Multitouch possible in theory?

[Guest Spooke]

Yap I made the post that single touch works a few secs before I realised that multitouch works too :P

So what are the current problems so far? Do u think u can get it fixed soon and ready to be used?

As in the picture it seems that the touch points are offset from ur fingers ;)

Still u done a great job, cant wait to use this ;) :)

[Guest MarcusHenrique]

And what are the chances that this BigBear MT ROM works on my NON-pulse U8220? :)

[Guest BigBearMDC]

So where do we go from here? What's left to do before we can see a multitouch ROM? :)

I'll see what I can do to add this to the KeyInputQueue.java.

That is the last step to real multitouch (backported from Android 2.0).

I'd then add ext3 and ext4 support to my kernel, add A2SD and dropbear (if I find out how) and then throw the whole stuff into my CROM ;)

Of course you will get the source and my boot.img until I finish my ROM :P

[Guest Spooke]

I'll see what I can do to add this to the KeyInputQueue.java.

That is the last step to real multitouch (backported from Android 2.0).

I'd then add ext3 and ext4 support to my kernel, add A2SD and dropbear (if I find out how) and then throw the whole stuff into my CROM :)

Of course you will get the source and my boot.img until I finish my ROM ;)

Sounds good :P, U should have a custom theme on it too :D, cant wait till u get it done ;) :P

[Guest BigBearMDC]

And what are the chances that this BigBear MT ROM works on my NON-pulse U8220? ;)

Unfortunately, I'm not sure if this will work on the non T-Mobile version B)

But I'll try it ;)

So what are the current problems so far? Do u think u can get it fixed soon and ready to be used?

As in the picture it seems that the touch points are offset from ur fingers :P

Still u done a great job, cant wait to use this B) :)

Yap if I touch the screen with 2 fingers, there is an offset.

But the red circle moves in relation to my fingers if I move them around on the screen, so that's not a very big problem :P

I think it has something to do with the procedure I use to pass the coordinates of the second touch point to the MTVisualizer.apk.

This was not only my effort!

Many people in this forum helped me, and Luke Hutchinson told me what I have to do to pass the coordinates through the KeyInputQueue.

At this point I thank everybody who helped me.

Thanks, really!

I would never have gotten so far without your help.

:D

Edited March 25, 2010 by BigBearMDC

[Guest MarcusHenrique]

Unfortunately, I'm not sure if this will work on the non T-Mobile version ;)

But I'll try it :)

I tought so, your effort is very appreciated!

Any chances Official Android update from huawei comes with MT support already?

[Guest BigBearMDC]

I tought so, your effort is very appreciated!

Any chances Official Android update from huawei comes with MT support already?

Well I'm not sure...

HTC can handle the sues by Apple - but can Huawei?

But even if not, we will port it, at least I'm not afraid of Steve Jobs :)

Edited March 25, 2010 by BigBearMDC

[Guest MarcusHenrique]

Well I'm not sure...

HTC can handle the sues by Apple - but can Huawei?

But even if not, we will port it, at least I'm not afraid of Steve Jobs :)

LOL, haven't tought about Apple's patent. That sucks.

Keep up the good work, you're close to be worshiped! hehe

[]'s

[Guest Blake Loring]

Wouldn't be to worried, apple haven't followed up any of the multi-touch patent infringements afaik - besides, huawei are a massive company. Though, I still doubt its a major priority for huawei development team when trying to roll out a new version of the kernel. - On a side note once you'v finished the ROM is there any chance we could get the Kernel source, Being able to add stuff like Netfiler and compache to a already good ROM would be a major plus for me.

Edited March 25, 2010 by Blake Loring

[Guest BigBearMDC]

Wouldn't be to worried, apple haven't followed up any of the multi-touch patent infringements afaik - besides, huawei are a massive company. Though, I still doubt its a major priority for huawei development team when trying to roll out a new version of the kernel. - On a side note once you'v finished the ROM is there any chance we could get the Kernel source, Being able to add stuff like Netfiler and compache to a already good ROM would be a major plus for me.

Sure :)

But I think it would be easier if I just upload the output of diff?

Greetings,

BigBear

[Guest Blake Loring]

Sure :)

But I think it would be easier if I just upload the output of diff?

Greetings,

BigBear

That, A git/svn repo, or a zip file would do me just fine =). (On a side note with a repo you could try and get a few people to contribute)

[Guest BigBearMDC]

Hi guys,

I pulled up a git repo @github

I'll update this source frequently :)

Greetings,

BigBear

[Guest BigBearMDC]

Hi guys

I patched the driver now to return ABS_HAT0X and ABS_HAT0Y.

Now we have to patch the Android Stack.

Have a look at "/frameworks/base/services/java/com/android/server/KeyInputQueue.java".

We'll have to path this to also recognize ABS_HAT0XY.

That's the difficult part.

Hopefully someone is willing to help me :)

BTW I get an internal compiler error when I try to compile android 1.5:

target thumb C++: libandroid_runtime <= frameworks/base/core/jni/ActivityManager.cpp

frameworks/base/core/jni/android/opengl/util.cpp: In function 'int android::register_android_opengl_classes(JNIEnv*)':

frameworks/base/core/jni/android/opengl/util.cpp:756: internal compiler error: in merge_alias_info, at tree-ssa-copy.c:250

Please submit a full bug report,

with preprocessed source if appropriate.

See <URL:http://gcc.gnu.org/bugs.html> for instructions.

make: *** [out/target/product/generic/obj/SHARED_LIBRARIES/libandroid_runtime_intermediates/android/opengl/util.o] Fehler 1

make: *** Warte auf noch nicht beendete Prozesse...

frameworks/base/core/jni/com_google_android_gles_jni_GLImpl.cpp: In function 'void android_glGenTextures__I_3II(JNIEnv*, _jobject*, jint, _jintArray*, jint)':

frameworks/base/core/jni/com_google_android_gles_jni_GLImpl.cpp:864: internal compiler error: in merge_alias_info, at tree-ssa-copy.c:250

Please submit a full bug report,

with preprocessed source if appropriate.

See <URL:http://gcc.gnu.org/bugs.html> for instructions.

make: *** [out/target/product/generic/obj/SHARED_LIBRARIES/libandroid_runtime_intermediates/com_google_android_gles_jni_GLImpl.o] Fehler 1[/code]


Has anyone seen this before?


Greetings,

BigBear

[Guest david_dawkins]

The internal compiler error could be caused by the optimizer; however if you haven't modified the compile flags, then it's more an indicator that you're not using the same toolchain as the devs who checked it in.

I had a quick look for the two CPP files to see what exactly it was barfing on, but it might be easier if you email me those files. In my experience though, those errors happen on innocent-looking sections of code. We might find something though.

Dave

[Guest BigBearMDC]

The internal compiler error could be caused by the optimizer; however if you haven't modified the compile flags, then it's more an indicator that you're not using the same toolchain as the devs who checked it in.

I had a quick look for the two CPP files to see what exactly it was barfing on, but it might be easier if you email me those files. In my experience though, those errors happen on innocent-looking sections of code. We might find something though.

Dave

Hmm I'll try it with gcc-4.4 first, if that doesn't work either, I'll have a look at the CPP files.

I first thought I get the errors because I used the -j4 flag, so that the build process runs with 4 treads.

But that wasn't the error.

I just took 4 times longer until I got the error =/

Greetings,

BigBear

[Guest BigBearMDC]

Okay gcc-4.4 didn't work either, I got an error regarding a variable that was declared in an incorrect scope or something like this...

I reinstalled gcc-4.3 and now I get another very weird error:

#

# An unexpected error has been detected by HotSpot Virtual Machine:

#

#  SIGSEGV (0xb) at pc=0x40612370, pid=27801, tid=1795521392

#

# Java VM: Java HotSpot(TM) Server VM (1.5.0_19-b02 mixed mode)

# Problematic frame:

# V  [libjvm.so+0x47e370]

#

# An error report file with more information is saved as hs_err_pid27801.log

#

# If you would like to submit a bug report, please visit:

#   http://java.sun.com/webapps/bugreport/crash.jsp

#

/bin/bash: Zeile 1: 27801 Aborted				 javac -J-Xmx512M -target 1.5 -Xmaxerrs 9999999 -encoding ascii -bootclasspath out/target/common/obj/JAVA_LIBRARIES/core_intermediates/classes.jar -g -extdirs "" -d out/target/common/obj/JAVA_LIBRARIES/core_intermediates/classes \@out/target/common/obj/JAVA_LIBRARIES/core_intermediates/classes/java-source-list-uniq

make: *** [out/target/common/obj/JAVA_LIBRARIES/core_intermediates/classes-full-debug.jar] Fehler 41

make: *** Warte auf noch nicht beendete Prozesse...

Note: Some input files use unchecked or unsafe operations.

Note: Recompile with -Xlint:unchecked for details.

Note: Some input files use unchecked or unsafe operations.

Note: Recompile with -Xlint:unchecked for details.[/code]


What's that???


I solved the kernel problem, no I can't compile Android anymore ... :)


Greetings,

BigBear

[Guest BigBearMDC]

/tmp/ccQTwYCG.s: Assembler messages:

/tmp/ccQTwYCG.s:51991: Error: unknown pseudo-op: `.lbb586238'

make: *** [out/host/linux-x86/obj/EXECUTABLES/make_cfst_intermediates/make_cfst.o] Fehler 1[/code]


???

[Guest Blake Loring]

/tmp/ccQTwYCG.s: Assembler messages:

/tmp/ccQTwYCG.s:51991: Error: unknown pseudo-op: `.lbb586238'

make: *** [out/host/linux-x86/obj/EXECUTABLES/make_cfst_intermediates/make_cfst.o] Fehler 1

???

Try a make clean if you haven't already. Also check that you have the correct version of binutils, I'v found other errors about pseudo instructions came down to that.

[Guest david_dawkins]

Right, that could be anything. SIGSEGV is basically dereferencing an address that doesn't exist in our process space; one of the common ones is dereferencing 0 (NULL), which is what would happen if malloc failed and it wasn't checked. So, how are you doing for RAM/swap on that machine?

[Guest BigBearMDC]

Try a make clean if you haven't already. Also check that you have the correct version of binutils, I'v found other errors about pseudo instructions came down to that.

I run "make clean" everytime I start compiling the source.

My version of binutils is 2.20.

But I'm not getting past the JAVA error anymore ....

I already compiled Android on my other dev environment without problems...

[Guest BigBearMDC]

Right, that could be anything. SIGSEGV is basically dereferencing an address that doesn't exist in our process space; one of the common ones is dereferencing 0 (NULL), which is what would happen if malloc failed and it wasn't checked. So, how are you doing for RAM/swap on that machine?

Oh no, hopefully my RAM isn't damaged :)

But if it tries to dereference a address that isn't allocated it would mean its a Java bug...

Edit:

Hold on, I'll heck my RAM

This is really getting annoying...

Edited March 27, 2010 by BigBearMDC

[Guest david_dawkins]

I mean swap, as in "running out of memory". Normally you'd get OutOfMemoryException from Java or .NET, but for C/CPP apps it depends on how well the app was written (whether they check return from malloc/calloc and report it as "out of memory" etc).

[Guest BigBearMDC]

I mean swap, as in "running out of memory". Normally you'd get OutOfMemoryException from Java or .NET, but for C/CPP apps it depends on how well the app was written (whether they check return from malloc/calloc and report it as "out of memory" etc).

I have 3GB RAM and 6GB swap.

I did nothing special, I just created a swap partition (its the second partition on my hdd). I compiled android successfully before on my other dev machine that has only 1GB RAM and a 4GB swap partition.

Greetings,

BigBear

[Guest Blake Loring]

I have 3GB RAM and 6GB swap.

I did nothing special, I just created a swap partition (its the second partition on my hdd). I compiled android successfully before on my other dev machine that has only 1GB RAM and a 4GB swap partition.

Greetings,

BigBear

Dang, so "sudo make" is a no go, have you tried uninstalling and reinstalling your JRE and JDK - Or maybe a different version? All we can really conclude from that error log is that libjvm or the JNI are crashing when you try to compile and that they are crashing natively, so a uninstall/different version may help

[Guest BigBearMDC]

Dang, so "sudo make" is a no go, have you tried uninstalling and reinstalling your JRE and JDK - Or maybe a different version? All we can really conclude from that error log is that libjvm or the JNI are crashing when you try to compile and that they are crashing natively, so a uninstall/different version may help

Thanks, but it was my RAM - it's demaged :)

I just replaced it.

Hopefully the errors now vanish...

Greetings,

BigBear