Pages:
Author

Topic: [KOBO] Kobocoin - An African peer to peer payment/micropayment system - page 44. (Read 127548 times)

legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
Nice. The green bar is glowing. Looks like it's going to explode Shocked



 Cheesy
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
I ve just updated my win64 client 72 hours ago ,,,,from v2.0 i think it was ,,
and yeah it's loads better with the reject's
now I get 0 in 24hrs
Glad to hear it
 Smiley
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
hero member
Activity: 528
Merit: 500
I ve just updated my win64 client 72 hours ago ,,,,from v2.0 i think it was ,,
and yeah it's loads better with the reject's
now I get 0 in 24hrs
hero member
Activity: 528
Merit: 500
Quote from: no141 on April 01, 2015, 09:57:37 PM
@TheTribesman, Have you looked at http://www.hashengineeringsolutions.com/forum/index.php?topic=598.0 or in general that whole site(leaving the forum)?

The issue with X15 is it's 15 rounds of scientific hashing functions (blake, bmw, groestl, jh, keccak, skein, luffa, cubehash, shavite, simd, echo, hamsi, fugue, shabal, whirlpool). And you don't have native Java fuctions for all of them. As this is a POS coin now, how important is the hashing algorithm? Do you know what exactly the transactions are encoded with? I'm thinking way out of the box now, but good luck to figuring things out.
Thanks. Yes, I've seen that thread some time ago. I've gotten past that stage now (the hashing calc and chain seems to be working). Getting the app to display was another issue.

What you've brought up about the hashing chain is what was causing an issue with the Abe explorer. The original x15_hash module had been modified so heavily that the chaining order was incorrect (I see you got it correct in your post  Cheesy) and therefore the final hash produced was incorrect. Once that chain order was aligned with the coin's chain order (you can get that from the coin's source code) then the hashing calc was corrected.

The same problem also existed with X15 on the Android wallet (re the link you posted)

As you said, because it's PoS now (and we have the 2 dns seeders as well) there's a possibility (it's not scientific - just a hunch) that the latest Schildbach wallet will do the job if I embed the seeders just like the desktop wallets. I've downloaded the latest Bitcoinj and Schildbach wallet (a new kobocoinj and kobocoin-wallet), and I've started the build. I'm just setting up the build path. Once the build path is working I'll make the necessary changes and cross my fingers.

I suppose in theory the way to look at the android wallet is as an app sitting on top of the blockchain, rather than associating it with the PoS functions of the coin itself. All it needs to do is receive, send, and store, just like any other app that will eventually sit on the blockchain.




I like the transparency atm,,,

so this issue  only related to the android wallet yeah ?
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
@TheTribesman, Have you looked at http://www.hashengineeringsolutions.com/forum/index.php?topic=598.0 or in general that whole site(leaving the forum)?

The issue with X15 is it's 15 rounds of scientific hashing functions (blake, bmw, groestl, jh, keccak, skein, luffa, cubehash, shavite, simd, echo, hamsi, fugue, shabal, whirlpool). And you don't have native Java fuctions for all of them. As this is a POS coin now, how important is the hashing algorithm? Do you know what exactly the transactions are encoded with? I'm thinking way out of the box now, but good luck to figuring things out.
Thanks. Yes, I've seen that thread some time ago. I've gotten past that stage now (the hashing calc and chain seems to be working). Getting the app to display was another issue.

What you've brought up about the hashing chain is what was causing an issue with the Abe explorer. The original x15_hash module had been modified so heavily that the chaining order was incorrect (I see you got it correct in your post  Cheesy) and therefore the final hash produced was incorrect. Once that chain order was aligned with the coin's chain order (you can get that from the coin's source code) then the hashing calc was corrected.

The same problem also existed with X15 on the Android wallet (re the link you posted)

As you said, because it's PoS now (and we have the 2 dns seeders as well) there's a possibility (it's not scientific - just a hunch) that the latest Schildbach wallet will do the job if I embed the seeders just like the desktop wallets. I've downloaded the latest Bitcoinj and Schildbach wallet (a new kobocoinj and kobocoin-wallet), and I've started the build. I'm just setting up the build path. Once the build path is working I'll make the necessary changes and cross my fingers.

I suppose in theory the way to look at the android wallet is as an app sitting on top of the blockchain, rather than associating it with the PoS functions of the coin itself. All it needs to do is receive, send, and store, just like any other app that will eventually sit on the blockchain.
sr. member
Activity: 294
Merit: 250
@TheTribesman, Have you looked at http://www.hashengineeringsolutions.com/forum/index.php?topic=598.0 or in general that whole site(leaving the forum)?

The issue with X15 is it's 15 rounds of scientific hashing functions (blake, bmw, groestl, jh, keccak, skein, luffa, cubehash, shavite, simd, echo, hamsi, fugue, shabal, whirlpool). And you don't have native Java fuctions for all of them. As this is a POS coin now, how important is the hashing algorithm? Do you know what exactly the transactions are encoded with? I'm thinking way out of the box now, but good luck to figuring things out.
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
One reason for the delay in writing the shared library is that I'm also investigating other solutions. I might be able to get away with building our android wallet based on the standard schildbach wallet (now that we have 2 dns seeders). It makes future updates a lot easier. I'm testing if it's possible.
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
@TheTribesman, Aren't you able to look at crash reports of the application?

What app can people download that shows a working example of what you are trying to do?

The adb logs show the crash reports (it's a realtime log of the kernel that's running on the phone). This is what's I'm seeing:

Code:
F/art     (26408): art/runtime/runtime.cc:284] "GCDaemon" prio=5 tid=9 Waiting
    F/art     (26408): art/runtime/runtime.cc:284]   | group="" sCount=0 dsCount=0 obj=0x12c3d200 self=0xb7817508
    F/art     (26408): art/runtime/runtime.cc:284]   | sysTid=26423 nice=0 cgrp=apps sched=0/0 handle=0xb7817ac8
    F/art     (26408): art/runtime/runtime.cc:284]   | state=S schedstat=( 0 0 0 ) utm=0 stm=0 core=0 HZ=100
    F/art     (26408): art/runtime/runtime.cc:284]   | stack=0xa73db000-0xa73dd000 stackSize=1036KB
    F/art     (26408): art/runtime/runtime.cc:284]   | held mutexes=
    F/art     (26408): art/runtime/runtime.cc:284]   kernel: futex_wait_queue_me+0xe0/0x108
    F/art     (26408): art/runtime/runtime.cc:284]   kernel: futex_wait+0x170/0x238
    F/art     (26408): art/runtime/runtime.cc:284]   kernel: do_futex+0xfc/0x9fc
    F/art     (26408): art/runtime/runtime.cc:284]   kernel: ret_fast_syscall+0x0/0x30
    F/art     (26408): art/runtime/runtime.cc:284]   native: #02 pc 0001f39b  /system/lib/libbinder.so (android::IPCThreadState::talkWithDriver(bool)+138)
    F/art     (26408): art/runtime/runtime.cc:284]   native: #04 pc 0001f8ed  /system/lib/libbinder.so (android::IPCThreadState::joinThreadPool(bool)+48)
    F/art     (26408): art/runtime/runtime.cc:284]   native: #06 pc 000104fd  /system/lib/libutils.so (android::Thread::_threadLoop(void*)+112)
    F/art     (26408): art/runtime/runtime.cc:284]   native: #08 pc 000100d3  /system/lib/libutils.so (???)
    F/art     (26408): art/runtime/runtime.cc:284]   native: #10 pc 00013e37  /system/lib/libc.so (__start_thread+6)
    F/art     (26408): art/runtime/runtime.cc:284]
    F/libc    (26408): Fatal signal 6 (SIGABRT), code -6 in tid 26408 (_android_wallet)
    I/DEBUG   (  271): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
    I/DEBUG   (  271): Build fingerprint: 'motorola/falcon_retgb/falcon_umts:5.0.2/LXB22.46-28.1/1:user/release-keys'
    I/DEBUG   (  271): Revision: 'p3c0'
    I/DEBUG   (  271): ABI: 'arm'
    I/DEBUG   (  271): pid: 26408, tid: 26408, name: _android_wallet  >>> com.machadocrypto.kobocoin_android_wallet <<<
    I/DEBUG   (  271): signal 6 (SIGABRT), code -6 (SI_TKILL), fault addr --------
    I/DEBUG   (  271): Abort message: 'art/runtime/check_jni.cc:65] JNI DETECTED ERROR IN APPLICATION: JNI RegisterNatives called with pending exception 'java.lang.ClassNotFoundException' thrown in unknown throw location'
    I/DEBUG   (  271):     r0 00000000  r1 00006728  r2 00000006  r3 00000000
    I/DEBUG   (  271):     r4 b6f66114  r5 00000006  r6 0000000b  r7 0000010c
    I/DEBUG   (  271):     r8 00000000  r9 b7604908  sl b760e4c0  fp 00000001
    I/DEBUG   (  271):     ip 00006728  sp beb9e9b0  lr b6f01679  pc b6f26c30  cpsr 60070010
    I/DEBUG   (  271):
    I/DEBUG   (  271): backtrace:
    I/DEBUG   (  271):     #00 pc 0003bc30  /system/lib/libc.so (tgkill+12)
    I/DEBUG   (  271):     #01 pc 00016675  /system/lib/libc.so (pthread_kill+52)
    I/DEBUG   (  271):     #02 pc 00017287  /system/lib/libc.so (raise+10)
    I/DEBUG   (  271):     #03 pc 00013ad9  /system/lib/libc.so (__libc_android_abort+36)
    I/DEBUG   (  271):     #04 pc 00012264  /system/lib/libc.so (abort+4)
    I/DEBUG   (  271):     #05 pc 0022f2e3  /system/lib/libart.so (art::Runtime::Abort()+170)
    I/DEBUG   (  271):     #06 pc 000a7109  /system/lib/libart.so (art::LogMessage::~LogMessage()+1360)
    I/DEBUG   (  271):     #07 pc 000b1217  /system/lib/libart.so (art::JniAbort(char const*, char const*)+1102)
    I/DEBUG   (  271):     #08 pc 000b1751  /system/lib/libart.so (art::JniAbortF(char const*, char const*, ...)+68)
    I/DEBUG   (  271):     #09 pc 000b49d7  /system/lib/libart.so (art::ScopedCheck::ScopedCheck(_JNIEnv*, int, char const*)+1322)
    I/DEBUG   (  271):     #10 pc 000bf6a5  /system/lib/libart.so (art::CheckJNI::RegisterNatives(_JNIEnv*, _jclass*, JNINativeMethod const*, int)+36)
    I/DEBUG   (  271):     #11 pc 00003637  /data/app/com.machadocrypto.kobocoin_android_wallet-1/lib/arm/libx15.so (JNI_OnLoad+42)
    I/DEBUG   (  271):     #12 pc 001e6757  /system/lib/libart.so (art::JavaVMExt::LoadNativeLibrary(std::__1::basic_string, std::__1::allocator > const&, art::Handle, std::__1::basic_string, std::__1::allocator >*)+1610)
    I/DEBUG   (  271):     #13 pc 0020e14f  /system/lib/libart.so (art::Runtime_nativeLoad(_JNIEnv*, _jclass*, _jstring*, _jobject*, _jstring*)+534)
    I/DEBUG   (  271):     #14 pc 00077c99  /system/framework/arm/boot.oat
    W/ActivityManager(  873): Process com.machadocrypto.kobocoin_android_wallet has crashed too many times: killing!
    I/DEBUG   (  271):
    I/DEBUG   (  271): Tombstone written to: /data/tombstones/tombstone_08
    I/BootReceiver(  873): Copying /data/tombstones/tombstone_08 to DropBox (SYSTEM_TOMBSTONE)
    W/ActivityManager(  873):   Force finishing activity com.machadocrypto.kobocoin_android_wallet/.WalletActivity
    W/WindowManager(  873): Failed looking up window
    W/WindowManager(  873): java.lang.IllegalArgumentException: Requested window android.view.ViewRootImpl$W@1a0a87af does not exist
    W/WindowManager(  873): at com.android.server.wm.WindowManagerService.windowForClientLocked(WindowManagerService.java:8457)
    W/WindowManager(  873): at com.android.server.wm.WindowManagerService.windowForClientLocked(WindowManagerService.java:8448)
    W/WindowManager(  873): at com.android.server.wm.WindowManagerService.removeWindow(WindowManagerService.java:2574)
    W/WindowManager(  873): at com.android.server.wm.Session.remove(Session.java:186)
    W/WindowManager(  873): at android.view.ViewRootImpl.dispatchDetachedFromWindow(ViewRootImpl.java:2959)
    W/WindowManager(  873): at android.view.ViewRootImpl.doDie(ViewRootImpl.java:5460)
    W/WindowManager(  873): at android.view.ViewRootImpl$ViewRootHandler.handleMessage(ViewRootImpl.java:3283)
    W/WindowManager(  873): at android.os.Handler.dispatchMessage(Handler.java:102)
    W/WindowManager(  873): at android.os.Looper.loop(Looper.java:135)
    W/WindowManager(  873): at android.os.HandlerThread.run(HandlerThread.java:61)
    W/WindowManager(  873): at com.android.server.ServiceThread.run(ServiceThread.java:46)
    W/ActivityManager(  873): Wtf, activity ActivityRecord{d0ea2ea u0 com.machadocrypto.kobocoin_android_wallet/.WalletActivity t108 f} in proc activity list not using proc ProcessRecord{3bd79516 26408:com.machadocrypto.kobocoin_android_wallet/u0a176}?!?
    E/lowmemorykiller(  256): Error writing /proc/26408/oom_score_adj; errno=22
    E/JavaBinder(  873): !!! FAILED BINDER TRANSACTION !!!
    E/JavaBinder(  873): !!! FAILED BINDER TRANSACTION !!!

This is the culprit:

Code:
    F/libc    (26408): Fatal signal 6 (SIGABRT), code -6 in tid 26408 (_android_wallet)
    I/DEBUG   (  271): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
    I/DEBUG   (  271): Build fingerprint: 'motorola/falcon_retgb/falcon_umts:5.0.2/LXB22.46-28.1/1:user/release-keys'
    I/DEBUG   (  271): Revision: 'p3c0'
    I/DEBUG   (  271): ABI: 'arm'
    I/DEBUG   (  271): pid: 26408, tid: 26408, name: _android_wallet  >>> com.machadocrypto.kobocoin_android_wallet <<<
    I/DEBUG   (  271): signal 6 (SIGABRT), code -6 (SI_TKILL), fault addr --------
    I/DEBUG   (  271): Abort message: 'art/runtime/check_jni.cc:65] JNI DETECTED ERROR IN APPLICATION: JNI RegisterNatives called with pending exception 'java.lang.ClassNotFoundException' thrown in unknown throw location'
  

This is what should be showing:

Code:
I/Adreno-EGL( 6219): : EGL 1.4 QUALCOMM build: AU_LINUX_ANDROID_LA.BF.1.1.04.04.02.162.107_msm8226_LA.BF.1.1__release_AU ()
I/Adreno-EGL( 6219): OpenGL ES Shader Compiler Version: E031.25.01.03
I/Adreno-EGL( 6219): Build Date: 10/28/14 Tue
I/Adreno-EGL( 6219): Local Branch:
I/Adreno-EGL( 6219): Remote Branch: quic/l_LNX.LA.3.6
I/Adreno-EGL( 6219): Local Patches: NONE
I/Adreno-EGL( 6219): Reconstruct Branch: AU_LINUX_ANDROID_LA.BF.1.1.04.04.02.162.107 + cb93e16 + f50fe49 + d7c18e6 + 5b9a565 + 0f3a25d + 607156e + 75511aa + e4d16c0 + 686f3eb + 211a271 + dd281ee +  NOTHING
I/Adreno Profiler( 6219): Using libraries from system partition
I/OpenGLRenderer( 6219): Initialized EGL, version 1.4
I/Adreno Profiler( 6219): Using libraries from system partition
I/Adreno Profiler( 6219): Using libraries from system partition
D/OpenGLRenderer( 6219): Enabling debug mode 0

I have to use an SDK, and an NDK (as opposed to an SDK - yes, a whole new world) to build an X15 compliant shared library to resolve this issue.

If you goto Google Play and download the Peercoin wallet (kobocoin wallet is based on Peercoin wallet), or any other crypto wallet you can see how they perform. It's just a shame that X15/PoS/Android hasn't been cracked yet. The search had been going on for months before I joined in. Cracking the X15 hashing issue on Abe was a major step forward.

Also here's one I was playing with earlier but unfortunately it's not compatible with the getvalidhashes call that is required for the X15/PoS wallet (allegedly):

https://bitcointalksearch.org/topic/m.10607718

It will be done, but I can only hope that the app jumps into life when I fix it, rather than just getting presented with the next challenge (as I suspect will happen).
 Smiley
sr. member
Activity: 294
Merit: 250
@TheTribesman, Aren't you able to look at crash reports of the application?

What app can people download that shows a working example of what you are trying to do?
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
Sorry for the lack of proper updates.

The guts of the android wallet is now running. The hashing is working correctly and the blockchain seems fine, but the wallet rendering engine keeps crashing. The wallet is a lot further forward in that all settings seem to be correct. The kobocoinj is no longer being worked on. It's doing what it needs to do.

The challenge now is to change the build engine and rebuild the wallet UI. It's not straightforward but it's in progress.

God bless you and help you Smiley
...and send me all the right software with all the dependencies already met!

Oh, the things I've seen. I can never be the same again. I feel lighter and more efficient (but dirtier and angrier at the same time). My brain now scrolls in adb log format (matrix scroll has nothing on adb logcat), and the tips of my fingers hurt!

I think I have to give some sort of measure of progress:

Where we were 5 days ago: If you run the kobocoin wallet on a phone, it crashes immediately.

Where we are now: If you run the kobocoin wallet, the app starts, you get the white background, the actual wallet is about to load (yes... yes...), and then it crashes (boooo...).

it's actually a lot of progress, but It's brutally slow and a lot more work than you can imagine, and as I said, each error message seems to open a whole new landscape of new software that takes hours just to install and get running. They're all easy enough to use when they're up and running, but you have to get them up and running first (and they will ALL defy your every move).

Oh yes, they know the websites you'll visit for answers. They know them well. They planted them. They're waiting for you to come back with your big bag of fresh answers. They got a plan... They have decided to take you, and your mind, and your big bag of answers, on a wild detour for the next 5-6 hours. You will never get a chance to open up your big bag of answers. The software doesn't even take you back to where you left off. When you start it, it won't even exhibit the same behaviour (but it was working... what happened?..) Oh no, there's a whole new world out there. They want you to visit this new wondrous painful land of no documentation. They want to watch you put a keyword into google and get 2 results that lead to pages of nothing but code. Solve that, they say, and we'll give you a chance to see if your big bag of answers solves this problem.

So you solve it, and the big bag of answers fails anyways (laughter all round).
 Cheesy

(NB - yeah, code can laugh)
legendary
Activity: 1470
Merit: 1024
Sorry for the lack of proper updates.

The guts of the android wallet is now running. The hashing is working correctly and the blockchain seems fine, but the wallet rendering engine keeps crashing. The wallet is a lot further forward in that all settings seem to be correct. The kobocoinj is no longer being worked on. It's doing what it needs to do.

The challenge now is to change the build engine and rebuild the wallet UI. It's not straightforward but it's in progress.

God bless you and help you Smiley
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
Sorry for the lack of proper updates.

The guts of the android wallet is now running. The hashing is working correctly and the blockchain seems fine, but the wallet rendering engine keeps crashing. The wallet is a lot further forward in that all settings seem to be correct. The kobocoinj is no longer being worked on. It's doing what it needs to do.

The challenge now is to change the build engine and rebuild the wallet UI. It's not straightforward but it's in progress.
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
So why (some may ask) does a middle-aged Celt who's never been within 500 miles of the 'dark continent' care about an essentially African coin?

- Especially when I'm so heavily involved with 2 other coins (see sig, I'm no pumper).

Well, it's quite simple really. As I went through the IBOGA ritual in 2005, (In Holland) I'm spiritually/technically BWITI.

Also, I'm shamed to the core of my being about what evils my race did to the people of Africa.

So, from the heart, my very best wishes for success in helping the peoples of that region to see the benefits of de-centralised financial freedom Grin
Thanks. I'm really hoping it will have major benefits, especially for the unbanked or non-creditworthy.
hero member
Activity: 938
Merit: 1000
Crypto-missionary to the Kingdom of Rhegged
So why (some may ask) does a middle-aged Celt who's never been within 500 miles of the 'dark continent' care about an essentially African coin?

- Especially when I'm so heavily involved with 2 other coins (see sig, I'm no pumper).

Well, it's quite simple really. As I went through the IBOGA ritual in 2005, (In Holland) I'm spiritually/technically BWITI.

Also, I'm shamed to the core of my being about what evils my race did to the people of Africa.

So, from the heart, my very best wishes for success in helping the peoples of that region to see the benefits of de-centralised financial freedom Grin
full member
Activity: 140
Merit: 100
pretty good looks..We waiting for Launch...  Wink
legendary
Activity: 1288
Merit: 1000
I'm glad I decided to hold this coin, the dev is doing really well, good things happening here Smiley
legendary
Activity: 1470
Merit: 1024
UPDATE:

There is no real update other than work is still going on with the Android wallet. A major hurdle has been cleared, but as long as it's not working, then there's no real news to report. The only update of any worth at this point is a screenshot of the application on an android phone. That is what I'm working towards. Once I have that, then testing can begin.

I know progress is slow, but it's the same for every team trying to improve their platform. There's a lot of collaboration going on, but it's global (and slow), and a lot of times it's about fixing issues that people haven't seen before.

I'll keep you updated.
 Smiley

we ll follow you Smiley  android wallet was late but we know that people wait to use it.
legendary
Activity: 1019
Merit: 1003
Kobocoin - Mobile Money for Africa
UPDATE:

There is no real update other than work is still going on with the Android wallet. A major hurdle has been cleared, but as long as it's not working, then there's no real news to report. The only update of any worth at this point is a screenshot of the application on an android phone. That is what I'm working towards. Once I have that, then testing can begin.

I know progress is slow, but it's the same for every team trying to improve their platform. There's a lot of collaboration going on, but it's global (and slow), and a lot of times it's about fixing issues that people haven't seen before.

I'll keep you updated.
 Smiley
Pages:
Jump to: