23 mickeyl: you are still sleeping by night ? You have enough time then ;) Oct 15 12:31:45 AinuFallback: yes, docs and tests are among the major goals for the remainder of this year Oct 15 12:31:51 [fso related] Oct 15 12:31:53 OdyX: heh Oct 15 12:32:03 mickeyl: I'm truly interested in helping with that too Oct 15 12:32:04 sleeping is overrated, i nkow Oct 15 12:32:11 and doesn't pay Oct 15 12:32:11 AinuFallback: great! Oct 15 12:32:17 As I would like to be helpful with frameworkd Oct 15 12:32:23 But right now I feel kind of ... well Oct 15 12:32:25 Feable :-) Oct 15 12:32:27 * OdyX slept from 24:00 to 04:30 this night Oct 15 12:32:44 My "mistake" from yesterday just proved me frameworkd reached a non return point in "complexity" or completness Oct 15 12:32:47 +e Oct 15 12:33:22 AinuFallback: yep, i see your point. reference docs are fine as they stand but all the subsystems are missing high level docs... "what is that thing doing and _how_" Oct 15 12:33:36 mickeyl: I was more thinking about howtos Oct 15 12:33:43 "how to write a new subsystem" Oct 15 12:33:47 Even if we won't do that in the end Oct 15 12:33:51 It will explain how to manage devices Oct 15 12:33:55 And the logic of configuration Oct 15 12:34:00 aah Oct 15 12:34:01 resource requests Oct 15 12:34:02 yes, that sounds good Oct 15 12:34:12 And then the meta objects Oct 15 12:34:20 The naming conventions Oct 15 12:34:33 And as you just told the architecture Oct 15 12:34:53 I was thinking more something like "why is frameworkd doing that this way" Oct 15 12:34:57 +about Oct 15 12:35:05 Rather than "what is doing that" Oct 15 12:35:21 i see. so more reasoning behind the architectural decisions Oct 15 12:35:30 If you can understand how some system was conceived it's just a matter of time before you can check sources and see what is doing what Oct 15 12:35:49 Yes Oct 15 12:36:03 And I still think oeventsd and sound management has nothing to do in frameworkd :-p Oct 15 12:36:22 hehe Oct 15 12:36:26 we agree to disagree there Oct 15 12:36:32 but you know... it's an offer Oct 15 12:36:35 you don't have to use it Oct 15 12:36:40 I agree with that Oct 15 12:36:48 But it should come as a frameworkd "plugin" Oct 15 12:37:13 Correct me if I'm wrong Oct 15 12:37:18 But in my head Oct 15 12:37:27 I like to compare frameworkd to some micro kernel Oct 15 12:37:28 you mean as in packaging or as in being able to disable? Oct 15 12:37:34 Some basic functionnalities are handled Oct 15 12:37:35 because the latter is already there Oct 15 12:38:03 we can eventually package the subsystem fine granular, if you think it's wise to do it Oct 15 12:38:04 And these functionnalities are kind of a bridge between device logic and user (functional) logic Oct 15 12:38:14 Latter is here I agree Oct 15 12:38:25 In the end it's just a matter of packaging I agree again Oct 15 12:38:38 But it just makes things confusing Oct 15 12:38:57 As I just told, I thought frameworkd was there to offer basic primitives to consider a phone as an abstract object with devices Oct 15 12:39:10 Actually *doing* things on events is a whole different thing altogether Oct 15 12:39:29 It must be done it 's obvious, but must that be done in frameworkd ? I don't know Oct 15 12:39:52 see I'm torn by this as well Oct 15 12:39:54 take a look into oeventsd Oct 15 12:39:59 there are things like actions/triggers Oct 15 12:40:03 but also fso-actions and fso-triggers Oct 15 12:40:09 to split the generic from the specific stuff Oct 15 12:40:14 we might take this a bit further Oct 15 12:40:19 and seperate applogic more Oct 15 12:40:22 Keep in mind though that I consider this kind of thought process to be intellectual masturbation :-) Oct 15 12:40:23 let me think about this for some time Oct 15 12:40:27 haha Oct 15 12:40:29 AS long as we can do stuff with frameworkd Oct 15 12:40:31 I'm happy Oct 15 12:40:57 I just like clean, well-designed and coherent architectures Oct 15 12:41:03 sure, me too. Oct 15 12:41:13 And it seems to me frameworkd was done this way except for oeventsd/opreferencesd Oct 15 12:41:15 definitely needs to be revisited Oct 15 12:41:39 problem is if you do something meta on that Oct 15 12:41:43 you will rebuild dbus :-) Oct 15 12:42:39 (as in writing abstract definitions for actions and enqueuing them in frameworkd through calls) Oct 15 12:43:21 yep. well at least we try hard to not introduce tight coupling. i force the guys into using dbus even for subsystem2subsystem communication, since i want all that stuff to be optional. if people think frameworkd is too heavy since it comes with all that stuff that they'd rather do in their app, then we have a problem. Oct 15 12:43:47 so perhaps we should move towards a frameworkd-core Oct 15 12:43:47 mickeyl: indeed Oct 15 12:43:52 and then packaging by subsystem Oct 15 12:43:53 for a start Oct 15 12:43:59 Could be a good idea :-) Oct 15 12:44:18 That way it would also cleanly separate responsabilities (commit rights ?) Oct 15 12:44:23 In any case Oct 15 12:44:33 Did you plan to do a full review of boot time for frameworkd ? Oct 15 12:44:49 Because I use it for a while now and it's clearly the slowest thing I know :-p Oct 15 12:45:18 sure Oct 15 12:45:21 but not before 0.9 or so Oct 15 12:45:27 i don't think it's worth atm. Oct 15 12:45:33 there are more pressuring issues Oct 15 12:45:41 i don't think it's that slow btw. Oct 15 12:45:43 turn off all debugging Oct 15 12:45:50 and see when it enters the mainloop Oct 15 12:45:53 Ah glad you remind me about that Oct 15 12:45:56 that goes pretty quickl Oct 15 12:46:06 I should think more about what I say before actually saying it Oct 15 12:46:30 I'm used to use it in full debug to have logs :-) Oct 15 12:46:39 yep, it's amazingly slow that way Oct 15 12:46:44 we burn tons of cycles in debug Oct 15 12:46:58 Hey Oct 15 12:47:06 quickdev: hey baby boy Oct 15 12:47:11 i love the logger infrastructure but we need to come up with a way to yank everything but WARNING and ERROR Oct 15 12:47:12 any news? Oct 15 12:47:18 and when i mean yanking i mean preprocessing Oct 15 12:47:26 quickdev: yes, market is crashing, sell your openmoko bonds Oct 15 12:47:27 so that the calls really vanish Oct 15 12:47:56 mickeyl: yep Oct 15 12:48:14 mickeyl: any way, if I now try to have a look at frameworkd and see what I would like to see ASAP Oct 15 12:48:21 I would obviously chose pim Oct 15 12:48:29 It's a must have for SHR and for users in general Oct 15 12:48:39 And as we have a flash-coder in SHR Oct 15 12:48:53 We would have a good pim management application in two seconds Oct 15 12:48:59 Except when he has to see his girlfriend :-) Oct 15 12:49:15 (But I just paid the cleaners, she shouldn't bother us anymore) Oct 15 12:49:48 hehe Oct 15 12:49:59 AinuFallback: charlie will add the first PIM stuff until end of this month, since he will leave the framework team by then. after that i want to hire the GSoC student that did the PIM project to continue it for FSO Oct 15 12:50:08 he's just missing in action atm. :/ Oct 15 12:50:11 [the gsoc student] Oct 15 12:50:17 abraxa ? Oct 15 12:50:19 yep Oct 15 12:50:33 mickeyl: ok i think i'm getting a hang of libgsmd now. it's pretty cool. :D Oct 15 12:50:33 What is he doing then ? :-) Oct 15 12:50:43 mickeyl: next stop is SHR Oct 15 12:50:54 trupheenix: SHR doesn't use libgsmd though :-) Oct 15 12:51:01 AinuFallback: then? Oct 15 12:51:08 then what ? Oct 15 12:51:10 AinuFallback: frameworkd? Oct 15 12:51:12 yep Oct 15 12:51:28 trupheenix: but if you want to update libgsmd in shr Oct 15 12:51:32 You're welcome to do it Oct 15 12:51:52 AinuFallback: hmmm i would like to... but i'm not clear on the exact pitfalls of using libgsmd. Oct 15 12:51:59 Though I'm sorry to say there's no use for that I think Oct 15 12:52:11 trupheenix: it's highly unstable, unmaintained Oct 15 12:52:14 And mostly buggy Oct 15 12:52:16 IIRC Oct 15 12:52:24 There were issues with suspend too Oct 15 12:52:31 And sometimes calls weren't received properly Oct 15 12:52:36 AinuFallback: yes that i have noticed.... Oct 15 12:52:51 trupheenix: what is your main problem with ogsmd (frameworkd) ? Oct 15 12:52:53 and it has no idea about multiplexing Oct 15 12:53:03 The fact that it's python ? Oct 15 12:53:09 AinuFallback: yes Oct 15 12:53:13 Why is this a problem ? Oct 15 12:53:44 AinuFallback: because it eats more cpu cycles. ultimately i think ogsmd should b entirely written in C. Oct 15 12:53:58 trupheenix: I'm sorry to say it's pretty quick here Oct 15 12:54:12 trupheenix: and all in all it's truly as efficient in term of speed as libgsmd Oct 15 12:54:15 (I used both) Oct 15 12:54:18 oh mickeyl by the way Oct 15 12:54:25 AinuFallback: battery life? python drains. i have tried writing python apps on N810. they just drain battery Oct 15 12:54:41 Could you please bug the hell out of raster to force him to implement this primitive to lock the screen in illume ? Oct 15 12:54:46 AinuFallback: then i would like to port ogsmd into C. Oct 15 12:54:47 :-p Oct 15 12:54:58 well Oct 15 12:55:03 if you really want to do that Oct 15 12:55:04 then do it in Vala Oct 15 12:55:05 trupheenix: porting it to vala would be wiser Oct 15 12:55:09 otherwise come back in 5 years Oct 15 12:55:13 when you have done it Oct 15 12:55:13 Damn, he stole my line :-) Oct 15 12:55:16 haha Oct 15 12:55:35 AinuFallback: i'll try to force him -- if not we'll just make a patch :) Oct 15 12:55:40 :-) Oct 15 12:55:49 I just think it's a pity that we can't fully lock the screen Oct 15 12:55:56 zhone screen locker is crap as it doesn't lock illume Oct 15 12:56:03 I'd welcome ogsmd in Vala, _but_ only if it's a drop-in replacement, i.e.s supports the same level of functionality Oct 15 12:56:23 Even if I have to write "I'm a dumbass" to some character device I would like any kind of good solution :-) Oct 15 12:56:27 AinuFallback: no one is paying me to do illume stuff Oct 15 12:56:35 unlike FSO Oct 15 12:56:35 raster: I know but I like to tease you :-) Oct 15 12:56:45 so u're unliekly to get response out of me by "bugging the hell out of me" Oct 15 12:56:46 :) Oct 15 12:56:47 raster: feel free to tell me to stop ;-) Oct 15 12:56:53 AinuFallback: tomorrow there will be one more such runtime and u would want to port it to that. what is wrong with plain old C? Oct 15 12:57:03 likely u're goign to end up on a blacklist of people i don't pay attention to :) Oct 15 12:57:07 trupheenix: I have no problems with C Oct 15 12:57:17 trupheenix: and I wrote lfg just for that issue in fact Oct 15 12:57:25 If we can use lfg right now to use python primitives Oct 15 12:57:37 When they'll be in vala or in brainfuck or in whatever language that could please the user Oct 15 12:57:43 We'll be able to use them the same way Oct 15 12:57:51 raster: meeeh :-> Oct 15 12:57:53 locking of screen is so hellishly unimportant Oct 15 12:58:01 AinuFallback: i'm not in support of such new runtimes because they are just mere conveniences. Oct 15 12:58:04 compared to everything else Oct 15 12:58:17 AinuFallback: it's OK for doing RAD or prototypes but not for a system level program Oct 15 12:58:17 raster: laugh if you feel that way but I think even if it's not the more important thing it's a bit important Oct 15 12:58:21 you can more than happily live without it Oct 15 12:58:22 Because no locking = input on screen Oct 15 12:58:26 Input on screen = no suspend Oct 15 12:58:33 And that's my issue :-/ Oct 15 12:58:46 trupheenix: what runtime are you talking of now? Oct 15 12:59:00 u will get input on screen reagrdless of locking Oct 15 12:59:09 because fso listens to the ts driver for input Oct 15 12:59:11 and that doesnt stop Oct 15 12:59:16 trupheenix: how many cpu cycles does the python implementation of ogsmd actually use over the course of, say, a day? Oct 15 12:59:16 mickeyl: Vala. Erlang. .NET blah blah... ppl just seem to b coming out with a new language or runtime EVERYDAY!! Oct 15 12:59:22 I would be fairly surprised if it was all that many. Oct 15 12:59:22 trupheenix: nope. Vala compiles to gobject-C, no runtime involved. Oct 15 12:59:23 raster: that's only one of my problems anyway Oct 15 12:59:25 trupheenix: Vala has no runtime, it is translated to C before compiling. Oct 15 12:59:30 pb_: i SPEAK ON THE PHONE Oct 15 12:59:34 raster: this morning my phone called two people randomly Oct 15 12:59:36 Viiru: yes.. i got that Oct 15 12:59:40 well Oct 15 12:59:43 raster: because I wasn't able to lock it automatically Oct 15 12:59:49 you are obviously not convincable Oct 15 12:59:52 so lets stop this discussion Oct 15 12:59:53 go for C Oct 15 12:59:58 and please come back when it's done. Oct 15 13:00:00 Viiru: tomorrow u are going to come up with one more abstraction and ask me to port stuff into that just for convenience? Oct 15 13:00:13 Viiru: i think its ok to use such frameworks for RAD or prototyping Oct 15 13:00:16 raster: as I told you before I'm quite open if you suggest me a good and usable solution Oct 15 13:00:26 trupheenix: No. You are asking people to port stuff to what you want it to be in. Oct 15 13:00:26 Viiru: but they aren't good solutions for production environments Oct 15 13:00:29 But grabbing X input doesn't look like one to me :-/ Oct 15 13:00:33 (plus i wouldn't know how to do it) Oct 15 13:00:39 trupheenix: right, and how many cycles does ogsmd use while you are doing that? Oct 15 13:00:41 Viiru: i'm porting it i'm not asking others to port it Oct 15 13:00:41 trupheenix: I'm not asking you to do or not do anything. And wont. Oct 15 13:00:51 pb_: haven't chekd... Oct 15 13:00:59 and the answer is Oct 15 13:01:00 0 Oct 15 13:01:04 it's stuck in a select call Oct 15 13:01:07 waiting for data from the modem Oct 15 13:01:15 trupheenix: Then port. Do not rant. Oct 15 13:01:16 trupheenix: it sounds a bit premature to be thinking of rewriting it in another language in order to save cycles, then Oct 15 13:01:59 pb__: well...!!!! there's intuition+common sense which tells u to write something in a base level lang. especially when it's something like handling a phonecall. Oct 15 13:02:43 trupheenix: possibly, but common sense should also tell you to gauge the size of the possible improvement before spending time working on it. Oct 15 13:03:27 also, intuition might tell you that an application whose task is basically string-oriented is perhaps not best done in C since other languages have better support for those kinds of operations. Oct 15 13:03:39 "Programmer time is expensive; conserve it in preference to machine time" Oct 15 13:03:39 pb__: there is definitely room for lot of improvement! Oct 15 13:04:00 "Prototype before polishing. Get it working before you optimize it" Oct 15 13:04:09 trupheenix: how can you assert that there is room for a lot of improvement if you haven't measured the starting point? Oct 15 13:04:10 pb__: who said C is not good at handling strings? C is inconvenient to handle strings. Oct 15 13:04:49 Hmm. A troll, perhaps? Oct 15 13:04:50 pb__: i write code. i find scope for improvement every time i take a look at it with an open mind. Oct 15 13:05:01 Or was it a trollop ? :-) Oct 15 13:05:03 heh Oct 15 13:05:16 oh well. like mickey said, go forth and reimplement. Oct 15 13:05:33 pb__: thx :) i would love to do that Oct 15 13:05:40 oh c'mon people. No need to get hostile here. Fight this flamewar another time... Oct 15 13:05:50 AinuFallback: ASU has aux for lock Oct 15 13:05:54 its a trivial lock mechanism Oct 15 13:06:00 my motorola doesnt auto-lock Oct 15 13:06:04 u need to manually lock it Oct 15 13:06:06 trupheenix: Saying "Everything written in Python is slow and uses up your battery" isn't exactly my definition of "with an open mind"... Oct 15 13:06:08 and its a released USABEL working product Oct 15 13:06:19 i just rememebr to flip the lock sqitch every time i finish with it Oct 15 13:06:38 you are making a big problem out of something existing manufacturers dont even SEE as a problem Oct 15 13:06:47 also the APP doesnt need control over it Oct 15 13:07:02 Viiru: there you may be right. but dude Py is not a solution for a production app! Oct 15 13:07:03 raster: please keep in mind I'm not trying to make a big problem out of it Oct 15 13:07:09 I just merely expressing my needs as an user Oct 15 13:07:16 I understand why you're telling as a developer Oct 15 13:07:19 s/why/what/ Oct 15 13:07:36 trupheenix: I think it is, and have produced a few production apps in Python.. Oct 15 13:07:43 trupheenix: you have no idea how many python software is in production. it rocks the internet, dude. Oct 15 13:07:48 *shrug* Oct 15 13:08:12 mickey|coffeebrk: Ha ! That's why internet is slow here then ! Oct 15 13:08:15 * AinuFallback flees Oct 15 13:08:18 mickeyl: Viiru: production app on a mobile device. not a server. Oct 15 13:08:19 auto-lock wont work thanks to it being in your pocket and the screen still being touched Oct 15 13:08:21 if its manual lock Oct 15 13:08:24 thats alreayd done Oct 15 13:08:31 e comes with a default key binding for it (aux) Oct 15 13:08:44 not my fault the xglamo build in oe doesnt export power or aux as keys Oct 15 13:08:45 :) Oct 15 13:08:52 Heh :-) Oct 15 13:09:06 trupheenix: I don't develop for mobiles atm, so I don't have any facts to contribute. Oct 15 13:09:12 is there any way to reenable it by the way raster ? Oct 15 13:09:19 Because here I don't have this lock menu anymore Oct 15 13:09:24 (which I was knowing about indeed) Oct 15 13:09:29 AinuFallback: probably Oct 15 13:09:32 I do intend to develop apps for the Openmoko, and I do intend to develop those with Python. Oct 15 13:09:35 checkl the patchjes to xglamo from om's tree Oct 15 13:09:39 see whats missing in oe's Oct 15 13:09:46 i havent had time to bother looking at it Oct 15 13:09:51 if u had that Oct 15 13:09:53 out of the box Oct 15 13:09:56 u coudl press aux Oct 15 13:09:58 and it's be locked Oct 15 13:10:04 press pwoer and it's instantly suspend Oct 15 13:10:08 or should Oct 15 13:10:14 that would be fine.. Oct 15 13:10:34 that would make the phone equally as usable as the motroola rokr e6 i own Oct 15 13:10:37 no not fantastic Oct 15 13:10:42 but a usable phone Oct 15 13:11:10 so from that view i still am wondering why u need detailed manual control over lock from apps so urgently Oct 15 13:11:11 :) Oct 15 13:11:19 if anything the things you want are Oct 15 13:11:23 because this idle signal "lock" bugs me :-) Oct 15 13:11:32 I want to do something with it and I can't :-p Oct 15 13:11:38 1. a hind on an app window that suggests to the wm "i am hyper important - dont lock the screen automatically no matter what" Oct 15 13:11:49 or Oct 15 13:12:01 AinuFallback, I think manual lock would be enough.. Oct 15 13:12:16 As for now I agree quickdev as I said before Oct 15 13:12:20 2. i am a hyper-sensitive app - if my widnwo exists at all i want you to lock the screen asap - or simply someohow deny access to a user to get to my window withotu some form of authentication" Oct 15 13:12:31 not detailed "we must lock now" calls Oct 15 13:12:36 AinuFallback, btw...my ophonekitd does not suspend, every status is 6, every! ** (process:1440): DEBUG: idle notifier state handler called, id 6 Oct 15 13:12:43 you want to do things as hints/suggestions on a high-semantic level Oct 15 13:12:46 quickdev: update lfg and frameworkd Oct 15 13:12:51 and for that you need high level use cases throught out Oct 15 13:12:54 :) Oct 15 13:13:10 use cases ftw! :) Oct 15 13:16:26 :) Oct 15 13:16:30 that's what i'm after Oct 15 13:16:32 as i said Oct 15 13:16:39 often peole say " i want X" Oct 15 13:16:57 but they decide they want X because they havent stood back and looked at the bigger picture of the problem Oct 15 13:17:04 and i always want the bigger pricture Oct 15 13:17:09 because more often than not Oct 15 13:17:11 the solution is not X Oct 15 13:17:13 but Y Oct 15 13:17:14 :) Oct 15 13:17:23 raster: or more often than not you need glasses :-p Oct 15 13:17:34 hence the bigger picture :-> Oct 15 13:17:41 here i seriously doubt detailed manual control over "must lock now" and "must disable lock now" is going to just cause ighting between processes Oct 15 13:17:45 as 1 process demands "lock now" Oct 15 13:17:51 andother demands "must not lock now" Oct 15 13:17:58 and they can go back and forth fighting Oct 15 13:18:06 thats what something like a wm is there ot arbitrate Oct 15 13:18:14 hints on widnwos are better Oct 15 13:18:31 as the wm can decide what to do based on its own idea of what the user is currently focused on doing Oct 15 13:18:36 and user preferences Oct 15 13:18:56 i say this Oct 15 13:19:04 because EXACTLY this happens with keyboard repeat\ Oct 15 13:19:08 and it causes no end of hassles Oct 15 13:19:15 same with the old xscreensaver controls in x for blanking Oct 15 13:19:19 apps fight over it Oct 15 13:19:23 the root window contents is another one Oct 15 13:19:32 as apps fight over who gets to draw on root Oct 15 13:19:35 or who sets the wallpaper Oct 15 13:19:50 i can continue naming exmaples of "apps fighting over a resource" Oct 15 13:19:54 when you give them explicit api control Oct 15 13:20:03 it almost always turns out bad Oct 15 13:20:04 :( Oct 15 13:22:15 raster: what if the app has no window ... Oct 15 13:22:37 * rwhitby suspects raster sees everything as windows ... Oct 15 13:23:03 rwhitby: it has no business dealing with screen locks :) Oct 15 13:23:07 rwhitby: that's until I complete my ncurses phone library Oct 15 13:23:09 as its not involved in the screen Oct 15 13:23:23 raster: bzzt Oct 15 13:23:47 i disagree Oct 15 13:23:58 give me an example Oct 15 13:24:03 my accel and sound senser which has no window, but is set up to lock the screen when it detects being put in a pocket Oct 15 13:24:14 (good use case indeed) Oct 15 13:24:27 Should have thought about that one Oct 15 13:24:55 hmm Oct 15 13:24:57 good point Oct 15 13:25:09 * AinuFallback laughs maniacaly Oct 15 13:25:38 Thanks to rwhitby, we drove raster mad ! :-) Oct 15 13:25:53 thought it still shouldnt have explicit control Oct 15 13:25:57 it should just be able to hint Oct 15 13:26:04 "i'd like to lock now" Oct 15 13:26:10 agreed, but not through a window Oct 15 13:26:19 raster: that's the thing I was looking for, no matter what you call it in fact Oct 15 13:26:28 raster: problem is, right now idle notifier is dumb on frameworkd Oct 15 13:26:41 Though if I wire him to some kind of idleness detector for all kind of inputs Oct 15 13:26:44 I would like to do that indeed Oct 15 13:26:52 By any mean Oct 15 13:27:13 that would be more to the point Oct 15 13:27:16 as i said before Oct 15 13:27:34 frameworkd should just advertise" system has gone idle - do what it is you do when we go into that state" Oct 15 13:27:38 if that is lock Oct 15 13:27:39 then lock Oct 15 13:27:44 if it is to dim the screen - then dim Oct 15 13:27:45 raster: it does that then Oct 15 13:27:47 itf its do nothnig Oct 15 13:27:50 then do nothing Oct 15 13:27:51 but it does it using its way Oct 15 13:27:53 (dbus signal) Oct 15 13:28:10 but explicit control over "display lock screen now no matter what" to me just is bad Oct 15 13:28:24 raster: in the end it just is a badly expressed need Oct 15 13:28:25 agreed Oct 15 13:28:34 Because the actual need is what you just described Oct 15 13:28:41 AinuFallback: already said this :) Oct 15 13:28:47 Yep Oct 15 13:28:48 fso shoudl not go dim the screne Oct 15 13:28:54 it should not turn of bl itself Oct 15 13:28:58 I agree with you on that Oct 15 13:29:06 Talked about that with mickey|coffeebrk Oct 15 13:29:06 it shoudl simply broadcast the "everythign i know about is idle now" Oct 15 13:29:13 raster: it does that Oct 15 13:29:18 it should have input providers that indicate something has gone idel Oct 15 13:29:22 but oeventsd acts given this signal Oct 15 13:29:27 when all providers (necessary) havs said they are idle Oct 15 13:29:28 this can be disabled Oct 15 13:29:36 then it tells everyone the whole thing is idle Oct 15 13:29:37 raster: I almost thought about building a shelf gadget to handle that in fact Oct 15 13:29:49 so frameworkd acts as a collector and arbitrator of inputs Oct 15 13:29:50 (i.e. idleness listener) Oct 15 13:29:59 it probably should not be wathcing touchscrene or buttosn or keys itself Oct 15 13:30:11 this is an interesting discussion but I really have to go Oct 15 13:30:20 Will look into what you say later on :-) Oct 15 13:30:30 as such - i will vet to doing that Oct 15 13:30:35 eventually Oct 15 13:30:42 but hey - it's not on a priority list Oct 15 13:30:46 patches accepted Oct 15 15:55:26 Hey Oct 15 15:56:03 Ainulindale? Oct 15 15:59:55 Yes ? Oct 15 16:02:03 what do we next? Oct 15 16:02:38 what's most important? Oct 15 16:13:29 an app for the settings Oct 15 16:13:42 and integrate navit into shr Oct 15 16:14:25 i'm working on a new edje for shr Oct 15 16:14:55 Ainulindale, what about a settings applications that changes the profiles for example? Oct 15 16:16:23 profiles, dimming, suspend, gps, bluetooth, wifi Oct 15 16:17:12 Damn Oct 15 16:17:15 Sorry for the delay Oct 15 16:17:25 This technician was talking and talking and talking Oct 15 16:17:33 "Linux is good before ble bleh bleeeeh" Oct 15 16:17:35 But now see Oct 15 16:17:38 My connection is stable Oct 15 16:17:52 It was this f@#:king alimentation Oct 15 16:18:03 quickdev: well regarding your question Oct 15 16:18:15 1st, this manual suspend thing Oct 15 16:18:26 2nd check how to reenable illume lock screen Oct 15 16:18:39 Then various modifications on the UIs Oct 15 16:18:51 Buttons, icons, and this thing to look more easily in contacts Oct 15 16:18:56 (really a hassle) Oct 15 16:19:03 Then as you just told, a settings application Oct 15 16:19:11 Which should be in lf-p Oct 15 16:19:44 Your opinion ? Oct 15 16:20:25 settings should be in libframeworkd-phonegui-efl? Oct 15 16:20:33 same opinion Oct 15 16:20:43 regarding 1/2 Oct 15 16:23:38 Ainulindale, shouldn't we make that settings application modular / extensible? Oct 15 16:46:42 quickdev: what do you mean ? Oct 15 16:47:16 Ainulindale, a settings manager that could easily be extend by a setting Oct 15 16:48:03 as settings are in frameworkd Oct 15 16:48:20 my guess was that doing that in lf-p should be a good thing Oct 15 19:04:38 Hire, you're working on a theme? Oct 15 19:04:46 yes Oct 15 19:05:14 it will based on the raster's edje Oct 15 19:05:34 a theme for illume then? Oct 15 19:05:43 yes Oct 15 19:06:29 next I wiil see what I can do for the gtk theme Oct 15 20:01:33 rwhitby, here? Oct 15 20:33:08 quickdev: here please Oct 15 20:33:23 yeah Oct 15 20:33:28 thought it was an oe issue Oct 15 20:33:34 because it failed on several boxes Oct 15 20:34:28 rwhitby, it's only a path to fix? Oct 15 20:35:10 quickdev: more than that Oct 15 20:35:16 I need to do it. Oct 15 20:35:44 ok, would be fine, thanks in advance Oct 15 20:37:01 cloning new repo now Oct 15 20:45:56 quickdev: should be ok now Oct 15 20:47:59 rwhitby, http://rafb.net/p/ooWkGp70.html - does this look correctly? Oct 15 20:51:22 quickdev: yes Oct 15 21:08:36 mickey|sofa: hey there Oct 15 21:08:45 What is the device I'm supposed to use for Device.Audio ? Oct 15 21:08:51 (maybe alphaone could answer that one also) Oct 15 21:10:00 (I would say /Device/Audio but just checking) Oct 15 21:12:27 by the way mickey|sofa Oct 15 21:12:33 there should be a way to identify scenarii Oct 15 21:12:39 i.e. "gsmhandset" "speakerout" Oct 15 21:12:48 I don't know if it has been implemented that way Oct 15 21:13:15 Ah I spoke too quickly Oct 15 21:13:28 I don't know if it has been normalized though Oct 15 21:32:24 quickdev: audio almost ended Oct 15 21:32:27 (partially though) Oct 15 21:32:34 Just for the parts useful to us Oct 15 21:32:40 fine :) Oct 15 21:34:32 Trying to recompile ophonekitd here just to see nothing is broken Oct 15 21:37:47 freesmartphone.org: 03ainulindale 07libframeworkd-glib * r7cef687754b8 10/src/ (8 files in 3 dirs): Oct 15 21:37:47 freesmartphone.org: Implemented Device.Audio. Oct 15 21:37:47 freesmartphone.org: TODO : Oct 15 21:37:47 freesmartphone.org: - Implement methods : GetSupportedFormats, GetAvailableScenarios (which should be scenarii), PlaySound, StopSound, StopAllSounds, GetScenario. Oct 15 21:37:50 freesmartphone.org: - Implement a handler for SoundStatus signal in FrameworkdHandlers Oct 15 21:37:52 freesmartphone.org: 03ainulindale 07libframeworkd-glib * ra2e6cb5d1788 10/src/odeviced/ (2 files): Merge ssh://git@git.freesmartphone.org/libframeworkd-glib Oct 15 21:38:08 Hmpf Oct 15 21:38:12 Dumb merge. Oct 15 21:38:36 quickdev: it's all yours Oct 15 21:39:18 why "dumb merge"? Oct 15 21:42:01 Nothing, just that merges appear in the commits Oct 15 21:44:03 I'm going to regenerate packages on buildhost Oct 15 21:46:03 Ainulind1le, don't think that it will work...seems that all packages are regenerated atm.*all* Oct 15 21:46:41 All ? Why is that ? Oct 15 21:49:02 quickdev: anyway I rebuilt ophonekitd and lfg by hand Oct 15 21:49:13 try it Oct 15 21:49:17 Already di Oct 15 21:49:18 +d Oct 15 21:51:50 Recompiled frameworkd too Oct 15 21:53:22 and kernel? ;) Oct 15 21:53:54 You didn't answer before, why is everything rebuilt ? Oct 15 21:55:27 quickdev ? Oct 15 21:56:55 because rwhithby corrected the error and I think that's why a new build is needed Oct 15 21:57:02 Which error ? Oct 15 21:57:07 In any case I rebuilt the package index Oct 15 21:57:28 So ophonekitd/frameworkd/lfg should be up to date if you update/upgrade Oct 15 21:57:32 ask him about further details, not me ;) Oct 15 21:59:12 quickdev: by the way, is there any buildhost with an up to date kernel ? Oct 15 22:00:17 there's uimages on the OM buildhost Oct 15 22:00:23 But I don't know how recent they are Oct 15 22:00:26 http://totalueberwachung.de/~alphaone/openembedded-build/glibc/images/neo1973/uImage-2.6.24+r8+gitr1+71f219082f5c04a80f9b143fc92a51a3de5c6683-r8-om-gta02.bin Oct 15 22:00:29 most recent kernel Oct 15 22:00:39 see here: http://git.openmoko.org/?p=kernel.git;a=shortlog;h=stable Oct 15 22:01:47 quickdev: ah could you also check why the icons aren't there for contacts/message ? Oct 15 22:03:30 no icon yet Oct 15 22:05:57 software icon on illume launcher Oct 15 22:06:07 I wasn't talking about software icon as in "call icon" Oct 15 22:06:11 Just the icons on the launcher Oct 15 22:07:17 yes, I'm talking about the same thing ;) Oct 15 22:07:44 (My update worked) Oct 15 22:08:52 hmmm quickdev no logs here for buttons Oct 15 22:09:06 read /dev/input/event* Oct 15 22:09:22 opkg info openmoko-linux Oct 15 22:09:26 Something something Oct 15 22:09:32 on aux Oct 15 22:09:37 sorry, opkg info kernel-2.6.24 Oct 15 22:09:46 mh? Oct 15 22:10:07 I flashed the kernel so will that information be relevant ? Oct 15 22:10:43 probably not, yeah Oct 15 22:10:58 Aaaaargh Oct 15 22:11:00 Arkanoid.sid ! Oct 15 22:11:09 Well in any case there's a problem here Oct 15 22:11:13 suspend while plugged in Oct 15 22:12:26 Meeh. Oct 15 22:12:29 Guess what ? :-) Oct 15 22:13:02 mh? Oct 15 22:13:04 ;) Oct 15 22:13:09 I re* (process:1360): DEBUG: power status: 4 Oct 15 22:13:14 (Critical) Oct 15 22:14:14 Someone lowercased the signals ! :-) Oct 15 22:14:21 power in france is not very reliable? see your internet connection? :) Oct 15 22:14:38 Who's this Michael Lauer guy anyway ? :-p Oct 15 22:14:48 Ainulind1le, do that else{ g_error() THING! } Oct 15 22:14:58 Won't right now Oct 15 22:15:04 argh ;) Oct 15 22:17:29 freesmartphone.org: 03ainulindale 07libframeworkd-glib * rc079416582d0 10/src/odeviced/frameworkd-glib-odeviced-powersupply.h: Oct 15 22:17:29 freesmartphone.org: Followed odeviced PowerSupply "lowercasing". Oct 15 22:17:29 freesmartphone.org: TODO : Listen to quickdev for a bit and do this g_error thing on unknown constants. Oct 15 22:18:26 Happy ? :-) Oct 15 22:18:51 yeah, now I can sleep well this night :) Oct 15 22:19:02 after I did the DTMF stuff, hehe Oct 15 22:19:07 80% done Oct 15 22:19:22 weird Oct 15 22:19:23 ** (process:1360): DEBUG: idle notifier state handler called, id 5 Oct 15 22:19:23 ** (process:1360): DEBUG: power status: 4 Oct 15 22:19:25 Didn't suspend. Oct 15 22:19:40 Something fishy in ophonekitd ? Oct 15 22:19:48 Fishy, fishy, fishy, fish ? Oct 15 22:21:21 rebitbaking package-index on buildhost Oct 15 22:22:27 yeah, should have suspended Oct 15 22:22:43 only thing I see Oct 15 22:22:48 either error or active_calls != 0 Oct 15 22:24:35 yeah Oct 15 22:24:55 Or two frameworkd Oct 15 22:24:57 Which was the case Oct 15 22:26:25 Ok well trying with a reboot :-) Oct 15 22:34:21 and? Oct 15 22:34:56 well Oct 15 22:35:01 won't come out of suspend Oct 15 22:35:26 kernel bug? Oct 15 22:35:39 I don't know Oct 15 22:36:03 <[Rui]> hi, does http://downloads.openmoko.org/daily/ boot to anyone today? Oct 15 22:40:32 well quickdev Oct 15 22:40:35 rebooted and now it works Oct 15 22:40:44 I'll try that tomorrow Oct 15 22:41:16 ok, fine Oct 15 22:47:37 Meanwhile good night :-) Oct 15 23:08:59 could anybody guide me in getting a branch off of the u-boot git-repository? Oct 15 23:09:21 I'm particularly interested in a commit from zeche about 7 days ago. Oct 15 23:09:42 but that commit has some changes to a file that doesn't seem to exist in the main repository Oct 15 23:10:40 I can see the file on docs.openmoko.org using gitweb, but the snapshot thingy doesn't seem to work. Oct 15 23:10:46 any suggestions? **** ENDING LOGGING AT Thu Oct 16 02:59:58 2008