2011 02 :
Deep Blue Lambda
Deep Blue Lambda
Sunday, 27 February 2011
Clojure for Android source published
Over the next few weeks, I will be publishing the source code for the Clojure REPL for Android in a few different instalments:
- Clojure for Android. a modified version of Clojure adapted to run on the Dalvik virtual machine;
- Clojure Android Toolkit. a library of utilities for Clojure developers working on Android; and
- Clojure REPL. the source code of the application itself.
I have now published the modifications to my source code in a repository available on GitHub. My work is based on the 1.2.x branch of the Clojure source code and is available in the android-1.2.x branch .
This post will document my goals for Clojure on Android, give an overview of the changes I have made, describe the current implementation of dynamic compilation, and present areas for future work.
The three primary goals of the Clojure for Android release are as follows:
- Create a version of Clojure that works for both the Java and Dalvik virtual machines in the hope that the changes can eventually be included in Clojure itself,
- Create a development version of Clojure that supports dynamic compilation to enable more rapid development of applications, and
- Create a lean Clojure runtime that will deliver acceptable performance on Android devices.
Overview of changes
There really are not many changes in this initial release of Clojure for Android. They fall into three categories: the addition of a Dalvik-specific dynamic class loader, some minor runtime changes, and an update to the build configuration to support Android.
New DynamicClassLoader hierarchy
The most significant change is to DynamicClassLoader. In the original implementation, this class manages a constant pool, maintains a cache of class definitions, provides (deprecated) class path alteration capability, and is in charge of turning compiled class byte codes into classes available within the virtual machine.
In my implementation, it retains all of those abilities save for the last one. It now is an abstract class that delegates class realisation to its subclasses, of which there are two:
- JvmDynamicClassLoader relies on Java’s standard ClassLoader.defineClass method.
- DalvikDynamicClassLoader uses a tortuous method, described later.
There are a few relatively minor runtime changes:
- The addition of a new var, clojure.core/*vm-type*. which will be set to either :dalvik-vm or :java vm at runtime.
- Choosing the correct DynamicClassLoader implementation depending on *vm-type*
- A workaround for a bug fixed in ‘FroYo’ where the context class loader is set to a system class loader instead of the application’s class loader
- The pre-emptive loading of clojure.set. clojure.xml. and clojure.zip is disabled on Dalvik.
Build system update
The build system has received somewhat more extensive changes. There are two basic scenarios:
- Building Clojure without Android support: This should work just fine. Just ant run as usual.
- Building Clojure with Android support: You will need to create a local.properties file with pointers to the Android SDK directory and SDK version you want to use. More documentation is available in readme.txt .
When building with Android support, the build file will create a stripped down version of the dx.jar file from the Android SDK. By default, this will do a simplistic removal of purely test classes. However, if you have ProGuard, it can do a more exhaustive shrinking. This is enabled by setting the proguard.jar property.
When Android is enabled, the build will create two additional JAR files:
- clojure-nosrc.jar. the opposite of clojure-slim.jar. a compiled-only version of Clojure. This JAR also contains the dx tool classes that are needed at runtime.
- clojure-dex.jar. a version of clojure-nosrc.jar where all of the classes have been compiled into a Dalvik executable. This file is suitable for loading by one of На s class loading mechanisms.
To illustrate how I implemented dynamic compilation in Clojure, I will first present the traditional path from compiled Java class to instantiated Dalvik class. Next I will show how the modified version of Clojure takes dynamically generated classes through the same process. Finally I will present the trade-offs involved in the current implementation and what you should keep in mind when using the dynamic compilation.
Traditional work flow
The following is a brief description of the path of a compiled class file from to build execution:
- At build time:
- Java files are compiled into Java classes made up of JVM byte codes.
- All of the classes are prepared into a Dalvik Executable (DEX file) by the dx tool. This file is called classes.dex .
- The DEX file is placed into the Android package.
- At install time:
- The installer reads the DEX file from the package.
- The DEX file is verified to remove illegal instructions and performs some computations to aid in garbage collection.
- The verified DEX data is then optimized, creating a hardware — and platform-specific version of the code. Some optimizations include replacing virtual method call resolution with indices in a vtable, inlining method calls, pruning empty methods, etc.
- The resulting optimised DEX file (ODEX file), is written to a special cache directory.
- At run time:
- The ODEX file is checked to make sure it is still valid. If not, then the original DEX file is again verified and optimised.
- The application loads its classes from the ODEX file.
Dynamic Clojure work flow
- The Clojure evaluator compiles a form into a class using the embedded ASM bytecode engineering library.
- The DalvikDynamicClassLoader processes the compiled byte code as follows:
- It uses the embedded dx tool to translate the JVM class into an in-memory DEX file.
- It writes the DEX file into a temporary JAR file in *compile-path* .
- It uses На s dalvik.system.DexFile to load the JAR file. In doing so, Android will create an ODEX file in *compile-path* .
- Loads the class from the DexFile object and returns it.
The main disadvantage to this form of dynamic compilation is that it is slow. It requires using the disk, as well as the performing all sorts of computations at runtime. Anyone who has used the Clojure REPL for Android can attest to its sluggishness.
Unfortunately, to the best of my knowledge, there are no other accessible APIs available for doing this better. Most of the work is done in native code, making it difficult to bundle it into Clojure. There is some hope that this may change in the future. From the Dalvik documentation :
Some languages and frameworks rely on the ability to generate bytecode and execute it. The rather heavy dexopt verification and optimization model doesn’t work well with that.
We intend to support this in a future release, but the exact method is to be determined. We may allow individual classes to be added or whole DEX files; may allow Java bytecode or Dalvik bytecode in instructions; may perform the usual set of optimizations, or use a separate interpreter that performs on-first-use optimizations directly on the bytecode
Until such an API is released, it is necessary to either take the slow simple but route, or to create a Clojure compiler for the Dalvik VM from scratch.
I think that dynamic compilation is of interest primarily to developers. Most applications will have no need for dynamic compilation as they can be AOT-compiled. As such, the slowness may well be acceptable. After all, waiting seconds for a function to recompile running in your application is much more tolerable than needing to go through a full compile-deploy cycle that may be measured in minutes.
There are two things to be aware of when using dynamic compilation:
- You will need to be sure to point *compile-path* to some place where your application has write access.
- Some forms may blow the stack during compilation, such as (for [x (range 5) y (range 5)] [x y]). This is a limitation of the runtime.
There is still much to be done. As the source now is released, I look forward to seeing what sorts of feedback and improvements will come from others. Given the three goals I stated above, I think much of the work is as follows:
Integration with upstream
Get feedback on how to best integrate these changes into the language from other Clojure developers in general and, I hope, the Clojure/core team itself. Of course, it will most likely take some time before these changes make it into a Clojure release.
Clojure for Android development
While not perfect, I think the current solution largely satisfies this goal. Writing a new compilation back-end may make things better, but I am not sure that it will provide as good as a return working on the third goal.
I think that any new development should follow the master branch of Clojure going forward. The patches to the code itself should be simple enough to manage. Porting the build system changes to Maven will be more cumbersome.
Lean Clojure runtime
This is the place where the most needs work to be done. There have already been some good ideas presented on how to improve this, such as:
- Eliminating metadata from compiled code to reduce memory footprint,
- Finding ways to cut down on the immense amount of object churn during bootstrap, and
- Generally finding ways to cut down on the amount of work done during the bootstrapping process.
My current idea is to find ways to modularise clojure/core.clj somewhat to be able to either completely eliminate some functionality (such as those for dynamic compilation) or at least delay loading it. Not every program makes use of every feature of the language. Some programs may never use one or more of: agents, futures, primitive vectors and arrays, etc. If there were some way to make some of these things load-on-demand, if only in an Android environment, that could significantly improve bootstrap times.
I look forward to the feedback from others and welcome any help in trying to get these things working. I think that it is quite possible to make Clojure a first-class development language for the Android platform.
Wednesday, 23 February 2011
Presenting Clojure REPL at February CHUG meeting
We’ll be this meeting Thursday at the studio (1024 Studewood), at 7pm. Daniel will be talking about his Clojure REPL for Android. Very cool!
I’ll order Pink’s pizza again, if people liked that last time. Please RSVP for pizza so I know how much to order. And bring $5 to donate to the cause if you’re having any. )
The usual instructions:
1024 Studewood is a block-ish south of 11th on Studewood, aka Studemont, aka Montrose. If you’re coming on north Studewood, it’s on your right just past Stella Sola (which is on the left). It’s a small green house, the only one that sticks out almost all the way to the road. It will make you think it’s a photography studio, because it is.
DO NOT PARK IN THE LOT NEXT DOOR. YOU MAY GET TOWED.
There are driveways immediately on either side of the house, sufficient for two cars each (unless you drive something big). Otherwise, park on a side street (10 1/2 St.), they’re easily accessible across Studewood, or curbside on Studewood. Just don’t park in the lot(s) right next to the studio, they are used as valet space for Stella Sola.
My plan is to give an informal presentation, primarily focusing on the changes I made to get Clojure with dynamic compilation working on the Dalvik VM. I will also demonstrate the VimClojure integration with a REPL running on a device. I plan to share my thoughts on the current state and future of expectations for Android development with Clojure.
It will be a fun evening filled with Clojure, Android, and pizza. I hope to see you there.
Monday, 21 February 2011
Creating Android applications with Clojure: Slimming things down with ProGuard
Last time, I described how to integrate Clojure into the standard Android build process. In this post, I will build on this by integrating ProGuard into the build process in an effort to make applications smaller and faster.
First, I will give a brief overview of ProGuard and describe how to enable the ProGuard capability of the Android build process. Next I will show how to configure ProGuard so that it can work with your Clojure code. Finally I will summarise my recommendations and let you know what to expect next in this series.
The sample application I will use is available on my Clojure-Android-Examples repository on GitHub under the slimmed directory.
As described on its web site, ‘ProGuard is a free Java class file shrinker, optimizer, obfuscator, and preverifier.’What does this mean?
Minimisation ProGuard can rid your application of unused classes, methods, and fields, leading to smaller application sizes and load times. Optimisation ProGuard can analyse your code and perform a large number of optimisations, such as removing unnecessary field accesses and method calls, merging identical code blocks, inlining method invocations, etc. This should make your code smaller and run faster (depending on VM implementation). Obfuscation ProGuard can remove debug information normally included class files and rename classes, methods, etc. to make them meaningless. The idea is to make it harder to reverse engineer a compiled application into source code. Preverification This is something specific to Java 6 or Java ME VMs. As a result, it does not apply to working with Android.
Now that you have an idea of what ProGuard can do, how can you start using it?
It is very simple to enable ProGuard when you are creating a release build of your application using ant release. All you have to do is edit the build.properties file created by Android and add the following line:
If you want to enable ProGuard for debug builds, you will also need to override the -debug-obfuscation-check target. I set up my target as follows:
Line two above ensures that the libraries you are using, such as Clojure, will be processed in addition to your application’s code.
These two changes will give you the ability to do the following:
- When running ant install or ant debug. your program will be built as usual.
- If you want to create a ProGuard-processed debug build, simply use ant -Dproguard.enabled=true install or ant -Dproguard.enabled=true debug .
If you want to run ProGuard for every build, you can simply add proguard.enabled = true to your build.file properties.
However, if you try to run ProGuard now, you will find your build fails. First, it has to be configured.
The ProGuard configuration bundled with the latest Android SDK is configured with some sane defaults. However, if you are doing Clojure development, you will need to make some adjustments in order for ProGuard to work correctly. The following sections should help guide you to use ProGuard with your application.
All of these changes will be made to the proguard.cfg file in the root directory of your application.
Shrinking your application with ProGuard
ProGuard’s default configuration will not even allow the build process to complete. To get proper shrinking, you must silence some warnings and ensure that all of your application’s code is not shrunk away. If you are following along, you may want to add the following two lines to your proguard.cfg before continuing any further:
Naturally, these disable ProGuard’s optimization and obfuscation functionality. By disabling them here, you can concentrate on getting one thing working at a time.
Silencing Clojure-related warnings and notes
Clojure refers to a number of classes that are not available on Android, particularly the Swing components from the used clojure.inspector and clojure.java.browse namespaces and some java.beans classes used by clojure.core/bean. ProGuard will complain about these missing referenced classes, but these warnings can be disabled by adding the following lines to proguard.cfg :
With these two lines in place, the sample application shrinks install from an size of 4.04 MB to only 460KB. Also, instead of taking 4.7 seconds to load, it crashes almost instantaneously. Why? It’s because most of Clojure has been stripped out.
The reason why Clojure is stripped out is because ProGuard does not understand how Clojure initialises a namespace. Briefly, when loading a namespace, Clojure tries to do one of two things:
- Read and evaluate the source file, if available.
- Check to see if the namespace was AOT-compiled by trying to load the namespace’s initialisation class and its corresponding load method.
For example, when trying to load the namespace org.deepbluelambda.example. Clojure will look for the file org/deepbluelambda/example.clj or the class org.deepbluelambda.example__init .
Therefore it is necessary to configure ProGuard to keep Clojure’s core initialization classes. This can be done by adding the following lines to proguard.cfg :
You must include these eight lines, otherwise your application will fail to run. If you use any other Clojure namespaces, you will need to add a corresponding line. For example, if you use clojure.set and clojure.contrib.string. you will need to add the following lines to your proguard.cfg :
Taking these steps will help ensure that Clojure itself loads up, but how about your application?
Keeping your Application
In addition to making sure all of Clojure’s namespaces load properly, you will need to ensure that all of your application’s namespaces load as well. The approach is the same; just not be sure to forget any gen-class ed namespaces.
Furthermore, if you expose superclass methods with gen-class. as below, you will need to ensure that ProGuard does not delete those methods.
To ensure that the above activity works fine, the following lines need to be added to the ProGuard configuration:
The first three lines should be familiar to you, the second three lines will preserve any public method, with any return and parameter types, so long as its name starts with ‘super’. Thus, it should be able to handle any number of :exposes-methods so long as the exposed method name starts with ‘super’.
Measuring the results
Now that you have successfully shrunk your application, what can you expect? The following table summarises the results from my sample application:
Короткий опис статті: classes dex Deep Blue Lambda Daniel is Solano Gomez’s web site about technical topics such as Clojure and Java development, Gentoo Linux, and his open source project sh-clojure. This page includes all entries from February 2011.
2011 02 :
Deep Blue Lambda