Can someone enlighten me on how to include an external java framework to android framework and compile it as a part of the custom android framework?
The idea is to make this external framework a part of my custom android framework and expose it as android service for other processes to use.
In essence what I did is -
checked for android port of the java framework
Copied the src dir (including all the external jar files) from the ported framework to framework/base/core/java/android/
Edited ActivityManager in the framework to call the copied library
run make
Is this the right approach to edit Android framework? Since, I am including the java files and associated external jars of the external project inside the Android framework I suspect make fails while converting the jars to dex.
Can somebody point me towards any available resources dealing with Android Framework internals and on how to edit Android framework?
I checked the Tutorial: Android Internals - Building a Custom ROM, Pt. 2 of 2 on youtube from this year's Marko Gargenta presentation in Beijing, China. But he didn't make it (i.e. adding external framework or jars) clear either.
First: I'm not very experienced in the Android build process, so there might be a better solution, but this one should work (at least for Android 2.2).
I assume, that you use Eclipse for the source code modification and already executed a successful make -j4...
Create a new source folder: external/<your-library>/src
Copy all Java source files into this folder.
Open the following make file: <android>/frameworks/base/Android.mk
Search for # Build ext.jar and add your library as follows:
[...]
ext_dirs := \
../../external/apache-http/src \
../../external/<your-library>/src \
../../external/gdata/src \
../../external/protobuf/src \
../../external/tagsoup/src
[...]
Now you should be able to access these libraries from the Android framework...
Finally execute make -j4 ext to build only the external module,
Or make -j4 to build all changed files (if a previous build exists in /out/).
README.txt from git repo in https://android.googlesource.com/platform/vendor/sample is a good example howto do it. It's good to clone it to see the examples.
Basically you need to create jar file and add it in filesystem, add xml configuration into /system/etc/permissions/ and declare <uses-library> in apk's AndroidManifest.xml.
In my case I had problem with adding <uses-library> on a wrong place (not in <application>).
So it's good to check the result with:
$ aapt dump badging Foo.apk | grep uses-library
uses-library:'android.foo'
Related
I am developing an Ionic Capacitor plugin which imports 2 iOs .framework files and a .bundle that refers to one of this .framework files. The thing is that no matter how I link/embed and point/copy this files on the plugin project, xcode claims, on the app project, that it cant find the module in the swift file.
I already tried to add the files to the project, used the "Embedded binaries" option, linked libraries, allow non-modular includes (on build options menu), add the files to the headers (on build phases), and so on....
The line that xcode point the error is:
import OneFramework
And xcode claims:
No such module 'OneFramework'
I was expecting that when I add the plugin to my app project via npm, and later running a "$ rm -rf ios && ionic capacitor run ios" to run the app, xcode find all the modules of the plugin that I am trying to do.
I found the solution. To achieve this the first thing to know is that when you do npx #capacitor/cli plugin:generate what the CLI do for you is the generation of a cocoa pod. The root of this pod is the generated folder itself.
With that in mind, the next thing to do is to learn how to make pods, but i'll sumarize the principal aspects that led me to the success.
-First of all you open the *.xcworkspace. Followed by that, click on the "Add Files to Pod..." option and add your files. Please ensure that the "Copy files if needed" option is marked. Please refer to the picture below.
-Now its nice to create a folder for your .framework and another for the .bundle (if there are any) files. Do this by right clicking the Pods project and select the option "New group". Select a name like that is different from the pattern of xcode, it is nice to know that this folders are created by you.
-If you done this right, the frameworks you recently added to the project will appear on the pods project like this:
-Now, for your swift implementation find your files, drag your .frameworks that are on the pods project for the "Frameworks, Libraries and Embedded content" of the plugin project. The result will be something like this:
-Ok, files included and linked. Now we should let our cocoa pod know about this and declare this files. The file "YourAwesomePlugin.podspec" (located at the root of the plugin project) is the main entrance of the pod. In this file you will declare which files (.frameworks, .bundle, etc) belong to your pod and consequently will belong to your plugin when you npm install it. To declare this you'll need three directives:
s.vendored_frameworks = 'ios/Pods/YourFrameworkFolder/**'
s.resource = 'ios/Pods/YourResourceFolder/YourBundle.bundle'
s.xcconfig = {'ENABLE_BITCODE' => 'NO'} #This is mandatory on my case, but you need to evaluate if this options applies to your plugin.
-Now we hit play on the plugin project. To test on your app if the plugin is ok, you need to add the path of the root of the plugin project on the podfile of the pods project of the APP project. Like this:
-To install it you can go on Yourproject/ios/App and run pod install.
Please note that:
To declare the existence of your recently created plugin you you need to do some declarations as well, but this part is easy and already documented on capacitor/plugin docs.
The installation method via pod install that I suggested is for testing. It would be nice if you pack your plugin using npm and npm install it like all other plugins.
I dont have much knowledge on cocoapods like I wish, but this works and I think that is a clean solution. If not, please let me know.
If this answer is useful for you, please thumbs it up, it is a week of research and trying that I am sharing, along the time to write it all down.
I just installed all the required tools for the new Sencha Cmd(Ruby, Compass, Ext Js 4.2 SDK) everything works fine, but I don't know how to obtain the unique obfuscated file from the .js sources in my Ext Js application.
Any help, please? Thanks in advance.
I tried the "sencha app build" command from all the folders in my application path, always get the "Please ensure this command was executed from a valid application directory. Unable to locate 'app.dir' config property from 'sencha.cfg'
After more reading about the subject I think before the "sencha app build" command to work I need to generate the bootstrap.js file manually because my application wasn't generated using the framework, but I'm stuck here also. Anyone ever did that? Please?
The best way to get things work if your have legacy project (when project structure is not generated by sencha cmd tool) is to generate empty project and the move all necessary files/folders in to your existing project and try to adjust all configurations. You will need to move:
.sencha folder
folder with extjs development classes
bootstrap.js, bootsrap.css, build.xml
modify you start page correspondingly (index.html, index.jsp, or whatever)
create application.js to start your application
bootstrap files are always generated automatically. To generate it normally you have to use sencha app refresh command.
Basically the most important changes have to be made in .sencha folder. Most of the config files in this folder have a lot of comments so this is not a problem to understand what to change. Also it's good idea to refer official doc page: Sencha CMD
I have gone through this document :Getting Started with the Facebook SDK for iOS v3.1
I have successfully configured my Xcode project by following it.
But I want to do it without adding the Facebook.framework.
As previously we used to only add the "src" folder which we used to clone from :Facebook SDK for iOS - Github. Why can't I just copy the "src" folder and configure my project.
I want to play around with FBFriendPickerViewController.m (dot m) file. Which I am not able to view if I follow the standard tutorial. But is available in the "src" folder.
Please how can I do that ? And it also seems that FbGraph.h file doesn't exist.
Please help.
If you want to play with the source, go to:
https://github.com/facebook/facebook-ios-sdk/
After getting the code, run this script to build the framework locally:
<Install Location>/scripts/build_framework.sh
This will drop the framework in the folder:
<Install Location>/build
Follow the same instructions in the Getting Started to include the Facebook SDK, however, you will get the framework from your local build folder.
Note: You'll likely have to modify the build search path in Xcode if you're swapping out, say one of the sample projects (that was looking for the Framework in a particular place).
I am building an iPhone application which connects to an LDAP server (no encryption). I am trying to use iOSPorts but I am having trouble following the documentation, I think it was written for Xcode 3, while I've only ever used Xcode 4.
All that is required is to do simple search queries of the server and returning the results.
Some of the trouble I am having:
For example in step 3 my iOSPorts/include folder is empty and only contains the .gitignore file.
For step 4 which files do I need to add just the ports folder containing database, devel, iOSPorts and security?
For step 6 the direct dependencies panel, is this the target dependencies of build phases , which libraries need to be selected?
I think step 7 is for Xcode 3 too,
Any help with this would be much appreciated!
I am building an iPhone application which connects to an LDAP server (no encryption). I am trying to use iOSPorts but I am having trouble following the documentation, I think it was written for Xcode 3, while I've only ever used Xcode 4.
The original instructions were created for Xcode 3 since Xcode 4 was still in early beta when I published the first release of iOS Ports. The instructions have been updated within the last month for use with Xcode 4. I try to use the name used by "Xcode 4 User Guide" when describing the steps, so I recommend reading the first few chapters of this document from Apple.
For example in step 3 my iOSPorts/include folder is empty and only contains the .gitignore file.
Initially the include directory will be empty. The individual ports will copy the header files to the include directory. This is to prevent the end developer from needing to add a path to OpenSSL, Cyrus SASL, and OpenLDAP in order to compile an app for OpenLDAP.
Any help with this would be much appreciated!
So here is a quick run down of how iOS Ports works.
Each port of an Open Source package has an Xcode project file. That Xcode project file lists other Xcode project files and Makefiles as dependencies. Here is the high level steps used by an iOS Ports Xcode project file to compile the port:
Execute the port's MakeFile
Makefile: Download the source archive from the Internet (fails if Internet is unavailable)
Makefile: Verifies the integrity of the source archive using md5sum.
Makefile: Unpacks the source archive.
Makefile: Applies any required patches to the source code.
Makefile: Copies headers to the iOS Ports include directory.
Build any libraries from other ports listed in the build settings' Target Dependencies (for instance, libldap.a in OpenLDAP requires libsasl2.a from Cyrus SASL).
Compiles requested library.
The README has more information on how to add the project files to your project and link against the libraries. There also a few example programs in the examples directory.
If you do continue to have problems, please let me know. I try my best to keep on top of requests for help regarding iOS Ports and my other projects.
I haven't gone through all the process, but I am pretty sure for step 3, you have to 'make' the project first. There is a Makefile in the top folder of the project.
The motivation for this question is me trying to get LDAP functions to work with an iPhone application which is a project I'm attempting for part of my dissertation.
When I was developing the application I used the ldap.framework framework that is part of Mac OS X. This works fine in the simulator, but when I try to now get the app on a device it tells me that I'm not allowed to use this framework.
After some research I found that I could build openldap using the arm architecture and add the static library to my application destined for my device.
I eventually managed to compile configure and build openldap by setting variables as mentioned here and using the following commands...
Ade$ ./configure CC=$DEVROOT/usr/bin/arm-apple-darwin9-gcc-4.0.1 \
LD=$DEVROOT/usr/bin/ld --host=arm-apple-darwin --with-yielding_select=yes
Ade$ make depend
Ade$ make
I was told that the file I'm looking for will have an extension of '.a' so I searched for a '.a' file that mentions ldap...
Ade$ sudo find / -name *ldap*.a
Password:
/Users/Ade/Desktop/openldap-2.4.16/libraries/libldap/.libs/libldap.a
/Users/Ade/Desktop/openldap-2.4.16/libraries/libldap_r/.libs/libldap_r.a
So I assume these are the files I need?
My question is what do I do next? I know I need to add the library to the Xcode project and probably add a load of '.h' files too?
If anyone can give me a pointer to documentation or shed any light on the next stage I would be really grateful.
Many thanks,
Ade
ps. I have also talked about this process on my blog at www.greenpasta.com.
I've done this same thing to build an LDAP client for the iPhoneOS 2.2. You just to drag the .a into the "link with libraries" build stage. I recommend using the regular (non _r) version of the library, unless you specifically need reentrancy on your ldap stuff (which I don't recommend). You can also add the .h's to your project, which is generally the easiest way to access them.
Simply drag the .a files into the Xcode project and choose "copy files into project". I'm not familiar with OpenLDAP but I think the _r version is just a threadsafe version. I would recommend using that and not copying the other. You should probably not copy both files into Xcode or you will get link errors.
Then do the same for the .h files that define the client APIs of OpenLDAP - again I'm not sure which these are but I'm sure you can find out easily.
I would advise organising the .a and .h files together in a Xcode group under resources.
Include the header files in your source and you should be good to go.
You may also need to add -lldap to your linker command (in the build settings pane).