**** BEGIN LOGGING AT Fri Dec 21 02:59:56 2007 Dec 21 04:17:58 anyone know if themes are set globally in android? Dec 21 04:18:06 or can you switch themes on the fly from within your applications? Dec 21 05:09:22 hrmm, what's the deal with TabHost/TabWidget, unusable? Dec 21 08:06:28 teehee, optimizing branch logic order is fun. Dec 21 08:06:45 and i enjoy that it matters in Android :P Dec 21 08:32:20 davidw: i saw that you released Hecl Dec 21 08:32:28 or rather, your interpreter. Dec 21 08:32:44 a very early alpha demo version to play with if you know what you're doing, yes Dec 21 08:33:10 well, is there any android software that doesn't fit that category? Dec 21 08:35:50 good point Dec 21 08:38:50 sleepy time, night folks Dec 21 08:42:06 later Dec 21 09:15:18 how can i remove the installed apps from the emulator? Dec 21 17:16:18 yawn Dec 21 17:20:54 * Lede yawns back Dec 21 17:21:55 i wish dpad up/down wrapped in the menu widget. Dec 21 17:43:47 lame, my android rss reader can use xkcd.org's rss feed, but the windows mobile one that i actually use daily can't :( Dec 21 17:43:59 it formats the images thinking its being clever and ends up making every comic strip unreadable Dec 21 17:59:57 lol thats kinda funny Dec 21 18:00:15 course even if it didnt junder android, its your software so you could fix it. yay! Dec 21 18:00:22 i love being a geek Dec 21 18:15:37 well, this wouldn't be easy to fix, since it would be webkit's bug. Dec 21 18:15:50 but fortunately no such bug exists, so... ;) Dec 21 18:34:26 this is the least productive work day of all time :) Dec 21 19:38:24 Is there anyway to load up a theme dynamically? or does teh theme have to be dumped out into xml format and compiled with the application before it can be used by the application? Dec 21 20:01:47 jerkface03: don't know, there probably is, but it is perhaps undocumented or hackish. Dec 21 20:02:07 the XML file is inflated at runtime, so there is obviously code switching it dynamically. Dec 21 20:08:48 i just had an interesting thought about testing Android on "real" handsets. Dec 21 20:09:24 what would be so difficult about creating a local networked KVM-type protocol between a real phone (say a WM smartphone) and an Android host running in the emulator? Dec 21 20:10:24 you could connect the WM phone to your machine and have it connect to it over the bridged ethernet interface, and write a program that simply proxies the display and keyboard from and to the emulated phone running matching server software. Dec 21 20:11:20 you wouldn't even need to do this in Java, accepting limitations of the Android stack. you could just hook into the framebuffer and keyboard devices and write a deamon on Linux. Dec 21 20:15:14 i have no idea about some of those things, but it sounds like a cool idea Dec 21 20:16:11 it wouldnt really be testing it on a real handset in terms of the technology or hardware but it would help you get an idea what it would look like in your hand, how good your interface really is etc Dec 21 20:19:59 cybereagle: and that is arguably the most important aspect to analyze. Dec 21 20:25:33 the hardest thing is knowing when your getting the scale right Dec 21 20:25:38 cos some screens are gonna be bigger than others Dec 21 20:25:51 and its hard to judge on a computer screen even in the emu Dec 21 20:29:57 agreed. Dec 21 20:30:21 i find that the emulator seems to look "large", but i think that is just because my monitor is 22". Dec 21 20:30:45 and in fact, it is. if i hold my 320x240 WM smartphone up to it, it seems almost twice as large. Dec 21 20:31:01 yeah Dec 21 20:31:37 also, i'm interested in the "feel" of my application shortcuts and stuff. Dec 21 20:31:43 i may actually do this hack. Dec 21 20:31:54 Does anyone know if the back button is programmable? will i get a KeyEvent when it's pressed? Dec 21 20:32:02 i have sufficient knowledge of the linux kernel, assuming that there isn't some mysterious google obscurity in there somewhere. Dec 21 20:35:46 actually, the biggest challenge would be writing the windows mobile part, since windows mobile is a freaking mess. Dec 21 20:56:31 first things first, i'm gonna try to compile fbgrab on Android :) Dec 21 21:04:20 i think the best way to do this would be to actually implement a vnc server. Dec 21 21:04:29 it would be a lot easier to test and do the client-side magic too Dec 21 21:17:25 yeah i was just thinking about that Dec 21 21:21:10 teehee, nice. Dec 21 21:21:17 $ adb shell data/hello Dec 21 21:21:17 depth: 16 Dec 21 21:21:17 width: 320 Dec 21 21:21:17 height: 240 Dec 21 21:21:26 successfully queried the framebuffer Dec 21 21:21:52 my only concern is how efficient i am required to be :\ Dec 21 21:22:16 because detecting drawing might be very tricky on the emulated host. Dec 21 21:22:23 the emulator runs slow as hell Dec 21 21:23:57 is there anyway to register a custom uri with the default browser on android? Dec 21 22:04:14 sweet, got fbgrab running :) Dec 21 22:17:25 now to move onto keyboard emulation. Dec 21 22:20:00 i sure hope the VNC protocol is simple and light weight, because i don't want to spend weeks working on this shit Dec 21 22:27:53 jasta: you want to write a vnc server for Android Dec 21 22:27:53 ? Dec 21 22:28:16 well, yes, but only as a development tool. Dec 21 22:28:41 get VNCj, hack to work, easier than reimplementing the RFB proto from scratch Dec 21 22:28:50 i have a VNC viewer on my windows mobile phone and i thought it might make sense to "test" applications on a mobile form factor by translating keyboard and video. Dec 21 22:29:29 i doubt vncj would work well. Dec 21 22:29:36 why's that? Dec 21 22:29:49 * Stephmw uses it Dec 21 22:30:01 depending on the complexity of the protocol, which i do not know yet, and the sophistication of the Android Java interfaces. Dec 21 22:30:21 my concern is that Android's java interfaces alone would be insufficient to support this effeciently. Dec 21 22:30:35 VNC isn't at all efficient ;) Dec 21 22:30:39 and therefore a server would need to exist at the framebuffer and keyboard device driver layer. Dec 21 22:31:11 Stephmw: I mean effeciency in this sense as critical. I feel it would be impossibly slow with the interfaces that Android provides. Dec 21 22:31:23 You would ideally need to hook into the SurfaceFlinger, which you cannot. Dec 21 22:31:36 so it would be the best alternative to go one step lower and try for the framebuffer device directly. Dec 21 22:31:36 jasta: VNC is quite bad at latency, regardless of where you hook in Dec 21 22:31:47 Stephmw: it isn't the protocol in particular that i'm concerned about. Dec 21 22:32:02 jasta: sure, but it'll be your limiting factor Dec 21 22:32:04 it's the screen grabbing. Dec 21 22:33:36 if the protocol is simple, it would be best, i think, to implement it from scratch as a daemon running atop the Linux kernel. Dec 21 22:34:04 I would be tempted to use something else as the transport - but feel free to have a play :) Dec 21 22:34:14 my thought is that an abstraction layer may be required with a second daemon that interprets the screen captures and decomposes them into simpler primitives for efficiency. Dec 21 22:34:30 jasta: if it's any help, I'm the guy behind OpenJDK's fbtoolkit project Dec 21 22:34:36 Stephmw: Well, the reason I'd use VNC is because, well, I don't want to write any WM C#.NET code. Dec 21 22:35:08 * Stephmw remembers that he needs to update the fbtoolkit.dev.java.net pages Dec 21 22:35:15 and i suppose that is a stupid reason if efficiency ends up being a bottleneck. Dec 21 22:36:23 i am only really interested in this project if i can finish it in the next week, too :) Dec 21 22:36:30 i don't want to spend a lot of time on a hack like this Dec 21 22:37:09 i thought it might be nice, though, to be able to plug my WM smartphone into my machine and virtually interact with the emulator and thus my other developing applications. Dec 21 22:37:45 I'll be rooting for you - it's doable timescale-wise Dec 21 22:37:58 dunno about whether Android will cope with it Dec 21 22:38:11 i don't think android will be the problem, but rather qemu :) Dec 21 22:38:20 qemu might be too sucky and inefficient to let this work. Dec 21 22:38:30 especially since my current strategy is to do this in C, as a daemon. Dec 21 22:38:52 i'd prefer not to get into the kernel, since i don't have a good dev platform set up for that. Dec 21 22:44:14 although i have to set one up anyway hehe, gotta get a keybaord driver in place Dec 21 22:46:12 does Android have /dev/input ? Dec 21 22:46:28 yeah Dec 21 22:46:55 i'm not familiar with that subsystem though. Dec 21 22:47:14 well, I know how to read it (it really isn't hard)... but I don't know if you can inject with it Dec 21 22:47:38 i wouldn't assume so Dec 21 22:51:05 well anyway, i'm building the kernel sources now Dec 21 22:53:08 cool :) I've moved off embedded for tonight, hacking on my photography :) Dec 21 22:53:49 is polling /dev/fb0 sensible or is there some other system that i should be looking at for capturing screen contents? Dec 21 22:54:12 I thought I saw a blog about missing /dev/fb0? Dec 21 22:54:14 i mean, ideally i'd like to have access to the SDL library that Google put together, but that's clearly not an option. Dec 21 22:54:28 but otherwise, sure, polling is a good first step Dec 21 22:54:30 Stephmw: well, it's called /dev/graphics/fb0, but it's there and usable. Dec 21 22:54:36 i already wrote a little test program and ran it in the emulator Dec 21 22:54:41 dumped the screen nicely Dec 21 22:54:47 cool, rgb565? Dec 21 22:54:52 yeah, natively. Dec 21 22:55:06 yeah, the fb kernel driver does the adaptation Dec 21 22:55:12 also supports rotation :) Dec 21 22:55:26 that kernel built remarkably fast hehe Dec 22 00:04:49 shit, the android kernel doesn't have module support Dec 22 00:05:02 ? Dec 22 00:05:08 like kernel modules Dec 22 00:05:45 really? thats a shame Dec 22 00:06:17 i wonder how esy it would be to replace the kernel on a real device Dec 22 00:06:29 probably not easy, but possible. Dec 22 00:06:40 you see when it says ANDROID with the flashing kernel? Dec 22 00:06:47 i bet there is some bootloader magic going on right there ;) Dec 22 00:06:51 yeah, they have the source they use for the kernel so you could modify it but keep it cmpatable Dec 22 00:06:52 but we don't have the source to it sooo Dec 22 00:06:56 it'd just be the delivery method Dec 22 00:07:03 anyone know how it boots? Dec 22 00:07:11 well i think they have the kernel source Dec 22 00:07:15 you mean what the bootloader is? no idea. Dec 22 00:07:33 yeah i know they have source, i'm rebuilding the whole kernel just to add 1 stinking module Dec 22 00:07:46 hehe Dec 22 00:07:57 what were you looking to add jasta? Dec 22 00:08:16 a custom serio port driver that will let me simulate scancodes Dec 22 00:08:36 well when we get the source or whatever else, the first thing i'm gonna look at doing is changing ANDROID_ to something else hehe Dec 22 00:09:54 i love the android bootloader image Dec 22 00:10:17 it ought to say loading though, for ages i thought the emulator wasnt working when i first started it up Dec 22 00:12:16 i hope they keep it Dec 22 00:12:22 they could improve the framebuffer bootloader though :) Dec 22 00:12:28 err bootsplash Dec 22 00:14:05 it probablyonly supports more primative graphics, but a nice animated droid waving his arms around might be fun lol Dec 22 00:20:09 i doubt that actually Dec 22 00:20:14 i bet its the full framebuffer glory Dec 22 00:20:25 minus hardware acceleration, though. Dec 22 00:20:28 although maybe Dec 22 00:21:43 ohhh program idea! Dec 22 00:21:52 ansi art simulator via the framebuffer Dec 22 00:21:54 hehehe Dec 22 00:22:17 you mean aalib? Dec 22 00:22:38 well that might be cool too Dec 22 00:22:53 but i meant more the color ansi art of the bbs ear Dec 22 00:22:56 *era Dec 22 00:22:57 i think you should focus on things perhaps more directly applicable to users. Dec 22 00:23:07 i'm a user, i want it ;) Dec 22 00:23:16 well, ok Dec 22 00:23:21 lol these are just silly ideas for my own amusement Dec 22 00:26:00 ahoy. will android allow my applications to react to events such as an incoming call? Dec 22 00:40:30 hey y. Dec 22 00:40:31 ;) Dec 22 00:44:36 haha. Dec 22 00:44:42 What are you doing in here? Dec 22 00:44:53 You an android developer? Dec 22 00:45:18 i'm getting there i haven't developed java in a couple years getting back into it Dec 22 00:45:48 also sorry about before i did mean it in good humor but i was kind of in a pissy mood on top of it. Dec 22 00:46:09 Yeah, that happens. Dec 22 00:46:25 Its totally cool. I could have been less of a dick. =) Dec 22 00:46:32 What do you want to do with Android? Dec 22 00:49:48 presence awareness Dec 22 00:50:13 i think what you looking for is intent in terms of reacting to an incoming call Dec 22 00:52:46 intent? Dec 22 00:55:38 http://code.google.com/android/reference/android/app/Instrumentation.html Dec 22 01:09:18 thanks Zer0Her0 Dec 22 01:11:45 welcome Dec 22 01:11:53 i'm not a complete idiot ;) Dec 22 01:12:44 Only about disk images in OS X. Hardly a fatal arena. haha Dec 22 01:19:13 no no i know bout that area actually i was just being an ass about it Dec 22 01:21:51 I know. and I am just hasslin' ya =) Dec 22 01:25:06 :-) aight that's only fair **** ENDING LOGGING AT Sat Dec 22 02:59:56 2007