Sent from my HTC smartphone on uhh,xuuc Now Networkcu from Sprint!
----- Reply message ----- From: android-developers@googlegroups.com To: "Digest Recipients" <android-developers@googlegroups.com> Subject: [android-developers] Digest for android-developers@googlegroups.com - 25 Messages in 12 Topics Date: Tue, Oct 2, 2012 4:48 pm Today's Topic Summary Group: http://groups.google.com/group/android-developers/topics Google has received a subpoena seeking information related to Android applications [2 Updates] Google cancelled this order. Reason: Other [6 Updates] how to make a video player. [1 Update] How can i use both Gallery and Grid views in one activity. [1 Update] Are onCreate and onRestoreInstanceState mutually exclusive? [1 Update] Dynamic .APK from GooglePlay store / Android [1 Update] width of my TextureView [2 Updates] Exclude dependencies from apk? [7 Updates] No response for click event on my custom button. [1 Update] IllegalStateException on dispatchKeyEvent on support library [1 Update] How does Android handle telephony? [1 Update] Testing In-app billing [1 Update] Google has received a subpoena seeking information related to Android applications coolbho3k <mike.g.hu...@gmail.com> Oct 02 02:24PM -0700 Unless Google got a legal gag order to keep quiet, they will ALWAYS let you know if they've given the government your personal information. The FBI told Google not to notify us but couldn't get a court order to do that. Google is generally very serious about your personal information. And Google will fight subpoenas on your behalf. On Monday, October 1, 2012 7:46:58 PM UTC-4, Tom wrote: Indicator Veritatis <mej1...@yahoo.com> Oct 02 04:39PM -0700 Good assumption, even if it turns out not to be true in this case. I would not, for example, have assumed that the phone number in the message was a genuine FBI office number unless independently verified. Why the FBI would need that much detailed information for the stated purposes is also suspicious. All they really need to know is whether or not the developer responsible for uploading to the alternative market is the same as the one uploading on Google. They do not need the developer's full contact info for that. If they were really concerned with fighting piracy rather than reaping domestic intelligence, they could have a third party compare contact info and let them know which are matches w/o revealing the data: kind of like a "contact information escrow agent". On Tuesday, October 2, 2012 8:15:05 AM UTC-7, Nathan wrote: Google cancelled this order. Reason: Other coolbho3k <mike.g.hu...@gmail.com> Oct 02 02:22PM -0700 I'm getting literally pages and pages of canceled orders with no explanation, just like this - any status from Google? On Thursday, September 27, 2012 4:42:44 PM UTC-4, Nathan wrote: TreKing <treking...@gmail.com> Oct 02 04:26PM -0500 > any status from Google? Is there ever? ------------------------------------------------------------------------------------------------- TreKing <http://sites.google.com/site/rezmobileapps/treking> - Chicago transit tracking app for Android-powered devices spocky12 <spock...@gmail.com> Oct 02 03:50PM -0700 I was about to post about the exact same problem when I saw this. Good news is I'm not alone. Bad news is we don't have any kind of information regarding this. My app uses in-app billing too. I used to have approximately 15% of canceled orders due to authorization refused and about 0 to 1% of "others" each day. Since last saturday, "authorization refused" didn't change but "others" is now about 12% of my commands :( "Authorization refused" was not so problematic because it's often related to bad account information. Most users usually corrected their account information and managed to finalize the command. This is unfortunately not true with "reason: other" as it seems that almost every user in that case can't manage to buy the app, no matter how many tries they make. This means I'm losing 12% of commands each day (hopefully users will try again in a few hours/days and won't abandon). @Nathan : I can't confirm about the "0" account age and didn't find any specific pattern. Can anybody confirm that removing/adding back the credit card is a successful workaround ? Kostya Vasilyev <kmans...@gmail.com> Oct 03 03:03AM +0400 2012/10/3 spocky12 <spock...@gmail.com> > I was about to post about the exact same problem when I saw this. Good > news is I'm not alone. Bad news is we don't have any kind of information > regarding this. This is the best we have at this time, I believe: http://support.google.com/googleplay/android-developer/bin/static.py?hl=en&page=known_issues.cs Purchasing Issues Problems with in-app billing and licensing server We are aware the reports regarding in-app billing and Google licensing server (LVL) issues, and are working on resolving these problems. Please stay tuned and thank you for your patience. Report this issue << > specific pattern. > Can anybody confirm that removing/adding back the credit card is a > successful workaround ? Since I was the one who mentioned that, I can confirm that sometimes it works and sometimes it doesn't, based on my limited interaction with users regarding failed orders (not everyone writes, and I usually set up a license in my own, back-up system, when Google's fails). Large scale, cloud-based computing sure is fun. Now imagine this in a different setting -- "Sorry Captain, the life support systems have shut down due to an unexpected thread deadlock". HAL 9000 wasn't malicious, just buggy. -- K spocky12 <spock...@gmail.com> Oct 02 04:22PM -0700 Thanks Kostya Vasilyev for your reply. I'm aware of the developer known issues. What is weird is that it's been written there for some time now and seems to affect people on a large scale only recently. This is just getting worse... I don't have enough experience on the maket for remembering a 25% canceled orders, but I find 12% cancelation rate is far too much already ;) On Wednesday, October 3, 2012 1:09:16 AM UTC+2, Kostya Vasilyev wrote: coolbho3k <mike.g.hu...@gmail.com> Oct 02 04:35PM -0700 My app uses neither LVL nor in-app billing, by the way. Still getting a ton of these canceled orders. On Thursday, September 27, 2012 4:42:44 PM UTC-4, Nathan wrote: how to make a video player. Indicator Veritatis <mej1...@yahoo.com> Oct 02 04:32PM -0700 Your question has a lot more to do with Flash than with Android. On Sunday, September 30, 2012 9:39:51 AM UTC-7, Sukhchain wrote: How can i use both Gallery and Grid views in one activity. TreKing <treking...@gmail.com> Oct 02 04:25PM -0500 > Hi friends,i am displaying images by using Grid and Gallery view with in > the same activity.Can any body help me with code. http://www.catb.org/esr/faqs/smart-questions.html ------------------------------------------------------------------------------------------------- TreKing <http://sites.google.com/site/rezmobileapps/treking> - Chicago transit tracking app for Android-powered devices Are onCreate and onRestoreInstanceState mutually exclusive? TreKing <treking...@gmail.com> Oct 02 04:25PM -0500 > onSaveInstanceState. > 1) where do these methods fit the activity lifecycle? I have read a lot of > documentation but there is no clear idea, except a broad statement that when the state of the activity is to be saved Labeled very clearly. http://developer.android.com/guide/components/activities.html#SavingActivityState > 2) Are onCreate and onRestoreInstanceState mutually exclusive? What does that mean? > 3) is onRestoreInstanceState called when the activity is destroyed? No, why would you restore the state of something that was just destroyed? > what does this mean? What does what mean? > an activity always destroyed except for scenarios when another activity is > floating on top of current. Huh? > 4) onRestoreInstanceState appears to be called only from instrumentation > in jelly bean. Is this no longer relevant to activity lifecycle? The state callbacks are technically not part of the Activity lifecycle. They're "special cases". > 5) One thing that keeps on being mentioned is that onRestoreInstanceState is > executed after start and onCreate before start. So if onCreate is called > everytime, then is onRestoreInstanceState redundant? No. onCreate means activity was created - do you init stuff. onRestore means activity was created *and* was previously in use - so try to restore it's state. ------------------------------------------------------------------------------------------------- TreKing <http://sites.google.com/site/rezmobileapps/treking> - Chicago transit tracking app for Android-powered devices Dynamic .APK from GooglePlay store / Android TreKing <treking...@gmail.com> Oct 02 04:15PM -0500 On Tue, Oct 2, 2012 at 8:20 AM, Marcelo Tomio Hama < > download and installs it > 4. After the installation, the app shows, as main page, the > http://www.somestore.com webpage I still don't understand the point of this, but this is not possible, AFAIK. ------------------------------------------------------------------------------------------------- TreKing <http://sites.google.com/site/rezmobileapps/treking> - Chicago transit tracking app for Android-powered devices width of my TextureView bob <b...@coolfone.comze.com> Oct 02 11:07AM -0700 Can someone help me see how to figure out the width of my TextureView? I tried this, but I got zeros (in the_width and the_height): *protected void onCreate(Bundle savedInstanceState) {* * * *mTextureView = new TextureView(this);* * * *setContentView(mTextureView);* * * *int the_width = mTextureView.getWidth();* *int the_height = mTextureView.getHeight();* TreKing <treking...@gmail.com> Oct 02 04:12PM -0500 > Can someone help me see how to figure out the width of my TextureView? This has come up before. Search this group for those two methods, I'm sure you'll find something useful. ------------------------------------------------------------------------------------------------- TreKing <http://sites.google.com/site/rezmobileapps/treking> - Chicago transit tracking app for Android-powered devices Exclude dependencies from apk? Lindley <lindl...@gmail.com> Oct 02 12:00PM -0700 I am using Eclipse with the ADT plugin, as well as ant. I am trying to design a plug-in architecture. I have three projects, call them Main, Plugin, and Impl. Main builds a service apk, Plugin is an Android Library. Plugin just defines two classes, AbstractPlugin and PluginFactory. AbstractPlugin is an interface, and PluginFactory does what is necessary to set up a DexClassLoader and instantiate by reflection a class from Impl which implements that interface. Main uses the resulting AbstractPlugin in some way. Therefore, the compile-time dependencies are Main -> Plugin and Impl -> Plugin. There is no compile time dependency between Main and Impl, since the class in Impl is instantiated by reflection. Therefore, when the Main.apk is built, classes from Plugin are included but classes from Impl are not. The problem is then how to get those dex'd classes in Impl into the phone in a place where DexClassLoader can get at them. There are two options. First, Impl can be built as an Android library, then the resulting classes.jar can be manually dex'd with dx, and then the resulting classes-dex.jar can be pushed into some predetermined directory on the phone. This works, but I am not entirely thrilled with it. For one thing, it prevents Impl from having any resources of its own, only Java code. Second, Impl can be built as an apk which just happens to contain neither a service nor an activity. This can be installed normally, and then PluginFactory can get the path to the apk from the PackageManager to pass it to DexClassLoader. The problem with this approach is that in each configuration I have tried so far, Impl.apk ends up pulling in the classes from Plugin as well since there is a compile time dependency. Therefore when DexClassLoader tries to resolve something, it realizes there are now two different AbstractPlugin implementations in the system (which are really the same one in two different apks), and bails out. What I would like to do is keep Plugin on the classpath during the build of Impl, but exclude its classes from the apk. Is there a way to set this up either in Eclipse or in Ant directly? (What about in maven?) This does not necessarily solve the resource problem since the active Context in Main would not necessarily know about the resources in Impl if it had any, but it would be a start. Kristopher Micinski <krismicin...@gmail.com> Oct 02 03:29PM -0400 So instead of doing this (ClassLoaders being bad...), I would use an AIDL service stub for the methods you needed... Alternatively, I would suggest you look at using an Android library project, if you can statically link. What kinds of things do you need. Multiple functional looks very much like you might want a single app, and not a collection of plugins. kris Lindley <lindl...@gmail.com> Oct 02 12:47PM -0700 For a variety of reasons, most notably because the plug-in portion is not supposed to be Android-specific, I don't want to use AIDL here. Static linking is a fallback option. I don't like it but I'll do it if that's the only way. There really ought to be a cleaner fix for this. On Tuesday, October 2, 2012 3:31:58 PM UTC-4, Kristopher Micinski wrote: Lindley <lindl...@gmail.com> Oct 02 12:51PM -0700 I'll give a little more detail. The Impl class is supposed to be one of several pure-java implementations of the service in Main that I might drop in. The purpose of Main is to abstract all Android-specific code away from this implementation. In particular, the service in Main *already* has an AIDL interface, so putting on in Impl would just be silly. On Tuesday, October 2, 2012 3:47:52 PM UTC-4, Lindley wrote: Kristopher Micinski <krismicin...@gmail.com> Oct 02 04:08PM -0400 So wait, I'm not clear, why can't this be statically linked, and why is this a plugin type architecture.. It seems like you're trying to get code reuse, but still, I'm not sure why something like a ClassLoader is required from. From where will the classes be loaded? kris Lindley <lindl...@gmail.com> Oct 02 01:45PM -0700 I would prefer not to make it statically linked because I am not the one writing the plugins. Each plugin is potentially going to be on a different branch of a repository, and I would like to code to continue working without undue hardship when I switch branches. Different versions of the project files specifying static dependencies on each branch is an option, but not a good one. Where the classes will be loaded from is precisely the issue. DexClassLoader can either take a jar containing a classes.dex or an apk. I mentioned both options in the original post. I'm fairly certain I can get things working using the jar approach, but it is inelegant in some respects. I prefer the paradigm of installing an apk to install a plugin, and therefore I need to figure out how to explicitly exclude certain dependencies from being included in the apk. On Tuesday, October 2, 2012 4:11:20 PM UTC-4, Kristopher Micinski wrote -- You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en
