**** BEGIN LOGGING AT Wed Aug 06 03:00:02 2008 Aug 06 04:09:03 what do I have to do to make my DatePicker 'active?' i click around and nothing happens Aug 06 04:15:36 Is only the Dialog version of it interactive? Aug 06 04:42:16 Oh. It only responds to direction input. lame. Aug 06 04:42:48 a shame theres no way to select the year easily either Aug 06 04:56:27 Has anyone else made a good date widget? specifically for inputting birthdate? Aug 06 04:59:01 dead tonight Aug 06 06:01:16 yawn Aug 06 06:03:10 howdy Aug 06 06:03:48 coding... ;) Aug 06 06:03:55 it's good to be back into it. Aug 06 06:04:02 i'm pleased to find that my inspiration has returned. Aug 06 06:04:08 taking a bit of a break.... i posted my messages apk Aug 06 06:04:16 getting ready to write a game Aug 06 06:04:21 very nice Aug 06 06:04:32 i'm writing my solitaire engine right now :) Aug 06 06:04:43 something simple.... i haven't messed with opengl on android yet Aug 06 06:04:59 sigh, i forgot watch expressions is broken :( Aug 06 06:06:05 ah the fun of an outdated sdk :) Aug 06 06:06:24 hehe, a lot of this old code is so funny to look at Aug 06 06:06:31 how badly i understood all this stuff hehe Aug 06 06:06:35 :) Aug 06 06:06:44 i'll be back tomorrow, going to relax a little bit Aug 06 06:06:58 enjoy :) Aug 06 06:07:51 later :) Aug 06 06:08:01 what timezone are you in? Aug 06 08:12:02 morning Aug 06 08:12:31 Morning. Aug 06 13:11:53 morning Aug 06 14:18:02 this is kinda cool: Aug 06 14:18:03 http://androidguys.com/2008/08/06/scripting-your-android-device/ Aug 06 14:18:39 might be a good way to create a plugin architecture Aug 06 14:26:46 is that real? Aug 06 14:30:06 looks pretty real Aug 06 14:34:45 zhobbs_: why would you want to use that to make a plugin architecture? Aug 06 14:35:18 Dalvik allegedly will be able to dynamically load dex files, which would be a much better way to implement a traditional plugin architecture with Java Aug 06 14:35:54 for sure Aug 06 14:37:00 "allegedly will be able to" vs actually works in m5 Aug 06 14:39:13 I'm disappointed that there's no access to the JS Engine in Webkit w/o instantiating a WebView. Aug 06 14:39:29 I would have been an excellent addition to support application scripting. Aug 06 14:40:22 zhobbs_: Well, I believe that claim simply on the basis that it would be trivial to support. Aug 06 14:41:43 jasta: I don't see that it's quite that trivial Aug 06 14:43:01 though most of the capability to launch dex files must be present to simply load apps Aug 06 14:43:47 didnt read the whole article, but does it depend on netbeans running? Aug 06 14:43:51 i mean eclipse Aug 06 14:43:57 uh ugly mistake Aug 06 14:44:14 tomgibara: That is what makes it trivial. Loading a dex file from some arbitrary stream is a necessity for Dalvik to even work. Aug 06 14:44:46 jasta: It's not necessarily trivial within the context of an executing application though is it. Aug 06 14:44:55 tomgibara: It is, certainly. Aug 06 14:45:17 jasta: :) You certainty amuses me Aug 06 14:45:42 jasta: First thing that comes to mind: what do you do with the constant pools? Aug 06 14:45:47 Dalvik would just need a new [small] set of instructions that allow whatever code path is taken to initially load classes and launch an application to be available to compiled dex code. Aug 06 14:45:50 jasta: Strings in particular? Aug 06 14:46:18 tomgibara: You treat them just as you treat any other class which must be initialized at runtime, and potentially destroyed later. Aug 06 14:46:59 Dalvik, as any sensible VM would, does not load every class the system has when your application starts. Aug 06 14:47:03 jasta: You are lamentably rash - String constants are privileged in Java. Aug 06 14:47:25 jasta: Privileged by being automatically interned. Aug 06 14:48:19 tomgibara: That is irrelevant, complexity introduced only by your inability to reason how a reasonable VM must work. Aug 06 14:49:00 This VM in particular was implemented to necessitate loading classes dynamically, which it does. It simply has no instruction set for doing that in compiled dex files. Aug 06 14:49:45 Or if it does have those instructions, they are buggy or haven't been enabled yet by the class to dex translation. Aug 06 14:50:00 jasta: Haha... I point out a complexity in something that you immediately assume to be trivial, and you say I'm unable to reason about VMs Aug 06 14:50:49 tomgibara: If you wish to explain why you believe this to be a problem, you're more than welcome. Aug 06 14:51:08 I have explained why I believe it not to be, after all. Aug 06 14:51:13 jasta: You say this VM in particular was implemented to necessitate loading classes dynamically - why? Aug 06 14:51:14 what's the class loader model in dalvik? Aug 06 14:51:37 muthu: A damn sight simpler than the one in Java I hope. Aug 06 14:51:39 tomgibara: Because I know that to be true from obversation. Classes load dynamically. Aug 06 14:51:56 they load when necessary Aug 06 14:52:03 they said they have optimized the dalvik vm Aug 06 14:52:08 its obvious from the log output of a starting apk Aug 06 14:52:12 jasta: That is different from having two independent sources of classes Aug 06 14:52:33 tomgibara: The source of the class is simply a software architecture problem, which is a simple one to solve at that. Aug 06 14:52:34 Optimization in this context might include assuming there was one dex file only Aug 06 14:52:53 jasta: Irrelevant Aug 06 14:53:32 jasta: I'm pointing out the additional complexity associated with introducing new classes and constants into an established address space. Aug 06 14:53:46 tomgibara: That is the exact argument I would give. If the source of the class data (dex file, in particular) is irrelevant, then loading classes dynamically would just be simply fitting it into the existing design. Aug 06 14:53:50 jasta: My string constant issue being just one simple possible example. Aug 06 14:53:53 i don't think its all that complex tomgibara Aug 06 14:53:59 tomgibara: And I'm telling you that Dalvik already loads classes dynamically. Aug 06 14:54:07 This is plain to see from observation alone. Aug 06 14:54:24 Again, I don't dispute it defers class resolution Aug 06 14:54:25 And because of this, it would be trivial to extend it to do this from an arbitrary dex stream instead of one elsewhere on the system. Aug 06 14:54:25 just open up logcat :) Aug 06 14:55:46 I have 10 yrs of Java development experience had once reported a cross ClassLoader security bug that took Sun over a year to fix. Aug 06 14:55:59 I just googled this and it looks like it's just a stupid bug. Dalvik already does support this, you just have to hack around it. Aug 06 14:56:22 namely, the classes.dex must be inside a jar file. I jar'd up a Aug 06 14:56:22 classes.dex file and fed it into the code for the osgi class loader Aug 06 14:56:22 (see http://blog.luminis.nl/luminis/entry/osgi_on_google_android_using Aug 06 14:56:23 ) and all was happy. Aug 06 14:56:32 this was a quote from a groups message Aug 06 14:56:51 Let me explain clearly: defering the resolution of classes is not necessarily the same thing as being able to accomodate all of the information in a dex file. Aug 06 14:57:23 My contention is that jasta said it is trivial, when there is no evidence to me that supports the assertion. Aug 06 14:58:01 Perhaps not, but I just demonstrated that this functionality already exists. Aug 06 14:58:15 I personally expect it to be straightforward because the application dex is probably loaded into a context that supports the system dex file already. Aug 06 14:58:53 This is a completely different matter than infering that because classes are lazily resolved, its easy to load a new dex file Aug 06 14:59:16 tomgibara: I fail to see why. Would you mind explaining some of those complexities? Aug 06 15:00:05 jasta: I can't because I don't know anything about the Dalvik VM - but I know that the complexities in Hotspot seem monsterous to me Aug 06 15:00:22 Such as? Aug 06 15:01:49 Hotspot can identify when there is only one implementation of a given interface and then optimize calls to the interface based on assuming that it must be the given class. When a new class is loaded into the VM it 'backs out' the optimization, possibly decompiling JITed code before recompiling it Aug 06 15:02:21 Similarly it will inline calls to a method that is not fil Aug 06 15:02:25 final Aug 06 15:02:27 Well, Dalvik doesn't use JIT code, so that complexity is gone. Aug 06 15:03:06 I know - I said I could only refer to Hotspot because i don't know much about Dalvik Aug 06 15:03:29 Backing out of optimizations is certainly not something I considered initially, however it would seem reckless of Google to create optimizations taht would be difficult to back out of when they attempt to support a 1.0 feature. Aug 06 15:03:33 But I repeat again - your belief that there's no complexity because you can't see it is laughable. It may be true, or it may be hidden. Aug 06 15:04:24 i still wouldn't call it a complexity Aug 06 15:04:51 tomgibara: Well, I actually meant to imply that I don't know about the complexity you were referring to, hence I eventually just asked because you would not otherwise explain it. Aug 06 15:05:15 Although I remain unconvinced that it would be truly nightmarish to support this functionality, given deferred class resolution. Aug 06 15:05:42 Nightmarish, perhaps, if you never saw the feature coming and designed an entire VM assuming it would not be possible. But then shame on you, right? :) Aug 06 15:05:46 I don't say it's nightmarish, but I would be suprised if it was wholy trivial either. Aug 06 15:06:41 tomgibara: Well, you mentioned a bug that took Sun over a year to fix. I assumed you meant that as an example of a "nightmare" ;) Aug 06 15:07:24 No, just a dumb mistake that allowed class domains to leak references - inducing a cross domain browser bug. Aug 06 15:07:33 And actually, I think in the case of optimizations here that it only makes sense to consider the optimizations non-trivial, not the dynamic class loading. Aug 06 15:10:36 whoa jasta, i forgot how quick it is to use svn when its not Google code servers :) Aug 06 15:10:38 Optimizations that must make arbitrary assumptions about the pattern of a program, and then must back out of them when those assumptions are shown to be false, would be complex in general. Aug 06 15:10:51 michaelnovakjr: lol Aug 06 15:10:58 michaelnovakjr: hey, i committed some code last night for my Solitaire game. Aug 06 15:11:09 it does fire up and draw, poorly. no game logic eyt, though. Aug 06 15:11:09 i saw that this morning Aug 06 15:11:19 it built cleanly for me Aug 06 15:11:31 btw, i'm on the east coast :) Aug 06 15:38:22 AttractiveApe: the Messages app is up on android-random Aug 06 15:38:45 Oh really! Aug 06 15:38:56 Is it awesome? Aug 06 15:47:05 i would hope so AttractiveApe :) Aug 06 15:50:31 AttractiveApe: i would at least expect you to find it better than the current SMS app there Aug 06 15:52:58 :p k Aug 06 16:37:19 yawn Aug 06 16:39:40 howdy Aug 06 16:40:39 good morning Aug 06 16:41:36 how's it going? Aug 06 16:42:27 evening Aug 06 16:45:00 how's it going? Aug 06 16:45:01 * jasta eats some yogurt for breakfast Aug 06 16:45:28 yogurt is so carzy to me Aug 06 16:45:59 yea? Aug 06 16:46:19 what other food do you eat that's still alive while you're eating it? Aug 06 16:46:55 I'm pretty sure we carry a bunch of weird stuff that is still alive Aug 06 16:47:12 i'm sure we do carry it, but i don't think we eat it regularly Aug 06 16:47:13 morning Aug 06 16:47:14 do we? Aug 06 16:47:34 i mean like isn't that the point of cooking raw meats well? so you definitely aren't eating anything alive? Aug 06 16:47:57 i could be wrong, i don't know that much about food science Aug 06 16:48:04 I meant my company carries stuff. Aug 06 16:48:07 but it's weird to me to think about eating an active culture Aug 06 16:48:33 "According to the currently adopted definition by FAO/WHO, probiotics are: ‘Live microorganisms which when administered in adequate amounts confer a health benefit on the host’." Aug 06 16:48:35 and that instead of being dangerous, it's good for you Aug 06 16:48:36 beer is often alive Aug 06 16:48:37 jasta: blue cheese? Aug 06 16:49:05 that HTC diamond looks like a slick phone Aug 06 16:49:16 yeah. many cheeses I believe. Aug 06 16:49:40 cheeses in the US are pasteurized by law. Aug 06 16:50:38 jasta: That's a bit sad. Aug 06 16:50:41 but i guess some stuff survived pasteurization, but not much. i'm sure it's nothing like yogurt when it's done though. Aug 06 16:51:17 even imported cheese?.. Aug 06 16:51:19 canned and bottled beer in the US is also pasteurized, so again, ... Aug 06 16:51:32 yakischloba: i believe in order for it to be imported it must have been pasteurized. beer included. Aug 06 16:51:47 guinness, for example, is made differently when exported to the US than other parts of the world. Aug 06 16:52:19 i believe that is because of pasteurization Aug 06 16:54:25 michaelnovakjr: here's hoping it's similar to the android device they're launching ;) Aug 06 16:55:01 interesting you bring that up because the layout looks a lot like the emulator when closed Aug 06 16:55:03 hmm. I was pretty sure you could have unpasteurized cheese as long as it was aged a certain amount or something Aug 06 16:55:10 that is the button icons below the screen :) Aug 06 16:55:11 best bet of android out on august? Aug 06 16:55:21 phone Aug 06 16:55:24 Hai-Fai: huh? Aug 06 16:55:33 Don't hold your breath Aug 06 16:55:37 Hai-Fai: lol, no ;) Aug 06 16:55:42 well, it was someone's guess that first phones might be out on august Aug 06 16:56:02 apparently someone who doesn't come around much :) Aug 06 16:56:10 Hai-Fai: here's a strategy. assume it will be launched on Jan 1st, 2009, then enjoy your pleasant surprise when it's any bit earlier. Aug 06 16:56:16 and it might have been 1 year ago :D Aug 06 16:56:18 :) Aug 06 16:56:19 and only mild frustration when it really launches in February. Aug 06 16:57:03 Hai-Fai: for your reading pleasure: http://gizmodo.com/5032966/htcs-android-phones-still-on-track-for-q4-2008 Aug 06 16:57:08 nah, I just wish it would be my great christmas present =) Aug 06 16:57:36 but at least new sdk would be great Aug 06 16:58:04 yes well, we're getting encouraging news that the SDK should be out much before phones, as they plan to release the SDK when they ship the final versions for testing/certification/etc Aug 06 16:58:18 but i believe we're still some ways off from that. Aug 06 16:58:31 indeed Aug 06 16:58:44 a month or more would not surprise me, for example. Aug 06 16:58:53 yep, it's sad Aug 06 16:59:00 stop speculating guys :) Aug 06 16:59:09 you're only hurting yourself Aug 06 16:59:13 Hai-Fai: well, it's sad and then it's also entirely expected, so... Aug 06 17:00:06 * jasta raises an eyebrow Aug 06 17:00:14 ...that article i just posted... Aug 06 17:00:27 "By Jason Chen"? Aug 06 17:00:33 no it's another Jason Chen Aug 06 17:00:40 i should hope so!!! Aug 06 17:00:43 inconvenient hah Aug 06 17:00:43 he (gizmodo's) said so when Android came out Aug 06 17:03:00 * jasta is enjoying designing a touch screen interface for a solitaire game Aug 06 17:03:21 i plan to support number pads, touch screen, and dpad... Aug 06 17:03:27 and multiple layouts, yay Aug 06 17:03:47 jasta: it's still a boring game :)) Aug 06 17:03:54 Freecell isn't boring Aug 06 17:03:58 you've just never played it obviously Aug 06 17:04:00 it is too Aug 06 17:04:15 Hey has anyone made another widget for date input? I found the DatePicker highly unsuitable for entering birth dates Aug 06 17:04:20 you really abuse that word, "obviously" Aug 06 17:04:34 yakischloba: next SDK has a much better one Aug 06 17:04:40 I hate it when mom and dad fight. Aug 06 17:04:55 romainguy_: that would work well for birth dates? Aug 06 17:05:01 yes Aug 06 17:05:04 ok Aug 06 17:05:10 I'll hold off on creating my own then -_- Aug 06 17:05:40 romainguy_: are you sure, not Klondike? Aug 06 17:05:47 i really enjoy Freecell :\ Aug 06 17:05:59 but yeah Klondike is basically the worst game ever Aug 06 17:07:30 Someday I hope jasta and romainguy_ get to share an office. They really deserve each other ;) Aug 06 17:07:30 I hope there's a good online multiplayer spades/hearts game Aug 06 17:07:46 quick pickup games like the ones that come with windows Aug 06 17:07:51 yakischloba: no way :p Aug 06 17:07:58 mm chess Aug 06 17:08:00 lol Aug 06 17:08:18 i would harrass him to a degree previously thought impossible. i am not above nerf guns. Aug 06 17:08:31 well Aug 06 17:08:35 romainguy_: don't be shy. your affection is poorly masked. Aug 06 17:08:40 at Sun I built my own office catapult using spare server racks Aug 06 17:08:56 so I'm not afraid of nerf guns Aug 06 17:09:01 romainguy_: pics or it didn't happen. Aug 06 17:09:11 amen Aug 06 17:09:26 lol Aug 06 17:09:28 the first version: http://www.flickr.com/photos/romainguy/82384700/ Aug 06 17:09:39 haha, he's got that on speed dial Aug 06 17:09:47 engineering schematics or it didn't happen. Aug 06 17:10:14 romainguy_: hah, that's actually not bad. Aug 06 17:10:36 i will out-do this and create a trebuchet that launches coffee mugs. Aug 06 17:10:39 haha Aug 06 17:10:45 I also had a nice collection of crossbows made out of various office items Aug 06 17:11:13 :) Aug 06 17:11:14 romainguy_: did you see that Mythbusters episode where they made paper crossbows and got one shot off with it that they concluded was potentially lethal. Aug 06 17:11:27 yes :) Aug 06 17:11:40 so basically i should call the ATF and have you arrested? :) Aug 06 17:11:42 see? Match made in heaven. Aug 06 17:11:47 yup Aug 06 17:13:13 romainguy_: I thought _I_ was anal about keeping my desk clear. Aug 06 17:13:31 my desk is an explosive mess of papers. Aug 06 17:13:38 most of them are even dirty and tea stained. Aug 06 17:14:07 yakischloba: I hate to have my desk messy Aug 06 17:14:23 yakischloba: I also can't go to sleep if the closet is open so... :)) Aug 06 17:14:28 but i don't know why it's that way, because i don't even use the papers on my desk. Aug 06 17:15:13 romainguy_: my girlfriend can't sleep if there is a light on in the hallway just barely coming through the side of the door. she actually wakes up in the middle of the night if someone turns the hallway light on. Aug 06 17:15:23 heh. i like it clean as well. I don't know about the closet thing though. Aug 06 17:24:54 jasta: you have paper on your desk? Aug 06 17:24:58 Bad sysadmin. BAD. Aug 06 17:25:04 i just leave it there for ages Aug 06 17:28:17 jasta your a sysadmin? Aug 06 17:28:31 i have a hybrid systems/engineering position Aug 06 17:28:43 so i wouldn't say it's a sysadmin job, but i wouldn't say it's a programming gig either. Aug 06 17:28:46 i had one of those last year... Aug 06 17:29:05 i enjoy it in principal, but i feel like i have done all that i can at my current company Aug 06 17:29:14 so i'm bored Aug 06 17:29:24 yea, i also know that feeling :) Aug 06 17:30:56 the more i do it, the more i'd like a job doing mobile programming Aug 06 17:31:06 its more interesting than i thought Aug 06 17:31:18 michaelnovakjr: because you're not doing J2ME :) Aug 06 17:31:23 or Java ME Aug 06 17:31:39 granted, Android is an excellent development platform :) Aug 06 17:32:00 even the little bit of iPhone development i've done so far has been interesting Aug 06 17:32:14 and yes, even though I don't like the iPhone model :) Aug 06 17:33:37 C# on win mobile wasn't too bad either Aug 06 17:34:11 romainguy_: i turned down a J2ME job just last week hehe Aug 06 17:34:47 michaelnovakjr: Don't you ever say that. Mobile development on WinMo is an abomination. Aug 06 17:35:09 its not that bad Aug 06 17:35:26 Have you tried doing anything non-trivial? Or deploying an app to support multiple resolutions/orientations? Aug 06 17:35:44 There's not even a standard list widget that works. You have to create it yourself, and even do all the drawing yourself. Aug 06 17:36:06 there's a list widget Aug 06 17:36:15 yeah, but have you tried using it? :) Aug 06 17:36:16 jasta what version of win mobile? Aug 06 17:36:19 yeo Aug 06 17:36:27 yea, i've used it in an app before Aug 06 17:36:37 i didn't say it was elegant.... but it wasn't bad Aug 06 17:36:42 If I had more patience to boot VMware, I'd send you some sample code I worked with which can trivially break it. Aug 06 17:37:00 I was developing .NET CF 2.0 on WinMo 5 Aug 06 17:37:17 i eventually gave up because it was so pathetic. Aug 06 17:37:25 i only had a winmo5 device for about 5 days, i didn't like it Aug 06 17:37:33 i did some stuff on win mobile 6 Aug 06 17:37:49 wm6 _may_ have .NET CF 3, but i doubt it. Aug 06 17:37:55 my favorite thing about deploying WinMo app is when you had to deploy them for 3 CPU architectures Aug 06 17:38:15 :) Aug 06 17:38:18 romainguy_: the screen problem is even worse! Aug 06 17:38:29 blech, it's bad. Aug 06 17:58:40 romainguy_: how do i create a tiling drawable? i have a background pixmap i want to apply to a view but it needs to tile. Aug 06 17:58:54 you wait for the next SDK :) Aug 06 17:58:58 wait, i think i know this already. i must use Shader.TileMode, huh? Aug 06 17:58:58 or you create your own Drawable Aug 06 17:59:01 yes Aug 06 17:59:22 yeah i've used that before, i'll look it up Aug 06 18:38:18 romainguy_: that was not so painful. i just use a PaintDrawable and set a tiling BitmapShader... Aug 06 18:38:22 used* Aug 06 18:38:38 hopefully that's not miserably inefficient or something Aug 06 18:38:43 herriojr: did you get your app working? Aug 06 18:39:07 jasta, not anymore inefficient than working on an old sdk :) Aug 06 18:39:33 but then again, who said anything about being efficient? :) Aug 06 18:39:45 i'm talking about efficient code Aug 06 18:39:47 and i care :) Aug 06 18:40:21 :) Aug 06 18:41:04 if i didn't care, i wouldn't be programming right now :) Aug 06 18:54:43 michaelnovakjr: so, i want to make it when you click on a card stack, it highlights it to indicate you have selected it Aug 06 18:54:50 with a simple orange glow around it or whatever Aug 06 18:55:11 do you think i should just adust all the layout/measure/drawing code to pad for this space? Aug 06 18:55:25 what was the link you guys gave me the other day to try and download an app to a kaiser running android? Aug 06 18:55:27 or do you think i should wrap it in another View which specifically does this padding/glow effect? Aug 06 18:55:53 can't you actually add a highlight to the layer you desire Aug 06 18:56:02 michaelnovakjr: what do you mean? Aug 06 18:56:07 herriojr: http://code.google.com/p/android-random Aug 06 18:56:14 thanks Aug 06 18:58:56 jasta, nevermind i was thinking of something else, i would say wrap it in another View, depending on how many you would have to adjust in the first option Aug 06 18:59:09 it might be the more efficient approach Aug 06 18:59:22 well it would certainly be the simplest, that way my CardStackView wouldn't need to hold all that logic Aug 06 18:59:23 herriojr: no problem Aug 06 18:59:42 it would just be a simple selectable view that when selected, draws a glow around whatever view it is contaning Aug 06 18:59:46 jasta, yea... plus that logic has to be computed.... which might bog it down Aug 06 19:02:59 it would be nice if the View class had a method like setSelector( Drawable d ) like in AbsListView Aug 06 19:03:12 or maybe not the View class, maybe ViewGroup Aug 06 19:03:23 that wouldn't help in this case Aug 06 19:03:29 no? Aug 06 19:03:40 the issue is that the glow is *around* the widget. outside of its drawing rectangle. Aug 06 19:03:49 ah, good point Aug 06 19:04:00 that's why i could either hack that padding into the view i'm using it for, or create a view that encapsulates another but adds padding for this glow. Aug 06 19:05:12 yea Aug 06 19:06:21 i'd say encapsulating might be the way to go, unless of course it has to redraw the whole screen anyway. Aug 06 19:06:44 is there a reference hardware implementation that run Andriod? Aug 06 19:06:50 linnuxxy: not officially Aug 06 19:09:57 jasta: i c.. thnx Aug 06 19:10:27 who build that touch.screen device that horovitz demoed in that first intro video? HTC? Aug 06 19:29:13 is anyone familiar with the kaiser bootkit for android? Aug 06 19:30:26 what's your issue? Aug 06 19:30:58 so, I'm trying to figure out where I need to install applications, on the android.bin (initrd image) or the android-data.bin Aug 06 19:31:16 android-data i would assume Aug 06 19:31:29 do you use a linux computer? Aug 06 19:31:36 yeah Aug 06 19:31:48 mount the android-data and put the apk in the app directory Aug 06 19:31:48 so, they both seem to pretty much be the same thing Aug 06 19:32:01 I can't mount it, I have to unpack it Aug 06 19:32:08 via cpio Aug 06 19:32:20 ok, do you see an app directory? Aug 06 19:32:30 yeah, I've tried that, but it never works Aug 06 19:32:35 under data/app Aug 06 19:32:45 in the android-data.bin Aug 06 19:32:58 are there other apk's in there? Aug 06 19:33:12 Sms.apk and one other Aug 06 19:33:31 that is where you should put your apk Aug 06 19:33:38 what happens when you do? Aug 06 19:34:04 one of two things....it either has a hell of a hard time booting up android, or it never shows up in the menu Aug 06 19:34:53 it could be the way you are repacking the android-data Aug 06 19:35:25 find ./ | cpio -H newc -o > ../new-data Aug 06 19:35:44 then I run gzip it up and rename it to android-data.bin Aug 06 19:36:41 when I try and install it in the initrd image through the same way, it won't boot android Aug 06 19:36:48 it freezes while loading /dev Aug 06 19:38:31 how are you running gzip? Aug 06 19:38:42 gzip new-data Aug 06 19:39:12 try gzip -9 Aug 06 19:39:19 gzip -9 new-data Aug 06 19:39:20 ok Aug 06 19:39:39 I think I may have found the problem, though....I don't see in initrd where android-data.bin is referenced Aug 06 19:41:56 why would that be? Aug 06 19:42:42 I don't completely understand how the kernel knows to load android-data.bin, let me check to see if it is compiled into the kernel Aug 06 19:44:41 I'm thinking that they may be using a file they don't actually need, which may be why I've been so confused why it hasn't been working Aug 06 19:44:48 at least I hope that's the issue Aug 06 19:46:37 yup, that's my issue Aug 06 19:46:53 they didn't document how android.bin and android-data.bin were used Aug 06 19:53:35 herriojr: i was thinking about answering you but you have me on ignore so meh. Aug 06 19:58:34 how do I have you on ignore? Aug 06 19:59:08 no, you weren't on ignore as far as I know Aug 06 19:59:24 maybe i'm thinking of someone else. Aug 06 19:59:43 android.bin seems to be an initrd, which is the responsibility of the bootloader. Aug 06 19:59:48 in this case, that would be haret. Aug 06 20:00:51 android-data.bin seems to just be an alternative initrd. Aug 06 20:01:15 doesn't seem to be much difference between the two. let me check diff. Aug 06 20:02:10 yeah, I already diffed them Aug 06 20:02:10 android-data is different. the differences seem to be in modem setup, and at commands Aug 06 20:02:13 not much different Aug 06 20:02:15 so it's likely that android-data is just for getting GPRS Aug 06 20:03:30 but again, haret loads these files. you instruct it which one through default.txt Aug 06 20:03:32 herriojr: hopefully you put the HttpDownloader app on there Aug 06 20:03:43 that way you don't have to do that again :) Aug 06 20:03:45 I will once I verify that everything works Aug 06 20:04:08 ;) Aug 06 20:04:44 haret is flaky on the kaiser, so it may take some time to verify :) Aug 06 20:05:56 herriojr: my mistake, i just look trhoguh my logs and it was indeed someone else :) Aug 06 20:06:08 some troll that came in just before you and i got the handles mixed up. Aug 06 20:06:39 michaelnovakjr: i utilized custom attributes in the Solitaire app Aug 06 20:06:43 ;) Aug 06 20:06:53 oh nice :) Aug 06 20:07:07 let's you control card orientation and visibility of the card stack view i have. Aug 06 20:09:23 sweet Aug 06 20:09:34 i plan to start looking more at all the graphics stuff tongith Aug 06 20:09:37 i'll commit it soon Aug 06 20:09:51 i'm going to write up a simple puzzle game Aug 06 20:09:54 i actually have a pretty solid handle of the graphics stuff i think. i spent a lot of time futzing with it (and romain *grin*) for five Aug 06 20:10:18 cool Aug 06 20:10:50 i was comtemplating writing up a simple little AppManager for the phone that basically wraps the HttpDownloader Aug 06 20:12:40 not a bad idea. Aug 06 20:13:06 something that could simply give an interface to search for the apk's and then install them. Aug 06 20:13:33 also manage updates as well so the user would have to look for updates Aug 06 20:13:40 nothing too overkill though Aug 06 20:14:32 i guess it could be the beginning of an open source apk manager Aug 06 20:15:05 you might want to look at something like SlideME. perhaps they already have an implementation? Aug 06 20:20:41 that seems interesting Aug 06 20:23:04 romainguy_: in attrs.xml, how do you specify a styleable attribute that uses string literals, but isn't a string format? Aug 06 20:23:23 for example, view visibility can be visible, invisible, or gone, but nothing else. Aug 06 20:23:30 how is that achieved? Aug 06 20:25:02 it's an enum Aug 06 20:25:06 you can also use flags Aug 06 20:25:20 Aug 06 20:25:20 Aug 06 20:25:21 Aug 06 20:25:21 Aug 06 20:25:21 Aug 06 20:25:23 Aug 06 20:25:25 Aug 06 20:25:26 ahh, excellent. Aug 06 20:25:32 i figured it would be something like that Aug 06 20:36:06 michaelnovakjr: it's hard for me to get in the habbit of prefixing the module name to the svn commit message :) Aug 06 20:36:29 yea, I forgot a few times myself Aug 06 20:42:05 android-random has like 7k loc in it these days Aug 06 20:43:31 michaelnovakjr: somehow our whole repository is locked right now Aug 06 20:43:33 that's cute Aug 06 20:43:47 at least my side thinks it is Aug 06 20:44:00 locked huh? Aug 06 20:45:14 you mean like you have a conflict? Aug 06 20:45:22 no, like svn status shows all is locked Aug 06 20:45:46 i just ran an svn update and then svn status, i don't get locks Aug 06 20:45:53 did you try to resolve them? Aug 06 20:46:45 meh, i'll deal with it later Aug 06 21:15:30 somehow my image is becoming corrupt when I package it...here's what I'm doing: Aug 06 21:15:30 su root Aug 06 21:15:30 cp android.bin android.gz && gunzip android.gz && mkdir filesystem && cd filesystem && cpio -i < ../android && cp ../Snake.apk data/app && find ./ | cpio -H newc -o > ../new_android && gzip ../new_android.bin Aug 06 21:15:51 I also tried it with gzip -9 btw Aug 06 21:16:17 are you root? Aug 06 21:16:26 also, that should be gzip > ... Aug 06 21:16:58 fyi, your first set of commands are unnecessary. gunzip -c ../android.bin | cpio ... will work. Aug 06 21:19:56 tomatoes, tomatoes....can't really do that without a voice :) Aug 06 21:20:38 I'm going to be pissed when it ends up being something really simple Aug 06 21:21:36 I think it must be something to do with how it was originally cpio -i Aug 06 21:22:12 or rather cpio -o Aug 06 21:22:20 hang on, i'll recall the commands i have used Aug 06 21:22:53 if I unpack the original, it unpacks subdirectories and files....if I unpack the one I create, it only unpacks those files and folders in the root directory Aug 06 21:23:09 gunzip -c ../initrd-android | cpio -id Aug 06 21:23:20 find ./ | cpio -H newc -o | gzip > ../initrd-android-custom Aug 06 21:23:31 these are both as root. Aug 06 21:23:34 ok, I'll try it, hopefully it works Aug 06 21:23:46 the find . expects cwd to be in the extracted root. Aug 06 21:24:23 cwd? Aug 06 21:24:35 current working directory Aug 06 21:24:39 ah ok Aug 06 21:24:42 yeah, I know that :) Aug 06 21:25:13 I'm not a total idiot :) Aug 06 21:25:28 just a partial one Aug 06 21:35:09 is it fine that I keep getting cpio: : No such file or directory when unpacking my package? Aug 06 21:35:25 "cpio: : No such file or directory" Aug 06 21:36:32 No. Aug 06 21:36:36 you very likely are not root. Aug 06 21:36:48 I'm root Aug 06 21:37:02 it still unpackages everything, but it throws that first Aug 06 21:37:18 only on the one I created Aug 06 21:37:22 extract and recombine without changing anything and then md5sum the two. that will tell you if you can ignore that warning. Aug 06 21:37:28 my guess is no. Aug 06 21:41:05 yeah, they're different Aug 06 21:41:44 hmm Aug 06 21:46:05 romainguy_: am i supposed to treat these attributes as integers? i.e., use StyledAttributes#getInt? Aug 06 21:47:43 it seems that getInt's default value is returned in every case. Aug 06 21:57:27 i just confirmed with getValue that it doesn't find a value. Aug 06 21:57:41 *confused* Aug 06 21:59:43 ok, I fixed the error, had to do "find * | cpio ...." Aug 06 21:59:50 let's see if it works now Aug 06 21:59:54 cd .. Aug 06 21:59:54 ls Aug 06 22:00:18 herriojr: find * will not catch hidden files, btw. Aug 06 22:00:44 shit Aug 06 22:00:48 ok Aug 06 22:00:51 hmm Aug 06 22:01:31 I know what I'll do....do a find, then remove all ./ on each line Aug 06 22:01:47 that won't be a problem. Aug 06 22:02:03 you probably just must remove '..' Aug 06 22:02:31 though i dont know why that would be there Aug 06 22:02:33 find doesn't find '..' Aug 06 22:02:39 I think it is the line './' Aug 06 22:02:51 cpio is taking as input a list of files to compress. the leading ./ should not affect them. Aug 06 22:03:05 but meh, i dont have time to look at this for you Aug 06 22:03:29 np, I'm checking it out Aug 06 22:05:11 yup, removing the ./ line got rid of the warning Aug 06 22:05:42 romainguy_: *poke*, i'm really confused here... Aug 06 22:19:02 jasta: enum is just a getInt() yes Aug 06 22:19:24 i'm implementing the constructo with Context, AttributeSet, and Map params... Aug 06 22:19:25 and invoking: Aug 06 22:19:31 Resources.StyledAttributes a = context.obtainStyledAttributes(attrs, Aug 06 22:19:31 R.styleable.CardStackView); Aug 06 22:19:31 TypedValue tv = new TypedValue(); Aug 06 22:19:32 boolean b = a.getValue(R.styleable.CardStackView_card_orientation, tv); Aug 06 22:19:37 in every case, b is false. Aug 06 22:19:45 i mean, for any value of my attribute i give. Aug 06 22:20:21 likewise, aapt does not error out when i write something like solitaire:card_orientation="foobarbaz" Aug 06 22:20:27 even though foobarbaz is of course not one of the possible enums Aug 06 22:21:27 doesn't even error out at runtime. Aug 06 22:22:06 neither does solitaire:foobar="baz" Aug 06 22:22:25 i take it this functionality is somehow not working in M5? Aug 06 22:23:39 it is working in M5 Aug 06 22:23:44 it's what we use everywhere Aug 06 22:23:49 like LinearLayout_orientation Aug 06 22:24:27 jasta: boolean b = a.getValue(R.styleable.CardStackView_card_orientation, tv); << you're not using getInt() here Aug 06 22:24:37 i was just testing. getInt yields the default always. Aug 06 22:25:08 getValue would not return false where getInt would return a non-default I assumed. Aug 06 22:25:42 the Home screen in M5 uses this Aug 06 22:25:44 and it works Aug 06 22:25:54 did you define the namespace correctly? Aug 06 22:26:27 i don't know. i just wrote xmlns:solitaire="http://wherever/whatever", i assumed it was not validating the DTD Aug 06 22:27:12 wait, i think i see the problem. weird that this was not an error. Aug 06 22:28:29 no that wasn't it. i had accidentally left off Aug 06 22:28:45 define your namespace this way: xmlns:launcher="http://schemas.android.com/apk/res/com.android.launcher" Aug 06 22:28:53 with the package name of your apk at the end Aug 06 22:29:09 xmlns:solitaire="http://schemas.devtcg.org/apk/res/org.devtcg.games.solitaire" Aug 06 22:29:13 i used that Aug 06 22:29:26 org.devtcg.games.solitaire is the package name from the manifest Aug 06 22:29:45 the whole thing is public if you want to just look Aug 06 22:31:22 http://code.google.com/p/android-random/source/browse/#svn/trunk/Solitaire Aug 06 22:31:54 res/layout/klondike.xml is the layout i'm using, src/.../view/CardStackView.java is the view in question which is supposed to honor these attributes Aug 06 22:41:56 good evening everyone. Aug 06 22:41:58 hey Dougie187. Aug 06 22:42:08 hows it going jasta> Aug 06 22:42:39 good, futzing with android Aug 06 22:42:43 but im stuck so, now im watching tv :) Aug 06 22:43:17 lol Aug 06 22:43:18 are you at home? Aug 06 22:43:20 hi Dougie187 Aug 06 22:43:24 hey yakischloba Aug 06 22:43:29 hows it going with you yakischloba? Aug 06 22:43:30 how goes it Aug 06 22:43:32 Dougie187: yeah, i have a vacation day today Aug 06 22:43:42 nice. well im gonna hop on if you wanna play some. Aug 06 22:43:45 eh alright. grinding out a boring rails application :/ Aug 06 22:45:00 that doesnt sound too fun. Aug 06 22:45:46 No it sucks. I've been putting it off for a long time. I have an interview for a new job tomorrow, and I feel obligated to finish this before I leave the company if I get the new job. So I figured I should start on it now. Aug 06 22:46:15 well that is rather nice of you. Aug 06 22:46:25 i think most people would say, f them. Aug 06 22:46:25 lol Aug 06 22:46:44 i have the same attitude as yakischloba Aug 06 22:46:56 hardly returning the favor. Probably 80% of the programming I've learned in the last year has been on the company's dime. Aug 06 22:47:01 lately i've been doing a lot of work to document and de-vodoo some of my stuff Aug 06 22:47:06 so they arent screwed when i leave Aug 06 22:47:31 Yeah. I'm going to have to do the same. I don't think anyone even knows what I do anymore. Aug 06 22:47:57 yakischloba: ditto. Aug 06 22:48:23 well thats cool Aug 06 22:49:47 I'm going to continue to provide help when necessary after I leave, and also continue to develop the software I've worked on here. I would really feel lame if I just ditched out after having the boss make such an investment in me and not get his return. Aug 06 22:51:57 I just want to be in an environment more conducive to learning...and get paid more :) Aug 06 22:54:26 me too! Aug 06 22:54:26 lol **** BEGIN LOGGING AT Wed Aug 06 22:57:16 2008 Aug 06 23:10:49 time to get out of this sweaty office and work in a coffee shop or something. back in a bit! Aug 07 00:55:49 michaelnovakjr: *poke* Aug 07 02:57:31 so, what's next? **** ENDING LOGGING AT Thu Aug 07 02:59:57 2008