Database Develop

[Q] [REQUEST] Fix for SSL Issue

Hi! In GB there is an issue about ssl. Google has accepted it and says that it will solved in a future release (2.3.4) but let us the code for a patch in custom rom.
Any developer could be interested in it? (please)
Here is the link: https://code.google.com/p/android/issues/detail?id=15356
(comment n. 6) It is described for nexus one, but i attest it on i9000
Thanks a lot and sorry for my poor english.

+1 looks like this is necessary if you're setting up ActiveSync and it's the fix is obvious:
diff --git a/patches/small_records.patch b/patches/small_records.patch
index f5f2286..a2ea51c 100644
--- a/patches/small_records.patch
+++ b/patches/small_records.patch
@@ -157,7 +157,7 @@
+ /* If we receive a valid record larger than the current buffer size,
+ * allocate some memory for it.
+ */
-+ if (rr->length > s->s3->rbuf.len - SSL3_RT_HEADER_LENGTH)
++ if (rr->length > s->s3->rbuf.len - SSL3_RT_HEADER_LENGTH - align)
+ {
+ if ((p=OPENSSL_realloc(s->s3->rbuf.buf, rr->length + SSL3_RT_HEADER_LENGTH + align))==NULL)
+ {
diff --git a/ssl/s3_pkt.c b/ssl/s3_pkt.c
index 51f21e0..c9ef2cd 100644
--- a/ssl/s3_pkt.c
+++ b/ssl/s3_pkt.c
@@ -359,7 +359,7 @@ fprintf(stderr, "Record type=%d, Length=%d\n", rr->type, rr->length);
/* If we receive a valid record larger than the current buffer size,
* allocate some memory for it.
*/
- if (rr->length > s->s3->rbuf.len - SSL3_RT_HEADER_LENGTH)
+ if (rr->length > s->s3->rbuf.len - SSL3_RT_HEADER_LENGTH - align)
{
if ((p=OPENSSL_realloc(s->s3->rbuf.buf, rr->length + SSL3_RT_HEADER_LENGTH + align))==NULL)
{

The simplest solution would be for someone to compile AOSP libssl.so with the patch and redistribute this library with a simple ADB script to get it running on your phone. Presumably you'll need root though. A complete custom ROM isn't really necessary.

That would be great! If not i must go 2.2.1 again. I only can acceda exchange work via activesync.
I think must be more people with this issue.
Enviado desde mi GT-I9000 usando Tapatalk

Development team V2

Hello everyone,
I want to start again with a development team. After miroslav decided to finish the project on his own, I want to try it again. But this time with a real team Only with people who want to participate, and not just type make. Who wants to join?

miroslav_mm said:
So, I opened your eyes and showed real face our "superstar"!?
Ok! Let's stop this stupid discussion and continuous develop ROM.
At this moment I have main part:
Torchscreen +
BT (BT-calling) +
WiFi (need to test) +
Sensor +
GSM (Waiting call) +
Mobile Data +
Audio +
Vibrate +
HW + 3D ??? (need new kernel or port new 3D kernel driver (kgsl) (in progress))
usb storage ??? (only with special app (native in progress))
Camera ??? (in progress)
Video codec ??? (in progress)
Anyone doing something? Maybe fix some stuff?
Click to expand...
Click to collapse
He just want to clean his topic. Send him a PM.

I understand that but there is still no team. He just made his project open for patches. And I think, personally, a team could help the speed on development of a ROM So please join
Sent from my GT540 using Tapatalk

So join his team , and start patching?

wwenigma said:
So join his team , and start patching?
Click to expand...
Click to collapse
Since when does he have a team? What I know about it, is that he just made some kind of open-source project where you can add your own updates and where he gets no credit.
Sent from my GT540 using Tapatalk

I can help you with Java stuff!

I think miroslav_mm wants to release ICS under name of SWIFT-TEAM, with developer names in credits.... may you ask him about it?
Do not split guys...

Miroslav said that it's only him on the team...so i don't see any team there..if we are more than 1 guy in a team we can get faster a daily ROM for our Phones

rexes13 said:
Miroslav said that it's only him on the team
Click to expand...
Click to collapse
Jep, because PCF+MGP join then leave and jasper580 cant make neccessary steps... ask him. All two, ask him about it.

wwenigma said:
Jep, because PCF+MGP join then leave and jasper580 cant make neccessary steps... ask him. All two, ask him about it.
Click to expand...
Click to collapse
I can make nesseccary steps if I want, but I don't want to. I want to start cleaned up. That's why I created this topic. So:
If you want to join: welcome.
If you don't want to: stop spamming
@rexes13 Welcome to the team!
Sent from my GT540 using Tapatalk

Good news.
Yesterday I successfully ported 3d kernel driver (kgsl) from 35, so now we have full functionality kernel with OpenGL ES-CM 1.1.
So HW, 3D and other stuff with support OpenGL working, not excellent but working, need improve source code Surfaceflinger and HardwareDisplay.
Maybe next weak I will make new thread and publish first public test version, but all development moments we can discus here.
Before I need redownload all sources CM9 again. Yesterday I had the same error like in wwenigma and aaa801
Code:
target thumb C: libhardware_legacy <= hardware/libhardware_legacy/power/power.c
arm-linux-androideabi-gcc: ": No such file or directory
<command-line>: warning: missing terminating " character
and I dont know WTF, so delete all source tree. Maybe today will try download again and build again.

hell yeah! im just another non-dev member LOL sry

Need help!
Look on code:
Code:
//static int32_t msm_tscal_scaler = 65536;
static int32_t msm_tscal_xscale = 70046;
static int32_t msm_tscal_xoffset = -4191987;
static int32_t msm_tscal_yscale = 71735;
static int32_t msm_tscal_yoffset = -3004437;
this pointers for torch 320*240. So when I using that I need tap on 2mm high on icon for open something. This pointer need for screen calibration. Need find right pointer.
I can't find a logic in this numbers.

miroslav_mm said:
Need help!
Look on code:
Code:
//static int32_t msm_tscal_scaler = 65536;
static int32_t msm_tscal_xscale = 70046;
static int32_t msm_tscal_xoffset = -4191987;
static int32_t msm_tscal_yscale = 71735;
static int32_t msm_tscal_yoffset = -3004437;
this pointers for torch 320*240. So when I using that I need tap on 2mm high on icon for open something. This pointer need for screen calibration. Need find right pointer.
I can't find a logic in this numbers.
Click to expand...
Click to collapse
The offsets appear to have a ~4:3 ratio. It doesn't make sense though why it's x:y=4:3, and not the other way around, since the screen is usually 3:4(e.g. 320x480 or 240x320). Also, considering our screen is 320x480, which is a ratio of 1,5 and not 1,(3), I think those would be totally wrong for our phone. Maybe try to modify one of the two to match 1,5 ratio. E.g. make xoffset=-4506655. I don't know. :/

Please discuss development on your page. I want to make a dev team here
Sent from my GT540 using Tapatalk

deleted
10 character

This numbers working with this java code together:
]
Code:
public void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
// Read current values
final int xoffset = (readValue("xoffset")*240+480)/960;
final int yoffset = (readValue("yoffset")*320+460)/920;
final int xscale = readValue("xscale");
final int yscale = readValue("yscale");
view = new View(this) {
private int step = 1;
private int rawx1;
private int rawy1;
private int rawx2;
private int rawy2;
private void writeValue(String parameter, int value) {
try {
FileOutputStream fos = new FileOutputStream(new File("/sys/module/msm_touch/parameters/tscal_" + parameter));
fos.write(String.valueOf(value).getBytes());
fos.flush();
fos.getFD().sync();
fos.close();
} catch (FileNotFoundException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
}
}
@Override
public boolean onTouchEvent(MotionEvent ev) {
if (ev.getAction() != MotionEvent.ACTION_UP)
return true;
if(step==1) {
rawx1 = ((int)ev.getRawX()*65536-xoffset+32768)/xscale;
rawy1 = ((int)ev.getRawY()*65536-yoffset+32768)/yscale;
if(rawx1<100 && rawy1<100)
step = 2;
} else if(step==2) {
rawx2 = ((int)ev.getRawX()*65536-xoffset+32768)/xscale;
rawy2 = ((int)ev.getRawY()*65536-yoffset+32768)/yscale;
if(rawx2>140 && rawy2>220)
step = 3;
} else {
int distx = rawx2 - rawx1;
int new_xscale = (140*65536 + distx/2)/distx;
int disty = rawy2 - rawy1;
int new_yscale = (220*65536 + disty/2)/disty;
int new_xoffset = 1;
new_xoffset += 50*65536-rawx1*new_xscale;
new_xoffset += 190*65536-rawx2*new_xscale;
new_xoffset /= 2;
new_xoffset = (new_xoffset*960+120)/240;
int new_yoffset = 1;
new_yoffset += 50*65536-rawy1*new_yscale;
new_yoffset += 270*65536-rawy2*new_yscale;
new_yoffset /= 2;
new_yoffset = (new_yoffset*920+160)/320;
// Pass new calibration to kernel
writeValue("xoffset", new_xoffset);
writeValue("yoffset", new_yoffset);
writeValue("xscale", new_xscale);
writeValue("yscale", new_yscale);
// Save calibraiton data to /data/system/pointercal
StringBuilder sb = new StringBuilder();
sb.append(new_xscale);
sb.append(" ");
sb.append(0);
sb.append(" ");
sb.append(new_xoffset);
sb.append(" ");
sb.append(0);
sb.append(" ");
sb.append(new_yscale);
sb.append(" ");
sb.append(new_yoffset);
sb.append(" ");
sb.append(65536);
App + ui input framework + kernel working together.
App - creation poinercal
ui input framework - reading pointercal send correct info to kernel
kernel touch driver - doing move.

Öhm, the touchscreen driver use a virtual square (320*320) to detect touch, stretched to a rectangle(320*480)?
mod: about calibration, http://code.google.com/p/openeve/issues/detail?id=91

jasper580 said:
Please discuss development on your page. I want to make a dev team here
Click to expand...
Click to collapse
I cant understand you! What do you mean? Do you wanna make new dev team? Working on my sources but without me?

I exactly want to ask that. Maybe he start from scratch, his github is near empty.

Theme Chooser themes, Windows, aapt, and uppercase letters in XML filenames

EDIT: A solution has been found, courtesy of XDA member titanic_fanatic, who compiled a Windows version of aapt that accepts uppercase letters in XML filenames! (Read his original message here.) I've attached it to this post. One of the biggest hurdles to effective theme creation in Windows is finally solved!
----------------------------------------------------------------------------------------​
Forgive me if this has already been asked, but I know that it's something that's annoyed theme developers under Windows for a while now. I seem to recall seeing hints about fixing this, but I've done search after search, and can't find it to save my life.
Theme Chooser theme developers under Windows are hobbled by the fact that XML filenames can't contain uppercase letters, which prevents us from theming certain applications (example: com_keramidas_TitaniumBackup.xml would be needed for Titanium Backup). I know that this is a problem related to aapt.exe. Has anyone come up with a workable solution for this issue? If not, I'd like to take a crack at it. I might not get anywhere, but I have nothing to lose.
Obviously, in order to do this, I'll need the source for it. I can do the heavy lifting regarding getting the source and working out how to compile it for Windows, but I need to know where the source is, what to get, and other requirements. If anyone can help and point me in the right direction(s), I'd really appreciate it!

MJPollard said:
Forgive me if this has already been asked, but I know that it's something that's annoyed theme developers under Windows for a while now. I seem to recall seeing hints about fixing this, but I've done search after search, and can't find it to save my life.
T-Mobile/CM7 theme developers under Windows are hobbled by the fact that XML filenames can't contain uppercase letters, which prevents us from theming certain applications (example: com_keramidas_TitaniumBackup.xml would be needed for Titanium Backup). I know that this is a problem related to aapt.exe. Has anyone come up with a workable solution for this issue? If not, I'd like to take a crack at it. I might not get anywhere, but I have nothing to lose.
Obviously, in order to do this, I'll need the source for it. I can do the heavy lifting regarding getting the source and working out how to compile it for Windows, but I need to know where the source is, what to get, and other requirements. If anyone can help and point me in the right direction(s), I'd really appreciate it!
Click to expand...
Click to collapse
Hmm .. + 1 with you .. but the only way is ... ask brut.all ....
Sent from my Galaxy Nexus using XDA App

qazsxdcfv said:
Hmm .. + 1 with you .. but the only way is ... ask brut.all ....
Click to expand...
Click to collapse
Thanks, I guess I'll have to after all (not that I mind doing it). Over 250 views and only one person to respond isn't giving me much hope that someone has an answer.

Have you found a solution for windows ?

Any solution?

Randomly came across this thread looking for info on compiling the aapt binary for Windows. I'm still looking for the answer for that but I do know the change that needs to be made and how to do that part, as I have already done this mod for aapt on Linux. You have to have a full build environment set up, and by that I mean set up to build a full ROM, like CM10 for example. The change to make is in the frameworks source code at frameworks/base/tools/aapt/Resource.cpp.
See this commit for the details:
http://review.cyanogenmod.org/#/c/24204/
Add the parts I highlighted in blue to the two lines below:
Code:
String16 baseName(it.getBaseName());
const char16_t* str = baseName.string();
const char16_t* const end = str + baseName.size();
while (str < end) {
if (!((*str >= 'a' && *str <= 'z')
[COLOR="RoyalBlue"][B]|| (*str >= 'A' && *str <= 'Z')[/B][/COLOR]
|| (*str >= '0' && *str <= '9')
|| *str == '_' || *str == '.')) {
fprintf(stderr, "%s: Invalid file name: must contain only [a-z[COLOR="RoyalBlue"][B]A-Z[/B][/COLOR]0-9_.]\n",
it.getPath().string());
hasErrors = true;
}
str++;
}

m4570d0n said:
Randomly came across this thread looking for info on compiling the aapt binary for Windows. I'm still looking for the answer for that but I do know the change that needs to be made and how to do that part, as I have already done this mod for aapt on Linux. You have to have a full build environment set up, and by that I mean set up to build a full ROM, like CM10 for example. The change to make is in the frameworks source code at frameworks/base/tools/aapt/Resource.cpp.
See this commit for the details:
http://review.cyanogenmod.org/#/c/24204/
Add the parts I highlighted in blue to the two lines below:
Code:
String16 baseName(it.getBaseName());
const char16_t* str = baseName.string();
const char16_t* const end = str + baseName.size();
while (str < end) {
if (!((*str >= 'a' && *str <= 'z')
[COLOR="RoyalBlue"][B]|| (*str >= 'A' && *str <= 'Z')[/B][/COLOR]
|| (*str >= '0' && *str <= '9')
|| *str == '_' || *str == '.')) {
fprintf(stderr, "%s: Invalid file name: must contain only [a-z[COLOR="RoyalBlue"][B]A-Z[/B][/COLOR]0-9_.]\n",
it.getPath().string());
hasErrors = true;
}
str++;
}
Click to expand...
Click to collapse
Thank you for your help but could you tell me where can I find frameworks/base/tools/aapt/Resource.cpp?
I don't understand where to find it?

Sorry to resurrect an old thread, but I've finally come across a Windows version of aapt.exe that accepts uppercase letters in XML filenames, courtesy of XDA user titanic_fanatic. I've modified the first post in the thread with the details. Happy theming!

MJPollard said:
Sorry to resurrect an old thread, but I've finally come across a Windows version of aapt.exe that accepts uppercase letters in XML filenames, courtesy of XDA user titanic_fanatic. I've modified the first post in the thread with the details. Happy theming!
Click to expand...
Click to collapse
That link does seem to work anymore (and the webpage gives malware warning in Norton).
Does anyone else have this file?

[Q] Impossible to hook a method in com.android.vending

Some week ago i posted into the request and index thread for set up a request to develop a module that will eliminate the need to have a certain amount of free space when you install app from the Play Store.
After some days i disassembled the play store apk [using APk_OneClick ] and i found the information and method i needed and wrote down into the forum, here my reply:
MonoS94 said:
Ok, i dig a bit into the play store apk for the 4.9.13 version in
Code:
com/google/android/finsky/download/
there is a class named
Code:
Storage
with two interesting method
Code:
dataPartitionAvailableSpace()
and
Code:
externalStorageAvailableSpace()
This two method return a long representing the available space into a particular partition using this formula [found in the partitionAvailable() method]
Code:
for (long l = localStatFs.getAvailableBytes(); ; l = localStatFs.getBlockSize() * localStatFs.getAvailableBlocks())
So hooking and replacing this two modules with a dummy version returning always full space we will defeat this annoying oddity, also being part of the play store apk i think that this will not affect any other app.
I hope that with this information someone will make a module.
Click to expand...
Click to collapse
I waited some days and about a week ago i download the ADT, set up the environment, learned a bit how to code for Android and Xposed and wrote down some code
Code:
package com.MonoS.FreeSpaceForPlayStore;
import static de.robv.android.xposed.XposedHelpers.findAndHookMethod;
import de.robv.android.xposed.IXposedHookLoadPackage;
import de.robv.android.xposed.XC_MethodReplacement;
import de.robv.android.xposed.XposedBridge;
import de.robv.android.xposed.callbacks.XC_LoadPackage.LoadPackageParam;
public class Main implements IXposedHookLoadPackage {
private static final String GOOGLE_PLAYSTORE = "com.android.vending";
public void handleLoadPackage(final LoadPackageParam lpparam) throws Throwable {
if (!lpparam.packageName.contains(GOOGLE_PLAYSTORE))
return;
XposedBridge.log("I'm in bro ;)" + lpparam.packageName);
findAndHookMethod("com.google.android.finsky.download.Storage", lpparam.classLoader, "dataPartitionAvailableSpace", XC_MethodReplacement.returnConstant(1073741824));
}
}
Compile the code, install on the phone, activate and reboot and the module don't work, it enter into the package but then throw me an ClassNotFoundError exception.
Then checked the name of all the method but they are right.
Then why can't my module hook the method??
Thanks for the attention

[DexKit] Xposed Module Jni Lib | An easy-to-use, high-performance dex deobfuscation library

Library: DexKit
Project Link: https://github.com/LuckyPray/DexKit
About: An easy-to-use, high-performance dex deobfuscation library. Easy to use your CMAKE/Android projects.
Highlight:
JNI multi-threaded processing Dex bytecode, more efficient than JVM implementation.
It single search is ms level, You can even inject the host application at runtime without causing an ANR, to the extent permitted.
You can use it to handle reinforced apps(use ClassLoader cookies).
API introduction​There are two APIs can meet most of your usage scenarios:
DexKit::BatchFindClassesUsingStrings
DexKit::BatchFindMethodsUsingStrings
Note: In all cases you should avoid searching for keywords that contain duplicate content, eg: {"key_word", "word"}, as this will cause tags to be overwritten, resulting in inaccurate search results. If there is such a need, open the advanced search mode as much as possible, and use the string to match the content exactly, for example, modify it to this: {"^key_word$", "^word$"}
Click to expand...
Click to collapse
And there are many other APIs:
DexKit::FindMethodCaller: find caller for specified method.
DexKit::FindMethodInvoking: find the called method
DexKit::FindMethodUsingField: Find method to get/set specified field
DexKit::FindMethodUsingString: find method used utf8 string
DexKit::FindMethod: find method by multiple conditions
DexKit::FindSubClasses: find all direct subclasses of the specified class
DexKit::FindMethodUsingOpPrefixSeq: find all method using opcode prefix sequence(op range: 0x00-0xFF)
DexKit::FindMethodUsingOpCodeSeq: find all method using opcode sequence(op range: 0x00-0xFF)
DexKit::GetMethodOpCodeSeq: get method opcode sequence(op range: 0x00-0xFF)
Note: At present, all instructions are only for standard dex instructions and do not include odex optimization instructions.
Click to expand...
Click to collapse
For more detailed instructions, please refer to dex_kit.h.
Quick start​However, this approach will import an extra so file. If you don't want to import an extra so file, please use the second/third method.
build.gradle:
Code:
allprojects {
repositories {
...
maven { url 'https://jitpack.io' }
}
}
app/build.gradle:
Code:
dependencies {
implementation 'com.github.LuckyPray:DexKit:<version>'
}
​JAVA Example​DexKitBridge provides 2 factory methods to create Dexkit instances:
DexKitBridge.create(apkPath): normally, please use it.
DexKitBridge.create(classLoader, true): for reinforced apps, used classLoader and set option useCookieDexFile to true.
PS: DexKitBridge.create(classLoader, useCookieDexFile = false) ≈ DexKitBridge.create(apkPath), but the former may contain part of the system dex.
Java:
import io.luckypry.dexkit.DexKitBridge;
// ...
public class DexUtil {
static {
System.loadLibrary("dexkit");
}
public static void findMethod() {
// for no-reinforced apps please use apkpath to load, because of the exist of dex2oat and CompactDex(cdex),
// dexkit currently only handles StandardDex.
String apkPath = application.applicationInfo.sourceDir
// try-with-resources, auto close DexKitBridge, no need to call DexKitBridge.close()
// if you don't use try-with-resources, be sure to manually call DexKitBridge.close() to release the jni memory
try (DexKitBridge dexKitBridge = DexKitBridge.create(apkPath)) {
if (dexKitBridge == null) {
Log.e("DexUtil", "DexKitBridge create failed");
return;
}
List<DexClassDescriptor> classes = dexKitBridge.findSubClasses("android.app.Activity", null);
for (DexClassDescriptor clazz : classes) {
String name = clazz.getName();
String simpleName = clazz.getSimpleName();
Class<?> clz = clazz.getClassInstance(hostClassLoader);
Log.i("DexUtil", "findSubClasses: " + clz);
}
} catch (Throwable e) {
Log.e("DexUtil", Log.getStackTraceString(e));
}
}
}
If you have problems using it, please create an issue on the github repo.

Reserved