Want to help with Android nightly builds?

by Mark Murphy » Sat, 11 Apr 2009 08:01:58 GMT


Sponsored Links
 A few folk are starting to build the infrastructure for a nightly build
farm, to extend Al Sutton's groundbreaking work on creating downloadable
SDKs out of the public Android repositories.

It will take some time for us to get everything together, including
working out the kinks in the Android development process itself -- see
David Turner's post on this thread:




Want to help with Android nightly builds?

by Disconnect » Sun, 12 Apr 2009 00:21:24 GMT


 re you doing dream builds or emulator builds?

FYI I haven't discontinued my dream builds - I've been focusing on the work
some people have done to get actual, working google services into master.
(And it seems like despite the opening of donut and the claimed 'real soon
now' release of cupcake, they have recently merged in a bunch of changes
that break basically everything. Audio, make/break calls, etc are all broken
in the newest trees.)

Do you have plans to integrate unapproved fixes? (Either fixes in the review
system that aren't pushed to the tree yet, or fixes that didn't make it
there to begin with.) That has been a fairly standard problem with these
builds - since most of the google devs don't use the external tree, it is
frequently out of sync or just flat broken. (EG the recent merge of opencore
2, when the internal cupcake tree is still written to use opencore 1.) And
if you are trying to do dream builds, its even worse - updated hw support
binaries, etc.

(For those few remaining people who don't know :) my g1 cupcake/master
builds are at http://andblogs.net/images/snapshots. 3/16 is reportedly quite
stable.)

On Fri, Apr 10, 2009 at 8:01 PM, Mark Murphy <mmur...@commonsware.com>wrote:


--~--~---------~--~----~------------~-------~--~----~


Sponsored Links


Want to help with Android nightly builds?

by Mark Murphy » Sun, 12 Apr 2009 01:10:15 GMT


 


Initially, I would think the focus would be on emulator builds.


Personally, my goal is 100% automation, meaning if it ain't in a tree
somewhere, it ain't getting built.

Now, since this is all git-backed, we could evolve where there are sets
of builds blending in other sources that are published via their own
private git repos.

As I noted, though, I'm a n00b on creating build farms, so if nobody
else decides to be buildmaster, I want to walk before I run.


Yup. I expect it will take quite some time for a build farm to generate
useful builds, and even after that, there will be lots of times when the
farm either can't build (branch head is broken) or generates useless
output (e.g., HOME button doesn't work). That's partly an issue of build
farm tech and partly an issue of the current Android OSP development
process.

Over the long haul, we should be able to get this to be as smooth as
silk; short-term, it'll be as smooth as a battlefield. But, ya gotta
start somewhere.

-- 
Mark Murphy (a Commons Guy)
 http://commonsware.com  |  http://twitter.com/commonsguy 

Android App Developer Training:  http://commonsware.com/training.html 

--~--~---------~--~----~------------~-------~--~----~



Want to help with Android nightly builds?

by Incognito » Sun, 12 Apr 2009 21:08:38 GMT


 The whole technical aspect sounds interesting. What are the actual programs 
that people have to run to accomplish the nightly builds? You have any links 
anywhere?





A few folk are starting to build the infrastructure for a nightly build
farm, to extend Al Sutton's groundbreaking work on creating downloadable
SDKs out of the public Android repositories.

It will take some time for us to get everything together, including
working out the kinks in the Android development process itself -- see
David Turner's post on this thread:




Want to help with Android nightly builds?

by Mark Murphy » Sun, 12 Apr 2009 23:07:13 GMT


 


I believe the build slaves are not running too much more than the
Android source and build scripts. Windows probably requires some sort of
service to allow it to communicate with the buildmaster.

BuildBot:  http://buildbot.net/trac 

Hudson:  http://wiki.hudson-ci.org/display/HUDSON/Meet +Hudson

-- 
Mark Murphy (a Commons Guy)
 http://commonsware.com  |  http://twitter.com/commonsguy 

Android App Developer Training:  http://commonsware.com/training.html 

--~--~---------~--~----~------------~-------~--~----~



Want to help with Android nightly builds?

by Fred Grott » Sun, 12 Apr 2009 23:13:18 GMT


 Hudson has a an Android client to monitor :)








--~--~---------~--~----~------------~-------~--~----~



Want to help with Android nightly builds?

by Fred Grott » Sun, 12 Apr 2009 23:50:30 GMT


 I may become involved..it depends upon success of an ongoing
business/project gig discussion  that I am currently in with a startup..
The start up  at some point would be relying upon those nightly builds.

How different is the build scripts/receipes from bitbake?








> 



Want to help with Android nightly builds?

by Mark L. Chang » Tue, 14 Apr 2009 01:01:05 GMT


 



Mark, I'm for sure down for doing #1. I can also run #2 if it makes
sense. I just have zero experience doing this, but have plenty of
boxen at my disposal to run stuff. We should chat offline. You have my
email.

--~--~---------~--~----~------------~-------~--~----~



Other Threads

1. Sqlite cursor closable problem

Hello,

I am receiving finalizing closable issues.
the snippet from the logcat as follows:

I/dalvikvm(  585): Ljava/lang/IllegalStateException;: Finalizing
cursor android.database.sqlite.sqlitecur...@437c0e60 on null that has
not been deactivated or closed

I/dalvikvm(  585):      at
android.database.sqlite.SQLiteCursor.finalize(SQLiteCursor.java:596)

I/dalvikvm(  585):      at dalvik.system.NativeStart.run(Native Method)

I/dalvikvm(  585): Uncaught exception thrown by finalizer (will be
discarded):
--
Here I do understand the problem is related to cursor that may not
have handled with in the code.
here pid 585 basically the main process ID. But I do I narrow down to
the code that particular cursor need to take care form this error in
logcat.

Please note I am using many cursors as I retrieving records from the
DB. And I have already taken care of all the cursor. But these error
forcing me to look that some cursor is overlooked.

Could anybody throw some light that would help me to track down the
particular cursor that are causing these trouble.

-- 

2. Activity Not coming to foreground

hi guys,

i'm getting the outgoing number using BroadCastReceiver and then
starts a Activity.
but my activity runs in the background???  how can i bring it to
foreground??

        public void onReceive(Context context, Intent intent) {
                // TODO Auto-generated method stub
                this.context = context;
                // this.intentttt = intent;
                // String s = intent.getAction();

                if (intent.getAction()
                                
.equalsIgnoreCase(Intent.ACTION_NEW_OUTGOING_CALL)) {
                        phonenbr = 
intent.getStringExtra(Intent.EXTRA_PHONE_NUMBER);
                        intent = new Intent(context, OutGoing.class);
                        intent.setFlags(Intent.FLAG_ACTIVITY_NEW_TASK);
                        intent.putExtra("Number", phonenbr);
                        context.startActivity(intent);
                } else {
                        telManager = (TelephonyManager) context
                                        
.getSystemService(Context.TELEPHONY_SERVICE);
                        telManager.listen(new StateListener(),
                                        PhoneStateListener.LISTEN_CALL_STATE);
                }
        }


regards,
Rrandika

-- 

3. desktop remote on phone

4. Cara Matikan data Service di Spica

5. Implement Camera Support for Android Emulator using Web Cam

6. Test Kecepatan PUSH MAIL dari 3 Provider pada 3 Handset Berbeda dengan 2 Operating Sytem yang berbeda pula.

7. Binding Json Array to Adapter