**** BEGIN LOGGING AT Fri Jun 12 02:59:56 2009 Jun 12 03:01:53 hmm now my volume keys don work as theyre caught by my onkeydown Jun 12 03:02:43 are you calling super.onkeydown? Jun 12 03:02:49 Gaz, you should only be returning true on keys that you want to catch. everything else should return super.onKeyDown(keyCode, event); Jun 12 03:06:20 ohh Jun 12 03:06:22 thx Jun 12 03:08:00 thats weird, so it worked but now i dont get the display showing volume going up and down Jun 12 03:10:00 oh it makes the beep but doesnt sem to change Jun 12 03:10:03 guess i can fix that tommorow Jun 12 03:10:32 so i just put this in onkeydown ? super.onKeyDown(keyCode, msg); Jun 12 03:10:42 ah return that Jun 12 03:11:18 Gaz: if you're not handling the keypress, then do return super.onKeyDown(keyCode, msg); if you are handling it, return true Jun 12 03:12:05 if you have handled it* Jun 12 03:12:12 right Jun 12 03:16:57 thx it works Jun 12 03:22:34 on my edittext why do i have to click "next" before i can click done, my other one isnt liek that Jun 12 03:22:41 on the virtual keyboardi mean Jun 12 03:25:28 i'd guess it depends on whether there's another edittext after it. i.e. for input in landscape mode, so you don't have to go back and forth, but that is a guess Jun 12 03:38:27 wow Jun 12 03:38:32 and youre absolutely right Jun 12 03:38:45 what an odd but exceedingly lovely feature Jun 12 03:56:20 Is anyone familiar with using HttpPost in android and could point me to a good tutorial for org.apache.http, because most of the tutorials seem to be for org.apache.common Jun 12 03:59:35 meanburrito920__, you have to read the docs for 4.x even though it's in beta Jun 12 03:59:44 most of the Google links take you to their stable 3.x branch Jun 12 04:03:12 hrm, everyone knows http://dl-ssl.google.com/android/eclipse/ is a 404 right? Jun 12 04:04:08 b0nn, yeah, you have to download the zip and install it that way Jun 12 04:04:43 k Jun 12 04:08:20 b0nn, try http://dl-ssl.google.com/android/eclipse/site.xml Jun 12 04:16:42 weird Jun 12 04:16:53 I install the ADT plugin + SDK Jun 12 04:16:58 that is wierd. Jun 12 04:17:12 but when I try to create a new project Android isnt listed Jun 12 04:24:51 Hello...want in my java-application to mount a partition or to write in a file.sh and chmod that to a+x (777)...is it possible? Jun 12 04:25:46 schash: possible.. you can run native executables.. Jun 12 04:26:54 wmealing;thanks....but even in adb-console I couldn't chmod 777 /path.sh a file... Jun 12 04:27:15 schasch, not all paths allow executions. Jun 12 04:27:23 -s Jun 12 04:27:26 the sdcard for example Jun 12 04:27:48 wmealing:I no execution an sd? Jun 12 04:28:03 ...on SD? Jun 12 04:28:35 wmealing: wich path is allowed the write a file and chmod that....? Jun 12 04:28:36 you can't execute scripts from the SD Jun 12 04:28:38 or at least i can't Jun 12 04:31:12 wmealing: run native executables....I did that...an "ls -l" is fine.... but at mount-order the parameters were not accepted... Jun 12 04:31:42 wmealing: run native executables...how can I mount a partitition? Jun 12 04:31:50 mount is a native executable Jun 12 04:32:09 please attempt to be a little more verbose and exact in your questions. Jun 12 04:34:16 I was a able to exec("ls -l")....that was no problem but with mount something was wrong...too many parameters O guess...so how can I mount with running the 'mount-native-exec'? Jun 12 04:37:16 you would do it the same way, buy my guess is that your application/user does not have sufficient privileges to mount. Jun 12 04:37:22 s/buy/but Jun 12 04:39:31 asking me how.. is going to open a whole new can of worms. Jun 12 04:39:37 most of which i dont have the answer to Jun 12 04:40:05 so with my *.apk it is not possible to mount or to create somewhere a file and to chmod that a+x-777? Jun 12 04:40:27 you can create a file in /data/data/yournamespace/ Jun 12 04:40:40 that can be chmod 777 (not sure thats a good idea, but.. your call) Jun 12 04:42:49 wmealing:thanls...if mounting is not possible I will try to create a file.sh in /data/data/yournamespace/ and chmod that to 777.....that should go? Jun 12 04:43:29 well, change yournamespace whatever, make sure the uid of the app can r/w/x from it. Jun 12 04:43:29 etc Jun 12 04:43:46 schasch, and why would you want to chmod 777 ? Jun 12 04:45:10 wmealing:I would write a mounter.sh with a mount order of my ext3 partitition and exec() that.... Jun 12 04:45:29 your app still wont have superuser privs Jun 12 04:45:36 bad idea? Jun 12 04:45:56 write mounter.sh with my *.apk.... Jun 12 04:45:56 well, it wont solve your problem Jun 12 04:46:08 yep, i understand where you are coming from Jun 12 04:46:09 i really do Jun 12 04:46:17 i'm struggling with this _exact_ problem right now Jun 12 04:46:20 well, last night Jun 12 04:46:37 why...not allowed to mount? Jun 12 04:46:50 you need superuser permissions to mount Jun 12 04:46:52 .-( Jun 12 04:46:55 your app does NOT run as superuser Jun 12 04:47:06 however, i believe if you have a hacked phone.. you can request su permissions Jun 12 04:47:59 so NO way to mount a user...? Jun 12 04:49:04 you need to "become" uid 0 Jun 12 04:49:07 root Jun 12 04:49:09 super user Jun 12 04:49:43 and i think it -can- be done.. i just am stuck requesting/getting the right permissions Jun 12 04:51:40 wmealing: ...."and i think it -can- be done.. i just am stuck requesting/getting the right permissions"....what do you mean? Jun 12 04:52:00 could it be done or not?....sorry....? Jun 12 04:53:01 schasch, lets take a step back. Jun 12 04:53:14 schasch, is your phone rooted, ie do you have root Jun 12 04:55:13 I haveroot with adb shell...or not...??? Jun 12 04:59:51 i dont know, do you do "id" and it says you have id 0 ? Jun 12 05:02:05 I will test... Jun 12 05:04:22 # id Jun 12 05:04:22 uid=0(root) gid=0(root) Jun 12 05:04:25 yes... Jun 12 05:04:44 ok, now run "id" from your app Jun 12 05:08:17 ...wait please......I am testing.... Jun 12 05:13:55 wmealing:uid=10022(app_22) gid=10022(app_22) Jun 12 05:15:32 can you 'su' id Jun 12 05:25:51 22 was in the emulator.exe.... uid=10028(app_28) gid=10022(app_28)...on my device... Jun 12 05:26:01 su id I am testing.... Jun 12 05:26:05 ok Jun 12 05:26:31 you see where i'm going with this righ t? Jun 12 05:30:02 I user "output_ = exec("/system/bin/id", "", "");"....I am testing to involve su...? Jun 12 05:30:15 s/I user/I used Jun 12 05:31:26 ...??? output_ = exec("su /system/bin/id", "", ""); ...?...am testing... Jun 12 05:32:32 I did not found /system/bin/su.... Jun 12 05:37:52 exec("/system/bin/id", "", "") returned "uid=10028(app_28) gid=10028(app_28)"....but.....exec("su /system/bin/id", "", "") returned nothing and didn't change reult of exec("/system/bin/id", "", "")..... Jun 12 05:38:07 or how do I have to use su? Jun 12 05:38:36 I can not use "find" on my device..... .-( Jun 12 05:41:05 how do you do rowspan with a tablerow in a tablelayout? Jun 12 05:45:46 thanks... layout_span child Jun 12 05:46:09 a little channel confusion, nyt? :) **** ENDING LOGGING AT Fri Jun 12 06:20:50 2009 **** BEGIN LOGGING AT Fri Jun 12 06:21:13 2009 Jun 12 07:27:31 Hello can I mount my 2nd partitition from the SD at boottime?I remounted my / changed the init.rc, synced....but after booting it seems to be overwritten... .-( Jun 12 07:52:41 MidWindowManager.java is much different from PhoneWindowManager.java, the softkeyboard causes the window display not normal Jun 12 09:08:42 What is the right way to determine the caller name from the phone number ? Jun 12 10:10:36 how would i go about changing spelling language / keyboard layout from an app ? Jun 12 10:56:20 can i output info to the eclipse consol to more easy check values of variables etc ? Jun 12 10:58:01 thats what the debugger is for Jun 12 11:01:45 i tried using i am to dumb to useit it would seem Jun 12 11:02:32 its one of easier ones out there :P Jun 12 11:02:40 and you can use logging too Jun 12 11:02:43 see the documentation Jun 12 11:29:25 deebo: thanks i got it Jun 12 11:32:59 hey guys, anyon notice their game slow down when its playing background music? Jun 12 11:56:14 ok finaly figured out how to display imput method selector. can i from a notification set the intent to display it ? EG: "mInputMethodManager.showInputMethodPicker();" Jun 12 12:00:34 Hey urgent help needed...when i am uploading the Apk file in market it is displaying the following error "The certificate that signed this apk is not valid until the future. Create a new certificate" Jun 12 12:00:49 whats the problem Jun 12 12:02:47 mib_dgbwsbk8: mm could it be more clear than that? :) Jun 12 12:03:35 i didnt get u deebo Jun 12 12:04:50 Guys Urgent..!!!:( Jun 12 12:05:58 can someone explain me make sure your system clock is set correctly mib_dgbwsbk8 Jun 12 12:07:50 with what i have to compare..!! Jun 12 12:10:15 not sure what you mean ... Jun 12 12:10:20 just sync with a timeserver Jun 12 14:27:03 mib_dgbwsbk8: you there? Jun 12 14:27:14 i had that issue Jun 12 14:28:08 you might have found my blog if you googled infact Jun 12 14:28:14 its 4th for that hit Jun 12 14:28:28 mib_dgbwsbk8: http://gazy.spaces.live.com/blog/cns!C6DAB7D1A0F14906!1478.entry Jun 12 14:34:48 Hey Gaz Thanks lol..!!!:) Jun 12 14:43:05 hey, if you want to develop native code android apps, do you need the dev phone or is the htc magic ok too? Jun 12 14:44:08 the magic is fine Jun 12 14:59:37 ok Jun 12 14:59:52 how does the compilation work, do i need to get the android OS on my pc for that? Jun 12 15:03:40 :/ Jun 12 15:04:04 any suggestions for quickly converting longs to mm/dd/yyyy? Jun 12 15:04:18 dateformatter seems to be chopping up my listview scrolling badly Jun 12 15:22:41 mikedg: is it allocating or something? Jun 12 15:27:06 mikedg: funny, i noticed yesterday as well that formatting dates is very slow... i was quite surprised. Jun 12 15:31:36 Hm, I seem to recall something about this... Jun 12 15:32:17 I think it was in one of the I/O sessions someone had mentioned it Jun 12 15:36:15 mikedg: What does your core look like? a = DateFormat.getDateInstance for (..) { a.format(whatever); } ? Jun 12 15:38:42 mikedg: you will have to do it the hard way i'm afraid. no allocation, so, ensure you have a stringbuffer of the right size and then modify the characters in place Jun 12 15:38:59 mikedg: bear in mind in many locales dates are written differently Jun 12 15:39:40 * TD & Jun 12 15:41:14 mikedg: Yeah, definitely use the StringBuffer method if you aren't already Jun 12 15:42:04 make that a StringBuilder even Jun 12 15:44:30 that, too Jun 12 15:45:20 wait, no Jun 12 15:49:31 romainguy: Is there some way to detect that the lock screen is obscuring your activity? Jun 12 15:51:38 * kRutOn checks hasWindowFocus Jun 12 15:54:55 just just onStop() Jun 12 15:57:41 Yeah, but you get onResume before the user unlocks Jun 12 16:03:34 thanks guys for the above Jun 12 16:03:38 im gonna look into it now Jun 12 16:03:42 ran out for lunch Jun 12 16:04:36 a project that uses Apache 2.0 stuff can become a GPL project right? Jun 12 16:04:58 how do i handle displaying the licenses? if I dont use any of the Apache 2.0 stuff without significant modification? Jun 12 16:05:15 Like I want to open source QuickCut, which uses a handful of stuff from Anycut (Apache 2.0) Jun 12 16:05:20 but I want quickcut to be gpl Jun 12 16:07:25 mikedg: http://www.apache.org/licenses/GPL-compatibility.html Jun 12 16:08:00 hey, I kinda want to make the layout of Amazon mp3, so when you go from portrait to landscape it goes from 1x4 buttons to 2x2. Is there any tutorial/guide that mentions this? I'm a little lost with relative layout Jun 12 16:09:39 ivan: you just make a new layout and put it into layout-land directory Jun 12 16:09:42 same name Jun 12 16:09:45 the os will handle the switching Jun 12 16:10:14 lemme know if you need more explanation Jun 12 16:11:29 blanu: i think that confused me even more :) Jun 12 16:11:32 but thanks Jun 12 16:13:01 I think the key part is probably "Apache 2 software can therefore be included in GPLv3 projects, because the GPLv3 license accepts our software into GPLv3 works." Jun 12 16:14:41 thanks Jun 12 16:15:28 mike.. I'm taking about the redistribution of the elements. Jun 12 16:15:54 I want to achieve a CSS float:left kind of thing Jun 12 16:15:59 IvanSF: he's saying that there are actually two different layouts, with the same elements displayed differently Jun 12 16:16:13 ohhh Jun 12 16:16:20 I'm not sure you can do CSS type of stuff Jun 12 16:17:12 didnt know about layout-land Jun 12 16:17:13 thanks Jun 12 16:17:22 yeah you basically just make sure everything that you reference has the same id's Jun 12 16:17:36 and you require no code changes Jun 12 16:18:27 does any one know how view the comments left about your app on the android market....without using a phone? Jun 12 16:18:39 nwmotogeek: www.cyrket.com Jun 12 16:18:40 cyrket.com Jun 12 16:20:06 awsome thanks! Jun 12 16:20:22 mikedg: if you use layout-port you _must_ user layout-land or it just defaults to layout/? Jun 12 16:20:25 i'm interested in putting better than an 'adb' frontend on junit testing.. does anyone know of a tool ? Jun 12 16:23:22 ivan: i dont know i just use layout and layout-land Jun 12 16:23:28 dont know how it defaultls otherwise, sorry Jun 12 16:28:49 wmealing_: I think they basically munge the output of the InstrumentTestRunner, but I agree with you. Jun 12 16:29:06 mikedg: it worked, however if you change orientation while displaying the activity it won't change the layout Jun 12 16:29:07 wmealing_: They also allow you to output EMMA coverage files Jun 12 16:30:08 kRutOn, i'm going to look into it.. i just feel as though its a little simplistic for what i need/want Jun 12 16:37:28 grrr, i dont udnderstand dateformat.format with string buffer Jun 12 16:41:16 mikedg: What's it doing unexpectedly? Jun 12 16:41:56 im getting nothing in the stringbuffer Jun 12 16:42:15 public StringBuffer format(Date date, StringBuffer toAppendTo, FieldPosition pos) Jun 12 16:42:17 i dont understand what fieldposition does Jun 12 16:42:20 so i set it to null Jun 12 16:47:01 i guess its not as big of a deal as I thought Jun 12 16:47:45 using Date.setTime instead of Calendar.settimeinmillis hen calendar.gettime is much much faster Jun 12 16:48:04 just have to worry about internationalization though since so much of Date has been deprecated Jun 12 16:48:34 oh yeah, the Calendar uses ICU Jun 12 16:49:37 i still overall have some shitty scrolling issues though :( Jun 12 16:53:55 maybe I can cache the formats, Jun 12 16:54:05 dont think that would help too much though bah Jun 12 16:55:47 yawn Jun 12 16:56:14 mikedg: whats the problem? Jun 12 16:56:37 i want to display 2 dates in each item of a gridview Jun 12 16:56:45 i have the dates in long Jun 12 16:56:56 its way to slow to scroll and do a dateformat imo Jun 12 16:57:07 well, create the format object only once Jun 12 16:57:09 first of all Jun 12 16:57:12 i did that Jun 12 16:57:20 private static final Jun 12 16:57:26 and store the formatted result as a string and re-use it. Jun 12 16:57:45 and have some way of flagging it dirty so you can still update entries if you need to Jun 12 16:58:01 are bitmaps involved btw? Jun 12 16:58:04 how can i reuse it since most of the dates are different Jun 12 16:58:11 no Jun 12 16:58:41 mikedg: the adapter has to hold all of your data objects already. you can also hold each data objects string representation Jun 12 16:58:53 if null, compute and store, if non-null, re-use. Jun 12 16:59:24 that said, i dont believe date formatting overhead is going to be enough to cause a gridview to scroll terribly slow. can you show me your code so i can prove you're not doing something else expensive? Jun 12 17:00:08 use traceview Jun 12 17:00:30 yes, also a good idea. use Debug.startMethodTracing/stopMethodTracing inside of getView() Jun 12 17:00:35 and run traceview on the resulting trace files Jun 12 17:00:49 but i bet if you paste your getView call we'll be able to identify a big problem easily Jun 12 17:00:50 theres 24 dateformats occurring per page Jun 12 17:00:54 i commented out the portion that formats the dates and it scrolls nicely Jun 12 17:00:55 actually might be up to 32 per page Jun 12 17:00:55 30 i mean Jun 12 17:00:55 ill pastebin it Jun 12 17:02:45 http://pastebin.com/m11012094 theres this, i made a custom TextView that takes longs and formats them Jun 12 17:02:49 this is the problematic method Jun 12 17:03:00 and if I uncomment the if (0==0) return it runs nicely Jun 12 17:03:10 or if I stick a constant into the super.settext Jun 12 17:04:05 im not sure why you extended TextView for this purpose Jun 12 17:04:12 a simple utility function would have been far more general purpose Jun 12 17:04:30 i had a reason at some point :/ i dont remember though Jun 12 17:05:11 can you show me your getView() call either way? Jun 12 17:06:15 mikedg: do you make database calls in getView/bindView? Jun 12 17:06:45 it's an extended cursoradaptor Jun 12 17:07:11 well, if you cant show me the method, just use debug tracing as was mentioned earlier. Jun 12 17:07:16 that will tell you what you need to know Jun 12 17:07:21 i have to run Jun 12 17:07:27 oh ok Jun 12 17:07:33 if your on later ill paste again Jun 12 17:07:39 i wont be Jun 12 17:07:43 thanks Jun 12 17:07:45 :( Jun 12 17:10:51 http://pastebin.com/m643981d8 Jun 12 17:10:58 if you wanna take a look herrio :) Jun 12 17:13:33 mikedg: What's your formatter initialized as Jun 12 17:13:51 private static final DateFormat mFormatter = new SimpleDateFormat(DATE_FORMAT); Jun 12 17:14:05 not worrying about internationalization now tho Jun 12 17:17:16 can someone comment on what's the correct way of doing that: Activities A -> B -> C. I want to remove A abd B from the back stack when launching C so that pressing back in C won't take the user back to A or B (but pressing back in B will take the user back to A) ? Jun 12 17:17:21 mikedg: i thought you said bitmaps were not involved? Jun 12 17:17:52 i thought you meant directly with the date stuff Jun 12 17:17:58 anyway, i still recommend that you store the string representation in the adapter Jun 12 17:17:58 you know, I remember at Google IO something about Dates having to figure out their timezone or whatever, which is a huge performance problem Jun 12 17:18:00 the IO Session "Debugging Arts of the Ninja Masters" actually has SimpleDateFormat format() as an example for some innocent looking function to be very time costly. they used it to illustrate the use of the traceview Jun 12 17:18:31 exactly what poke said ;) Jun 12 17:18:34 i missed that session :( Jun 12 17:18:42 apparently it made an impression on everyone Jun 12 17:18:42 lol Jun 12 17:18:43 mikedg: well go read it :) Jun 12 17:18:48 http://code.google.com/events/io/sessions/DebuggingArtsNinjaMasters.html Jun 12 17:19:16 you do not want to know how I currently handle this little problem :P Jun 12 17:19:55 their solution was to put of the calculating the date to a thread and then fill in later Jun 12 17:20:10 can sqlite3 do anything nice to help? Jun 12 17:20:32 poke: ahhhh thats a good plan, similar to how the messaging app does that crap Jun 12 17:21:49 should it not be posable to change the language in serrings -> locale & text ->touch input -> keyboard language, i cant seem to find this setting in the api referance. Jun 12 17:22:10 cause the user really cant read the dates while they are scrolling in my app anyway Jun 12 17:27:20 thanks for the think tank guys Jun 12 17:31:15 is there a way to get the current visiblity of the window's progress bar? (basically, I'm looking for isProgressBarIndeterminateVisible() to complement setProgressBarIndeterminateVisibility() ) Jun 12 17:40:03 I created an sd card and I'm trying to load the avd but it gives me a warning: emulator: WARNING: ignoring locked SD Card image at /dump/128megs.sdcard Jun 12 17:41:27 did one of your emulators die oddly? Jun 12 17:44:04 No I never ran this emu. android create avd --name avd_128megs --target 2 --sdcard /dump/128megs.sdcard Jun 12 17:47:14 anxiolytic_1: is the sd card mounted or being used by anything else? Jun 12 17:47:53 does the emulator binary have privs to read and write the sdcard file? Jun 12 17:48:13 The card isn't mounted (I mounted it, added files, unmounted). permissions are 777 Jun 12 17:48:59 does the emulator have rx permissions on /dump? Jun 12 17:49:17 rwx Jun 12 17:49:43 oh aye :) Jun 12 17:49:54 I got it working. I moved the file to another partition which is ext3. dump was fat32 Jun 12 17:52:13 hello,how can I change /init.rc in koolu-7-beta?.......my changes are lost after reboot... Jun 12 17:57:38 Has anybody had any luck getting the android source code lately? I'm getting an error: "fatal: Unable to look up android.git.kernel.org (port 9418) (Name or service not known)" Jun 12 17:58:18 name lookup works for me right now. sounds like a DNS problem in your direction. Jun 12 17:59:07 I've been unable to get a maps API key the entire week. The form to submit your MD5 cert is a 404 Jun 12 17:59:07 morrildl, ping Jun 12 17:59:18 anxiolytic_1, there's another link, one second Jun 12 17:59:21 that one is old Jun 12 18:00:07 anxiolytic_1, http://code.google.com/android/maps-api-signup.html Jun 12 18:00:09 that's the link you need Jun 12 18:00:30 KNY: pong? Jun 12 18:01:04 morrildl, I was told to talk to you or jasonchen about ADC2 questions Jun 12 18:01:17 ctate, yeah, it looks like a DNS problem, though I am able to download the manifest OK. Jun 12 18:01:20 I need to find out what an actual ICMP echo response looks like so I can quote that when people ping me Jun 12 18:01:24 KNY: okay Jun 12 18:01:39 morrildl, most IRCds use ping/pong, IIRC Jun 12 18:03:13 KNY: thanks Jun 12 18:03:26 anxiolytic_1, no problem, took me like 30 minutes to find it last night haha Jun 12 18:03:49 Strange. I'm able to download a branch, but not the whole project from git. I guess the cupcake branch it will have to be. Jun 12 18:20:13 morrildl, did you disappear or did you not get my /msg? Jun 12 18:25:27 kny: what questions do you have about adc2 Jun 12 18:26:16 mikedg, private questions, hence the /msg Jun 12 18:26:21 regarding app ideas Jun 12 18:26:37 oh Jun 12 18:26:45 cheater! Jun 12 18:26:48 haha Jun 12 18:27:10 well, I can tell you the quick summary from my chat with him--ADC2 apps do not need to comply with the DDA Jun 12 18:27:32 dda Jun 12 18:27:32 ? Jun 12 18:28:04 http://www.android.com/us/developer-distribution-agreement.html Jun 12 18:28:34 did you ask if we can submit multiple apps Jun 12 18:28:37 can someone look at this and tell me if they notice anything wrong http://pastebin.com/d639cd175 Jun 12 18:28:49 its nice that they dont have to comply with DDA Jun 12 18:29:29 so we can make dirty dirty apps that offer alternative markets! Jun 12 18:29:38 right? Jun 12 18:29:43 the alarm is in queue and goes off... I can see it in adb, but it never reaches AlarmService.java Jun 12 18:30:40 mikedg, yeah, I suppose you could but then you couldn't distribute it after the comp was over Jun 12 18:31:09 who cares if you win $200k for your porn market Jun 12 18:33:52 lalala Jun 12 18:33:57 here's the service it calls http://pastebin.com/df4356ff Jun 12 18:34:36 it never even steps into onCreate... ideas? Jun 12 18:36:11 did you know this page? you can see how the ranking of every app has changed over time in the market place: http://androidstats.com/ranking/application/8110 Jun 12 18:37:30 Hmm, I wonder if anyone got experience with using proguard with android apps? Jun 12 18:37:44 How to finetune the optimization phase so that dx won't barf? Jun 12 18:38:26 mpardo: when you say it goes off, what do you mean? Jun 12 18:38:32 your code receives the alarm? Jun 12 18:39:06 mpardo: also, just FYI, if you're setting a once-a-day alarm, you might want to think about using setInexactRepeating() instead Jun 12 18:39:11 the alarm is "triggered" and then rescheduled (repeating). I can see using "adb shell dumpsys alarm" Jun 12 18:39:40 but it's not delivered? or what? Jun 12 18:40:08 yacc, it's barfing? gimme a sec, I'll find out what settings I'm using Jun 12 18:40:33 ctate, exactly. I'm debugging... and the code should step into the service, but it doesn't Jun 12 18:40:43 tauno, mine is barfing about: [apply] com.android.dx.cf.code.SimException: local variable type mismatch: attempt to set or access a value of type int using a local variable of type java.lang.Object. This is symptomatic of .class transformation tools that ignore local variable information. Jun 12 18:40:48 is it launching the service at all? Jun 12 18:40:52 nope Jun 12 18:41:09 tauno, adding -dontoptimize makes that go away, but it also turns my <10KB app into a 50KB app. Jun 12 18:41:21 what is AlarmService ? Jun 12 18:41:29 yacc, ah yeah, I'm also using -dontoptimize :( Jun 12 18:41:41 its my class that extends Service Jun 12 18:42:07 tauno, and I do not think that there are more detailed options to specify what it is allowed to do, ... Jun 12 18:42:14 yacc: did you try with --no-locals Jun 12 18:42:22 in the log, do you see it even *trying* to launch the service? Jun 12 18:42:22 I don't remmeber why but it might be the same reason that it was acting silly on me when it was turned on :( Jun 12 18:42:27 i'm not sure setInexactRepeating will work because the time to go off is a preference setting and the user will expect the app to do what it wants Jun 12 18:42:31 kRutOn, --no-locals for dx or for proguard? Jun 12 18:42:32 also, if you just fire that PendingIntent yourself, does it start the service? Jun 12 18:42:36 yacc: d Jun 12 18:42:38 yacc: dx Jun 12 18:42:40 let me check Jun 12 18:42:49 mpardo: if the exact fire time is important, then yeah, inexact is not for you Jun 12 18:43:21 ctate, you mean in the logcat view? Jun 12 18:43:28 yes Jun 12 18:43:41 let me set run the test again and take a look Jun 12 18:44:17 kRutOn, not that 50K is that when I think, similiar programs that have to carry the same data seem to be >25KB even for Java programs anyway. Jun 12 18:44:24 can I ask why the droid compiler isn't optimizing the code as much as proguard? <10KB to 50KB seems like a lot of room to improve things there Jun 12 18:44:24 kRutOn, what does --no-local do? Jun 12 18:44:45 tauno, Well, the comparision is unfair. Jun 12 18:45:09 tauno, 10KB was without data tables, ... ;) Jun 12 18:45:25 ah ok, I understood you compared the proguard optimized vs non-optimized .apk size :) Jun 12 18:46:03 yacc: I forget exactly, but I had to search through the source code for a while to find out. Jun 12 18:46:53 tauno, but with dx --no-locals and proguard running with optimizations I'm getting still 39K to 49K before. Jun 12 18:47:16 http://drunkhobo.com/~sean/wasted_again.mp3 Jun 12 18:47:16 tauno, well, yes the app was somewhat slimmer without the data it needs to know about ;) Jun 12 18:47:22 ctate, "Unable to start service Intent... blah... : not found" Jun 12 18:47:31 what kind of crap is that? Jun 12 18:47:58 yacc, so --no-locals worked? Jun 12 18:48:05 tauno, yeah. Jun 12 18:48:11 lol hold on a sec Jun 12 18:48:16 great, have to try that too then :) Jun 12 18:48:26 ctate, manifest Jun 12 18:48:48 if danfuzz beats you on the head for using --no-locals, it's not my fault :-) Jun 12 18:49:09 it does something.. bad? :P Jun 12 18:49:16 The irritating thing are all these undocumented thingas like --no-locals. Jun 12 18:49:26 Or how adb ppp is supposed to work, ... Jun 12 18:50:04 kRutOn, well, I've got his permission to use that option I think, as a workaround for a bug I reported, anyway ;) Jun 12 18:50:22 woot! that worked Jun 12 18:50:38 yacc: Actually, it appears Dan replied to your message with this before. Jun 12 18:50:47 yeah, that's what I just ran into Jun 12 18:51:02 I was trying to find any instances where he mentioned no-locals and its use Jun 12 18:51:38 Apparently it turns off some error checking Jun 12 18:52:07 He requests any .class file that causes an error with dx that is fixed by --no-locals to be submitted as a bug to be triages Jun 12 18:52:10 kRutOn, well, the optimize proguard problem is new. Jun 12 18:52:10 triaged Jun 12 18:52:45 kRutOn, well, the error message is quite informative. Jun 12 18:53:05 kRutOn, I mean, proguard is probably being clever and tries to use one local variable for different types, ... Jun 12 18:53:13 kRutOn, curious that JVMs allow this. Jun 12 18:53:47 kRutOn, dx OTOH has this "clean-room-reimplementation" feel ;) Jun 12 18:54:50 kRutOn, I've got the "--no-locals" permission because scala match statements seem to create local variables for overlapping bytecode regions, ... Jun 12 18:56:20 yacc: yeah, a lot of tools do things like that Jun 12 18:57:27 kRutOn, well, not sure what the Google stance on this is. Jun 12 18:58:00 While being holy and following the specs seems sane, it throws out a number of interesting stuff in the Java space, ... Jun 12 18:58:03 mpardo: belated, the problem was that you hadn't declared the service in your manifest? i'm guessing? Jun 12 18:59:46 ctate, yep Jun 12 18:59:58 the small crap we often forget Jun 12 19:02:45 Does android.net.http.RequestQueue exist? I have seen a few posts mentioning that class Jun 12 19:03:04 but couldn't find it Jun 12 19:09:03 skyred: not in the SDK Jun 12 19:11:07 i am still unabel to locate how to change settings -> locale & text ->touch input -> keyboard language, i have tried google it and of cause also http://developer.android.com/reference/packages.html *cry* Jun 12 19:20:21 Is it possible to add a new TableRow to a TableLayout? Jun 12 19:23:41