**** BEGIN LOGGING AT Wed Jan 28 02:59:57 2009 Jan 28 03:00:00 rotfl Jan 28 03:00:11 when the bytes in da var, cast it like it's hooooooooooooot Jan 28 03:01:43 ¬_¬ frank... Jan 28 03:02:02 the superstar is the winner of the cast(ing )show ... Jan 28 03:02:27 heh Jan 28 03:24:09 so what is the best way to change the type of a var? Jan 28 03:24:27 depends, plus be more clear Jan 28 03:24:43 implicit/explicit casting? Jan 28 03:25:06 Dave: beats me, i'm kind of a programming noob Jan 28 03:25:27 Dave: i can create a code example of what i want to do real quick thou Jan 28 03:25:54 there some good c tutorials in the internet Jan 28 03:31:34 cool :) Jan 28 03:31:40 http://pastebin.com/m61472e26 i want to achieve something like this Jan 28 03:31:42 pwgen, that's highly ambiguous :p Jan 28 03:32:36 oh, but i would like to be able to make any un-needed var# "NULL" in the function call Jan 28 03:32:43 if at all possible Jan 28 03:33:28 yes Jan 28 03:33:54 so is there a "best" way to do that? and do you know of an example? Jan 28 03:36:46 do i have to wait until after the framework says the sim is ready before calling the gsm.network.register method? Jan 28 03:38:49 AntonTakk: varargs Jan 28 03:39:05 you really want farargs Jan 28 03:39:09 err varargs Jan 28 03:39:22 man va-start Jan 28 03:39:23 err Jan 28 03:39:26 man va_start Jan 28 03:39:26 ok, i'll look that up, thanks Jan 28 03:39:36 trhat lets you have a variable number of arbuments to a function Jan 28 03:39:42 and you do need to walk the args Jan 28 03:39:44 and check them Jan 28 03:39:51 but then u may do something like Jan 28 03:40:14 myfunc("Message1", INT, 1, INT 2, STR, "Hello", END); Jan 28 03:40:17 for example Jan 28 03:40:27 and #define INT, STR, END etc to ints Jan 28 03:40:41 and thus u know that every 2nd arg is an int - then some argument Jan 28 03:42:13 hrm Jan 28 03:42:25 that would be ok, i suppose Jan 28 03:42:57 trying to make my self a very simple set of functions for talking to fso Jan 28 03:44:02 ideally, i have functions with names like usage_method("Method", arg1, arg2, etc) Jan 28 03:45:51 oh no, it's old man raster! Jan 28 03:47:03 and have it setup signals so when a (set of) value(s) are returned, it comes back packed in an associative array, so i can get it as simply as var = array[ValueName]; :) Jan 28 03:56:53 raster: looks good, and i assume it can be used with any variable type? Jan 28 04:15:32 i think it's gonna register to the network this time :) Jan 28 04:18:38 yay Jan 28 04:18:43 Go anton go :D Jan 28 04:21:23 nope, no beans Jan 28 04:24:33 beans? Jan 28 04:24:42 and here I was hoping for sandwiches Jan 28 04:31:18 * raster slaps Dave Jan 28 04:31:23 old man... pffft Jan 28 04:31:33 that's SEEDY old man to you! Jan 28 04:31:35 :) Jan 28 04:31:47 :D Jan 28 04:31:51 Yes, milord. Jan 28 04:32:42 thats better Jan 28 04:36:19 :) Jan 28 04:45:44 is anyone working on an elementary-based wifi connection manager? Jan 28 04:50:32 not that i know of Jan 28 04:52:04 Would anyone in their right mind want to? Jan 28 04:52:17 hehe, why not? Jan 28 04:52:29 :p Jan 28 04:52:58 that reminds me Jan 28 04:53:08 i shoudl have some kind of "signal" widget Jan 28 04:53:09 well I'm left-handed so I'm technically in my right mind...so yes Jan 28 04:55:38 the problem is more the back end Jan 28 04:55:39 working Jan 28 04:55:42 gui is not hard Jan 28 04:56:12 you want a back-end that works reliably Jan 28 04:56:14 why not just make it a front end for NetworkManager? Jan 28 04:56:18 scans for essid's regularly Jan 28 04:56:24 joins the ones u have configured automatically Jan 28 04:56:24 seems reasonably reliable Jan 28 04:56:27 and reports status Jan 28 04:56:43 a ui to set up your ap's, keys, list/browse them etc is easy Jan 28 04:56:52 * mwester-laptop continues to wonder at the need for "GUI-first" development that seems to plague the Om company, and to a lesser degree, community Jan 28 04:56:53 AntonTakk: nm keeps changing api Jan 28 04:57:08 raster: ah, that sucks Jan 28 04:57:41 mwester-laptop: i don't want bash to be the official ui for my phone ;) Jan 28 04:57:48 I was thinking it could be adapted from wpa_gui Jan 28 04:58:08 just as a client for wpa_supplicant Jan 28 04:58:58 mwester-laptop: for someone who doesnt do gui stuff.. thats pretty biased. Jan 28 04:59:03 the fact is u need both Jan 28 04:59:10 AntonTakk: And I don't need yet another gui with nothing but segv, missing code, and bugs behind itQ! Jan 28 04:59:11 frankly gta02 is usless without a gui Jan 28 04:59:38 Amen! Jan 28 04:59:38 but the gui is fairly pointless without a back-end that works Jan 28 04:59:52 raster: Look at the latest kernel commits, and try to tell me that again. It ******STILL************ cannot suspend and resume! Jan 28 04:59:57 thinking one-sided be it back or front end is going to get you no love. Jan 28 05:00:19 mwester-laptop: suspend/resume is a broken paradigm entirely - but yes. Jan 28 05:00:26 note that they arent paying kernel coders to write guis Jan 28 05:00:36 tho as such this SHOULd have been solved long ago Jan 28 05:00:39 long long long ago Jan 28 05:00:49 also rememebr much kernel dev is done withotu agui Jan 28 05:00:52 The most useless, misfocused development for a product ever done... "Let's put out a device, who cares if it works, then let's go chase the ever-moving latest kernel and to hell with making it ever work!" Jan 28 05:00:55 the gui is never tested Jan 28 05:01:07 i rememebr andy being proud of his custom userspace that just boots to a propt Jan 28 05:01:15 i'm not surprisied things dont work Jan 28 05:01:24 nothing it pumpign stuff to glamo etc. during his tests Jan 28 05:01:30 as gfx is idle Jan 28 05:01:53 a big problem is the isolation in which stuff is done Jan 28 05:01:58 Yep. Jan 28 05:02:08 also as i rememebr there was a big barrier to working together Jan 28 05:02:15 u had to do everything in your own dev Jan 28 05:02:27 then push it and wait for someone to integrate it into the mainline Jan 28 05:02:43 then wait yet longer for that to be built and finally get to other people Jan 28 05:02:47 Andy isn't building a phone. Andy is enjoying doing kernel work, which is going to be never-ending, since new kernel code is ALWAYS going to be available from upstream. That's the problem, and Jan 28 05:02:49 a change u make can take weeks to propagate Jan 28 05:02:51 its silly Jan 28 05:03:03 when u are working that way everyone works on the bleeding edge Jan 28 05:03:09 if Om doesn't see it, then they're blind. If they can see it, then they're stupid. Jan 28 05:03:11 u resolve problems as soon as they turn up Jan 28 05:03:14 u deal with the pain Jan 28 05:03:25 and soon enough u learn to either be really fast with fixes Jan 28 05:03:29 or careful in what you do Jan 28 05:03:47 Well, I gave up. Jan 28 05:03:54 hehehe Jan 28 05:03:58 Seems that none of the users REALLY care Jan 28 05:04:04 i dont pay attention either Jan 28 05:04:30 I set out to offer a stable kernel, albeit a bit older, and all lI ever heard was "when are you going to switch to andy-f**(**-tracking"??" Jan 28 05:04:34 Well, now they have it. Jan 28 05:04:39 Enjoy! Jan 28 05:04:55 mwester-laptop: you're no longer maintaining your 2.6.24 "knife"? Jan 28 05:04:59 Nope.l Jan 28 05:05:03 Nobody wants it. Jan 28 05:05:17 mwester-laptop: is andy's 2.6.28? Jan 28 05:05:29 (now) Jan 28 05:05:32 And nobody was willing to step up to do the same with the newer 2.6.28 stuff, so it's pretty dead. Jan 28 05:05:56 wasn't there something on the ML about them just rebasing to 2.6.29? ugh Jan 28 05:06:01 raster: Nope, andy just rebased onto 2.6.29-rc0.000001-beta-certain-to-buggier-than-lousiana-in-the-summerr Jan 28 05:06:21 oooh Jan 28 05:06:23 Like I said - kernel-chasing. Jan 28 05:06:26 A never-ending sport. Jan 28 05:06:30 And a total waste of time. Jan 28 05:06:50 And Werner hops on the band-wagon encouraging this, and they all decide that they need to rebase MORE OFTEN. Jan 28 05:06:51 oh well - 2.6.8 panics nicely for me Jan 28 05:06:55 instantly into boot Jan 28 05:07:04 beats me why - too lazy to fig out the debug board Jan 28 05:07:09 So for those who though andy-tracking was "stable" --- they're all in for a hell of a shock. Jan 28 05:07:12 got enough other working devices here not to care enough Jan 28 05:07:13 :) Jan 28 05:07:31 hi Sup3rkiddo Jan 28 05:07:36 That's the problem. Without a stable base, developers like you go elsewhere, raster. Jan 28 05:07:47 playya_, good morning :) Jan 28 05:07:50 mwester-laptop: true indeed Jan 28 05:08:07 THAT's why we don't need GUI-first developement -- we need a stable API, and a stable kernel, so that developers feel that it's SAFE to invest their time. Jan 28 05:08:38 there's a coffee in my hand but i didn't wake up this morning. so no good morning :( Jan 28 05:09:32 (I just rebased and partially rewrote my GSM debug code for 2.6.28 -- so we could try to track down the serial port GPS problem... but that effort is now 6 hours of wasted time that I'll never get back.) Jan 28 05:09:35 playya_, oh.. :I Jan 28 05:09:40 mwester-laptop: you need several thnigs top happen at once - that means stable kernel/drivers, stable baseband, stable x, then a good ui environment on top of x Jan 28 05:09:49 how things are going with fsod Sup3rkiddo ? Jan 28 05:09:50 ie enouhg futzing over what font to use Jan 28 05:09:58 just so all the components are there and work and in their right slots Jan 28 05:10:04 waiting for async dbus servers in vala? Jan 28 05:10:21 playya_, not much. am stuck with the final semester project for my bachelors till march Jan 28 05:10:56 I'm still search for a thesis :( Jan 28 05:11:43 playya_, blehmingo.. final thesis? Jan 28 05:12:19 bachelor thesis yes Jan 28 05:12:42 atleast you can choose the domain of your choice.. :/ Jan 28 05:12:44 what the Jan 28 05:12:53 what about a navigator for OM using E Jan 28 05:13:38 anyone interested in this idea :P Jan 28 05:13:44 playya_, you mean a file manager? Jan 28 05:14:02 no. gps navigator software Jan 28 05:14:05 * Sup3rkiddo is looking forward to use the guitar tuning app for the Freerunner Jan 28 05:14:09 playya_, oh.. doh! Jan 28 05:14:18 playya_, rip apart TangoGPS :D Jan 28 05:15:00 :p Jan 28 05:15:16 GuitarRunner, anyone? Jan 28 05:18:04 Dave: I've been dreaming with something like a GuitarRunner (About seeing it, not about coding it :-) ). Jan 28 05:18:14 haha yeah :D Jan 28 05:18:16 same here Jan 28 05:23:38 gah Jan 28 05:23:40 kernel bugs Jan 28 05:23:56 kernel build bugs Jan 28 05:26:33 raster: which kernel? latest andy-tracking? Jan 28 05:27:03 eo-dev Jan 28 05:27:06 2.6.28 Jan 28 05:28:10 raster: Ah, OK. Right now the latest andy-tracking in git is 2.6.29 .... and it might not build with an old config ... Jan 28 05:29:10 raster: I don't know how the new stable will be forked, but I guess we will be told soon... Jan 28 05:29:59 Yep, that's how you create new stable releases, for sure! Jan 28 05:33:56 mwester-laptop: it's rather amazing how long its taking to get just a core os "Working" Jan 28 05:34:21 i have to say - it'd be working much better... if certain hardware choices had not been made Jan 28 05:34:22 :) Jan 28 05:34:32 ie it'd have been mucho easier Jan 28 05:36:10 :) I'll add that it'd be working much better... if openmoko had stuck with 2.6.26 until they got everything functional on that OS release. Jan 28 05:36:31 hmmm Jan 28 05:36:42 the new sdio stack has nuked many bugs Jan 28 05:36:48 so there is some method to the madness Jan 28 05:36:52 as best i can see Jan 28 05:37:02 but to me - thats the "kernel team" work Jan 28 05:37:13 and i should work on things other than that Jan 28 05:37:25 then again Jan 28 05:37:35 Sure, there are always a few gains to be had in a new baseline. But then we went *backwards* in terms of the WSoD with the Glamo for 6 months. Jan 28 05:37:44 i dont get paid to even care about the kernel Jan 28 05:37:45 :) Jan 28 05:37:58 aaah yeah Jan 28 05:38:02 is that now resolved? Jan 28 05:38:14 i got bored of seeing the trac/emails Jan 28 05:38:23 So did we *really* net gain anything? I submit that the glamo breakage in 2.6.28 for so many months did more to discourage developers than the wifi not working well. Jan 28 05:38:46 glamo not working is a show stopper Jan 28 05:38:59 if u cant (reliably) see the output of your stuff (the screen) Jan 28 05:38:59 raster: No, it's not resolved. Someone found a workaround, and Andy put that in, and everyone has been happy to shove that problem under the rug. Jan 28 05:39:02 thats just bad Jan 28 05:39:52 So no, so to speak, we are all standing on a very lumpy rug, hoping to ourselves that the glamo problem doesn't pop back out at some point and cause more trouble... Jan 28 05:40:14 it wont Jan 28 05:40:19 because gta03 will come... Jan 28 05:40:24 and that will get rid of that Jan 28 05:40:25 :) Jan 28 05:41:30 But, heck, we all know what the truth is -- the gta02 is going the way of the gta01; the push to the newer kernel is all about Ben Dooks' work, and Om has stuck it's head in the sand (or somewhere dark) and is trying to ignore the fact that pissing off the gta02 owners may hurt sales of the gta03. Jan 28 05:42:21 i was being negative Jan 28 05:42:23 i admit Jan 28 05:42:32 i hope they do not abandon gta02 Jan 28 05:42:43 or even reduce focus on it to make it a solid platform Jan 28 05:42:51 it has enough ram and flash and all the devices Jan 28 05:42:55 sure the gfx is slow Jan 28 05:42:58 So in the end, it's all about the gta03. end of story. that's why we're sitting here with a gta02 where the kernel won't even build, and when that gets fixed, it won't *really* be any better than 2.6.24 was... Jan 28 05:43:00 but it is capable of working Jan 28 05:43:08 Yep, it's capable. Jan 28 05:43:19 but glamo makes that pain very high Jan 28 05:43:19 But Om is desperate now. Jan 28 05:43:33 So what is openmoko responsible for these days? Only the kernel? Are they working on anything userspace? Jan 28 05:43:57 Om gets no income by making the gta02 work -- just cost. From a business point of view, there is nothing in it for them to worry about the gta02 any more. Jan 28 05:44:15 mwester-laptop: i am not sure... but my guess is that gta03 is not going to be popular Jan 28 05:44:46 mwester-laptop: in the bigger scheme there is Jan 28 05:44:48 its lost PR Jan 28 05:44:52 lots customer relations Jan 28 05:44:52 I can't imagine who they would sell it to... they have no user-space. Qt Extended has stalled, Android hasn't gone anywhere. Jan 28 05:44:55 lost Jan 28 05:44:57 i mean Jan 28 05:45:20 Om puts no store in customer relations, that's clear. Jan 28 05:45:24 gta02 really hit a problem with the delays Jan 28 05:45:34 Yes. Jan 28 05:45:34 if it had come out when it was meant to Jan 28 05:45:41 like 2007 Jan 28 05:45:47 it was an "ok" bit of hw Jan 28 05:45:56 but delayed enough it was far behind the curve in terms of hw Jan 28 05:46:05 (no 3g, gfx, cpu etc.) Jan 28 05:46:20 and probably should have been ditched for a newer hw base Jan 28 05:46:32 otoh, it's a cell phone, nota next gen gaming console Jan 28 05:46:43 given everything so far - that would have been faster and cheaper Jan 28 05:46:44 :) Jan 28 05:47:06 AntonTakk: its competitors.. cellphones ARE next gen gaming consoles Jan 28 05:47:09 as well as phones Jan 28 05:47:11 Had they sold it as a development platform and made sure that the base SW was stable enough to support user-space development, then maybe right now we would have a proper user-space, and a stable kernel, ready for the GTA03. Jan 28 05:47:24 g1, iphone(2g and 3g), palm pre ... i can go on Jan 28 05:47:43 As it is, the GTA03 will be out, and it will not have anything more than an OS -- it won't even be a phone, much less a gaming device. Jan 28 05:48:53 Hey J Jan 28 05:49:35 i mean, i can still go out and buy the model of smart phone i have brand new for $400, 200 MHz CPU, 32 MB ram, 64 mb flash, no gps, no accelerometers - you don't have to be cutting edge hardware to be successfull Jan 28 05:49:36 mwester-laptop: let's see how it goes Jan 28 05:50:16 AntonTakk: which one? Jan 28 05:50:26 raster: tmobile wing Jan 28 05:50:42 aka htc herald Jan 28 05:50:55 aaaah Jan 28 05:51:13 * mwester-laptop hasn't seen one of those. Jan 28 05:51:13 slight difference Jan 28 05:51:16 it came OUT in 2006 Jan 28 05:51:40 gta02 came out 2008 Jan 28 05:51:46 vast difference Jan 28 05:51:49 raster: true, but there is aparently still a market for it, it wouldn't be in the display shelves if there wasnt Jan 28 05:52:01 and at 400 thats overpriced Jan 28 05:52:46 raster: duh, but i wasn't about to buy a crackberry and pay extra for e-mail on top of my data plan Jan 28 05:53:24 hehehehe Jan 28 05:53:26 fair enough Jan 28 05:55:10 i think if OM gets their act together and can put a decent enough face on FSO, while mickey and co get the basic necessities solid in fso (calls, sms, data connections, pim) and OM learns from it's hardware mistakes, the gta03 could do quite well Jan 28 05:55:41 * Dave refrains from commenting. Jan 28 05:56:49 and tbh, i hope their case redesign is a move in the right direction at least Jan 28 05:57:07 uggh, yes. Jan 28 05:57:14 case redesign == good Jan 28 05:57:23 no kiddin' Jan 28 05:57:38 i don't want another phone that would just need a bubble on top to look like an apple mouse Jan 28 05:58:35 haha yeah Jan 28 05:58:43 AntonTakk, a apple mouse has less buttons Jan 28 05:58:52 OWNED Jan 28 05:58:57 playya_: actually, it has more Jan 28 05:59:02 AND a scroll wheel Jan 28 05:59:25 right click, left click, scroll ball click, and scrolling Jan 28 05:59:51 and imho that's just fing sad Jan 28 06:00:28 * Dave pictures the Freerunner with a scroll wheel Jan 28 06:00:49 this is supposed to be an open hardware smart phone with *nix os, not a suppository with a touch screen Jan 28 06:01:45 Dave: and it wouldn't even need a scroll wheel, per se, but some kind of buttons for scrolling/navigation/selection Jan 28 06:02:12 not one of those scroll ball cursor dealies that the new blackberries have :( Jan 28 06:02:22 even just a dpad with an enter "key" in the middle Jan 28 06:02:24 though some people like 'em Jan 28 06:02:35 haha, yeah, that would be sweet, Anton Jan 28 06:02:40 some people like CDE, too Jan 28 06:02:43 I'll just cut it out of one of my old LG's :p Jan 28 06:02:56 CDE? Jan 28 06:04:09 Dave: old gui env used by sun, one sec Jan 28 06:04:47 http://www.guidebookgallery.org/screenshots/cde15solaris9 Jan 28 06:05:00 that brings back memories :) Jan 28 06:06:31 oh gawd Jan 28 06:06:42 I haven't quite figured you out yet, Anton... Jan 28 06:06:52 Methinks he has a Sun fetish. Jan 28 06:08:03 OH GOD Jan 28 06:08:11 * AntonTakk hides the sparcs and the "reserved for RS/6000 sticker Jan 28 06:08:11 WHY DID YOU MAKE ME REMEMBER THAT HORRIBLE ENVIRONMENT Jan 28 06:08:25 >:( Jan 28 06:08:41 what? me? no....i'm not a wierdo...... Jan 28 06:08:57 :P Jan 28 06:09:07 thanks a lot though Jan 28 06:09:14 heh Jan 28 06:09:18 Just for that, I quoted you in the wiki :) Jan 28 06:10:06 it's like the whole ww2/germany thing, if we forget, we run the risk of making the mistake again :) Jan 28 06:10:54 actually, i think my solaris 8 or 9 CDs come with enlightenment 14, maybe 15 Jan 28 06:11:43 * mwester-laptop still has his Solaris 2.0 CD Jan 28 06:11:52 nice Jan 28 06:11:59 The ones personally signed by Bill Joy. Jan 28 06:12:02 :) Jan 28 06:12:18 * AntonTakk flashes his SparcStation10 at mwester-laptop ;) Jan 28 06:12:40 :P I only had an IPC Jan 28 06:13:22 i relatively recently aquired an E450, thou i was horibly lied to about the specs Jan 28 06:15:44 anyways Jan 28 06:15:52 back to more recent things Jan 28 06:16:18 FSO really isn't liking my attempts to register Jan 28 07:20:49 hmm. ok. shr has nice building directions. Jan 28 07:22:02 Huh? Jan 28 07:22:11 Is that ironic or did someone work on that? =) Jan 28 07:23:57 well, they exist, and seem pretty easy to read. if they break, i'll complain, and edit the wiki, assuming i can get an account. same as all my other projects. :) Jan 28 07:24:23 Well no problem, we need that kind of thing :-) Jan 28 07:25:09 you need more than that. ;) Jan 28 07:28:13 Indeed. Jan 28 07:42:39 Ainulindale: you have worked with communicating with the fso, right? Jan 28 07:46:59 raster: btw. CPU problem exists also with current EFL from yesterday Jan 28 07:47:37 * mrmoku_ likes the new sexy labels in sysaction menu though :D Jan 28 07:47:56 mrmoku_: i need to know why the anim continues Jan 28 07:48:12 i need to at least know the trigger Jan 28 07:48:24 and be able to reproduce Jan 28 07:48:58 AntonTakk: huh? Jan 28 07:49:04 raster: do you still have some gta02 lying around? Jan 28 07:49:17 yes Jan 28 07:49:21 but dead Jan 28 07:49:26 raster: well it seems to stop when we change the theme by the way Jan 28 07:49:31 oe dev build broken in 2.6.28 kernel Jan 28 07:49:33 and it's there every boot Jan 28 07:49:42 soemoen missing a space in the -m MODE option Jan 28 07:49:49 raster: well you could install SHR then Jan 28 07:50:01 Ainulindale: i need something i can build and test Jan 28 07:50:10 and that i can rebuild and add debugging to Jan 28 07:50:29 You can build and test SHR Jan 28 07:50:43 We're on the FSO branch of OE Jan 28 07:50:49 So no breakage Jan 28 07:51:08 it will take a lot of time to get that working Jan 28 07:51:33 Well Jan 28 07:51:46 We could give you access to the buildhost if that's necessary Jan 28 07:51:49 Would that be helpful? Jan 28 07:51:51 whats the exact branch? Jan 28 07:52:06 and is it hosted on oe's git? Jan 28 07:52:13 yep Jan 28 07:52:15 no Jan 28 07:52:18 i need it working here Jan 28 07:52:27 whats the branch then exactly? Jan 28 07:52:42 git-branch ? Jan 28 07:52:43 :) Jan 28 07:52:44 Looking for it Jan 28 07:53:43 damn can't see it locally Jan 28 07:53:48 mrmoku: didn't you change it? Jan 28 07:54:28 (and gitweb doesn't worK...) Jan 28 07:55:43 (thus you see my problems :)) Jan 28 07:55:55 by the time i ask 10 peolpe what branch it is.. exactly Jan 28 07:56:01 get 5 different answers Jan 28 07:56:07 try all of them Jan 28 07:56:09 fail Jan 28 07:56:14 finally get a 6th answer thats right Jan 28 07:56:19 and have sunk 2 weeks into it Jan 28 07:56:23 i might have a build Jan 28 07:56:25 :) Jan 28 07:56:44 git-branch -a should list them all Jan 28 07:57:20 any known breakage in fso testing? Jan 28 07:57:21 origin/fso/milestone5 Jan 28 07:57:25 is the likely candidate Jan 28 07:57:35 in the list i see Jan 28 07:58:14 Ainulindale: shr/oe-revision Jan 28 07:58:39 not on oe's dev git Jan 28 07:58:51 so where is it? Jan 28 07:58:52 :) Jan 28 07:58:53 raster: sorry, it's a file on the buildhost :-) Jan 28 07:59:00 the branch is named fso/milestone5 Jan 28 08:01:26 whats the image? Jan 28 08:02:36 fso-image ? Jan 28 08:02:42 fso-illume-image? Jan 28 08:02:43 heh, why is it the number of bugs left to fix in an fso milestone always jumps sharply just before a scheduled release? :) Jan 28 08:02:52 done see any shr stuff Jan 28 08:03:45 AntonTakk, thats because ppl start testing more Jan 28 08:04:07 yeah, makes sense Jan 28 08:04:39 freesmartphone.org: 03timo.lindfors 07framework * rc53125a8f181 10/framework/config.py: Jan 28 08:04:39 freesmartphone.org: config: do not send timestamps to syslog since syslog already adds time Jan 28 08:04:39 freesmartphone.org: This patch makes frameworkd not log timestamps when it is Jan 28 08:04:39 freesmartphone.org: logging to syslog. Syslog already adds timestamps so extra timestamp Jan 28 08:04:40 freesmartphone.org: is just making the log lines longer for no good reason. Jan 28 08:05:01 raster: sorry, I'm working I'm a bit busy, please ask to mrmoku :-) Jan 28 08:06:19 raster, in that branch we also reverted the fastboot stuff Jan 28 08:06:30 it broke everything except gta02 Jan 28 08:06:42 shoragan: the fastboot patches? Jan 28 08:06:51 That's weird we apply them on GTA01 and have no problem IIRC Jan 28 08:06:55 openmoko did some initscript optimizations Jan 28 08:06:58 hmm Jan 28 08:06:59 Yes I know Jan 28 08:07:03 stefan look at it Jan 28 08:07:04 But their distro conf was borked Jan 28 08:07:10 DISTRO_INITSCRIPTS = instead of ?= Jan 28 08:07:23 We corrected that and used our initscripts and it worked Jan 28 08:07:26 shoragan: which fastboot stuff? Jan 28 08:07:29 hmm Jan 28 08:07:30 (That or I'm hallucinating again) Jan 28 08:07:40 mrmoku: could you confirm? Jan 28 08:07:47 Ainulindale, is your git tree public? Jan 28 08:07:49 Yep Jan 28 08:07:54 git.shr-project.org Jan 28 08:08:13 But be wary Jan 28 08:08:24 The conf modification has been made dirtily on the buildhost :-) Jan 28 08:08:45 http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=366bf6043abfbbbfa98168548f42aa9aaf6c28b9 Jan 28 08:09:03 Well anyway, coffee break Jan 28 08:09:23 ah, you have an overlay and not a branch Jan 28 08:21:55 Ainulindale, ist there any specific reason to use an overlay instead of a git branch? Jan 28 08:22:18 otherwise i could simply merge some of your stuff into oe Jan 28 08:24:49 freesmartphone.org: 03jluebbe 07zhone * r55dddb8a9a58 10/src/zhone: correctly handle a missing "peer" field in the CallStatus message Jan 28 08:25:01 shoragan: We plan to do that Jan 28 08:25:06 shoragan: as soon as it's stable enough Jan 28 08:25:07 great :) Jan 28 08:25:50 morning Jan 28 08:27:34 freesmartphone.org: 03jluebbe 07zhone * r61b2a414faed 10/data/locale/ (ru/LC_MESSAGES/zhone.mo ru/LC_MESSAGES/zhone.po zhone.pot): run data/locale/update.sh Jan 28 08:31:35 shoragan: anyway I have commit access and I already talked about that with mickeyl so it won't be long Jan 28 08:31:45 ah, ok then :) Jan 28 08:34:29 raster: sorry, me has also to do some daywork :( Jan 28 08:35:13 The easiest way to build SHR is the Makefile from build.shr-project.org Jan 28 08:36:35 wget build.shr-project.org/Makefile Jan 28 08:36:38 make setup Jan 28 08:36:49 echo fso/milestone5 > shr/oe-revision Jan 28 08:36:50 oh ALSO an overlay Jan 28 08:36:52 make update Jan 28 08:36:59 raster: yep :-) Jan 28 08:37:08 thuis why i didnt see the image file Jan 28 08:37:09 cd shr-unstable Jan 28 08:37:10 make image Jan 28 08:37:33 raster: we're working to get the stuff into OE Jan 28 08:37:43 right now it's still overlay Jan 28 08:38:24 Qi doesn't yet support DFU right? Is there any other way to flash with Qi? Planned? Jan 28 08:38:48 Qi doesn't support usb at al Jan 28 08:39:43 shoragan: There's still NOR, I guess, but I was just wondering :) Jan 28 08:41:40 this.. will take a while Jan 28 08:42:07 raster: hehe, yep :-) Jan 28 08:45:58 Hej, I got trac magic powers. Cool. /me starts closing bugs :-) Jan 28 08:46:53 spaetz: yeah - clos em all ;) Jan 28 08:48:42 hehe Jan 28 08:49:50 mrmoku: nb Jan 28 08:49:55 echo origin/fso/milestone5 > shr/oe-revision Jan 28 08:49:58 wrong instructions Jan 28 08:49:59 :) Jan 28 08:50:10 raster: did no say origin ;) Jan 28 08:50:22 no Jan 28 08:50:25 thats what it SHOULD be Jan 28 08:50:31 7:41PM ~/shr > echo fso/milestone5 > shr/oe-revision Jan 28 08:50:31 ... Jan 28 08:50:37 raster, depends on if you have a lokal branch Jan 28 08:50:38 error: pathspec 'fso/milestone5' did not match any file(s) known to git. Jan 28 08:50:38 Did you forget to 'git add'? Jan 28 08:51:16 theres no shr-unstable Jan 28 08:51:25 just shr Jan 28 08:51:39 and no MAkefile in shr Jan 28 08:51:42 :) Jan 28 08:52:14 let me do this again wit the right branch Jan 28 08:52:26 i suspect the makefile fucked up as the branch didnt work Jan 28 08:53:04 mickey|zzZZzz: wake up. Is frameworkd goig to handle long (ie split SMS)? :) Jan 28 08:53:38 raster: no Makefile in shr is correct... no shr-unstable / shr-testing subdirs not so Jan 28 08:53:53 after make setup you should have them Jan 28 08:54:11 nup Jan 28 08:54:17 hmmmm Jan 28 08:54:54 spaetz, not in ogsmd Jan 28 08:55:28 because the parts could arrive over a longer time, so you need to keep a lot of state Jan 28 08:55:33 even over reboots Jan 28 09:00:46 raster: just tested... If I do 'wget build.shr-project.org/Makefile; make setup' Jan 28 09:00:53 i have all subdirs needed... ???? Jan 28 09:01:49 with origin you're right though... Jan 28 09:02:20 shoragan: ok, a trac ticket on trac.shr... suggests that this should be handled by frameworkd rather than by the sms showing app Jan 28 09:03:44 spaetz: what? Jan 28 09:04:32 Split SMSes? Jan 28 09:04:41 Ainulindale: Multi-part SMS... Jan 28 09:04:48 (how's your head btw.)? Jan 28 09:05:28 Better Jan 28 09:07:07 long sms'es split over several messages Jan 28 09:07:18 Just browsing through trac. Jan 28 09:07:50 How tricky is upgrading libglib-2.0-0_2.16.4-r1_armv4t.ipk to to glib2.0-2.16.5-1 (debian numbering)? Jan 28 09:08:02 would fix the reported midori crash bug Jan 28 09:37:28 has dos1 commit rights to [shr-overlay.git] / openembedded / packages / shr-settings / shr-settings_0.1.0_git.bb ? Jan 28 09:38:42 yes Jan 28 09:39:01 ok, I just assigned the ticket to him, so he can add python-wifi as a dependency Jan 28 09:53:45 mrmoku: are you "quickdev" or is that a different person? Jan 28 09:54:25 I guess he's a different person, though Jan 28 09:56:16 and why would anyone want to have the "old" original om2007.2 boot splash in SHR? Are there plans to use usplash? Jan 28 09:56:51 splash from 2007.2 whose nice Jan 28 09:57:33 I've never seen it. So the first issue is to use usplash at all before requesting a specific splash screen :) Jan 28 10:02:15 spaetz: hehe, no definitely not... we're quite different :D Jan 28 10:03:38 k, you just happened to comment on the same bugs, so I was a bit confused :) Jan 28 10:11:38 * PaulFertser_ can't modify ticket on FSO's trac due to some incompatibility with emacs-w3m :( Jan 28 10:12:34 Or he can, but the result is not obvious... Jan 28 10:20:36 raster: I see that you had a discussion about "wrap=word" vs "wrap=char" in edje TEXTBLOCKS earlier last year Jan 28 10:21:00 is there no way to make words longer than a line be "char" wrapped and the rest word wrapped? Jan 28 10:23:39 * mrmoku thinks that would be a sane default for wrapping, if wrapping is turned on :-) Jan 28 10:23:49 discussion was here. http://ibot.rikers.org/%23openmoko/20080430.html.gz and it ended basically with: Jan 28 10:24:21 raster: eah i dont have a fallback if it doesnt fit. its kind of an unhandled case. i need to set aside some time to go into the code again and find that case Jan 28 10:24:27 mrmoku: I agree Jan 28 10:25:05 Currently SMS with long words will simply go off-screen, it seems Jan 28 10:41:32 Ainulindale: do you know why quickdev does threading in the efl client? Jan 28 10:43:38 spaetz: btw. good work you're doing :-) Jan 28 10:47:35 mrmoku: just cleaning up a bit :) Jan 28 10:50:17 spaetz: yeah, but it's an important work... to not get lost :-) Jan 28 10:55:21 lindi-: i think you should use org.freesmartphone.ogsmd bus to call GSM methods. It was different in MS4. Jan 28 10:56:06 PaulFertser_: thanks! Jan 28 10:58:51 DocScrutinizer2: ping Jan 28 11:09:03 Is dialig #793# the same as sending USSD messages? Jan 28 11:09:14 I am a noob when it comes to GSM technicalities Jan 28 11:09:23 anyway, away for lunch now. Jan 28 11:16:54 * rwhitby wonders if the new cellinfo screen in fso can also recognise the cell broadcast for the name of the current locality Jan 28 11:17:04 s/fso/zhone/ Jan 28 11:17:16 rwhitby, no Jan 28 11:17:35 it's not using any cell broadcast Jan 28 11:17:47 haha :) eleGlade v0.7 release :) http://d3p.pl/fb/?&tag=eleGlade&id=363&subPage=detale Jan 28 11:18:14 shoragan: are there any framework interfaces for cell broadcast yet? Jan 28 11:18:38 http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.GSM.CB.html;hb=HEAD Jan 28 11:28:23 mrmoku: he has to Jan 28 11:28:34 mrmoku: because of the glib mainloop Jan 28 11:33:25 Has anyone build software with qtextended for neo? Jan 28 11:38:24 Ainulindale: on the other hand he has to take care to have all e stuff within one thread... Jan 28 11:51:12 mrmoku: even with origin/... still no shr-testing Jan 28 11:51:16 shr/ hae no Makefile Jan 28 11:51:19 something is up here Jan 28 11:52:01 spaetz: aaah textblock wrapping Jan 28 11:52:05 right now its very explicit Jan 28 11:52:15 if a word doesnt fit - it wont get wrapped Jan 28 11:52:19 min size will be the longest word Jan 28 11:52:25 wrods are stuff between whitespace Jan 28 11:52:40 so languages like chinese/japanese etc. that dont use spaces.. cant word-wrap Jan 28 11:52:47 they need to use char wrap specially Jan 28 12:13:15 raster: don't understand that... what are you doing? Jan 28 12:15:01 can someone add to python-elementary to ./elementary/elementary.c_elementary.pyx ~line 388 add property clicked: Jan 28 12:15:01 def __set__(self, value): Jan 28 12:15:01 self._callback_add("clicked",value) Jan 28 12:15:51 raster: shr-unstable/shr-testing is in parallel to shr Jan 28 12:15:54 10:50PM ~/shr > ls Jan 28 12:15:54 total 16K Jan 28 12:15:54 0 Makefile@ 4.0K bitbake/ 4.0K common/ 4.0K openembedded/ 4.0K shr/ Jan 28 12:17:02 hmm looks like you called make setup... the Makefile is a link (to common/Makefile me hopes) Jan 28 12:18:01 yup Jan 28 12:18:06 as you instructed Jan 28 12:18:06 :) Jan 28 12:18:34 32 wget http://build.shr-project.org/Makefile Jan 28 12:18:34 33 make setup Jan 28 12:18:34 34 echo origin/fso/milestone5 > shr/oe-revision Jan 28 12:18:34 35 make update Jan 28 12:18:54 (without origin/ the branch fetch fails) Jan 28 12:20:24 raster: heh, perfectly correct... and after make setup you should have those damn subdirs... Jan 28 12:21:19 raster: do you have some MAKEFLAGS set (to build in parallel)? Jan 28 12:22:00 yup Jan 28 12:22:05 (this is what makes *me* do that two times always :) Jan 28 12:22:06 so its not parallel safe? Jan 28 12:22:15 raster: not make setup Jan 28 12:22:49 * mrmoku wonders if he should stuff some unset MAKEFLAGS into the makefile... Jan 28 12:23:12 :) Jan 28 12:23:46 welcome to my environment Jan 28 12:23:51 it has lots of.. env vars Jan 28 12:23:51 I also have MAKEFLAGS set in my env... and *everytime* I do make setup I forget about it :-) Jan 28 12:23:51 :) Jan 28 12:24:12 so I always end up doing it two times Jan 28 12:24:12 at least with all my code... it works Jan 28 12:24:15 :) Jan 28 12:24:22 the makefiles are parallel-build safe Jan 28 12:24:22 :) Jan 28 12:25:04 for make setup there is not much one could do in parallel... most steps depend on the prior ones Jan 28 12:25:28 then the makefile should have the deps right :) Jan 28 12:25:33 anyway Jan 28 12:25:37 not to worry Jan 28 12:26:12 :-) (inherited that from the fso Makefile though ;) Jan 28 12:26:14 mrmoku: which bits of make setup are not parallel safe? Jan 28 12:26:23 the deps should serialise it all Jan 28 12:26:49 * mrmoku is taking a look Jan 28 12:27:12 * rwhitby never uses parallel, but is usually careful on target deps Jan 28 12:29:05 i abuse my cores Jan 28 12:29:35 raster: my build machine only has one Jan 28 12:30:33 rwhitby: setup depends on setup-shr _and_ setup-shr-unstable Jan 28 12:30:49 though the link for shr-unstable can be done after cloning the shr-overlay Jan 28 12:30:58 just checked the fso Makefile... its the same there Jan 28 12:31:22 s/can/must/ Jan 28 12:31:24 rwhitby: lazy core :) Jan 28 12:32:09 mrmoku: ah - I didn't add setup-shr ... Jan 28 12:32:34 what do you see in fso makefile? Jan 28 12:32:41 (nobody but me has touched that one) Jan 28 12:33:04 rwhitby: uuhh, now I see... yep, me is an idiot :-) Jan 28 12:33:12 rwhitby: wrong :-p Jan 28 12:33:19 rwhitby: I borked it once Jan 28 12:33:39 Ainulindale: not the one on downloads.freesmartphone.org Jan 28 12:33:52 Ah, indeed :-) Jan 28 12:34:02 But I thought it redownloaded the git one Jan 28 12:35:52 mrmoku: all the deps for shr-%/.configured on common/.git/config, bitbake/.svn/entries and openembedded/.git/config is what makes it parallel safe Jan 28 12:36:22 hhahaha multi objects eleGlade in one function :) http://d3p.pl/fb/?&tag=eleGlade&id=364&subPage=detale Jan 28 12:36:32 and it has a dep for shr/.git/config on that line Jan 28 12:36:47 so I don't understand what's making it break for you yet ... Jan 28 12:37:12 hi Jan 28 12:38:14 * mrmoku is trying make setup with the fso Makefile... to see if that works :-) Jan 28 12:39:40 mrmoku: I'm interested in any problems you find in it. Jan 28 12:43:02 has anybody tried google maps with the new tangogps? 0.9.5? Jan 28 12:44:21 rwhitby: I end up with: Jan 28 12:44:21 bitbake common Makefile openembedded Jan 28 12:44:28 when doing make setup Jan 28 12:44:41 with MAKEFLAGS=j8 Jan 28 12:44:49 there is no more option to check inverse Jan 28 12:45:00 * mrmoku wants his cores to glow ;) Jan 28 12:45:14 mrmoku: you have acces to 8core machine ? o.O Jan 28 12:45:35 yoyoek1: nop :-) Jan 28 12:45:44 just a stupid dual core Jan 28 12:45:45 cool Jan 28 12:46:08 mrmoku: I have quad core and on j4 it is unusable during compilation Jan 28 12:46:19 I/o wait is kiling all fun Jan 28 12:46:22 mrmoku: so the first line of setup-% does a recursive make which then depends on a .configured target which depends on all the other setup targets Jan 28 12:47:02 mrmoku: can you pastebin the log of the parallel make setup? Jan 28 12:48:08 rwhitby: http://pastebin.com/d5d4bcc5b Jan 28 12:48:16 brb Jan 28 12:50:40 mrmoku: I see the problem - seems that the .configured deps of the recursive make don't realise that the other parallel paths of the top-level make are doing the same thing, and they start clobbering each other. Jan 28 12:53:03 rwhitby: is it solvable? Jan 28 12:53:32 if not we could just unset MAKEFLAGS :-) Jan 28 12:54:35 mrmoku: try putting '-j 1' after ${MAKE} in the first line of setup-% and setup-fso-* targets Jan 28 12:55:08 (and anywhere else that a recursive call to ${MAKE} is done in a setup target dependency Jan 28 12:55:35 mrmoku: as per http://www.gnu.org/software/automake/manual/make/Options_002fRecursion.html#Options_002fRecursion Jan 28 13:01:34 rwhitby: http://pastebin.com/d4a7929eb Jan 28 13:02:00 think as soon it linked the Makefile to the downloaded one it bails Jan 28 13:03:18 what does # Jan 28 13:03:19 make[1]: Warnung: -jN in »make«-Verarbeitungszweig erzwungen: Jan 28 13:03:19 # Jan 28 13:03:19 Jobserver-Modus nicht verfügbar. mean in English? Jan 28 13:04:25 * mrmoku should have set locale to C :D Jan 28 13:06:38 make[1]: warning: -jN forced in submake: disabling jobserver mode. Jan 28 13:08:50 the problem is that the top-level make starts the setup-common target and the setup-fso-% target at the same time, and then the setup-fso-% target does a recursive make to a dep which also depends on setup-common, and it then tries to do the same git checkout but the common directory already exists since the very first setup-common target is still running but hasn't finished yet. Jan 28 13:10:07 according to the make manual, passing -j1 down should stop the descendent target from calling setup-common until the top-level one is finished. Jan 28 13:10:19 but it's obviously not doing that. Jan 28 13:12:20 rwhitby: after cloning common and linking the Makefile to it... my -j1 changes are gone :-) Jan 28 13:15:32 mrmoku: doesn't matter - the makefile is only read once at the top-level, so it was there for the recursive calls Jan 28 13:17:07 ahh, ok Jan 28 13:54:39 mmh, where is the code for the GSM (illume) Shelf Gadget? Jan 28 13:55:02 I'd look into fixing the "icon should indicate if GSM is disabled/SIM card missing" Jan 28 13:55:19 but I'd need a pointer to where that app is (and the source :-)) Jan 28 13:55:56 I thought it might have been: [shr.git] / openmoko-panel-gsm Jan 28 13:56:16 but that uses completely different icons, so it's unlikely Jan 28 13:57:57 ahh, googling leads to http://svn.projects.openmoko.org/svnroot/illume/src/e_mod_gad_gsm.c Jan 28 13:58:07 that might be it. Jan 28 13:58:22 spaetz: though it is in e svn I think Jan 28 14:02:25 ok, will have a look there. Still unfamiliar with what is where. Jan 28 14:02:28 thanks Jan 28 14:02:45 http://trac.enlightenment.org/e/browser/trunk/e/src/modules/illume/e_mod_gad_gsm.c Jan 28 14:05:05 freesmartphone.org: 03mickey 07framework * rc39b8e68712b 10/etc/freesmartphone/opim/csv-contacts.txt: add fso team as test data Jan 28 14:06:24 yep, think I found it. thanks Jan 28 14:07:01 the changed signal handler just asks for "org.freesmartphone.ogsmd", "/org/freesmartphone/GSM/Device", "org.freesmartphone.GSM.Network", Jan 28 14:07:24 but that call will probably fail if the GSM is turned off Jan 28 14:45:33 Qi for GTA02 can use initramfs? Jan 28 14:49:27 hm.. looks like it can't Jan 28 14:52:26 I don't think so hiciu. At least it states that it looks directly for the uImages Jan 28 14:53:00 I checked source, it can't Jan 28 14:53:22 hiciu: wasn't the idea that Qi could use initramfs to kexec the real kernel? Jan 28 14:53:50 PaulFertser: ? Jan 28 14:55:23 I don't know. I see there is part for loading initramfs, but there is no hardcoded path to initramfs (only in code for GTA03, in comments ;]) Jan 28 14:56:07 freesmartphone.org: 03mickey 07framework * rae581fa712cb 10/framework/subsystems/opimd/ (5 files): opimd: raise proper dbus errors, remove some redundancy (more to come), styling fixes Jan 28 14:56:19 hiciu: < PaulFertser> lindi-: you either have to hack on u-boot to add support for larger images. Or (much better!) you can dump u-boot altogether, prepare an initramfs you like (that kexec's kernel fetched over tftp or whatever) and append it to any working kernel. Flash Qi and the kernel and you're all set. Much more flexible. Jan 28 14:57:21 ok, thanks Jan 28 15:11:23 what value shoud by set for bubble in elementary in corner_set ? Jan 28 15:54:19 Ainulindale: sorry you'll have gotten a whole bunch of trac emails today. I won't be that active every day :-). Jan 28 15:56:07 hiciu: Do you know that initramfs can be easily appended to the kernel image? Jan 28 16:01:32 PaulFertser: ah, that needs quite special distro support? Jan 28 16:02:49 lindi-: i don't think so. During the kernel configuration a path to initramfs image can be specified. The kernel will use automatically. As far as i understand. I'll look for the specific config variable now. Jan 28 16:05:06 lindi-: The 2.6 kernel build process always creates a gzipped cpio format initramfs archive and links it into the resulting kernel binary. Look for CONFIG_INITRAMFS_SOURCE in linux-2.6/Documentation/filesystems/ramfs-rootfs-initramfs.txt Jan 28 16:06:15 PaulFertser: but initramfs is generated when the kernel package is installed Jan 28 16:06:22 PaulFertser: not when the kernel package is built Jan 28 16:07:43 lindi-: i think that one possibility (that's employed by most distros) is to store initramfs (with necessary modules and stuff) in a file and configure a bootloader to use it. And i think that this same image can be integrated in the kernel image itself, using CONFIG_INITRAMFS_SOURCE. Jan 28 16:08:53 PaulFertser: yes but can it be integrated without having to recompile the kernel? Jan 28 16:09:54 lindi-: i think some hack can be used to do that without recompiling but i see no reason why wouldn't initramfs developer compile his own kernel. Jan 28 16:17:21 PaulFertser: developer surely but of course we want a solution that works for everyone Jan 28 16:19:21 lindi-: I think that can easily be handled by the distribution's autobuilders, that is, to build several kernels with different variants of initramfs. Jan 28 16:22:58 lindi-: And Qi seem to support initramfs in external file for every device, just the functionality is not used for GTA0[12]. Jan 28 16:23:25 Greetings Paul Jan 28 16:24:52 Dave: Hello :) Jan 28 16:25:41 sean morphisa ? Jan 28 16:27:57 Let the Qi be with you... Jan 28 16:30:09 have someone seen morphis? Jan 28 16:36:34 hi, all.. its been a long long while since I had taken up my Freerunner, so missed lots of fun. So, in SHR unstable from 23rd Jan. mofi segfaults saying eth0 couldn't support scanning. Some known kernel issue? Jan 28 16:38:47 lindi-: after looking at System.map i think that in-kernel initramfs can be changed easily without recompiling, but it's impossible to do in a sane way without relinking :) Jan 28 16:39:39 PaulFertser: but initramfs _needs_ to be generated at install time Jan 28 16:40:02 PaulFertser: since end-user can configure it via /etc/initramfs-tools/initramfs.conf Jan 28 16:41:46 lindi-: it's the case when the idea is to support different hardware targets (to include only necessary modules, e.g.). But if someone writes a cool initramfs with menu it will work on any GTA02 out there, no need for user-configurability. Jan 28 16:43:10 PaulFertser: but the whole point of initramfs would be user-configurability :) Jan 28 16:43:24 PaulFertser: maybe you could have two initramfs's Jan 28 16:43:33 PaulFertser: one for menu and one for the actual distro Jan 28 16:44:43 lindi-: yes, you can. The files from external initramfs are combined with in-kernel initramfs. And Qi supports external initramfs's (i didn't try, but the code is there). Jan 28 16:46:00 lindi-: or one can design an initramfs that will read its parameters from some predefined location (SD card or NAND), that will allow user-configurability without recompiling. Jan 28 16:46:25 PaulFertser: debian also lets packages register hooks to the initramfs Jan 28 16:47:09 PaulFertser: for example i made one package that lets you have ssh access to the system while it is still running fsck Jan 28 16:47:38 PaulFertser: (it runs separate dropbear instance in a tmpfs and can operate even if the system hasn't managed to mount rootfs yet) Jan 28 16:48:01 lindi-: I think the bootloader is not the problem here, it allows to hook up an external initramfs (just no specific configuration for GTA02 yet). So, there's plenty of possibilities but all those people who talk about menu doesn't seem to do anything. Jan 28 16:48:48 PaulFertser: yeah at least i'm too busy right now :/ Jan 28 16:49:29 lindi-: i know, good luck with your tasks :) Jan 28 17:03:58 someone here who knows git a bit? Jan 28 17:04:44 I want to checkout andy-tracking to do modifications, but clone + checkout origin/andy-tracking creates an unnamed branch Jan 28 17:05:04 how do i make my local master tracking origin/andy-tracking? Jan 28 17:05:19 sleipnir: git checkout -b andy-tracking origin/andy-tracking Jan 28 17:05:37 creates a new local andy-tracking branch that tracks andy-tracking :-) Jan 28 17:05:50 hmm i did this, then branched from andy-tracking and did some modifications Jan 28 17:06:10 however i could not pull from upstream because i had local modifications Jan 28 17:06:19 so what is the point here? Jan 28 17:06:20 no need to branch from andy-tracking... its *your* local andy-tracking Jan 28 17:06:42 Hire: grown up... finally :P Jan 28 17:07:09 mrmoku: yeah i know, but i want to modify some sources and rebase them on andy-tracking from time to time Jan 28 17:07:26 yeah Jan 28 17:07:29 I think Jan 28 17:08:08 sleipnir: so you want both... andy-tracking and sleipnir-andy-tracking... Jan 28 17:08:17 just create it the same way Jan 28 17:08:17 exactly Jan 28 17:08:30 git checkout -b sleipnir andy-tracking Jan 28 17:08:55 and then from time to time... git rebase andy-tracking (while your on the sleipnir branch) Jan 28 17:09:01 mrmoku: do you mind in giving a step-by-step guide? Jan 28 17:09:13 mrmoku: git clone git://openmoko... Jan 28 17:09:24 yep, thats step #1 Jan 28 17:09:30 mrmoku: git checkout --track -b andy-tracking origin/andy-tracking Jan 28 17:09:43 yep, #2 Jan 28 17:09:43 mrmoku: git branch mytree Jan 28 17:09:55 mrmoku: git checkout mytree Jan 28 17:10:06 yep Jan 28 17:10:17 mrmoku: modify code (meanwhile andy-tracking changed) Jan 28 17:10:24 git checkout andy-tracking Jan 28 17:10:28 git pull Jan 28 17:10:32 and error Jan 28 17:11:02 hmm, when you're on andy-tracking and didn't modify it, it should not give you an error Jan 28 17:11:06 what error is it? Jan 28 17:11:16 http://pastebin.com/m6927eb4f Jan 28 17:11:52 error: Entry 'arch/arm/mach-s3c2440/Kconfig' not uptodate. Cannot merge. Jan 28 17:11:58 says it all :-) Jan 28 17:12:04 Kconfig is locally modified Jan 28 17:12:16 yes, that's what my branch is about ;-) Jan 28 17:12:36 sleipnir: you have to commit your changes into your branch Jan 28 17:12:42 mrmoku: how can i resolve this (sorry i'm a complete git noob) Jan 28 17:12:54 git checkout mytree Jan 28 17:12:57 ahhh now i see Jan 28 17:13:01 git gui (commit the stuff) Jan 28 17:13:02 then git add Jan 28 17:13:05 git checkout andy-tracking Jan 28 17:13:10 git commit Jan 28 17:13:17 and then update andy-tracking Jan 28 17:13:21 sleipnir: I strongly recommend git gui... its very nice for comitting Jan 28 17:13:25 +m Jan 28 17:13:29 yep Jan 28 17:13:32 ahh cool Jan 28 17:13:45 will try this out... brb Jan 28 17:13:47 yeah, git is *cool* :D Jan 28 17:19:57 Dave: hey, how's Barakomania coming along :D Jan 28 17:24:36 mrmoku: seems to work... do i also need to pull master when updating andy-tracking? Jan 28 17:25:47 sleipnir: only if you're interested in it... git pull fetches all new stuff into all remote branches (by default config) Jan 28 17:26:02 if you checkout a branch that is behind the remote branch git will tell you so Jan 28 17:26:54 mrmoku: ok. now what remains to be done is git checkout build-fixes; git rebase andy-tracking Jan 28 17:26:59 mrmoku: is this correct? Jan 28 17:27:03 yep Jan 28 17:27:27 depending on what changed in andy-tracking this could get worksome ;) Jan 28 17:27:43 but if you do just small changes it should just rebase me thinks Jan 28 17:27:52 mrmoku: it's only 4 small changes Jan 28 17:28:08 mrmoku: will give it a try... Jan 28 17:28:25 ok, have to go for an hour now... bbl Jan 28 17:28:44 mrmoku: thx! for your help! Jan 28 17:28:50 ainulindale: how much time do you spend for your ibanez? Jan 28 17:31:48 sleipnir: np... i love git :-) Jan 28 17:31:55 urgs Jan 28 17:32:03 one additional thing Jan 28 17:32:16 how do i see my patches after the rebase? Jan 28 17:32:26 git diff andy-tracking gives nothing Jan 28 17:32:48 one nice thing to see what's going on is 'gitk --all' Jan 28 17:34:17 mrmoku: hmm have git without tcl/tk :-( Jan 28 17:34:43 uuhh, I could not survive without gitk... Jan 28 17:35:11 git diff andy-tracking..mytree ? Jan 28 17:35:43 shows nothing Jan 28 17:36:07 sleipnir: maybe you're fixes are in andy-tracking also? Jan 28 17:36:32 s/you're/your/ Jan 28 17:37:14 but it tells me: Your branch is ahead of 'origin/andy-tracking' by 3 commits. Jan 28 17:37:36 git diff origin/andy-tracking..mytree Jan 28 17:38:11 sleipnir: really have to go now... pick up my son... bbl Jan 28 17:38:17 kk Jan 28 17:38:52 that did the trick Jan 28 17:38:54 thx again Jan 28 17:45:23 any suggestions for a client to use gmail on the moko with? Jan 28 17:45:59 mrmoku, annoying Jan 28 17:46:17 mrmoku, ironically however, some of you foreigners seem more enthused than us :P Jan 28 17:46:25 In any case, it's good, but annoying. :D Jan 28 17:46:34 Overzealous greenpeace hippies. Jan 28 17:54:27 oops Jan 28 18:01:54 I create folder on desktop using right button and contekst menu Jan 28 18:02:10 now elementary is crashin all time how can I delete the folder ? Jan 28 18:02:15 awesome Jan 28 18:04:30 brbrbrbrbrb halo Jan 28 18:09:15 hmm, anyone have any idea about how long it takes to build shr? Jan 28 18:09:25 FOREVER Jan 28 18:09:31 Thanks Dave! Jan 28 18:09:38 is that a long time? Jan 28 18:10:05 2 times bigger than VERY LONG Jan 28 18:10:22 so about 1/2 a parsec then? Jan 28 18:16:16 :) Jan 28 18:39:27 toggles_w: it took my good old pc 38hrs (celeron 500, 256mb RAM) Jan 28 18:41:51 wafl Jan 28 18:44:18 sicu: damn... Jan 28 18:44:43 so what you guys are saying is i should just download it ;-) Jan 28 18:45:46 duh :P Jan 28 18:45:56 * Dave pets his dual quad-core... Jan 28 18:46:15 sicu: eat my digital dust ;p Jan 28 18:46:26 man, i bet gentoo installs quick on that thing Jan 28 18:47:05 fuck gentoo? :) Jan 28 18:47:16 but yes :p Jan 28 18:47:33 i thought human penguin interaction was ilegal? Jan 28 18:49:16 i think it takes just as long to dfu shr as to build it. Jan 28 18:51:01 hahah Jan 28 18:51:22 * toggles_w twiddles thumbs Jan 28 18:51:37 toggles, go dfu yourself :P Jan 28 18:51:46 haha, I like how that sounds wrong. Jan 28 18:51:56 i haven't had a working phone for 24hrs... damn droiddidn't live upto expectations Jan 28 18:52:00 lol Jan 28 18:52:19 HEY, dont' blame me! Jan 28 18:52:21 ¬_¬ Jan 28 18:52:27 * Dave glares at toggles Jan 28 18:52:28 it was your keyboard! Jan 28 18:52:37 * Dave beats toggles Jan 28 19:06:33 I've compiled rootfs "from scratch" with melas, but when I want to boot it kernel says: "Failed to execute /sbin/init" Jan 28 19:07:19 I chrooted myself to my rootfs from other distro and everything works as it should.. Jan 28 19:10:33 rootfs and kernel must be built with one toolchain? Jan 28 19:11:23 hiciu: they should use the same ABI and be compiled for armv4. Jan 28 19:12:01 uh.. so I should rebuild kernel too.. Jan 28 19:13:50 hiciu: But how did you build the kernel? For armv5 or what? Building kernel doesn't require even arm4 libgcc. Jan 28 19:14:41 PaulFertser: I used precompiled kernel with my rootfs Jan 28 19:14:57 hiciu: in fact, i compiled the kernel with "armv5" toolchain and it worked, there's no reason why it shouldn't Jan 28 19:15:07 hiciu: so, no, i don't think you need to rebuild the kernel. Jan 28 19:15:34 hiciu: how do you compile your rootfs? Jan 28 19:15:46 PaulFertser: with "Melas" tool Jan 28 19:15:56 hiciu: what's that? Jan 28 19:15:59 PaulFertser: I'm trying to build this: http://www.igelle.org/ Jan 28 19:16:21 PaulFertser: Melas is here: http://melas.ossproject.org/ Jan 28 19:17:43 hiciu: are you sure you have /sbin/init in your rootfs? Jan 28 19:18:05 yes :) Jan 28 19:18:19 :) Jan 28 19:18:27 * Dave pokes Hire Jan 28 19:20:43 hiciu: then i think either you've incorrectly transfered rootfs to device or used some wrong kernel parameters. How do you boot, where's your rootfs? Jan 28 19:21:54 PaulFertser: my current rootfs is on first partition on sd card, I booted it with 'root=/dev/mmcblk0p1 rootdelay=2 loglevel=0 init=/sbin/init' Jan 28 19:22:23 it is here: http://hiciu.rootnode.net/download/roofs-with-upstart.tar.gz Jan 28 19:22:38 hiciu: are you using Qi? Jan 28 19:22:42 yes Jan 28 19:23:06 hiciu: then add a space on the string after /sbin/init Jan 28 19:23:18 Hey Jan 28 19:23:23 DocScrutinizer2, here? :) Jan 28 19:23:27 hiciu: It's possible that the newline will be appended to the filename and the kernel won't recognize it. Jan 28 19:23:39 I will try.. Jan 28 19:24:06 quickdev: a minute Jan 28 19:24:19 hiciu: why do you think init= parameter is needed at all? Jan 28 19:24:26 DocScrutinizer2, in a minute? or only a minute left? :) Jan 28 19:24:29 hiciu: why do you specify loglevel=0? Jan 28 19:24:43 only 50sec left ;D Jan 28 19:24:52 oh, I mean loglevel=9 Jan 28 19:25:13 whatsup? Jan 28 19:25:57 DocScrutinizer2, then I'd appreciate if you can give me a hint at some other time....I'd like to connect the neo's aux button to 12 V trigger....which is triggered in the home of my grandparents if there's emergency - thus the neo should call some ppls :) Jan 28 19:26:24 no problem Jan 28 19:26:24 for you it should be very easy - just a transistor or something like that :) Jan 28 19:27:03 ok, looking forward to :) Jan 28 19:27:03 I'll think about it and come to you with some suggestions eventually Jan 28 19:28:29 quickdev: first question: why AUX? isn't HOLD more convenient?! Jan 28 19:28:38 PaulFertser: thanks, it works now! :) Jan 28 19:28:47 hiciu: what was wrong then? Jan 28 19:28:47 DocScrutinizer2, it's very simple. Actually it's a wireless ringing-system. People can carry a wireless trigger (sender) and use it if there's emergency. The receiver then emits a loud tone. Additionally it's possible to connect another ring-system - that's where I'd like to connect the neo :) Jan 28 19:29:10 PaulFertser: I had to add a space after init=/sbin/init :) Jan 28 19:29:24 DocScrutinizer2, if AUX or not....what I'd need is a way to execute something, if the trigger is there Jan 28 19:29:45 I thought AUX is the easiest one to connect with two cables? Jan 28 19:30:02 hiciu: somebody's gotta fix this bug :-/ Jan 28 19:30:29 quickdev: toss me pointer to datasheet or electrical spec of that "12V" ringer output Jan 28 19:30:59 wtf? Jan 28 19:31:11 quickdev: no, HOLD is. not even open case of FR :) Jan 28 19:31:21 yay Jan 28 19:31:38 DocScrutinizer2, where is HOLD? Jan 28 19:31:45 headset Jan 28 19:31:55 ah, great! :) Jan 28 19:32:00 didn't even know about it Jan 28 19:32:03 What kind of naughtiness are you attempting to implement here, quickdev? :P Jan 28 19:32:18 quote me occasionally ;-) Jan 28 19:32:55 sth like "(C) jOERG" ;D Jan 28 19:33:49 Dave: big red button (for Obama) XD Jan 28 19:34:32 DocScrutinizer2, I'm going to quote you ;) currently uploading the spec Jan 28 19:34:57 heheh okay! :) Jan 28 19:35:19 DocScrutinizer2: Hello :) can i have your opinion about irq line from calypso? I suspect something wrong is going there. I've seen my FR not resume from calling it, though calypso was obviously alive (i've heard calling tones, not "out of coverage"). And it resumes every time with POWER button or usb insert. Jan 28 19:36:42 PaulFertser: yup, that's either IRQ line from calypso not asserted, or CPU discarding it Jan 28 19:37:15 DocScrutinizer2: i guess there's no testpoint or other way to assert it manually when it happens, is there? Jan 28 19:37:25 or FR hanging *after* accepting IRQ, somewhere in suspend Jan 28 19:37:45 DocScrutinizer2: and i don't have a slightest idea when calypso is supposed to assert IRQ. Jan 28 19:37:45 DocScrutinizer2, http://img515.imageshack.us/img515/496/dsc00977ad9.jpg Jan 28 19:37:53 no way to look up this right now, sorry Jan 28 19:38:19 quickdev: send me a mail please Jan 28 19:38:23 DocScrutinizer2: it was resumed by power button press, so that's not hang on resume. Jan 28 19:38:39 PaulFertser: depends Jan 28 19:38:51 DocScrutinizer2, joerg@openmoko.org ? Jan 28 19:39:08 it might hang in servicing gsm-uart e.g. Jan 28 19:39:16 quickdev: yup Jan 28 19:39:50 DocScrutinizer2: after resume by power button it started to ring and so on, so gsm-uart communication was working. Jan 28 19:40:25 hmm, I have wrap my head around this Jan 28 19:40:42 http://i40.tinypic.com/53w7dv.jpg Jan 28 19:42:19 quickdev: are you sure that your photo shows a connection to external ringer, not to external button? The second image shows single push-button connected to device, doesn't make much sense with your description. Jan 28 19:42:31 DocScrutinizer2, just saw that it was the spec of the sender - I need the spec of the receiver. I'm going to investigate it before mailing you.. Jan 28 19:42:49 quickdev: ha, and i was wondering :) Jan 28 19:43:05 PaulFertser, you're right, my fault Jan 28 19:43:06 :) Jan 28 19:44:44 Quickdev, what are you making? Jan 28 19:44:51 a door chime? Jan 28 19:47:11 Anybody want to get me a job in Netherlands, Belgium or Switzerland? Jan 28 19:47:24 (Doing Java or C/C++ development) Jan 28 19:48:35 wurp2: How's your iApplication? Jan 28 19:48:49 Dave, read backlog :) Jan 28 19:49:17 WURP Jan 28 19:49:28 Wurp, how desperate are you? :P Jan 28 19:49:32 :O Jan 28 19:49:39 Well, just the fact you're asking here says enough, I think :p Jan 28 19:49:40 hah Jan 28 19:51:28 Dave: :x Jan 28 19:52:10 :) Jan 28 19:55:24 Joerg, I'm very pleased with some of my fun and radical designs for case mods Jan 28 19:55:32 and for my adapted SD card reader Jan 28 19:55:44 we'll see, I'm going to try to send you postage :) Jan 28 19:55:52 PaulFertser: Going great Jan 28 19:56:02 I am waiting for the updated icon, then submiting Jan 28 19:56:05 iApplication?? Jan 28 19:56:24 Dave: Well, last time I applied for work in the UK, I sent 30 resumes and got no bites Jan 28 19:56:28 Dave: Yeah, I'm whoring Jan 28 19:56:33 ouch Jan 28 19:56:46 wurp, you defecting? :[ Jan 28 19:56:48 Ouch will be if I don't make money on it :-) Jan 28 19:56:55 I don't even own an iPhone, actually Jan 28 19:57:11 But I have a MacBook with a simulator, and another guy on the project has an iPhone Jan 28 19:57:54 Anyway, you guys are the EU peeps I know (other than a Swedish friend I need to hit up), so I'm fishing here :-) Jan 28 19:57:57 haha i'm the cooler version Jan 28 19:59:23 hahaha ... a new Dave version Jan 28 19:59:44 * iDave is much hipper than the regular Jan 28 19:59:56 <\marco> hi to all! Jan 28 20:00:25 iHateIcrap Jan 28 20:00:28 <\marco> An info...where's the sources of elementary-test? Jan 28 20:00:49 <\marco> I'm looking inside shr-image but i can't find it.. :/ Jan 28 20:01:42 Anyway, I have lots of bites in the US, but I want to move to NL, BE or CH, so hence the asking here Jan 28 20:02:49 iDave: Nobody is as cool as I am with my iMoko :P Jan 28 20:03:39 <\marco> ehm.. sorry I've found it! ^^ Jan 28 20:09:11 Fishing unsuccessfully Jan 28 20:09:46 DocScrutinizer2: Is Y-cable your invention? Were you the first to actually try it? Jan 28 20:10:14 PaulFertser: yes. no. Jan 28 20:11:23 Ainulindale: look at scap Jan 28 20:12:00 Ainulindale: should i commit this loading screen to illume-shr theme? Jan 28 20:12:36 Yep ! Jan 28 20:12:43 Please do Jan 28 20:14:16 Ainulindale: ok ;D Jan 28 20:14:53 Ainulindale: i can commit only from freerunner, but i will try ;) Jan 28 20:22:26 DocScrutinizer2: mentioned the Y-cable inventor on the wiki. BTW, i'm going to build one myself, but it'd be cool to find an elegant solution to be able to turn of the R's, so the same cable can be used in both powered and unpowered modes. Jan 28 20:22:31 though i'm sure Ainulindale would like to spiff it up a notch with some bright sparkling colours ;] Jan 28 20:23:53 iLoader Jan 28 20:25:26 PaulFertser: you mean USB-host and USB-device (logical). R's aren't related to the power. Yeah me too I wondered for a long time if there's an easy way to implement this little extra functionality. No results though :( So I'd dare to assume it's just impossible without adding active components Jan 28 20:27:37 DocScrutinizer2: did you notice my comment on the audio recording issue? (capturehandset.state works but I want to record from mic, not GSM) Jan 28 20:28:11 jOERG_zzZZzz: no, i meant Y-cable is designed to allow FR to act as a host while still be able to charge. As far as i understand R's are needed because when it doesn't provide power over usb, the built in R's are deactivated. Do i miss anything? I'm thinking about using Y-cable even when external power is not available. Jan 28 20:28:37 PaulFertser: modifying a hub sounds like a more compact solution btw Jan 28 20:29:09 lindi-: how a hub could be more compact than Y-cable? Jan 28 20:29:13 lindi-: duh, no. But I also dn't exactly understand your point. capturehandset.state *should* do *both* Jan 28 20:29:47 jOERG_zzZZzz: i just get 0x80 0x80 0x80 ... no matter how loudly i speak to the mic Jan 28 20:30:06 lindi-: PaulFertser : well modding a hub really is somewhat more self-comprised Jan 28 20:31:06 jOERG_zzZZzz: what selects whether it is recording from gsm or mic? isn't that one of the alsa controls and the state has to choose which one? Jan 28 20:31:28 lindi-: wait a few hours, or check the backlog of #openmoko. I did records with a capturehandset.state file via ssh even. it worked Jan 28 20:32:02 DocScrutinizer: i'll read the backlog. i read the ssh comment already though Jan 28 20:32:19 lindi-: I'd expect it to record inbound audio one channel, and outbound the other channel Jan 28 20:32:32 sounds cool Jan 28 20:33:02 Jan 28 22:33:46 ginger user.debug ogsmd.channel DEBUG : COMPLETED 'ATA' => ['ERROR'] Jan 28 20:33:14 nota bene I never checked this. I was happy with recording mic andpiping the raw audio out to a usb-host to listen to it Jan 28 20:33:50 (completely unrelated bug, i need to track it down first.. the syslog/strace/kernel ring buffer has limited size) Jan 28 20:34:16 (thus eliminating the annoying gsm-network influence on test results ;) Jan 28 20:34:35 Lots of Joergs here Jan 28 20:34:54 and Daves Jan 28 20:34:57 lindi-: lol, isn't that the definition of a ringbufer? Jan 28 20:35:11 DocScrutinizer: yep :) Jan 28 20:35:36 Dave: *no* joerg around here ;D Jan 28 20:35:37 * Dave ships iDave off to Soviet Russia Jan 28 20:35:44 DocScrutinizer :P Jan 28 20:36:23 Toggles: In soviet russia, jobs... just bite :p Jan 28 20:36:40 the same is true in soviet pennsylvania Jan 28 20:37:26 or the little US village Russia" Jan 28 20:37:56 mickey|zzZZzz: 'ATA' => ['+CRING: VOICE', 'OK'] sounds like an AT parser bug, right? Jan 28 20:39:14 haha Jan 28 20:39:22 Toggles, I was just there. Jan 28 20:39:34 Fireworks were cheap :) Jan 28 20:43:04 and ilegal!! yay! Jan 28 20:43:14 hardly Jan 28 20:43:24 in pa they are Jan 28 20:43:30 The sale of is not prohibited, anyway. Jan 28 20:43:51 Using them (without a special PA permit), on the other hand, may be :) Jan 28 20:44:06 hmm.. maybe county by county, at least here they are Jan 28 20:44:21 well except for the little tinyones you get in the grocery Jan 28 20:44:27 haha hell no Jan 28 20:44:43 I just crossed the boarder on 78 and the first things were about seven major warehouses of them Jan 28 20:44:46 wholesale fireworks Jan 28 20:44:48 it's big business :P Jan 28 20:44:57 people come from out-of-state to pick them up, of course Jan 28 20:45:05 since it's illegal here in NY/NJ Jan 28 20:45:06 well things musta chnaged, we used to go to ohio to get them Jan 28 20:45:11 hahaha yeah I know :P Jan 28 20:45:16 I remember doing that myself :D Jan 28 20:45:38 Toggles, from what they told me, the sale of is legal, but you're not allowed to set them off in PA *wink-wink* Jan 28 20:45:48 hmm Jan 28 20:45:50 The whole thing is shifty (a grey area) Jan 28 20:45:55 i am, i do it professionally ;-) Jan 28 20:46:01 Apparently, however, people with special permits can set them off legally Jan 28 20:46:25 So it seemed like if you apply for a permit you can do so legally in PA (you must be a resident, however) -- makes sense Jan 28 20:46:29 reminded me of similar gun laws Jan 28 20:46:38 which PA is also much more laxed about Jan 28 20:46:42 compared to the tri-state area :D Jan 28 20:47:39 weird how that happens Jan 28 20:47:56 but similarly, I prefer to pick them up off the streets Jan 28 20:48:04 makes the whole process a lot easier on the consumer :P Jan 28 20:48:07 little bitty baby ones ;-) Jan 28 20:48:27 so my app works fine, even causes a re-registration if zhone is already registered to the gsm network Jan 28 20:48:40 Anton, sneaky. Jan 28 20:48:43 I hope it doesn't crash! Jan 28 20:50:10 is there something else that needs to be done for gsm registration than request gsm, setantennapower true, and register (with no options) ? Jan 28 20:50:41 Anton where does your setup stand with respect to #1024? :x Jan 28 20:51:00 dunno, what is #1024? Jan 28 20:51:05 ¬_¬ Jan 28 20:51:08 oh geeze Jan 28 20:51:12 poor, poor man Jan 28 20:51:14 s^10 it is Jan 28 20:51:18 2 Jan 28 20:51:18 or lucky, depending on how you look at it :D Jan 28 20:51:32 ? Jan 28 20:52:07 AntonTakk1: one of our favorite numbers Jan 28 20:52:15 DocScrutinizer: ok, http://trac.freesmartphone.org/ticket/334 reported. now i can play with arecord again. Jan 28 20:52:33 DocScrutinizer2: Is Y-cable your invention? --> http://lists.openmoko.org/pipermail/community/2008-March/013978.html Jan 28 20:52:42 DocScrutinizer: after 'alsactl -f capturehandset.state restore' the command 'arecord | hd' really returns just 80 80 80 ... Jan 28 20:52:50 Blueslee! Jan 28 20:53:11 Dave!! Jan 28 20:53:18 Hallo :D Jan 28 20:53:18 so what does #1024 do? Jan 28 20:53:30 uggh Jan 28 20:53:33 IT MAKES ME WANT TO KILL PEOPLE Jan 28 20:53:38 * Dave sits down Jan 28 20:53:39 AntonTakk1: bugs.openmoko.org please Jan 28 20:53:50 lindi-: which capturehandset.state U r using? Jan 28 20:54:01 DocScrutinizer: 'b6deeb8f362cdce1ec5eaadbd0a471d8 capturehandset.state' Jan 28 20:54:18 DocScrutinizer: it came from FSO distro to debian i think Jan 28 20:54:29 DocScrutinizer: do you have a different file? Jan 28 20:54:30 hmmm, sorry I'm a n00b, don't undertsand this Jan 28 20:54:59 DocScrutinizer: and that was md5sum, not git revision :P Jan 28 20:55:12 Behold! http://docs.openmoko.org/trac/ticket/1024 Jan 28 20:55:20 Not so sure you want to look at it, Anton :) Jan 28 20:55:21 DocScrutinizer: can you put your capturehandset.state online? Jan 28 20:55:27 If you have no need, and want to keep your sanity :p Jan 28 20:55:34 That reminds me, I should check Wester's site Jan 28 20:55:51 Dave: i can't answer your question w/olooking at it Jan 28 20:56:28 Don't bother then if it works fine :) Jan 28 20:56:39 #1024 is the de/re-registration issue Jan 28 20:56:46 'deep sleep' gsm vooodoo Jan 28 20:57:12 Dave: i was not here, are there any news concerning #1024 or why it came up? Jan 28 20:57:31 Blueslee, I just was curious about anton's app and so mentioned it Jan 28 20:57:34 No new news, no Jan 28 20:57:39 Maybe I shouldn't have mentioned it :) Jan 28 20:57:45 I don't want to start a holy war :) Jan 28 20:58:09 AntonTakk1: what kind of app? Jan 28 20:58:15 lindi-: see capturehandset_GTA02_080730.state in people/joerg Jan 28 20:58:39 Oh, joerg I had a question regarding the gsm fw Jan 28 20:58:47 BluesLee: combined today/calendar/sms/contacts/dialer Jan 28 20:58:53 DocScrutinizer: it seems to be identical to mine Jan 28 20:59:17 DocScrutinizer: does 'arecord | hexdump -C' really produce something else than a stream of 80 80 80 for you? Jan 28 21:00:34 Dave: seems like a problem in gsmd or lower, how could my app have anything to do with it? Jan 28 21:01:10 AntonTakk1: really? screenshots? Jan 28 21:01:10 lindi-: hmmm, maybe I tweaked sth with alsamixer. sth like "DAI mode" comes to mind Jan 28 21:01:21 DocScrutinizer: hmm Jan 28 21:01:42 DocScrutinizer: any details would be appreciated ;) Jan 28 21:01:42 Anton vs opimd? Jan 28 21:01:55 AntonTakk1: this one http://scap.linuxtogo.org/files/de49f0af695190a37b49e60ec5cb1e26.png Jan 28 21:02:01 you see the date in filename? ;-) Jan 28 21:02:04 BluesLee: i have been thinking of making a small demo vid, but i don't really have any screenies yet Jan 28 21:03:22 finish the app first Jan 28 21:03:24 :p Jan 28 21:03:24 Dave: that was the (new) plan Jan 28 21:03:31 lindi-: that's why I asked you to wait a few hours. I'm just about to test all this again, need it for buzz rework QA Jan 28 21:03:34 (opimd) Jan 28 21:03:43 Ah, good :) Jan 28 21:03:48 DocScrutinizer: ok :) Jan 28 21:04:09 DocScrutinizer: i can record audio if i start arecord first and only then switch to voice-recording.state Jan 28 21:05:32 Dave: currently about two seconds fromtyping the command and hitting enter till the gui ison screen and available :) Jan 28 21:06:00 YAY Jan 28 21:06:10 Now just let us all into your system... so we can play with it :P Jan 28 21:06:19 hehe Jan 28 21:06:45 i can drop pre-compiled binaries on my web server :) Jan 28 21:06:58 AntonTakk1: what app are you making? Jan 28 21:07:01 (not ipk, dunno how to do that) Jan 28 21:07:31 lindi-: combo today/calendar/sms/contacts/dialer Jan 28 21:08:02 AntonTakk1: why do those all in one app? Jan 28 21:08:11 lindi-: speed Jan 28 21:08:15 AntonTakk1: is it for org.freesmartphone? Jan 28 21:08:18 AntonTakk1: :) Jan 28 21:08:19 yes Jan 28 21:08:26 AntonTakk1: gtk or e? Jan 28 21:08:37 efl all the way :) Jan 28 21:08:47 AntonTakk1: you have som dbus service to store communicate send, recive data ? Jan 28 21:08:51 AntonTakk1: ah :( Jan 28 21:09:21 yoyoek1: no, just learning dbus Jan 28 21:10:10 AntonTakk1: I yesturday wrote simple deamon base on sqlite communicate and spred data about contacts using dbus Jan 28 21:10:46 AntonTakk1: you will have a dbus for communication with diffrent apps ? Jan 28 21:11:47 hi, i updated tangogps in shr/openembedded, how can i send a patch? i found only make patch-openembedded, but this does apparently nothing Jan 28 21:12:29 # mdbus -s org.freesmartphone.opimd Jan 28 21:12:29 seeed Jan 28 21:13:43 yoyoek1: it's possible, but not one of my focuses Jan 28 21:14:33 AntonTakk1: you will have applitaction with today/calendar/sms/contacts/dialer and it will by one and second will by your phone book ? Jan 28 21:14:51 lindi-: atm it does look very 2007.2-esque, but with edje, that is a very simple thing to change Jan 28 21:15:38 AntonTakk1: i might still learn something even though kapula has a much smaller set of features surely Jan 28 21:15:52 Anton, yes it is! :D Jan 28 21:16:03 though don't hate on 2007, it's awesome! Jan 28 21:16:18 yeah, i kinda liked that ui Jan 28 21:16:54 lindi-: (alsactl first) the settings inside mixer are "sticky". So can you start arecord a *second* time? or do you have to restore state before? or does it work only once, no matter what you do? Jan 28 21:18:32 DocScrutinizer2: i need to switch back to stereoout.state before i kill arecord or i get oops Jan 28 21:19:43 DocScrutinizer2: (the bug report should describe this but the above is a short version to save your time :P) Jan 28 21:19:49 :D great - honestly! I'd expect it's that DAI modee Jan 28 21:20:48 i just wish i could figure out why the framework ignores my register method call Jan 28 21:22:40 AntonTakk1: where's the source? Jan 28 21:23:14 lindi-: willpaste in a sec, running another test Jan 28 21:23:27 cool Jan 28 21:27:34 moin Jan 28 21:28:32 Nobody96: g'd night :P Jan 28 21:28:36 :D Jan 28 21:28:38 ja Jan 28 21:28:56 :) Jan 28 21:29:29 lindi-: http://pastebin.com/m58145aac Jan 28 21:29:57 silly Duitsers Jan 28 21:32:45 AntonTakk1: shouldn't you handle the reply to request resource? Jan 28 21:32:52 AntonTakk1: instead of magic 20 second wai? Jan 28 21:35:39 lindi-: i'll be working on that after i figure out why it isn't registering to the network Jan 28 21:35:51 thou that might help me figure out why Jan 28 21:36:22 also, i'm not quite sure how :) Jan 28 21:36:54 AntonTakk1: the register reply would have the error message Jan 28 21:37:36 so just fprintf(stderr, variable); ? Jan 28 21:39:24 probably not, nothing else in dbus is that easy :)\ Jan 28 21:45:00 lindi-: i have been watching the system bus with dbus-monitor, requesting gsm and powering the antenna goes just fine Jan 28 21:46:54 but... Jan 28 21:47:28 bit it's almostlikr framework ignores register Jan 28 21:48:41 even thou the same request "works" from mdbus -s (returns "Unauthorized") Jan 28 21:49:20 but if i run my test app after zhone registers, itworks just fine and triggers a re-register Jan 28 21:49:22 raster: morning Jan 28 21:51:51 raster: did it build? :-P Jan 28 21:52:25 hi Jan 28 21:55:07 mrmoku: qemu failed Jan 28 21:55:11 overnight Jan 28 21:55:23 i'm one of those that get the asm build error with qemu svn Jan 28 21:55:37 so i've had to bump qemu defaulr_preference down to -99 Jan 28 21:55:47 building now... Jan 28 21:55:50 slowly Jan 28 21:58:15 raster: does qemu use host c compiler then, or internal one? Jan 28 22:00:01 i have never checked actually Jan 28 22:00:15 i think it may use host as its now buildign the gcc-cross Jan 28 22:00:17 i didnt check Jan 28 22:01:53 raster: that is a nasty bug if you ask me Jan 28 22:02:00 and not strange you hit assembler errors Jan 28 22:02:14 qemu should be built with gcc-native, and not the host c-compiler Jan 28 22:02:17 it's an internal asm error of some sort Jan 28 22:02:27 raster: you have ubuntu or debian host? Jan 28 22:03:45 yup Jan 28 22:03:48 built with host gcc Jan 28 22:04:17 gcc-3.4 -Wall -O2 -g -fno-strict-aliasing -fno-reorder-blocks -fno-gcse -fno-optimize-sibling-calls -fno-crossjumping -fno-align-labels -fno-align-jumps -fno-align-functions -mpreferred-stack-boundary=2 -fomit-frame-pointer -m32 -I. -I.. -I/home/raster/shr/shr-testing/tmp... Jan 28 22:04:29 that'd be why i see it Jan 28 22:04:32 and others don't Jan 28 22:05:18 for some reason it picked gcc-3.4 Jan 28 22:05:22 tho i do have 4.1 too Jan 28 22:05:34 and 4.2 Jan 28 22:05:38 and gcc by default is 4.2 Jan 28 22:05:46 ubuntu Jan 28 22:06:24 raster: after sending a callback with edbus, do i just pass the right args to e_dbus_callback_call to get the results? Jan 28 22:06:36 raster: I mostly only seen debian/ubuntu people getting ICE on qemu compiling Jan 28 22:06:54 er Jan 28 22:06:58 sending a method call Jan 28 22:07:00 raster: which if you ask me, affect a lot of users, and makes me vote for using gcc-native instead Jan 28 22:07:20 10 Jan 28 22:07:24 raster: qemu-0.9 (non-svn version) requires gcc 3.x in order to compile Jan 28 22:07:36 svn version can compile with gcc 4.x on some platforms Jan 28 22:07:47 (all that have the new framework implemented ready) Jan 28 22:08:20 the platforms that aren't ready might not even compile longer with gcc 3.x aswell Jan 28 22:09:16 AntonTakk1: u dont think i have ever used e_dbus_callback_call - thats used internally. i let edbus call whatevere callbacks are needed when the replies/signals/errors come in Jan 28 22:09:35 mw|: gcc-native SHOULD be used Jan 28 22:09:41 just as a mater of principle Jan 28 22:09:43 sorry for the "10" :S Jan 28 22:09:52 OE bootstraps its own environment Jan 28 22:10:01 so you KNOW you have a fixed given environment to work in Jan 28 22:10:23 it needs a native gcc of course to build its own gcc - a requriement of bootstrapping Jan 28 22:10:33 but after that it has its own known gcc version it can use Jan 28 22:10:50 raster: see if OE has a bugtrac on it? Jan 28 22:18:03 Wester :) Jan 28 22:18:09 Raster :) Jan 28 22:18:33 moo Jan 28 22:18:34 Dave :-) Jan 28 22:18:51 Hei mw Jan 28 22:19:35 god aften Jan 28 22:22:09 raster: one thing... you're building shr-testing? Jan 28 22:22:10 :] Jan 28 22:24:57 http://bugs.openembedded.net/show_bug.cgi?id=4814 Jan 28 22:25:18 http://bugs.openembedded.net/show_bug.cgi?id=3282 Jan 28 22:25:21 also seems similar Jan 28 22:25:30 (at least i guess is a symptom of the same issue) Jan 28 22:27:25 raster: both the same issue, just different error-messages due to slight different gcc versions on the host Jan 28 22:27:35 leave a note about why not using gcc-native instead ? Jan 28 22:28:33 i am Jan 28 22:29:05 *wierd* If I add breakpoint in ophonekitd calling SendAuth, (making it delay a bit) it works Jan 28 22:29:14 there - added Jan 28 22:29:15 If I don't have it, it doesn't work Jan 28 22:29:31 (only thing I do in the breakpoint is c and enter in gdb) Jan 28 22:29:38 * mw| hates timer-related bugs Jan 28 22:30:01 I can even see ophonekitd dbus debug output that the reply is received, but it never makes it to "reply" chain Jan 28 22:32:01 raster: you should build unstable though... testing is doomed to not work due to OE changes (frameworkd and kernel mainly). And the e problems we have with unstable Jan 28 22:32:48 and works when using valgrind (it slows everything down) Jan 28 22:33:26 raster: did we update E revision now? if so, did it remove the 30% cpu load Jan 28 22:33:27 oh fun. Jan 28 22:33:49 * raster ctrl-c's Jan 28 22:34:07 :D Jan 28 22:35:45 raster: sorry... for not making it obvious Jan 28 22:36:04 * raster pats his poor little cpu Jan 28 22:36:08 poor bugger Jan 28 22:36:24 Hey guys Jan 28 22:36:35 Howdy friend :) Jan 28 22:37:12 Spent the evening with ptitjes Jan 28 22:37:20 Talking about stuff Jan 28 22:38:48 Ainulindale: fruitfull talks me hopes :-) Jan 28 22:39:11 Yep Jan 28 22:39:21 He has a challenge Jan 28 22:39:35 I.e. migrate ophonekitd to vala this week. Jan 28 22:39:47 In exchange I'll package XSB and flora for ologicd. Jan 28 22:40:08 vala... does ophonekitd have a lot of gobjects ? Jan 28 22:40:11 Vala? why? Jan 28 22:40:16 Ainulindale: uhh, why that? Jan 28 22:40:24 To stop these memory issues. Jan 28 22:40:35 To be more "readable" for newbies Jan 28 22:40:48 to make debugging hell Jan 28 22:40:52 (I saw one potential memory leak in the code when I peak now in an error-path exit) Jan 28 22:40:54 admiral0: Naaaah :-) Jan 28 22:41:07 Ainulindale, where are you supposed to edit gprs settings? Jan 28 22:41:14 mw|: it's been a while since I checked ophonekitd code for leaks Jan 28 22:41:17 gobjects and gobdules? Jan 28 22:41:23 KaZeR: no idea Jan 28 22:42:09 Ainulindale, ah ok thanks.. (i was talking of settings in SHR,of course) Jan 28 22:42:33 it has 'INTERNET' by default but i can't figure out where it get them from.. Jan 28 22:46:43 No idea either Jan 28 22:46:45 I don't use GPRS Jan 28 22:47:55 * mrmoku wonders if now would be a good moment to take a closer look at vala... Jan 28 22:48:51 Heh Jan 28 22:48:54 :) Jan 28 22:49:06 mrmoku: The point is that it's easier to use for such a thing Jan 28 22:49:15 ophonekitd's code is far too long for what it does Jan 28 22:49:22 And far too complex Jan 28 22:49:30 scary Jan 28 22:49:45 At least that's my opinion Jan 28 22:50:21 * admiral0 thinks that plain C is the best thing for the freerunner Jan 28 22:50:31 admiral0: Vala produces plain C Jan 28 22:50:41 i know Jan 28 22:50:51 And FYI Jan 28 22:50:56 I'll do stress tests Jan 28 22:51:07 I'll check every part of that thing Jan 28 22:51:16 Just to see if it's better or worse Jan 28 22:51:29 But IMHO, it can't be worse Jan 28 22:51:44 try, but after that forget gdb Jan 28 22:51:51 Ainulindale: what does that mean for the phonegui frontends? Jan 28 22:51:57 mrmoku: nothing Jan 28 22:52:02 admiral0: already did Jan 28 22:52:14 mrmoku: this is just the first step towars full dbus integration Jan 28 22:52:18 i.e. dbus calls to trigger screens Jan 28 22:52:23 s/towars/towards/ Jan 28 22:52:39 Ainulindale, mw|: btw. I just found the new message segfault :-) Jan 28 22:52:50 What was it? Jan 28 22:53:15 mrmoku: cool Jan 28 22:53:15 I Jan 28 22:53:18 the order in which the widgets are created... have to create the scroller first and then the entry Jan 28 22:53:36 Ok then we can update ? :-) Jan 28 22:53:37 I'm stuck we PIN code auth not working on the host when running fast, but works when ran in gdb with breakpoints or valgrind Jan 28 22:53:52 something very odd is going on... the callback is never called Jan 28 22:54:24 *waiting for dbus-glib 0.78 to compile* Jan 28 22:59:46 same issue Jan 28 22:59:48 *grrr* Jan 28 22:59:57 Got to get my hands dirty then Jan 28 23:00:13 whee Jan 28 23:00:15 playtime! Jan 28 23:03:02 Ainulindale, no problems, thanks anyway. if you know who i should ask, let me know :) Jan 28 23:05:41 hi. I'm on shr-unstable and I discovered a scary thing... Jan 28 23:06:09 when I boot with framerowkd logging set to INFO, GSM does not work Jan 28 23:06:52 that is: I'm asked, pin, it says pin correct, the antenna indicator displays all the bars (strongest signal), but the text above is stuck with "No Service" Jan 28 23:07:06 and indeed I'm not connected, as calling to that # verifies Jan 28 23:08:15 gsm-using apps (openmoko-messages for example) die with "Failed to handle dbus error: Resource is not enabled, current status is 'disabled', 71..." Jan 28 23:08:45 but what's puzzling: if I change frameworkd.conf so that loglevel is DEBUG and reboot, GSM works well Jan 28 23:09:11 I'll upload logs somewhere, currently I can't get my FR online :-/ Jan 28 23:09:44 anyway looking at the frameworkd.log with loglevel INFO (when the problem manifests) doesn't reveal anything obviously wrong, at least to me Jan 28 23:11:22 so my qyuestion is... anybody noticed similar behaviour? I guess you all keep logging set to DEBUG... Jan 28 23:12:25 dent: I have loglevel INFO... without PIN though Jan 28 23:13:01 I've suffered this unability to get connected for few days, but was too busy to investigate Jan 28 23:13:13 and now noticed that switching log to DEBUG magically fixes that Jan 28 23:13:33 (it even made me revive my old dead phone... shame on me :) Jan 28 23:13:50 <\marco> hi to all Jan 28 23:13:51 <\marco> :) Jan 28 23:14:24 btw. I tested this with two sims, both requiring a pin though Jan 28 23:15:06 <\marco> there's a way to have the window-id of the window created? Jan 28 23:15:09 dent: have you checked ophonekitd logs? Jan 28 23:15:24 <\marco> I'm writing with c/elementary Jan 28 23:15:39 dent: I have that aswell Jan 28 23:15:52 sometimes text stucks on "no service" for ever, sometimes, SOS calls only Jan 28 23:16:03 and some times, it displays the name of my service provider Jan 28 23:16:04 hiciu: uf! I looked @ frameworkd only, let me see Jan 28 23:16:23 <\marco> I've to use some obscure Xfunction or there are some vars in elementary/window struct? Jan 28 23:17:23 aaaha... Unknown SIM error: Antenna powered off or SIM not unlocked yet, 77 (dbus-glib-error-quark), code 32 Jan 28 23:18:02 dent: http://shr-project.org/trac/ticket/232 Jan 28 23:18:11 dent: there is patch to fix it Jan 28 23:18:24 not good, but works Jan 28 23:18:36 I see! Jan 28 23:18:43 I've found a pattern for when things fail.... (PIN auth not working) Jan 28 23:18:46 timing issue, as I thought Jan 28 23:19:32 _dbus_pending_call_set_reply_unlocked (internal call in dbus called when reply is ready), is called with pending->function NULL, things freezes due to reply never reach our calling function waiting for callback Jan 28 23:20:32 anyone here knows something about upstart? I'm trying to run it on freerunner, but it hangs here: http://pastebin.com/d42ab12d (line 207) Jan 28 23:22:42 dbus_pending_call_set_notify is called AFTER reply has reach us... that doesn't look good SOMEBODY on the dbus/dbus-glib team should be spanked Jan 28 23:28:21 #¤%%#)(%#¤ dbus_g_proxy_begin_call_internal in dbus-glib has nasty timing-bug Jan 28 23:30:09 :D Jan 28 23:30:12 scary Jan 28 23:32:41 which again uses dbus_connection_send_with_reply which opens this timing issue Jan 28 23:33:59 and dbus_connection_send_with_reply_and_block has the same issue, using dbus_connection_send_with_reply first, and then try to block afterwards, which might be too late Jan 28 23:35:14 This bugzilla report shall be funny Jan 28 23:36:15 back home Jan 28 23:37:09 BUGS Jan 28 23:37:33 Dave: BUGS is not here Jan 28 23:37:36 :D Jan 28 23:45:26 can anyone tell me what neod is, and why it might break my device (not a openmoko phone) on suspend? Jan 28 23:45:35 :p Jan 28 23:46:06 oliford: isn't neod a thing from the past ? Jan 28 23:46:23 i've managed to isolate it as the culprit by gradually running things. everything else is now running and the whole openmoko system seems to be there, so i don't really feel i need it Jan 28 23:46:49 dunno, but its run from the end of in /etc/matchbox/session Jan 28 23:47:15 actually, i havn't quite figured out if you guys are still using openembedded to build for the openmoko phones Jan 28 23:48:05 oliford: Om200x, FSO and SHR all use OE for builds Jan 28 23:48:21 Debian, Android and OpenWrt don't. Jan 28 23:48:26 i've come at this from a funny angle. I'm building the openmoko-image from OE for an iPAQ and assuming its the latest openmoko build without really spending any time finding out how the whole openmoko thing works Jan 28 23:49:23 the last 3 not using it make sense to me. Jan 28 23:49:24 http://bugs.freedesktop.org/show_bug.cgi?id=19796 Jan 28 23:49:28 what are the first 3 Jan 28 23:49:31 now, somebody probably starts to hate me Jan 28 23:50:16 oliford: the first three are Openmoko distributions Jan 28 23:50:32 Om200x is the current base distro from Openmoko Jan 28 23:50:49 FSO is the freesmartphone.org distro, which the next Openmoko release is going to be based on. Jan 28 23:51:24 SHR is the shr-project.org distro, which is also based on FSO. Jan 28 23:52:19 thanks Jan 28 23:54:38 mw|: sorry but shouldn't the set_notify() be called prior to the send_with_reply() call ? Jan 28 23:55:07 arf not it's a handle! Jan 28 23:55:13 wow bad design!! Jan 28 23:55:35 ok, so how do the images get built? do you have another oe git tree or anything? What i'm after is working out the equivilent one I get by building 'openmoko-image' from the current (2 days ago) openembedded tree Jan 28 23:55:51 ptitjes: that is what I called foobar design, somebody should be spanked Jan 28 23:56:00 LOL Jan 28 23:56:10 even the function which is supposed to block fails to block fast enough randomly Jan 28 23:56:23 But that brings me to one thing Jan 28 23:56:31 *spank* *spank* *spank* Jan 28 23:56:34 NAUGHTY PROGRAMMER Jan 28 23:56:38 ;D Jan 28 23:56:38 If we see random dbus calls hang forever, that might be the reason Jan 28 23:56:46 since we use dbus async all over the place Jan 28 23:56:56 but strange that this hasn't been seen anywhere else Jan 28 23:56:58 did you experience this ? Jan 28 23:57:15 mw, where is this happening? Jan 28 23:57:19 or people doesn't use async much, due to the api for enabling it in glib is a bit hidden away Jan 28 23:57:43 Dave: I see this all the time when I use ophonekitd on my host against my frameworkd-fake Jan 28 23:57:54 :[ Jan 28 23:58:01 since my frameworkd-fake replies VERY fast (straight away almost Jan 28 23:58:02 and whose brilliant idea was this? Jan 28 23:58:14 mw|: I don't think being a method of dbusconnection what we could cal being hiden... Jan 28 23:58:52 hidden away as in public API to use, but documentation on glib-dbus is a bit... incomplete Jan 28 23:59:22 (more like read the source in order to figure out how it works and base everything on the few examples given) Jan 29 00:00:34 don't understand why this didn't took a good old struct with function pointers in it... Jan 29 00:01:27 yet a simple dbus_pending_call_new() could heal that design Jan 29 00:01:28 somebody should be spanked Jan 29 00:05:03 I need to sleep soon Jan 29 00:05:15 but atleast I found out why things didn't work yesterday evening Jan 29 00:05:37 *try to write a dbus/dbus-glib patch tomorrow then if I get no reply on the bugreport* Jan 29 00:06:06 mw|zZzZ: you'll have highest priority to use the spank tool!! :p Jan 29 00:06:36 goodnight Jan 29 00:06:46 night night Jan 29 00:09:18 No Jan 29 00:09:21 You cannot go, mw. Jan 29 00:09:25 no? Jan 29 00:09:32 you need a bedtime story? Jan 29 00:09:56 yes yes yes!! :D Jan 29 00:10:16 :D Jan 29 00:10:22 The little freerunner that could ^_^ Jan 29 00:10:25 hehe Jan 29 00:10:33 :D Jan 29 00:10:33 any big issues you guys are stuck at? Jan 29 00:10:47 Yes... Jan 29 00:11:04 * mwester falls over laughing Jan 29 00:11:22 :) Jan 29 00:11:28 mwester: due to the report, or peoples talk of spanking? Jan 29 00:11:37 your question Jan 29 00:11:46 heh Jan 29 00:12:18 he's mw|debuggor Jan 29 00:12:21 :p Jan 29 00:12:42 I'm good at finding bugs Jan 29 00:12:57 * mw|zZzZ and the bugs communicate in inhuman ways Jan 29 00:13:24 wow this might be rather dangerous... take care! Jan 29 00:14:14 "gdb - The GNU Debugger" :-D Niiiiiiiiice tool Jan 29 00:14:54 mw|zZzZ: as you are in the dbus api... what can be put in an error message Jan 29 00:15:05 ptitjes: any string you want Jan 29 00:15:19 can't you put somthing else like in any message ? Jan 29 00:15:40 the caller usually just checks the error-number if it wants to do action, and use the message if logging or displaying things to the end-user Jan 29 00:15:59 so there is an error number... Jan 29 00:16:17 can someone explain me why errors in frameworkd are strings... Jan 29 00:17:10 ptitjes: : error = g_error_new (ORG_FREESMARTPHONE_OBJECT_ERROR, ORG_FREESMARTPHONE_OBJECT_ERROR_PolicyUnknown, "Resource %s has an invalid policy.", name); Jan 29 00:17:33 ptitjes: the first to numbers are the "uniqe" numbers that identify the error Jan 29 00:18:09 first number is a domain/quark (uniqe) and second one is a local number inside the domain Jan 29 00:18:26 rest is a text message that can be used for logging / display to end users Jan 29 00:18:44 mw|zZzZ: arf when you don't use localization... Jan 29 00:18:45 that is the general design of GError Jan 29 00:19:04 ptitjes: you can localize the message Jan 29 00:19:15 but why does libframeworkd-glib compares string to find out what frameworkd error it is ? Jan 29 00:19:34 since dbus remote errors can't preserve the quark/error number pair Jan 29 00:19:47 (as far as I know) Jan 29 00:21:16 dbus spec says : An ERROR may have any arguments, but if the first argument is a STRING, it must be an error message. The error message may be logged or shown to the user in some way. Jan 29 00:21:24 * Dave gives mw a drink Jan 29 00:21:53 so we may transport what we needs in the error Jan 29 00:22:02 s/needs/need/ Jan 29 00:22:02 ptitjes meant: so we may transport what we need in the error Jan 29 00:22:08 * ptitjes is too tired Jan 29 00:22:27 Look-up at GError Jan 29 00:23:16 so frameworkd is not dbus but dbus-glib Jan 29 00:23:33 ptitjes: for a dbus remote error, domain=DBUS_GERROR, code=DBUS_GERROR_REMOTE_EXCEPTION, and the only spare field left is gchar *message; Jan 29 00:24:00 But that is completely stupid! Jan 29 00:24:15 heere again someone have to be spanked! Jan 29 00:24:18 if GError had supported nested errors, dbus-glib might had preserved more data Jan 29 00:24:39 GError is glib design, and has worked for 99.9999% of other uses Jan 29 00:24:58 and dbus can't send over quark/domain, since dbus isn't GError aware Jan 29 00:25:13 so dbus sends over only the error-message when peer sends an error-reply Jan 29 00:25:30 perhaps a better standard would be to send domain and code aswell Jan 29 00:25:34 yeah but they should not have transported technical error number but rather the application error numbers Jan 29 00:26:06 but again... the server-side that wants to send an error in return can write whatever text they want in the message, and can send numbers as text there if they really wanted to Jan 29 00:26:17 mw|zZzZ: you can put whatever you want in an dbus error message Jan 29 00:26:26 sending plain text allows for anything to be sent... it's up to the user to serialize whatever data they want Jan 29 00:26:55 ptitjes: but you can't unpack any data into a fixed designed GError Jan 29 00:27:06 mw|zZzZ: sending plain text needs matching in the receive part and then recoding to have meaning full error message for every language Jan 29 00:27:23 but dbus-glib is a binding over dbus Jan 29 00:27:42 they should not have restricted dbus use Jan 29 00:28:03 because someone using dbus-glib can't talk to every dbus server stuff Jan 29 00:28:19 this is bad Jan 29 00:28:30 I think they should.. that is my view (after working with serialization in systems before).... errors should have a fixed design, so that you can use the same error-report routines from multiple places Jan 29 00:28:37 LIBC uses errno for everything Jan 29 00:28:51 php and python uses exception for most stuff Jan 29 00:28:58 dbus-glib uses text-string Jan 29 00:29:06 yeah but you can put whatever you want in exceptions Jan 29 00:29:32 ptitjes: many languages have restrictions in exception data Jan 29 00:29:32 and exceptions are families (dependent classes) Jan 29 00:30:00 C for instances which glib and dbus internally is based on doesn't support exception Jan 29 00:30:19 * Dave shares his cookie with mw Jan 29 00:30:23 and having support for VARIANT data all over the place is messy and bloats the size of code Jan 29 00:30:27 I know but you can have structs to represent errors information Jan 29 00:30:59 and what happens if you don't know what format the error-reply comes in when you code in C Jan 29 00:31:05 VARIA Jan 29 00:31:18 anyway dbus-glib doesn't even support complex data structures as argument of method calls Jan 29 00:31:24 VARIANT is messy in C... works nice in generic languages where you have .ToString() and .Serialize() Jan 29 00:31:38 mw|zZzZ: this API specific Jan 29 00:31:45 this is Jan 29 00:31:57 ptitjes: that might be... haven't reached that need yet (but will soon) Jan 29 00:32:27 libframeworkd-glib is bloated with strcmp() calls! Jan 29 00:32:34 too bad Jan 29 00:32:45 and so highly efficient ;) Jan 29 00:33:01 (problem with complex data structures in C, is that struct layout can't be known in a generic way, they are compiler-ABI dependent.. C is messy here) Jan 29 00:33:29 strcmp aren't that slow, but slower than integer compares yes Jan 29 00:33:40 Vala guys are rewriting their dbus support to remove completely dbus-glib code Jan 29 00:33:50 they go for straight dbus and I understand why now Jan 29 00:34:00 that sounds nice Jan 29 00:34:16 dbus-glib lacks a lot of stuff Jan 29 00:34:41 does the python binding frameworkd uses is glib based ? Jan 29 00:34:52 like proper documentation, prototyping of callbacks when it generates code, a way to support structured data... Jan 29 00:35:17 if not they should quickly change this error propagation scheme Jan 29 00:35:22 structured data could have been resolved with using an array of struct containing an enum Jan 29 00:35:52 GError is choosed due to that beeing glib internal use Jan 29 00:36:17 but they could have choosen another design, but hard to find an easy design that is efficent Jan 29 00:36:35 mysql approach to VARIANT works to some degree Jan 29 00:36:48 they always just work with arrays of strings Jan 29 00:36:50 error code first - precision information next - is the best design Jan 29 00:38:34 dbus design probably works best in dynamic languages like python, php .net Jan 29 00:38:53 I have to rewrite ophonekitd in Vala but I'm so despaired to see the error matching and stupid string constants I'll have to port... Jan 29 00:38:58 if they hadn't had support for structures, and array of structures, things could have been a bit easier to implement Jan 29 00:39:45 frameworkd returns errors, and not numbers as far as I know ? Jan 29 00:40:12 when I think about it, it returns ErrorDomain which is text, followed by text message Jan 29 00:40:12 mw|zZzZ: yeah this is my problem... Jan 29 00:40:29 (Domain Quark can be translated back into a text) Jan 29 00:40:51 ptitjes: the good about that design is that you never will have issues about overlapping numbers Jan 29 00:41:23 X is designed this way, due to the events available is not a fixed design Jan 29 00:41:26 oh there are many numbers in 32 bits :) Jan 29 00:41:34 that is true Jan 29 00:41:42 I'm much frightened by typos in strings Jan 29 00:41:48 :D Jan 29 00:41:54 LIBC uses a half-way design... Jan 29 00:42:12 fixed names, but they resolv to numbers in header files, which might be different from system to system Jan 29 00:42:53 mw|zZzZ: this is like the status of a call being a string Jan 29 00:43:23 The idea with text instead of numbers is that if you add number, and a dude over in norway adds an errornumber, it makes sure you didn't use the same number Jan 29 00:43:33 I can still remember when people were wondering if it was "release" or "released" Jan 29 00:43:34 but that is only an issue in larger projects Jan 29 00:44:00 or if you can receive errors from multiple places Jan 29 00:44:19 like if you call libX, and libY failes.. libX just forwards the error from libY Jan 29 00:44:33 I understand it Jan 29 00:44:39 and libX and libY don't cooperate in error-numbers Jan 29 00:44:46 (though I'm used too Java's nested exception...) Jan 29 00:45:05 but all our errors-id's are fixed to a small set Jan 29 00:45:22 I have played with .net nested exception Jan 29 00:45:26 mw|zZzZ: one could use two numbers kind of namespaces Jan 29 00:46:15 first integer = namespace (or lib id if you want) and second = cause (lib-specific error number) Jan 29 00:46:30 though, nested exception there fails to work in a) invokes (no reason for it not working... just bad design of the invoke function by M$) b) remoting, due to M$ not having a defined API for serialize and transfer the exception data Jan 29 00:47:04 mw|zZzZ: because exceptions are not objects in .net ?? Jan 29 00:47:09 ptitjes: that is the current design of glib, only that it has a database for mapping strings-Id's into dynamic integer id's Jan 29 00:47:19 ptitjes: they are Jan 29 00:47:39 ptitjes: but in remoting, that would cause a remote-reference, which is not always a good thing Jan 29 00:48:00 in Java errors are all serializable Jan 29 00:48:11 ptitjes: but they could have cloned it... either way... when I played with it, you only got exception: "remote exception" Jan 29 00:48:19 you can serialize error that nests errors that nests.... Jan 29 00:48:52 brrr M$ are the one that spank themselves!! Jan 29 00:48:58 heheh Jan 29 00:49:13 .NET is in many cases a nice tool to work in Jan 29 00:49:29 WTF? Jan 29 00:49:29 fast to design in, a bit slow when you need performance Jan 29 00:49:36 and eats way to much memory Jan 29 00:49:47 mickey|zzZZzz: http://downloads.freesmartphone.org/fso-milestone5/feeds/ is being populated now Jan 29 00:49:58 and hard to find memory-leaks (.NET that is) Jan 29 00:50:11 but now I really need to sleep Jan 29 00:50:15 0145 here Jan 29 00:50:38 night night :-) ptitjes nice talking generic design with you :-D Jan 29 00:51:04 yeah we are on the same spindle Jan 29 00:51:07 (I probably are a bit stubborn) Jan 29 00:51:08 have a nice sleep Jan 29 00:51:29 no don't think that Jan 29 00:51:51 :p Jan 29 00:52:25 going to sleep too Jan 29 00:55:42 * Dave prods mw Jan 29 01:34:12 is there any problem with nesting case statements in c other than the possibility of ending up with enormous nested case statements? Jan 29 01:35:15 nope Jan 29 01:35:24 nested conditionals? Jan 29 01:35:30 you just end up with a big mess :D Jan 29 01:35:35 but necessary logic sometimes Jan 29 01:36:25 well,i can either do one enormous one, or three smaller ones that will probably end up being complex anyways Jan 29 01:37:14 trying to create a greatly simplified interface to dbus/fso Jan 29 01:38:10 yeah yeah :) Jan 29 01:38:59 like fso_method("ousaged", "usage", "usage", "RequestResource", "GSM"); to request gsm resource, forexample Jan 29 01:39:47 Using one case per function sometimes makes it easier to "grok" if the case statements are multi-page... Jan 29 01:40:30 (the compiler will optimize it away, mostly if not entirely) Jan 29 01:45:45 yeah Jan 29 01:47:18 but the logic for the path and interface will have to be nested anyways, ie, what ogsmd and ogpsd both have a "Device" path? Jan 29 01:48:46 but yeah, there is no reason to try tomake it all a single ginormous case :) Jan 29 01:57:29 would also *like* to have that one function call do the signal handlers (highly generic handeling, anyways) Jan 29 02:02:05 ~_~ Jan 29 02:07:42 Dave: ? Jan 29 02:10:12 moko moko moko moko Jan 29 02:10:21 the code is everywhere! Jan 29 02:11:43 there goes dave with the four-letter words again! Jan 29 02:13:24 :) Jan 29 02:13:53 Hi all Jan 29 02:14:06 oh great Jan 29 02:14:11 it's pwgen ¬_¬ Jan 29 02:18:01 can someone tell me how fso milestone5 can work without busybox ... ? Jan 29 02:18:57 e/illume Jan 29 02:19:16 I tiried to build it ... and busybox is missing on rootfs.jffs2 * Jan 29 02:19:28 no Jan 29 02:19:37 e/illume replaces busybox Jan 29 02:19:51 oh Jan 29 02:19:54 wait Jan 29 02:20:02 i'm thinking matchbox Jan 29 02:20:08 yeah :p Jan 29 02:20:10 nvm :) Jan 29 02:20:10 silly Anton :) Jan 29 02:20:12 sorry Jan 29 02:20:22 oh crap ... i came from a party an are i am drunk ... Jan 29 02:21:09 :D Jan 29 02:22:24 i tried to build milestone5 ... my image is only 43 mb in size and missing busybox Jan 29 02:22:41 sweet Jan 29 02:23:21 .. sweet party or missing busybox .. ?? Jan 29 02:24:58 PAAAAAAAARTAAAAAAAAAY! Jan 29 02:25:15 *FH* Jan 29 02:25:22 * Dave high-fives pwgen Jan 29 02:26:25 clap in high-fives Jan 29 02:28:21 damm mickeyl si still zzzzzzzzzzzz Jan 29 02:34:06 yeah Jan 29 02:34:09 he likes teh snoozies Jan 29 02:38:19 does someone know in what TZ mickeyl lives ? Jan 29 02:38:37 moon ? .. Jan 29 02:38:42 *G* Jan 29 02:39:22 asia? Jan 29 02:41:31 oh .. then i have to get rid of sleeping at night .. and change night and day ... Jan 29 02:41:56 spooky Jan 29 02:43:44 actually i live in the north ..(long nights short days ) so the LCD display will simulate the daylight .. Jan 29 02:44:56 O_O Jan 29 02:47:09 * Dave beers pwgen Jan 29 02:48:32 grmpfl now my lat build image of milestone5 image has busybox .. lets try .. ( beers of diffrent type on an good bye party for an friend ) Jan 29 02:49:19 :] Jan 29 02:49:20 s/lat/latest/ Jan 29 02:52:04 yeah, I got that Jan 29 02:54:41 i have done some mods to one battery, so i will do some power consumption measurements for my neo tomorrow .. and see whats is possible to the apm Jan 29 02:56:10 Dave will you be at FOSDEM ? Jan 29 02:56:15 sweet Jan 29 02:56:20 nah, fuck fosdem Jan 29 02:56:44 you couldn't drag me to another one, even if they paid me double Jan 29 02:56:58 :p Jan 29 02:57:24 .. cheap belgium beer .. nice people ... why not ... Jan 29 02:57:35 ¬_¬ Jan 29 02:57:38 If only you knew... Jan 29 02:57:43 raster: hey Jan 29 02:58:17 paroli! **** ENDING LOGGING AT Thu Jan 29 02:59:57 2009