**** BEGIN LOGGING AT Mon Sep 22 02:59:57 2008 Sep 22 03:27:49 mwester: monday is cold and wet here :-( Sep 22 03:28:23 ? Sep 22 03:28:31 must be an adelaide thing Sep 22 03:28:32 nice here Sep 22 03:28:33 btw Sep 22 03:28:40 anyone ever tried using at on oe? Sep 22 03:28:45 its nicely broken Sep 22 03:28:55 and doesnt wake things from sleep by setting an rtc wake time Sep 22 03:29:02 in fact - it doesnt even get that far Sep 22 03:29:08 but doenst have code to do it anyway Sep 22 03:31:03 Yep, there should be a bug (now closed, IIRC) on that. Sep 22 03:31:15 Apparently it's not supposed to work. (IIRC) Sep 22 03:31:34 rwhitby: But it's getting warmer for you folks every day now... Sep 22 03:32:07 mwester: I'm camping this weekend, so I hope it's warmer then Sep 22 03:32:29 :) Cool, hope you have fun on the trip! Sep 22 03:33:31 it's actually a training course on how to take scouts camping. you have to go camping in course groups to pass the course :-) Sep 22 03:35:00 That could be fun -- as long as they don't spend most of the time doing classroom type stuff on legal liabilities, policies, procedures, etc. (That's probably what they do here) Sep 22 03:35:40 no, that was a different course, held in a residential setting Sep 22 03:36:13 mwester: its not meant to work" wtf? Sep 22 03:36:29 at not meant to? Sep 22 03:36:35 thats.. bizarre? Sep 22 03:36:42 so what is the alternative for doign things like Sep 22 03:36:58 ":please wake up he system from sleep and sound an alarm at time x"? Sep 22 03:37:04 dotn tell me i have to go write it Sep 22 03:37:06 raster: I might be wrong. But there was something quite bizarre like that, and it sticks in my mind that it was the "at" command. Sep 22 03:37:23 * raster prepares to look up the /dev/rtc ioctl magic for that baby Sep 22 03:37:37 well at is quite broken in oe (fso) Sep 22 03:37:49 no compile (as it finds no mta to use on compile this breaks) Sep 22 03:37:51 fixed that Sep 22 03:37:53 raster: You'll probably not only have to write it, but debug it as well. There are also reports that wakeup alarms from the RTC are unreliable. Sep 22 03:38:04 then runtime - atrm is broken ":permission denined" to unlink jobs Sep 22 03:38:12 whihc is bizarre.. as ti runs as root Sep 22 03:38:20 and atd just seems to exit on its own Sep 22 03:38:23 and not keep running Sep 22 03:38:34 bugger Sep 22 03:38:38 fugger Sep 22 03:38:39 ok Sep 22 03:38:50 we do need a system service that lets us wake up using an rtc Sep 22 03:38:57 especially fore alarm clocks for example Sep 22 03:39:06 iun fact anything that may want to poll while suspended needs it too Sep 22 03:39:20 eg every 30 secs wake up and check wifi ap's Sep 22 03:39:31 at only has 1min granularity Sep 22 03:39:38 whihc isnt great for sub-1min wakeups Sep 22 03:39:39 Yes. We also need a kernel service that does the same, to check on things (like the battery) periodically, and then go back to sleep -- without the cost in battery current of bringing up the display, audio, and all that. Sep 22 03:40:27 FR has a low-batt IRQ, but not the neo, and as you observe, we need it for other things too. Sep 22 03:40:27 technically that can be done without the kernel Sep 22 03:40:33 if device stuff was done in userspace Sep 22 03:40:38 ie wakeup didnt bring display on Sep 22 03:40:40 etc. Sep 22 03:40:49 wakupe just turned cpu, ram and disk on Sep 22 03:41:01 it then is userspace's job to do the rest Sep 22 03:41:30 a "wake up every 10 mines to check battery level" from userspace i think is ok Sep 22 03:41:40 as long as the kernel doesnt go bringin everythign up Sep 22 03:41:45 right now for backlight its ok Sep 22 03:41:48 hats the worst offender Sep 22 03:41:54 if it was off on suspend Sep 22 03:41:56 its off on resume Sep 22 03:42:09 that way u can wake up without being noticed Sep 22 03:42:13 quickyl poll/check your stuff Sep 22 03:42:23 instnantly go "ok boys = job done. sleep!" Sep 22 03:42:32 though i'd much rather vote for zero-clock... Sep 22 03:42:41 and never suspend Sep 22 03:43:19 The audio still clicks on suspend/resume, though -- so the backlight is not sufficient, really. Sep 22 03:44:02 true Sep 22 03:44:17 * mwester imagines a meeting with an executive -- there is silence in the room as the speaker takes the podium, except for an annoying "click!" from mwester's pocket... every 30 seconds... Sep 22 03:44:21 i'd be happy if we had to bring audio up in userspace Sep 22 03:44:33 It would crash the kernel less. Sep 22 03:45:13 if we move mroe of this to userspace- kernel work is simpler Sep 22 03:45:16 more reliable Sep 22 03:45:27 it takes less work to get things runnign on a new system/soc Sep 22 03:45:42 if we do zero-clock Sep 22 03:46:00 then its onyl a matter of how many devices we can then reliably bring up/down at will (from userspace) Sep 22 03:46:07 I don't know what Andy is working on for the new version of the kernel; but I thought he was writing something that would manage the suspend/resume sequence more closely. Sep 22 03:46:09 and u can knockt them over one by one Sep 22 03:46:25 so as u get more of them to turn off/go int low power mode the better the battery life Sep 22 03:46:37 instead of having to manage to get them ALL to suspend/resume reliably all at once Sep 22 03:46:58 so as such lets say we have nothing Sep 22 03:47:05 we save some power by zero-clocking cpu/ram Sep 22 03:47:14 but everything else remains 100% fully powered up Sep 22 03:47:23 And yes, a zero-clock is ideal -- but has always been more expensive in terms of hardware. Which gets to another topic near and dear to my heart, which is that hardware people are d*** good at excel spreadsheets with BOMs and costs, and software people can't seem to convey the cost of sofware to financial people. Sep 22 03:47:37 then one thing at a time (backlignt, then glamo, then gsm then audio etc.) we manage to bring up/down reliably Sep 22 03:47:46 and so add it into the userspace powermanagement Sep 22 03:49:31 Along with a SoC that supports zero-clock, it would be nice if the drivers could change other things as well -- e.g. do we really need 640x480 resolution all the time -- do we save power if the normal phone display (dialer, status page) runs at a lower resolution? Sep 22 03:50:44 I think theres so much that can yet be done, but we cannot do it if we have to let the existing "almost random" suspend/resume logic remain in charge. JMO. :) Sep 22 03:50:57 mwester: then you loose flesh tones on the today screen wallpaper ;-) Sep 22 03:51:27 hmm Sep 22 03:51:29 * mwester thinks about that for a few seconds... ;) Sep 22 03:51:31 changing res is possible Sep 22 03:51:41 but a lot more complicated as its not just the app affected Sep 22 03:51:43 its everyt app Sep 22 03:51:46 they whole screen Sep 22 03:51:52 u dont just lose fidelity Sep 22 03:52:01 layout details likely change Sep 22 03:52:18 but thats another matter Sep 22 03:52:28 entirely Sep 22 03:52:40 most userpsac esoftware just isnt good at all at dynamic resolution changes Sep 22 03:52:43 Hmm.. ok, so probably not a good tradeoff. But can we do anything to lower the clock speed to save power on the Glamo? Sep 22 03:53:01 on the glamo - no idea Sep 22 03:53:08 though likely any saving is not worht it Sep 22 03:53:22 i think the glamo is like <100mw Sep 22 03:53:26 like mroe the 50mw range Sep 22 03:53:31 socs is 200+ Sep 22 03:53:36 * mwester thinks of maybe a mono screen, or a sloppy screeen, or even a flickering screen, that would become the normal resolution upon first tap on the touchscreeen. Sep 22 03:53:38 better to focus on soc Sep 22 03:53:47 backlight is trivial and eats a lot of power Sep 22 03:54:05 "dminishing returns" Sep 22 03:54:08 :) Sep 22 03:54:18 i would worry about things that would apply to future cpus/soc's Sep 22 03:54:26 at least in principle Sep 22 03:56:35 I'm like a pit-crew mechanic; I worry about winning *this* race. Sep 22 03:56:53 hehehehe Sep 22 03:56:59 i worry about winning the wra Sep 22 03:57:02 not the battle :) Sep 22 03:57:14 but you need both types to win in the end Sep 22 03:58:03 That line has always bothered me, because it's something that only the general from the rear of the brigade can say -- the guy carrying the machine gun in the mud is VERY focused on winning the battle... Sep 22 04:15:35 i'm not sure there needs to be a difference Sep 22 04:15:43 some peolpe prefetr to focus on the here and now issues Sep 22 04:15:53 other on the logner term ones and not the here-and-now ones Sep 22 04:16:10 btw anyone here tried om/oe/angstorm on a treo650? Sep 22 04:16:23 * rwhitby is waiting to do so ... Sep 22 04:16:43 rwhitby: on yours? Sep 22 04:17:01 zedstar is sending me an old one of his, so i can use it to develop on. Sep 22 04:17:19 couple of weeks time probably Sep 22 04:18:11 looks like flash is not accesisble currently Sep 22 04:18:26 that disk-onchip thing Sep 22 04:18:35 i've seen them before Sep 22 04:18:38 * mwester checked eBay; still pretty pricey. Sep 22 04:18:50 mwester: theres one on ebay.au Sep 22 04:19:00 right now its going for $1 Sep 22 04:19:01 :) Sep 22 04:19:47 raster: you'd want to boot from SD card anyway due to space considerations Sep 22 04:20:00 rwhitby: i know Sep 22 04:20:12 and currently u have to boot via palmos with cocoboot anyway Sep 22 04:20:19 Hmm.. so is the internal flash then left alone (i.e. PalmOS)? Sep 22 04:20:19 so DoC flash acessibility is not too relevant Sep 22 04:20:20 so u'd need a working palmos Sep 22 04:20:30 mwester: yeah Sep 22 04:20:37 that makes it quite sane to hack with Sep 22 04:20:40 and use as a phone Sep 22 04:20:41 Ok. SD cards are cheap. Sep 22 04:20:46 just har-dreset to go back to palmos Sep 22 04:21:00 phone calls even work Sep 22 04:21:08 tho seems audio is a bit fucked (one-sided calls) Sep 22 04:21:26 i was more looking at it as somethnig to play with for phone ui stuff Sep 22 04:21:30 raster: going price on ebay.au seems to be around $200 for the times I've looked recently Sep 22 04:21:31 that has touchscreen AND a keypad Sep 22 04:21:44 rwhitby: yeah. is ee ones "for sale" for $170 Sep 22 04:21:48 (and up) Sep 22 04:22:05 rwhitby: i'll see how that auction goes Sep 22 04:22:58 hmm Sep 22 04:23:04 another one for $99 Sep 22 04:23:06 You can buy a new-in-box one from a reputable dealer (physical storefront) for $150 Sep 22 04:23:21 http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=3476985&Sku=P175-1074 Sep 22 04:23:28 Not sure if they ship to au though Sep 22 04:24:26 hmm Sep 22 04:24:28 i choudl check Sep 22 04:25:02 I'm sure we could find a shipping agent ;-) in the US to buy it for you ... Sep 22 04:25:31 hehe Sep 22 04:26:30 "Received this item in record time. Only took 6 days for arrival to an APO box overseas." - one of the reviews Sep 22 04:31:05 http://cgi.ebay.com.au/NEW-PALM-TREO-650-UNLOCKED-BLUETOOTH-CAMERA-GSM-PHONE_W0QQitemZ170263709408QQihZ007QQcategoryZ1503QQssPageNameZWDVWQQrdZ1QQcmdZViewItem Sep 22 04:31:15 $162 Sep 22 04:31:23 if i feel lazy Sep 22 04:31:31 I was hoping that since I already have all the accessories (and a spare battery) for the 650, I could snag a cheap one that was missing that stuff from eBay, but even those are pricey - one closed at $70 for just the phone, and the seller wouldn't even guarantee it could place a call. Sep 22 04:32:26 interesting Sep 22 04:32:29 so they stil go for a bit Sep 22 04:33:01 http://cgi.ebay.com.au/Palm-Treo-650-unlocked-GSM-with-accessories_W0QQitemZ270278284872QQihZ017QQcategoryZ161653QQssPageNameZWDVWQQrdZ1QQcmdZViewItem Sep 22 04:33:01 it's a bloody good phone, that's why. Sep 22 04:33:04 loosk good Sep 22 04:33:06 $0.99! Sep 22 04:33:07 :) Sep 22 04:33:23 rwhitby: it is? Sep 22 04:33:47 yep. I haven't found an open or closed source phone better than the Treo650. Sep 22 04:34:59 I had the Sprint version of it from my last employer; when I moved to my current job, I ended up buying the GSM version of it out of my own pocket -- no regrets at all. Sep 22 04:35:03 you obviously havent joine the "let me suck steve jobs's testicles" fan club Sep 22 04:35:25 so the hw is solid? Sep 22 04:35:34 just a question of making linux work... Sep 22 04:35:37 yep Sep 22 04:35:48 good Sep 22 04:35:57 i like it if i know the hw is MEANT to work Sep 22 04:36:00 my phone has lots of paint scratched off it from lots of use :-) Sep 22 04:36:02 and KNONW to Sep 22 04:36:03 :) Sep 22 04:36:29 and it's been dropped onto carpet many times Sep 22 04:37:09 * mwester confesses to dropping it onto a concrete sidewalk -- more than once. Sep 22 04:37:33 oh, the connector on the base doesn't work anymore for data transers, but that's just cause it's all corroded on the connector. still works for power, and I hotsync via bluetooth anyway Sep 22 04:38:03 * mwester wonders if that's what the problem is on his connector. Sep 22 04:38:08 methnks i want my data connector to work for hackin' Sep 22 04:38:40 No salt water near me; my connector is clean. Sep 22 04:39:07 raster: note that the connector worked fine for over 3 years :-) Sep 22 04:39:18 and I could probably clean it if I bothered Sep 22 04:39:25 thats a good life for a phone Sep 22 04:39:27 so, running straight qtopia, what can I use with python to talk to the display? Sep 22 04:39:34 i think the regular lifespan is about 1-1.5 Sep 22 04:40:12 I thought pygame would work but seems no Sep 22 04:40:23 I've had a Treo of some sort for 6 years now Sep 22 04:41:14 poor palm Sep 22 04:41:19 180, 270, 600, 650. Sep 22 04:41:21 so early into the game Sep 22 04:41:30 and yet... now so fucked Sep 22 04:41:31 :( Sep 22 04:41:39 opportunity lost :( Sep 22 04:41:49 * raster ponders why with the zillions fo palm apps... Sep 22 04:41:53 If they had a Treo650 with wifi and 3G, and no other changes (including software), I'd buy it in a minute. Sep 22 04:42:08 well ok not lost Sep 22 04:42:16 but definitely not king of the hill by a country mile Sep 22 04:42:21 as they should have been Sep 22 04:42:29 a poor fourth or lower by now Sep 22 04:42:38 yeah Sep 22 04:42:40 what killed them Sep 22 04:42:43 the os/dev env? Sep 22 04:42:48 or ugliness? Sep 22 04:42:57 they always have simpl3e and functional stuff Sep 22 04:43:05 palmos was painful to code for tho Sep 22 04:43:14 i dug into it when i had a clie nx80 Sep 22 04:43:21 man was that not easy Sep 22 04:43:31 Treo Chatter Email is still the best email app on any mobile device Sep 22 04:43:33 everythnig was static, IIRC Sep 22 04:43:33 especially since u couldnt do a 100% native arm app Sep 22 04:43:52 u nedt o bootstrap with 68k then do subroutines in arm with magic jump goop Sep 22 04:44:11 mwester: yeah - that was an interesting adaption to their original hw Sep 22 04:44:15 but didnt scale Sep 22 04:44:15 1 click to open and get new mail, 1 click to delete message and read next, Sep 22 04:44:29 a lot fo crap was bolted onto palmos to make it look vaguely modern Sep 22 04:45:07 But for a phone/palmtop, it worked well on some pretty small hardware. Sep 22 04:45:19 sure Sep 22 04:45:25 considering where it came grom Sep 22 04:45:30 ie 68k Sep 22 04:45:34 with no disk Sep 22 04:45:39 * mwester owned the original Palm Sep 22 04:45:43 (simple rom only - everythign else in ram) Sep 22 04:45:59 for a while plam totally owned the space Sep 22 04:46:06 * rwhitby implemented the synthesizable 68k core on the PalmV and most of the Clie range while working at Mot Sep 22 04:46:12 again - i wodner what the fatal flaw was Sep 22 04:46:15 just not being sexy? Sep 22 04:46:42 rwhitby: what that the dragonball, then? Sep 22 04:46:46 yep Sep 22 04:46:55 cool Sep 22 04:46:58 * raster has fond memories of his 68k asm days Sep 22 04:47:06 man that was a nice cpu arch Sep 22 04:47:50 I led a team of 5 which took the 4micron custom 68K core and turned it into a 0.25micron synthesizable fully-static and fully-scan implementation Sep 22 04:48:13 Austin? Sep 22 04:48:19 Adelaide Sep 22 04:48:22 Ah. Sep 22 04:48:29 working for Austin Sep 22 04:48:34 da 'laid Sep 22 04:49:03 I still have the *original* 1970-vintage microcode flowcharts for the 68k :-) Sep 22 04:50:49 rwhitby: your starting to scare me o_0 Sep 22 04:51:19 kgoetz: I wasn't there in 1970's - the flowcharts were "handed down" to us to assist in the reimplementation Sep 22 04:51:46 rwhitby: ah, that explains it Sep 22 04:51:49 we did the project in 95 Sep 22 04:52:34 hmm Sep 22 04:52:42 no /sys/class/rtc ... bugger Sep 22 08:03:08 hi guys Sep 22 08:04:06 how do we handle getopt() when porting to ARM? Sep 22 08:04:11 any ideas and tips Sep 22 08:05:47 ? Sep 22 08:05:55 wqhy should that be an issue? Sep 22 08:07:01 because it messes my output with utilities that use it Sep 22 08:07:20 in what way? Sep 22 08:08:00 it returns an extra -? for all inputs Sep 22 08:10:22 odd Sep 22 08:10:26 glibc? Sep 22 08:10:28 or uclibc? Sep 22 08:10:33 raster: yes Sep 22 08:10:38 glibc Sep 22 08:11:25 brb Sep 22 08:14:18 strange Sep 22 08:14:30 * raster doesnt use getops... much easier just to walk the argv... Sep 22 08:51:07 raster: Guess you are right :) Sep 22 08:51:54 But what do we do with all the shit that needs to be handled Sep 22 08:51:57 ? Sep 22 09:52:12 shaz_: sounds like a bug... Sep 22 10:51:56 Weiss: I think its the difference in the way char data type is handled by ARM and x86. Sep 22 10:52:20 hmm? Sep 22 10:52:26 char declaration means unsigned char on ARM and signed on x86 Sep 22 10:53:16 so a comparison resulted in that silly output Sep 22 10:53:48 I did the way raster told me to walk the arguments and it worked out fine for a start :) Sep 22 10:57:13 hm.. but if the character codes are low, then it shouldn't make any difference (sign bit will always be unset) Sep 22 11:11:00 i dont see any reason arguments shouldnt be low Sep 22 11:11:03 ie normal ascii Sep 22 11:11:39 u dont make cmd-line arguments like -金魚 Sep 22 11:12:01 or -œøb Sep 22 11:12:02 :) Sep 22 11:12:32 raster: :-) Sep 22 11:21:04 * rwhitby wonders if anyone has tried qemu lately, cause both fso-testing and openmoko daily gta01 images fail to get past u-boot ... Sep 22 11:23:12 alphaone: boo! Sep 22 11:24:33 raster: heya Sep 22 11:26:27 do u guys have anything in fso so far for scheduled jobs Sep 22 11:26:28 ie at? Sep 22 11:28:02 no. everything time/scheduled events/wakeup is still missing. Sep 22 11:28:12 bbiab, dinner Sep 22 11:30:43 ok Sep 22 11:30:51 well looks like yet another thing i am busy implementing Sep 22 11:30:57 if u want it... i have it Sep 22 11:30:58 or the start Sep 22 11:31:00 api is done Sep 22 11:34:32 raster: Sounds nice. How is the API? Sep 22 11:35:07 really simple Sep 22 11:35:09 dbus Sep 22 11:35:23 e_dbus_interface_method_add(iface, "List", "", "a(i)", Sep 22 11:35:23 request_method_List); Sep 22 11:35:23 e_dbus_interface_method_add(iface, "Add", "iiss", "", Sep 22 11:35:23 request_method_Add); Sep 22 11:35:23 e_dbus_interface_method_add(iface, "Del", "i", "", Sep 22 11:35:24 request_method_Add); Sep 22 11:35:44 list just lists job id #'s Sep 22 11:35:49 u delete by job id # Sep 22 11:36:21 i assume if u added a job u know its job id # (you stored it - return from Add) Sep 22 11:36:22 is there a race between list and delete? Sep 22 11:36:27 (oops missed in return type list) Sep 22 11:36:55 lindi-: someone else deletes it before you do Sep 22 11:36:56 nb Sep 22 11:37:03 job #'s wont change once assigned Sep 22 11:37:11 i'll just make them monotonically increasing Sep 22 11:37:55 so u'd get a list like: 7, 24, 6778, 6779, 7032, 9434 Sep 22 11:38:02 for example Sep 22 11:43:24 any more input Sep 22 11:43:26 throw it now Sep 22 11:43:28 i'm building it Sep 22 11:43:31 i need it Sep 22 11:43:34 at doesnt do what i need Sep 22 11:43:57 (no rtc wakup support, broken on oe/om anyway, also only has minute resolution, not seconds) Sep 22 11:51:08 ok Sep 22 11:51:12 more funeral work... Sep 22 12:38:52 Hmm - so I try out the testing image from downloads.openmoko.org/daily on a GTA01, and usb networking doesn't work for some reason. Of course there is no terminal application installed, and the stupid boots stop you seeing any errors from the kernel during boot, so you have absolutely no way of diagnosing or fixing the problem. Sep 22 12:40:19 and /etc/resolv.conf is missing, so you can't even use the installer out of the box. Sep 22 12:40:51 d'oh Sep 22 12:42:32 ok, looks like a reboot fixes the usb problem. Sep 22 12:43:42 now in the installer program, the phone goes into suspend in the middle of updating the package lists Sep 22 12:43:56 and then you have applications who's names are so long that the extend off the edge of the screen Sep 22 12:44:22 so you have three apps called "qtopia-phone-x11-..." and don't know what each one really is Sep 22 12:45:02 well, after 5 minutes with om2008.9, all it does is reaffirm why I use fso-testing ... Sep 22 12:46:21 heh, now it suspends in the middle of installing an application ... Sep 22 12:47:21 cool, Categories/Miscellaneous has no less than 6 applications all called "qtopia-phone-x11-"(edge of screen) Sep 22 12:49:27 then packagekitd preventing running opkg manually until you kill it is just completely idiotic Sep 22 12:51:21 and of course the classic "Install a terminal program, but you don't get a virtual keyboard with that ..." Sep 22 12:52:22 and suspending when you're ssh'd into the phone via usb Sep 22 12:53:21 * rwhitby keeps tapping the screen, so that opkg can complete successfully Sep 22 13:20:07 Hey Sep 22 13:24:15 Ainulindale, here? :) Sep 22 13:47:00 quickdev: yes Sep 22 13:47:05 I was looking for you Sep 22 13:47:27 fine..I've got some questions Sep 22 13:47:48 Sorry for the delay I've been quite busy this week end Sep 22 13:48:00 no problem..sorry for disturbing that often Sep 22 13:48:39 I'd like to extend libframeworkd-glib for led interface usage..for example the vibrator Sep 22 13:49:13 atm the files are named wrong in my opionion. frameworkd-glib-device.c should be frameworkd-glib-gsm-device.c, shoudn't it? Sep 22 13:49:36 Yes it should, I thought about that in the beginning of the project Sep 22 13:49:41 But I was far too lazy Sep 22 13:50:07 ok Sep 22 13:50:14 I'd rather name them by their correspondent interface Sep 22 13:50:16 how about my patch? reviewed? if not...no problem ;) Sep 22 13:50:27 i.e. ogsmd/frameworkd-glib-ogsmd-device.c Sep 22 13:50:37 and the same thing for the headers of course Sep 22 13:50:43 yeah.. Sep 22 13:50:48 but it's not a real urgent matter Sep 22 13:50:58 yes, it isn't Sep 22 13:51:28 Your patch is good, no problem with that indeed Sep 22 13:52:02 could you apply it? Sep 22 13:52:18 I could but I won't, just to annoy you Sep 22 13:52:19 :-) Sep 22 13:52:59 where's the prob? ;) Sep 22 13:53:24 Weird, I have fuzz Sep 22 13:54:58 fuzz? what does that mean? Sep 22 13:55:16 Use of uninitialized value $module in concatenation (.) or string at /usr/local/bin/ciabot.pl line 182. Sep 22 13:55:19 Weird Sep 22 13:55:20 quickdev: applying the patch Sep 22 13:55:54 ciabot? mh? ;) Sep 22 13:56:08 mickey|dinner: problem with freesmartphone git Sep 22 13:56:20 just ignore that message Sep 22 13:56:24 freesmartphone.org: 03ainulindale 07libframeworkd-glib * r7b5afc2fc30a 10/src/ (frameworkd-glib-dbus.c frameworkd-glib-dbus.h): Applied quickdev error handling patch for dbus internal errors. Thanks quickdev ! Sep 22 13:56:24 freesmartphone.org: 03ainulindale 07libframeworkd-glib * r7b5afc2fc30a 10/src/ (frameworkd-glib-dbus.c frameworkd-glib-dbus.h): Applied quickdev error handling patch for dbus internal errors. Thanks quickdev ! Sep 22 13:56:53 Ainulindale, fine, thanks Sep 22 13:57:41 mickey|tw: how's taipei ? Sep 22 13:57:52 warm Sep 22 13:57:54 just arrived today Sep 22 13:57:56 pretty tired Sep 22 13:57:59 going to bed soon Sep 22 13:58:02 I guess so Sep 22 13:58:17 mickey|tw, how long did you fly? Sep 22 13:58:35 quickdev: ~12.5 hours Sep 22 13:58:49 mickey|tw: found anything interesting in the log?;) Sep 22 13:59:11 yes Sep 22 13:59:24 big thing i omitted Sep 22 13:59:31 dunno how you could trigger the state Sep 22 13:59:37 but i need to add it Sep 22 13:59:37 :) Sep 22 14:00:09 right, nice, thanks; I'll try to set up frameworkd from git in the mean time, then Sep 22 14:03:36 i completely forgot one state in my call state handler :) Sep 22 14:03:49 will add that this week Sep 22 14:21:25 mickey|tw: spotted you on linkedin :-) Sep 22 14:21:35 and rwhitby Sep 22 14:23:10 Ainulindale, is it just another community? Sep 22 14:25:23 it's a social network for professionals Sep 22 14:25:32 It may or may not be useful, it depends Sep 22 14:25:50 It is useful to me as the identity & access management field is pretty tiny these days, but growing and growing Sep 22 14:40:48 Ainulindale, in order to send the PIN from phonegui to ophonekitd I implemented a phonegui_provide_pin(char* pin) method. That's ok, isn't it? (Considering the whole concept) Sep 22 14:41:08 For this kind of thing, I don't think so Sep 22 14:41:25 It's another layer, could be useful, yes, considering what I told you before Sep 22 14:41:27 Ainulindale, how to do it otherwise? Sep 22 14:41:29 But for this kind of calls Sep 22 14:41:35 I think it is just some overhead Sep 22 14:41:41 I think you should use directly lfg Sep 22 14:41:43 do it directly? Sep 22 14:41:48 ok Sep 22 14:41:55 That's my opinion and just mine Sep 22 14:42:07 I think things have to go through ophonekitd for non user generated events Sep 22 14:42:14 Such as "incoming call" Sep 22 14:42:22 But that's all Sep 22 14:42:28 Doing more is doing overhead, to my eyes Sep 22 14:42:40 ophonekitd is there just to link the UI to dbus and manage events Sep 22 14:42:47 the ui do its stuff internally as it wants Sep 22 14:42:51 +oes Sep 22 14:42:59 Man my english is awful today Sep 22 14:43:30 Ainulindale, and the ui application has to handle wrong sim aswell? I mean...it isn't redisplayed if PIN is wrong, right? Sep 22 14:44:30