**** BEGIN LOGGING AT Wed Jan 30 02:59:56 2008 Jan 30 02:59:59 yeah, I kinda assumed sane caching Jan 30 03:00:05 pjut the right layers of software in and it becomes a moot discussion Jan 30 03:00:26 RAM is getting cheaper all the time, too, so probably caching can be done without trouble. Jan 30 03:00:55 * mjr is most interested though at the moment in getting that damn gprs multiplexing, so I could actually keep the phone net-connected Jan 30 03:05:48 i want a toy on hand, so i have search this kind of stuff, i found n810 and moko1973 almost have the some price, which one is most funny? Jan 30 03:06:20 I think n810 is more funny:) Jan 30 03:06:39 most software is ready, bigger screen. Jan 30 03:06:58 * mjr doesn't like the nokia stuff for being too proprietary. Certainly, it's a finished product if you're into that sort of thing... Jan 30 03:07:02 but we can only modify its software. Jan 30 03:07:10 but there is no phone function, while neo1973's phone function is not ready yet;) Jan 30 03:07:36 well, the phone function seems to work decently (unless you get a bad build :] ) Jan 30 03:07:39 can you modify neo's hardware? Jan 30 03:07:58 i don't know Jan 30 03:08:00 mjr: event on the gta01? Jan 30 03:08:16 yangh, why not? Jan 30 03:08:37 but it looks like you can add some new stuff to neo1973 Jan 30 03:08:47 I haven't try the recent build yet. Jan 30 03:08:48 well, SpeedEvil is adding some stuff Jan 30 03:09:25 what kinda stuff bomb? Jan 30 03:09:38 so sure you can interface some additions on it if you're handy Jan 30 03:09:59 I think he's been talking wlan and accelerometers, don't know how the progress goes Jan 30 03:10:41 n810 has skype and voip which can certainly make phone calls Jan 30 03:10:55 ljp always with the irrelevancies Jan 30 03:11:16 but you can't leave your building with skype Jan 30 03:11:23 mjr: no, someone mentioned n810 has no phone functions, which is wrong Jan 30 03:11:25 (also skype is part of the nokia deal one spesifically doesn't like) Jan 30 03:11:30 3G, etc Jan 30 03:11:39 or plain gprs Jan 30 03:11:43 billtang: I have a 3g phoen that says differently Jan 30 03:12:01 actually, skype over gprs doesn't sound sane Jan 30 03:12:08 for sure, but i don't -_-! Jan 30 03:12:08 ljp, only it's obvious what he meant (though maybe not to you, admittedly) Jan 30 03:12:27 vininim, indeed Jan 30 03:12:33 ljp: when we mean "phone function", that means there is a sim card slot, gsm attena... Jan 30 03:13:11 can we port android to neo1937, i think that ui is looks better Jan 30 03:13:15 then you need to specify gsm phone Jan 30 03:13:33 you should ask google about that Jan 30 03:13:40 android is vapoeware so far Jan 30 03:13:42 that's neo Jan 30 03:14:02 if android ever gets actually freed, "we" presumably can if "we" want to Jan 30 03:14:07 so as gta2 Jan 30 03:15:11 I read somewhere that tomorrow is a decision day wrt to hardware revision. Jan 30 03:15:19 somewhere = the wiki Jan 30 03:15:31 Btw, does OpenMoko only use C, or also C++? Jan 30 03:15:39 I wonder what kind of offerings the hardware deities like Jan 30 03:15:52 Tronic: gcc Jan 30 03:15:54 hence, g++ Jan 30 03:16:19 ewon: That's the compiler, but what language is the current software written in? Jan 30 03:16:27 i want use python, i am learning it now Jan 30 03:16:35 I believe the base stuff is in C Jan 30 03:16:37 a mix of c, c++ and python I would say Jan 30 03:16:45 python is my preference as well Jan 30 03:16:47 Python is way too slow even on PC, so I can't see how it could work on a mobile phone. Jan 30 03:17:03 there's a lot you don't see, yes Jan 30 03:17:09 and there is that gnome yet another language mentioned at the planet sometimes Jan 30 03:17:09 (yes, I do know that they have ports for Symbian and whatnot, but that doesn't mean it being usable) Jan 30 03:17:15 but it's running on n800 Jan 30 03:17:25 Vala, I think Jan 30 03:17:34 yah, that one Jan 30 03:18:26 anyway, certainly there are limits to what kinds of things you want to do with python on the neo, but there's a wide range of stuff that's useful and doesn't require maximum processing efficiency Jan 30 03:18:39 If neo is based on Intel's x86... Jan 30 03:18:45 isn't Jan 30 03:18:55 python is fine for high level gluing blobs together with logic and flow Jan 30 03:19:02 u would NOT write a game enigne in python Jan 30 03:19:05 or a rendering engine Jan 30 03:19:14 or anything seriously compute-intensivre Jan 30 03:19:15 It'll be more easier to development software. Jan 30 03:19:18 in any way Jan 30 03:19:52 for instance, I'd beg to tell me why my python remote control simply doesn't work on a mobile phone Jan 30 03:19:55 ye i know, i just like the Interpreter, it get me a different way to think Jan 30 03:20:46 Even if you use Python only for logic in places where speed does not matter, you need to load the interpreter and the libraries that you use. Jan 30 03:21:14 correct Jan 30 03:21:29 andthat is a price you pay for saving writing lots of extra housekeeping code in c/c++ Jan 30 03:21:32 If you have to mix Python with C++ (as you would do the parts that need to do run fast in C++, right?), things get really messy. Jan 30 03:21:37 yes,so what, is that very hard for neo1973? Jan 30 03:21:48 I gather that you can call C code just fine, but interfacing with C++ is trickier. Jan 30 03:22:00 i PERSONALLY will kep it all in c as its easier for me Jan 30 03:22:02 i'm a c coder Jan 30 03:22:06 just extern C it so it doesn't get munged Jan 30 03:22:12 i have no qualms doing it all in c Jan 30 03:22:21 its easy and fast for me Jan 30 03:22:24 others do not Jan 30 03:22:25 the hell I would do it in C++ either :] Jan 30 03:22:26 to each their own Jan 30 03:22:41 raster: You don't need any housekeeping code (you mean memory management, right) in C++. Jan 30 03:22:51 Tronic: yes u do Jan 30 03:22:52 Only in C. Jan 30 03:23:49 What C++ does worse than Python is the syntax. Python has piles of nice syntactic sugar on it, making the code look nicer (less unneeded characters). Jan 30 03:24:23 raster: I can see that you are a C coder :P Jan 30 03:25:06 My C++ programs typically don't have any delete's or other explicit memory management anywhere. It just works automatically. Jan 30 03:25:10 Tronic: i can ss you are a c++ coder. Jan 30 03:25:16 fact is housekeeping code exists in c++ Jan 30 03:25:22 u have new and delete methods Jan 30 03:25:28 if u have any special requirements Jan 30 03:25:29 E.g. this game http://ultrastar-ng.sourceforge.net/ Jan 30 03:25:35 like data sharing with ref counting etc. Jan 30 03:25:38 it exists Jan 30 03:25:43 u pit is in a different place Jan 30 03:25:45 what kind of job for c coder, embed stuff? Jan 30 03:25:45 but its there Jan 30 03:25:50 Zero deletes in there. Jan 30 03:25:52 in c i can do the same things as c++ Jan 30 03:25:54 i can do OO Jan 30 03:26:04 i can do inheritance Jan 30 03:26:05 classes Jan 30 03:26:07 c++ is more maintainable C Jan 30 03:26:08 etc. etc. etc. Jan 30 03:26:12 there, I said it. Jan 30 03:26:15 raster: can you do templates? Jan 30 03:26:30 That program also uses no ref counting and in fact I haven't find any situation where that would be useful, yet. Jan 30 03:26:34 *found Jan 30 03:26:57 But when I do, I'll just use boost::shared_ptr and have it handled automagically, too. Jan 30 03:27:17 raster: i hate it when i see f1(double x), f2(int x), f3(X x), .. all doing the same thing Jan 30 03:27:58 raster: OOP w/ virtual functions you can do in C, even though the syntax is quite horrible. Jan 30 03:27:59 borg_: partially. depends on the situation Jan 30 03:28:01 well actually Jan 30 03:28:04 it is possible Jan 30 03:28:10 in the end it just is more code Jan 30 03:28:16 One thing that you cannot do in standard C is automatic module registration on linkage. Jan 30 03:28:22 Tronic: i dont mind the syntax Jan 30 03:28:28 i find c++ syntax quite horrible Jan 30 03:28:29 (but you can do it in C++) Jan 30 03:28:38 its a personal preference Jan 30 03:28:57 and no - in c its not automatic Jan 30 03:29:03 but u can make it work Jan 30 03:29:22 dlopen() is your friend Jan 30 03:29:25 (and fridns) Jan 30 03:29:30 for doign all sorts of evil Jan 30 03:30:44 E.g. that game uses audio output modules that are detected automatically. You can try to compile them all, then link only those that compiled successfully and the program will know which modules got linked it. Also works for dynamically loaded libs (just dlopen them, no need for dlsym), but obviously the loading part is not standard C++ anymore. Jan 30 03:30:47 as for ref counting Jan 30 03:30:49 i live on it Jan 30 03:30:56 i dont know how i would survivie without Jan 30 03:31:00 i dont duplicate data Jan 30 03:31:05 i just ref it many times Jan 30 03:31:18 maybe that dictates a differnec in performance Jan 30 03:31:30 and maybe why c programmers tend to have more optimal c programs Jan 30 03:31:36 they have to do the dirty work - and they tend to care Jan 30 03:31:42 I put the data on a higher level (on whatever owns all the things that uses that data). Jan 30 03:31:44 and thus they tend to pay atention to detail Jan 30 03:32:00 Thus it will have lifetime greater than the users. Jan 30 03:32:23 that would simply be impossible for most of what i do Jan 30 03:32:25 s/uses/use/ Jan 30 03:33:13 it would also bloat out memory use and cpu overhead like no business Jan 30 03:33:13 I have it hard to believe that it would be impossible in "most that you do". Jan 30 03:33:21 it would Jan 30 03:33:34 As I personally haven't yet became across a single situation where it wouldn't be possible. Jan 30 03:33:43 i write libraires, image processing, redienrm, toolkits widgets, gui apps etc. Jan 30 03:33:54 data comes from the lower levels Jan 30 03:33:57 not fed in Jan 30 03:34:02 And I have worked with quite many different kinds of projects. Jan 30 03:34:19 if the same data is "loaded" again 0 simply a 2nd reference to it is returned avoiding a load and decode Jan 30 03:34:21 and another copy Jan 30 03:35:14 most objects runtime when refrerneces - get refcountied Jan 30 03:35:30 to make sure all handles work even after delete until veryone gives up0 their refs Jan 30 03:36:03 So you put the data on the higher level object and have it only loaded once. If you really want to defer the loading until needed, put a boost::scoped_ptr inside that high level object and load the resource only when requested. Jan 30 03:36:37 the data is not under control of the higher layer Jan 30 03:36:49 Only if you are very short on RAM, but have fast loading times and because of that want to unload the resource instantly when no-one uses it, you will need ref counting. Jan 30 03:36:57 as completely independnant systems/code bases do the same work Jan 30 03:37:05 use the same lower levels to create their objects and manage Jan 30 03:37:11 they CANT share data otherwise Jan 30 03:37:18 these are not monloithic programs Jan 30 03:37:49 they are highly modular, well layers libraries, modules and independant code blocks Jan 30 03:38:11 Should every openmoko source file have the gpl copyright header at the top? Jan 30 03:38:15 they cant manage data on a higher level withotu it being private to the higher level and thus needing duplication between non-sharing code blocks Jan 30 03:38:22 A non-trivial library should always have some kind of master object that can be instantiated in main() or whatever and all objects used would use that as a master object. Jan 30 03:38:52 The library initialization should be done in the constructor of that and the deinit in the destructor. Resources can be stored there nicely, too. Jan 30 03:38:53 the point of it thus they SHARE all data via their common layer Jan 30 03:39:04 the common layer handles everythng with refcounting Jan 30 03:39:18 yes Jan 30 03:39:21 A lot of libs try to do this in static context, but that is problematic as the programmer no longer has control on when the library is running. Jan 30 03:39:24 and if multiple systems init and destroy Jan 30 03:39:27 u need to refocunt Jan 30 03:39:39 i know Jan 30 03:39:47 thats why i am telling you refcounting is needed Jan 30 03:39:55 been doing this kind of stuff for a decade Jan 30 03:40:14 Different instantiations of the master object are fully separate, thus no need for refcounting. Jan 30 03:40:45 If you wish to make them shared, you will use static context, but I really don't like that approach. Jan 30 03:41:20 i dont think you are thinking the same thing Jan 30 03:41:31 Another problem with that is having to do synchronization on all library functions that access the shared state. Jan 30 03:41:42 the data is all dynamic Jan 30 03:41:44 all of iut Jan 30 03:41:49 As the same library may easily be used from multiple threads. Jan 30 03:42:06 Image *img = load_ijmage("/path/to/my.png"); Jan 30 03:42:15 then some completely separate piece of code does Jan 30 03:42:21 Image *img2 = load_ijmage("/path/to/my.png"); Jan 30 03:42:29 in a totally unrelated bit of code in another toolkit Jan 30 03:42:36 BOTH load_image calls return the same pointer Jan 30 03:42:43 same reference to the same Image object Jan 30 03:43:05 in order to make sure that ALL loads release their handle with an unref Jan 30 03:43:15 u need to refcount it to make sure it is deleted only when the last reference goes Jan 30 03:43:23 u CANT do that without refcounting Jan 30 03:43:30 True. Jan 30 03:43:31 UNLSs u do a complte new Image * object reference Jan 30 03:43:47 in whcih case its a duplicate of the data and multiple loads of the same data on disk Jan 30 03:43:53 thus consuming mountains of cpu and ram Jan 30 03:44:00 But frankly, I have never came across the need for such an optimization (even using a same bitmap in multiple completely unrelated places). Jan 30 03:44:21 If the places need to share the same bitmap, they are likely also related... Jan 30 03:44:53 (btw, in C++ you would do Image img("/path/to/foo.png"); instead of exposing any pointers to the user) Jan 30 03:45:52 well then - i'd be wary of using any programs you write Jan 30 03:46:15 as i can tell u statistically references to images in my code are in the 204digit range Jan 30 03:46:17 err Jan 30 03:46:21 The Image object would then contain a boost::shared_ptr m_data that its constructor gets somehow from the static context. Jan 30 03:46:21 2-4digit Jan 30 03:46:34 10-1000+ refs are normal Jan 30 03:46:54 doing them your way would probably mean memroy and load blot of several dozen if not 100 times Jan 30 03:47:03 Still, no need for deletes or destructors. Everything works automatically, except for fetching the resource, of course. Jan 30 03:47:11 I can't see how Python can be better in this sense. Jan 30 03:47:12 if it wrre not for doing shared caching and refocunting most of hat i do would not be possible Jan 30 03:47:36 dunno about python Jan 30 03:47:39 dont really care :) Jan 30 03:47:47 but refcounting solves all this Jan 30 03:47:52 saves duplicating objects Jan 30 03:48:08 Well, in C you need to do the cleanup manually, which may lead to errors (leaks or double-frees) very easily. Jan 30 03:48:13 image data is allocated and loaded dnyamically from on-dsik data Jan 30 03:48:17 or somewhere else Jan 30 03:48:30 and in c yes - u have to unref (free) the object Jan 30 03:48:47 but that just unrefs and at 0 ref dumps back into the speculative cache Jan 30 03:48:53 that cdoe actually doesnt have bugs Jan 30 03:49:00 there are no leaks Jan 30 03:49:13 :) Jan 30 03:49:13 no handles are lost Jan 30 03:49:18 as the object handles are tacked by canvas objects Jan 30 03:49:27 canvas objects are tracked by the canvas wrapper subsystem Jan 30 03:49:37 a non-deleted object is a visible entitiy Jan 30 03:49:42 on screen Jan 30 03:49:42 I mean if someone using the resource forgets to call the release function, your ref count will be too high and the resource will never get released. Jan 30 03:49:47 if its still there u see it Jan 30 03:49:59 and its just incorrect functioning of the program, not a leak Jan 30 03:50:00 :) Jan 30 03:50:05 Even worse, if someone accidentally calls it twice, the resource will be freed, even though others still have pointers to it => segfault. Jan 30 03:50:20 and yes - if they call it twice it may segv Jan 30 03:50:29 in practice that doesnt happen as all objects have magic numbrs Jan 30 03:50:31 ref counts Jan 30 03:50:44 and deletes are eve deferred by 2 render cycles for status caluclation purposes Jan 30 03:50:51 u can check magic numbers and delete flags Jan 30 03:51:15 the code even has segv catching so if it segv's during an object access it'll trap, save. back out safely and complain Jan 30 03:52:14 if u destroy the owning canvas all objects in it are freed automatcally Jan 30 03:52:29 in practice - it's not a problem that crops up Jan 30 03:52:45 frankly the probems are logic ons where u just get behavior wrong Jan 30 03:52:59 segv's and such from forgetting to unref (free) basically dont hapen Jan 30 03:56:39 Wait, you said there was nothing shared, but they belong to a canvas after all? :) Jan 30 03:58:53 DYNAMIC ENTRY Jan 30 03:59:16 Tronic: they are within the same canvas but know nothing of eachtoher Jan 30 03:59:47 eg in one corner i have a vidoe playback widget with conrols that happen rto use the same images as a playlist widget.object that happens tio share imagery Jan 30 03:59:54 they are independent objects Jan 30 03:59:58 How are they linked to that canvas? Jan 30 04:00:06 written separarely even in different libraries for example Jan 30 04:00:18 they happen to use the same canvas api to create/destory and manipulato objects Jan 30 04:00:36 Ah, so you only have one canvas at all times? Jan 30 04:00:40 (i.e. static context) Jan 30 04:01:06 I thought canvas as an object that you can create (as a part of a GUI, for instance). Jan 30 04:01:26 no Jan 30 04:01:31 u may have multiple canvases Jan 30 04:01:40 those are alco created and destroyd on the fly as needed Jan 30 04:01:52 Then how does load_image(...) know which canvas to link with? Jan 30 04:02:01 that was an example Jan 30 04:02:07 if u want the actual code i can copy and paste it Jan 30 04:02:13 Not really :P Jan 30 04:02:27 its a multi-stage setup Jan 30 04:02:30 u create empty image Jan 30 04:02:33 then u set the file source Jan 30 04:02:43 creation is relative to a parent canvas Jan 30 04:08:33 To summarize the discussion: (1) resources need to be ultimately owned by something, be it (a) static context, (b) some explicitly constructed object, or (c) garbage collector (not discussed here). (2) freeing a resource can be done (a) when deleted explicitly (if the programmer thinks that it is no longer needed, e.g. boost::scoped_ptr), (b) when the containing object goes out of scope (regular local or member variables, ... Jan 30 04:08:39 ... boost::scoped_ptr) or (c) when there are no more references to it (ref counting, GC). Jan 30 04:10:17 You prefer 1a + 2c, I prefer 1b + 2b (rarely _also_ 2a, and 2c when I come across a situation where it is useful). Jan 30 07:43:15 has anyone here played around with gllin? Jan 30 07:43:23 bit, why Jan 30 07:43:27 I'm trying to figure out why it wont boot on startup of the phone Jan 30 07:43:40 what's starting it? Jan 30 07:43:52 I've linked the init.d script to the rc* folders but no luck so far Jan 30 07:43:59 thw script itself works fine though Jan 30 07:44:10 strace -o foo script Jan 30 07:44:16 maybe? Jan 30 07:49:01 jobs: http://www.demonware.net/jobs.html Jan 30 07:51:28 maybe Jan 30 08:01:54 interesting, I get even more accurate GPS co-ordinates on my Neo rather than my Garmin Jan 30 08:01:57 impressive Jan 30 08:05:21 what do you mean by more accurate? Jan 30 08:08:17 SpeedEvil: not that googlemaps is perfect but by comparing the co-ordinated I pulled out of my garmin compared with the ones I pulled from my Neo is more accurate mapped Jan 30 08:08:25 haha. Jan 30 08:08:27 coordinates Jan 30 08:08:29 bla Jan 30 08:08:36 google-maps is not accurate to that level often Jan 30 08:08:57 Alas. Jan 30 08:08:58 SpeedEvil: nah, but I like it in lala-land Jan 30 08:14:29 unfortunately, the only practical way to do this is to A) find a survey point surveyed at better than 10cm Jan 30 08:14:48 B) average for 48 hours in one spot with a good sky view Jan 30 08:17:23 SpeedEvil, even B doesn't work if your equipment has systematic flaws... Jan 30 08:17:55 No, but most GPSs are not vulnerable to that sort of errors so much Jan 30 08:18:03 with a good sky view that is Jan 30 08:18:11 I guess that's probably true nowadays Jan 30 08:18:58 I do remember the embarrasing errors in internal coordinate system transformations in some previous generations.... Jan 30 08:19:37 those lead to a systematic error of tens of meters Jan 30 08:19:53 Yeah - that's a point. Jan 30 08:20:14 I was more meaning the internal representation of the position Jan 30 08:20:29 converting to an external coordinate system can be screwy Jan 30 08:20:44 no shit, it was a stupid idea to start with Jan 30 08:20:56 what was? Jan 30 08:21:10 converting internally Jan 30 08:21:27 You mean it should output in the native GPS format? Jan 30 08:21:48 I was just talking about the coordinate sytem transformations Jan 30 08:22:10 good morning Jan 30 08:23:01 hi Jan 30 08:44:42 openmoko: 03chris * r3981 10/trunk/src/target/OM-2007.2/applications/openmoko-dialer2/ (ChangeLog src/phone-kit/moko-network.c): Jan 30 08:44:42 openmoko: * src/phone-kit/moko-network.c: (on_network_registered): Jan 30 08:44:42 openmoko: Unset operator name when unregistered from network Jan 30 10:06:00 * * OM Bug 1202 has been created by tick(AT)openmoko.com Jan 30 10:06:01 * * Packagekit may crash while installing Jan 30 10:06:02 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1202 Jan 30 10:06:10 * * OM Bug 1203 has been created by chris(AT)openedhand.com Jan 30 10:06:12 * * neod emits far too much debugging output Jan 30 10:06:14 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1203 Jan 30 10:07:01 * * OM Bug 1204 has been created by plr(AT)udgaard.com Jan 30 10:07:02 * * Mobile keypad won't go away when selecting qwerty keyboard Jan 30 10:07:03 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1204 Jan 30 10:07:11 * * OM Bug 1205 has been created by plr(AT)udgaard.com Jan 30 10:07:13 * * Qwerty keyboard covers button when screen in locked mode Jan 30 10:07:15 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1205 Jan 30 10:07:23 * * OM Bug 1206 has been created by plr(AT)udgaard.com Jan 30 10:07:25 * * Even when "Dim only, don't lock" slected, screen is locked Jan 30 10:07:27 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1206 Jan 30 10:26:38 openmoko: 03werner * r3982 10/trunk/src/target/u-boot/patches/ (console-ansi.patch gta02-bootmenu.patch): (log message trimmed) Jan 30 10:26:38 openmoko: Bring the boot menu also to GTA02. Since the Glamo is slower than a dumb frame Jan 30 10:26:38 openmoko: buffer, we also need to accelerate the clearing of the screen. Jan 30 10:26:38 openmoko: console-ansi.patch: Jan 30 10:26:38 openmoko: - drivers/video/cfb_console.c (process_sequence): to clear the screen just Jan 30 10:26:42 openmoko: clear the screen, instead of slowly scrolling away its content Jan 30 10:26:44 openmoko: gta02-bootmenu.patch: Jan 30 10:33:30 openmoko: 03werner * r3983 10/developers/werner/uboot-glamo-accel-broken.patch: Jan 30 10:33:30 openmoko: Failed attempt to add Glamo hardware acceleration to u-boot. Jan 30 10:33:30 openmoko: Note that everybody else (Kdrive, etc.) uses command queues, not Jan 30 10:33:30 openmoko: MMIO. So it may well be that this hardware interface just doesn't Jan 30 10:33:30 openmoko: work. Jan 30 10:55:00 * * OM Bug 1207 has been created by garnser(AT)gmail.com Jan 30 10:55:01 * * Unable to unlock phone, OpenMoko logo unmovable Jan 30 10:55:02 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1207 Jan 30 11:08:01 hmm Jan 30 11:08:15 why are people too lazy to check for existing bugs? Jan 30 11:08:59 * Stephmw believes it's genetic Jan 30 11:09:14 * mickeyl does the "Resolve-as-duplicate" thing Jan 30 11:09:51 *blink* Jan 30 11:22:58 mickeyl: I assume the duplication is obvious, and a degree in software development is not required to see the duplication? :p Jan 30 11:27:35 openmoko Jan 30 11:28:12 what is that ? Jan 30 11:28:23 openmoko Jan 30 11:28:39 what is that ? Jan 30 11:28:48 openmoko Jan 30 11:28:56 tevar Jan 30 11:29:02 wibbit: of course Jan 30 11:29:03 openmoko Jan 30 11:29:11 morthor is a bot := Jan 30 11:29:13 morthor is a bot :) Jan 30 11:29:16 is he? Jan 30 11:29:33 for sure :) Jan 30 11:30:39 why Jan 30 11:33:02 openmoko: 03mickey * r3984 10/trunk/src/target/OM-2007.2/daemons/neod/src/buttonactions.c: neod: silence debug output if not wanted (set MOKO_DEBUG environment variable). closes #1203 Jan 30 11:33:20 i'm not Jan 30 11:41:34 omg ... Egypt's Internet backbone is almost broken Jan 30 11:41:39 why Jan 30 11:42:38 mummy pr0n Jan 30 11:43:03 why Jan 30 11:43:13 CVirus: edge routers between egypt and europe are dropping 50% pacjets Jan 30 11:43:17 why Jan 30 11:48:08 why Jan 30 11:49:38 <[PI]Sami> we has moar bandwidth then :D Jan 30 11:49:51 why Jan 30 11:50:05 LOL Jan 30 11:50:17 why Jan 30 11:51:08 morthor: stfu Jan 30 11:52:07 * ScaredyCat looks for a chanop Jan 30 11:55:01 why Jan 30 12:03:49 'morning Jan 30 12:04:56 why Jan 30 12:06:01 I ASKED WHY Jan 30 12:06:59 because Jan 30 12:07:15 mickeyl: ping Jan 30 12:07:45 pong Jan 30 12:07:49 ping pong Jan 30 12:07:52 ping Jan 30 12:07:54 pong Jan 30 12:07:58 ping pong Jan 30 12:08:01 ping Jan 30 12:08:03 pong Jan 30 12:08:09 ping pong Jan 30 12:10:06 at last Jan 30 12:10:27 has it been noisy in here, Stephmw ? Jan 30 12:11:00 tonyg: just some idiot that finally disconnected Jan 30 12:24:28 Stephmw: pong Jan 30 12:25:28 afaik it was to kick morthor Jan 30 12:25:54 aye Jan 30 13:34:23 abraxa__: ++ Jan 30 14:00:45 n Jan 30 14:15:10 Lo ppl Jan 30 14:36:57 mickeyl: Why aren't you set as the owner of the openmoko-terminal component in bugzilla? I have a patch in there for you. It is for a very low-priority bug and I just did it to try the vala language, but I assume it wouldn't take you much time to have a look at it. Jan 30 14:37:12 But feel free to wait if you have to do stuff with higher priority. :-) Jan 30 14:40:03 sounds like historical reaosons Jan 30 14:40:09 which number? Jan 30 14:42:12 #1191 Jan 30 14:43:12 k, thanks. assigning that to me. Jan 30 15:39:00 ScaredyCat: Was that a ping? ;) Jan 30 15:39:11 no Jan 30 15:39:26 karma Jan 30 15:39:35 ~karma abraxa__ Jan 30 15:39:36 abraxa__ has neutral karma Jan 30 15:39:46