Closed Bug 752537 Opened 12 years ago Closed 12 years ago

Test web-downloaded APK strategy

Categories

(Release Engineering :: General, defect)

Product:
Release Engineering
Component:
General
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(firefox14 fixed, blocking-fennec1.0 betaN+)

Status:
RESOLVED FIXED
Tracking Flags:
Tracking Status
firefox14 --- fixed
blocking-fennec1.0 --- betaN+

People

(Reporter: akeybl, Assigned: mozilla)

References

Details

(Whiteboard: [no-code]) 

We need to test our web-downloaded APK strategy. That'll require:

1) An en-US test APK uploaded to the store (we already have this)
2) Another locale's APK with the same product name and versioncode offset available
3) A third en-US test build with the highest versioncode uploaded to the store

Our testing will be to install #1 from the store, then install #2 from FTP (ensuring it's picked up by Google Play but does not update to #1), and then we'll upload #3 to the store and try updating (testing our ability to recombine the populations if desired in the future).

If we can already ensure that #2 has an AUS channel specified and active, then it'd be great to have a #2.5 build that has a higher versioncode so that we can try using the update service to install before moving on to build #3. If that's not easily done, we'll separate that test out.
Assignee: nobody → aki
I think I may have ran something wrong...
This is the results that I got on the Samsung Galaxy S II for locale Twitter running the crossbrowser video tool (modified):

Aurora version		End Throbber Time 	Elapsed	onload
locale53		2973.21428571429	3196.07142857143
locale13		2861.66666666667	3091.77777777778
single- 05/07/2012	2975.55555555556	3195.11111111111

https://docs.google.com/spreadsheet/ccc?key=0Arku3jleCA0UdGFUZnZwaTNiT3ptVkYybVhtYkhlc0E#gid=1

Something doesn't seem right with the values so I am going to be running it again.
From mfinkle's email:

[Dev]
Brad to follow up on obtaining startup perf numbers (geoff, raw numbers he is getting for throbber stop testing) with the support for 13 and 60 languages 

I took 3 builds (single en-US Aurora, Aki's 13 locale Aurora and Aki's 53 locale Aurora) and ran simple "how long does it take to get to the last Throbber Start" test. I started the builds via the commandline and loaded a 50KB textfile. I used a single core Nexus S using Gingerbread. The results of the startup times are below (average of 5 runs):

Baseline (Aurora en-US): 1882ms
13 Locale Aurora:        2053ms (9% slower)
53 Locale Aurora:        2116ms (12% slower)

The 53 locale build is not as slow as I would have thought.
For those interested in the crossbrowser_video_tool, the original is listed here : https://github.com/ctalbert/crossbrowser_video_tool

prereq is that the phone has to be rooted; it doesn't use the SUTagent, it does use the android SDK and also tries to kill the process (that's why the phone has to be rooted)

I modded it locally and changed it so I can use aurora as well running a cmdline of :
python runstartuptest.py  --sdk=/Users/nhirata/android-sdk-macosx --browser=aurora --pause=20 --preptime=10 --read_length=10

I also capture logcat in another terminal window so I can get the times more easily:
adb logcat -v time > (filename)

more gs2_single_take1.log |grep "Starting: Intent { act=android.intent.action.VIEW dat=file:///mnt/sdcard/startup/m.twitter.com/"|awk '{print $2}'

more gs2_single_take1.log |grep "Throbber start"|awk '{print $2}'
first of 4

more gs2_single_take1.log |grep "Throbber stop"|awk '{print $2}'
last of two
Side note: The update notification for aurora did show for the 13 locale build.  When I tried to update however, the screen went completely black.  I did have aurora open in the background.  The only way to recover was to hit the back button even after waiting.

After upgrading it caused unresponsiveness to changing websites.
(In reply to Alex Keybl [:akeybl] from comment #0)
> We need to test our web-downloaded APK strategy. That'll require:
> 
> 1) An en-US test APK uploaded to the store (we already have this)
> 2) Another locale's APK with the same product name and versioncode offset
> available
> 3) A third en-US test build with the highest versioncode uploaded to the
> store

For speed purposes, I'm planning on:

* kicking off an en-US android native build + repack
* kicking off a 2nd en-US android native build

The former will have a larger versioncode than the apk on the store, so should fulfill (2).
The latter will have a larger versioncode than both (1) and (2), so should fulfill (3).

If we specifically want to test the extra large versioncode (2112...) then I'll have to find a way to shoehorn in some code to do that, which will add time.
(In reply to Aki Sasaki [:aki] from comment #5)
> If we specifically want to test the extra large versioncode (2112...) then
> I'll have to find a way to shoehorn in some code to do that, which will add
> time.

I don't think that's necessary - thanks for fast tracking the builds Aki.
Multilocale hacks:

* go into built objdir
* rm -rf dist/bin/chrome/*/locale/*/global-platform
* remove global-platform lines from dist/bin/chrome/*.manifest
* make package again

This results in a 53 locale apk that's 8mb smaller, but throws errors when trying to load webpages.

* remove all directories from all locale repos other than mobile/android/base
* build multilocale

This results in a broken multilocale step, and no multilocale apk.

I'm going to put my hacky multilocale investigation aside and proceed with comment 5.
10:35:13     INFO -  make realchrome
10:35:13     INFO -  make[4]: Entering directory `/builds/slave/m-aurora-andrd-ntly/build/obj-firefox/toolkit/locales'
10:35:13     INFO -  /tools/python/bin/python2.5 /builds/slave/m-aurora-andrd-ntly/build/config/JarMaker.py \
10:35:13     INFO -  	   -j ../../dist/bin/chrome \
10:35:13     INFO -  	  -t /builds/slave/m-aurora-andrd-ntly/build -f flat  -c /builds/slave/m-aurora-andrd-ntly/./build/obj-firefox/merged/toolkit -c /builds/slave/m-aurora-andrd-ntly/mozilla-aurora/zh-TW/toolkit -c /builds/slave/m-aurora-andrd-ntly/build/toolkit/locales/en-US -I/builds/slave/m-aurora-andrd-ntly/mozilla-aurora/zh-TW/toolkit/defines.inc -DNDEBUG -DTRIMMED -DAB_CD=zh-TW -DOSTYPE=\"Linux\" -DOSARCH=Linux -DANDROID=1 -DANDROID_VERSION=5 -DCROSS_COMPILE=1 -DMOZ_THUMB2=1 -DHAVE_ARM_SIMD=1 -DHAVE_ARM_NEON=1 -DMOZ_ENABLE_PROFILER_SPS=1 -DMOZILLA_VERSION=\"14.0a2\" -DMOZILLA_VERSION_U=14.0a2 -DMOZILLA_UAVERSION=\"14.0\" -DNO_PW_GECOS=1 -DMOZ_LINKER=1 -DD_INO=d_ino -DSTDC_HEADERS=1 -DHAVE_SSIZE_T=1 -DHAVE_ST_BLKSIZE=1 -DHAVE_SIGINFO_T=1 -DHAVE_UINT=1 -DHAVE_UINT_T=1 -DHAVE_UNAME_DOMAINNAME_FIELD=1 -DHAVE_VISIBILITY_HIDDEN_ATTRIBUTE=1 -DHAVE_VISIBILITY_ATTRIBUTE=1 -DHAVE_DIRENT_H=1 -DHAVE_GETOPT_H=1 -DHAVE_MEMORY_H=1 -DHAVE_UNISTD_H=1 -DHAVE_MALLOC_H=1 -DHAVE_SYS_STATFS_H=1 -DHAVE_SYS_SYSMACROS_H=1 -DHAVE_LINUX_QUOTA_H=1 -DHAVE_MMINTRIN_H=1 -DHAVE_SYS_CDEFS_H=1 -DHAVE_DLOPEN=1 -DHAVE_MEMMEM=1 -DNO_X11=1 -DHAVE_STRERROR=1 -DHAVE_LCHOWN=1 -DHAVE_FCHMOD=1 -DHAVE_SNPRINTF=1 -DHAVE_MEMMOVE=1 -DHAVE_RINT=1 -DHAVE_SETBUF=1 -DHAVE_ISATTY=1 -DHAVE_FLOCKFILE=1 -DHAVE_LOCALTIME_R=1 -DHAVE_STRTOK_R=1 -DHAVE_CLOCK_MONOTONIC=1 -DMALLOC_H=\<malloc.h\> -DHAVE_STRNDUP=1 -DHAVE_MEMALIGN=1 -DHAVE_VALLOC=1 -DHAVE_I18N_LC_MESSAGES=1 -DNS_ALWAYS_INLINE=__attribute__\(\(always_inline\)\) -DNS_ATTR_MALLOC=__attribute__\(\(malloc\)\) -DNS_WARN_UNUSED_RESULT=__attribute__\(\(warn_unused_result\)\) -DMOZ_BUILD_APP=mobile/android -DMOZ_WIDGET_ANDROID=1 -DMOZ_PDF_PRINTING=1 -DMOZ_INSTRUMENT_EVENT_LOOP=1 -DUSE_ARM_KUSER=1 -DMOZ_DISTRIBUTION_ID=\"org.mozilla\" -DMOZ_ANDROID_HISTORY=1 -DMOZ_JAVA_COMPOSITOR=1 -DIBMBIDI=1 -DACCESSIBILITY=1 -DNS_PRINTING=1 -DNS_PRINT_PREVIEW=1 -DMOZ_RAW=1 -DMOZ_OGG=1 -DATTRIBUTE_ALIGNED_MAX=64 -DMOZ_WEBM=1 -DVPX_ARM_ASM=1 -DMOZ_WAVE=1 -DMOZ_SYDNEYAUDIO=1 -DMOZ_MEDIA=1 -DMOZ_TREMOR=1 -DMOZ_XTF=1 -DENABLE_SYSTEM_EXTENSION_DIRS=1 -DMOZ_CRASHREPORTER=1 -DMOZ_CRASHREPORTER_ENABLE_PERCENT=100 -DLIBJPEG_TURBO_ARM_ASM=1 -DMOZ_USE_NATIVE_POPUP_WINDOWS=1 -DMOZ_TREE_FREETYPE=1 -DHAVE_FT_BITMAP_SIZE_Y_PPEM=1 -DHAVE_FT_GLYPHSLOT_EMBOLDEN=1 -DHAVE_FT_LOAD_SFNT_TABLE=1 -DMOZ_UPDATER=1 -DMOZ_UPDATE_CHANNEL=aurora -DMOZ_DISABLE_DOMCRYPTO=1 -DMOZ_FEEDS=1 -DMOZ_GFX_OPTIMIZE_MOBILE=1 -DMOZ_DEBUG_SYMBOLS=1 -DMOZ_LOGGING=1 -DSIZEOF_INT_P=4 -DMOZ_MEMORY_SIZEOF_PTR_2POW=2 -DMOZ_MEMORY=1 -DMOZ_MEMORY_LINUX=1 -DMOZ_MEMORY_ANDROID=1 -DJSGC_INCREMENTAL=1 -DHAVE__UNWIND_BACKTRACE=1 -DJS_DEFAULT_JITREPORT_GRANULARITY=3 -DMOZ_OMNIJAR=1 -DMOZ_USER_DIR=\".mozilla\" -DMOZ_STATIC_JS=1 -DHAVE_STDINT_H=1 -DHAVE_INTTYPES_H=1 -DMOZ_TREE_CAIRO=1 -DHAVE_UINT64_T=1 -DMOZ_TREE_PIXMAN=1 -DMOZ_GRAPHITE=1 -DMOZ_ENABLE_SKIA=1 -DMOZ_XUL=1 -DMOZ_PROFILELOCKING=1 -DBUILD_CTYPES=1 -DMOZ_PLACES=1 -DMOZ_APP_COMPONENT_INCLUDE=\"nsBrowserComponents.h\" -DMOZ_APP_UA_NAME=\"Firefox\" -DMOZ_APP_UA_VERSION=\"14.0a2\" -DMOZ_UA_FIREFOX_VERSION=\"14.0a2\" -DFIREFOX_VERSION=14.0a2 -DMOZ_UA_BUILDID=\"\" -DMOZ_DLL_SUFFIX=\".so\" -DXP_UNIX=1 -DUNIX_ASYNC_DNS=1  \
10:35:13     INFO -  	  /builds/slave/m-aurora-andrd-ntly/build/toolkit/locales/jar.mn
10:35:13     INFO -  Traceback (most recent call last):
10:35:13     INFO -    File "/builds/slave/m-aurora-andrd-ntly/build/config/JarMaker.py", line 516, in <module>
10:35:13     INFO -      main()
10:35:13     INFO -    File "/builds/slave/m-aurora-andrd-ntly/build/config/JarMaker.py", line 486, in main
10:35:13     INFO -      jm.processIncludes(options.I)
10:35:13     INFO -    File "/builds/slave/m-aurora-andrd-ntly/build/config/JarMaker.py", line 164, in processIncludes
10:35:13     INFO -      self.pp.do_include(inc)
10:35:13     INFO -    File "/builds/slave/m-aurora-andrd-ntly/build/config/Preprocessor.py", line 430, in do_include
10:35:13     INFO -      raise Preprocessor.Error(self, 'FILE_NOT_FOUND', str(args))
10:35:13     INFO -  Preprocessor.Error: ('', 0, 'FILE_NOT_FOUND', '/builds/slave/m-aurora-andrd-ntly/mozilla-aurora/zh-TW/toolkit/defines.inc')
10:35:13    ERROR -  make[4]: *** [realchrome] Error 1
10:35:13     INFO -  make[4]: Leaving directory `/builds/slave/m-aurora-andrd-ntly/build/obj-firefox/toolkit/locales'
10:35:13    ERROR -  make[3]: *** [chrome] Error 2
10:35:13     INFO -  make[3]: Leaving directory `/builds/slave/m-aurora-andrd-ntly/build/obj-firefox/toolkit/locales'
10:35:13    ERROR -  make[2]: *** [chrome-zh-TW] Error 2
10:35:13     INFO -  make[2]: Leaving directory `/builds/slave/m-aurora-andrd-ntly/build/obj-firefox/toolkit/locales'
10:35:13    ERROR -  make[1]: *** [chrome-zh-TW] Error 2
10:35:13     INFO -  make[1]: Leaving directory `/builds/slave/m-aurora-andrd-ntly/build/obj-firefox/mobile/locales'
10:35:13     INFO -  make: *** [chrome-zh-TW] Error 2
10:35:13    ERROR - Return code: 2
10:35:13    ERROR - Failed to add locale zh-TW!
version			End Throbber Time 	Elapsed	onload
locale53		2973.21428571429	3196.07142857143
locale13		2861.66666666667	3091.77777777778
single / 05/07/2012	2975.55555555556	3195.11111111111
single / 05/08/2012	2783.68421052632	3017.21052631579

Numbers look more reasonable for the 5/8 build.  Not sure what happened with the 5/7.  Was there some sort of regression/fix?  like the throbber?  Not sure without looking into things.
(In reply to Aki Sasaki [:aki] from comment #5)
> (In reply to Alex Keybl [:akeybl] from comment #0)
> > 2) Another locale's APK with the same product name and versioncode offset
> > available

> For speed purposes, I'm planning on:
> 
> * kicking off an en-US android native build + repack

> The former will have a larger versioncode than the apk on the store, so
> should fulfill (2).

This is http://people.mozilla.org/~asasaki/single_fr_newer.apk .
(In reply to Aki Sasaki [:aki] from comment #10)
> This is http://people.mozilla.org/~asasaki/single_fr_newer.apk .

Crap, removing this. This still has an appname of org.mozilla.firefox_aki2.
Fixing and respinning.
(In reply to Alex Keybl [:akeybl] from comment #0)
> Our testing will be to install #1 from the store, then install #2 from FTP
> (ensuring it's picked up by Google Play but does not update to #1), and then
> we'll upload #3 to the store and try updating (testing our ability to
> recombine the populations if desired in the future).


QA:

* Build #1 is currently in the store under https://play.google.com/store/apps/details?id=org.mozilla.firefox_aki&feature=search_result#?t=W251bGwsMSwxLDEsIm9yZy5tb3ppbGxhLmZpcmVmb3hfYWtpIl0.
* Build #2 is ready to be installed from http://people.mozilla.org/~asasaki/single_fr_newer.apk
* Once we've installed build #1 from the store and build #2 from the web across a couple of devices, let Aki know and he'll be push the third build up
Are we testing on tablets and phones or tablets only?
akeybl answered in irc as we are only testing phones.  (meant phones in comment 14 not tablets)

Also from email from gbrown:
I repeated Mark's tests on a Galaxy S (Froyo) and found similar results:

Baseline (Aurora en-US): 1970ms 
13 Locale Aurora: 2076ms (5% slower) 
53 Locale Aurora: 2218ms (13% slower) 

I repeated Mark's tests on a Galaxy Nexus (ICS) and found something rather different:

Baseline (Aurora en-US): 1548ms 
13 Locale Aurora: 1469ms (5% FASTER) 
53 Locale Aurora: 1622ms (5% slower) 

 - Geoff
(In reply to Alex Keybl [:akeybl] from comment #13)
> * Build #2 is ready to be installed from
> http://people.mozilla.org/~asasaki/single_fr_newer.apk
> * Once we've installed build #1 from the store and build #2 from the web
> across a couple of devices, let Aki know and he'll be push the third build up

I'm crashing on startup of build #2.  Occured on a Galaxy Nexus and Desire HD. 

05-08 23:53:13.154: ERROR/AndroidRuntime(9845): FATAL EXCEPTION: main
05-08 23:53:13.154: ERROR/AndroidRuntime(9845): java.lang.RuntimeException: Unable to get provider org.mozilla.firefox_aki.db.TabsProvider: java.lang.ClassNotFoundException: org.mozilla.firefox_aki.db.TabsProvider
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread.installProvider(ActivityThread.java:4240)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread.installContentProviders(ActivityThread.java:3992)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread.handleBindApplication(ActivityThread.java:3946)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread.access$1300(ActivityThread.java:123)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread$H.handleMessage(ActivityThread.java:1185)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.os.Handler.dispatchMessage(Handler.java:99)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.os.Looper.loop(Looper.java:137)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread.main(ActivityThread.java:4424)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at java.lang.reflect.Method.invokeNative(Native Method)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at java.lang.reflect.Method.invoke(Method.java:511)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:784)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:551)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at dalvik.system.NativeStart.main(Native Method)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845): Caused by: java.lang.ClassNotFoundException: org.mozilla.firefox_aki.db.TabsProvider
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at dalvik.system.BaseDexClassLoader.findClass(BaseDexClassLoader.java:61)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at java.lang.ClassLoader.loadClass(ClassLoader.java:501)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at java.lang.ClassLoader.loadClass(ClassLoader.java:461)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     at android.app.ActivityThread.installProvider(ActivityThread.java:4225)
05-08 23:53:13.154: ERROR/AndroidRuntime(9845):     ... 12 more
05-08 23:53:22.505: WARN/ActivityManager(201): Activity destroy timeout for ActivityRecord{41a324d0 org.mozilla.firefox_aki/.App}
05-08 23:53:33.122: WARN/ActivityManager(201): Timeout executing service: ServiceRecord{426a9bc0 org.mozilla.firefox_aki/org.mozilla.gecko.sync.syncadapter.SyncService}
05-08 23:53:33.146: INFO/ActivityManager(201): Crashing app skipping ANR: ProcessRecord{41a904d0 9845:org.mozilla.firefox_aki/10072} Executing service org.mozilla.firefox_aki/org.mozilla.gecko.sync.syncadapter.SyncService
05-08 23:54:24.521: DEBUG/dalvikvm(201): GC_CONCURRENT freed 2312K, 24% free 24331K/31943K, paused 3ms+6ms
05-08 23:54:28.560: INFO/Process(9845): Sending signal. PID: 9845 SIG: 9
05-08 23:54:28.576: INFO/ActivityManager(201): Process org.mozilla.firefox_aki (pid 9845) has died.
05-08 23:54:28.576: WARN/ActivityManager(201): Service crashed 2 times, stopping: ServiceRecord{426a9bc0 org.mozilla.firefox_aki/org.mozilla.gecko.sync.syncadapter.SyncService}
05-08 23:54:28.583: WARN/InputManagerService(201): Window already focused, ignoring focus gain of: com.android.internal.view.IInputMethodClient$Stub$Proxy@41d37f60
05-08 23:54:34.107: INFO/ActivityManager(201): START {act=android.intent.action.MAIN cat=[android.intent.category.LAUNCHER] flg=0x10200000 cmp=org.mozilla.firefox_aki/.App} from pid 444
05-08 23:54:34.193: INFO/dalvikvm(9865): Turning on JNI app bug workarounds for target SDK version 11...
05-08 23:54:34.193: INFO/ActivityManager(201): Start proc org.mozilla.firefox_aki for activity org.mozilla.firefox_aki/.App: pid=9865 uid=10072 gids={3003, 1015, 1006}
05-08 23:54:34.216: DEBUG/OpenGLRenderer(444): Flushing caches (mode 1)
05-08 23:54:34.247: INFO/ActivityThread(9865): Pub org.mozilla.firefox_aki.db.tabs: org.mozilla.firefox_aki.db.TabsProvider
05-08 23:54:34.247: DEBUG/AndroidRuntime(9865): Shutting down VM
05-08 23:54:34.247: WARN/dalvikvm(9865): threadid=1: thread exiting with uncaught exception (group=0x40a3f1f8)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865): FATAL EXCEPTION: main
05-08 23:54:34.247: ERROR/AndroidRuntime(9865): java.lang.RuntimeException: Unable to get provider org.mozilla.firefox_aki.db.TabsProvider: java.lang.ClassNotFoundException: org.mozilla.firefox_aki.db.TabsProvider
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread.installProvider(ActivityThread.java:4240)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread.installContentProviders(ActivityThread.java:3992)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread.handleBindApplication(ActivityThread.java:3946)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread.access$1300(ActivityThread.java:123)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread$H.handleMessage(ActivityThread.java:1185)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.os.Handler.dispatchMessage(Handler.java:99)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.os.Looper.loop(Looper.java:137)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread.main(ActivityThread.java:4424)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at java.lang.reflect.Method.invokeNative(Native Method)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at java.lang.reflect.Method.invoke(Method.java:511)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:784)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:551)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at dalvik.system.NativeStart.main(Native Method)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865): Caused by: java.lang.ClassNotFoundException: org.mozilla.firefox_aki.db.TabsProvider
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at dalvik.system.BaseDexClassLoader.findClass(BaseDexClassLoader.java:61)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at java.lang.ClassLoader.loadClass(ClassLoader.java:501)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at java.lang.ClassLoader.loadClass(ClassLoader.java:461)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     at android.app.ActivityThread.installProvider(ActivityThread.java:4225)
05-08 23:54:34.247: ERROR/AndroidRuntime(9865):     ... 12 more
05-08 23:54:34.255: WARN/ActivityManager(201):   Force finishing activity org.mozilla.firefox_aki/.App
05-08 23:54:34.302: DEBUG/OpenGLRenderer(444): Flushing caches (mode 0)
What was your system language?
Does it crash if you change it to French?
(In reply to Aki Sasaki [:aki] from comment #18)
> What was your system language?
> Does it crash if you change it to French?

Unfortunately, this phone (HTC Desire HD), doenst have a french system language option installed.
- both builds from comment 13 do update to the single pack English only if the locale is set to English. This is due to the "partial locale blocking but not enabling multiple locales to be uploaded" from earlier
- The French build was crashing; see comment 17; regardless of the crash, you can still update the build to what's on the market.
- Multilocale pack does not seem to be too much of a performance hit. (see comment 15, comment 9, and comment 2)
- Update overall seems to be ok 
- selecting the update button from about:firefox will update firefox to 5/9 build of aurora after updating from play.
Note: 
Even with the french language for the android OS, the http://people.mozilla.org/~asasaki/single_fr_newer.apk build is still crashing.  I had installed that version, and then updated with the google play single locale english after setting the Android OS to English and the build worked fine.
(In reply to Naoki Hirata :nhirata from comment #20)
> - both builds from comment 13 do update to the single pack English only if
> the locale is set to English. This is due to the "partial locale blocking
> but not enabling multiple locales to be uploaded" from earlier

This is OK as long as we're able to push up a multi-locale as build #3 and have all single-locale localization installs update to it. Could we push one of the multi-locale builds (with a higher versioncode than build #2) to the store?

Then we'd set the system localization to French, install build #2 (French) from the web, upload the multi-locale APK to Google Play, and verify that the multi-locale APK is offered.
Yes, we would need to test this; I believe Aki would have to create a new build and push it as a higher build id.

If Aki pushes the newer build when it's done, I'll put it as high priority to get it tested right away.
(In reply to Naoki Hirata :nhirata from comment #24)
> Yes, we would need to test this; I believe Aki would have to create a new
> build and push it as a higher build id.
> 
> If Aki pushes the newer build when it's done, I'll put it as high priority
> to get it tested right away.

Aki's respin of the single locale FR build is here: http://people.mozilla.org/~asasaki/fr.apk

I've confirmed it launches in FR and doesnt crash anymore on startup.  HTC Desire HD.
I'd still like to see the testing in Comment 23 to complete our understanding here, but this bug is no longer a blocker for beta based upon our go/no-go meeting today.
(In reply to Alex Keybl [:akeybl] from comment #26)
> I'd still like to see the testing in Comment 23 to complete our
> understanding here, but this bug is no longer a blocker for beta based upon
> our go/no-go meeting today.

Yes, this is the testing I was referring to in comment 24... I am glad tchung covered the testing of the respin of the single locale FR build.
blocking-fennec1.0: --- → betaN+
Whiteboard: [no-code]
Test multilocale build is building.
http://people.mozilla.org/~asasaki/fennec-14.0b1.multi.android-arm.apk

I've uploaded it to the test market account, but I haven't activated the new test multi apk yet.  Naoki has login details and can activate it when he's ready to test.
I tested on 3 different devices set to varying languages :
https://docs.google.com/spreadsheet/ccc?key=0Arku3jleCA0UdFpCOVEweFV6TVVQVHZzaUdLOUg3RXc#gid=1


Notes:
1) I believe Spanish isn't included in this list of multilocale?  It's translated to English for that language.  
2) French for the phone found "l10n test fr" instead of "l10n test"

Otherwise the builds seem ok.  No major issues as of yet?
Status: NEW → RESOLVED
Closed: 12 years ago
status-firefox14: --- → fixed
Resolution: --- → FIXED
It's fr and zh-TW only.
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.