The ProGuard tool shrinks, optimizes, and obfuscates your code by removing unused code and
renaming classes, fields, and methods with semantically obscure names. The result is a smaller
sized .apk
file that is more difficult to reverse engineer. Because ProGuard makes your
application harder to reverse engineer, it is important that you use it
when your application utilizes features that are sensitive to security like when you are
Licensing Your Applications.
ProGuard is integrated into the Android build system, so you do not have to invoke it manually. ProGuard runs only when you build your application in release mode, so you do not have to deal with obfuscated code when you build your application in debug mode. Having ProGuard run is completely optional, but highly recommended.
This document describes how to enable and configure ProGuard as well as use the
retrace
tool to decode obfuscated stack traces.
When you create an Android project, a proguard.cfg
file is automatically
generated in the root directory of the project. This file defines how ProGuard optimizes and
obfuscates your code, so it is very important that you understand how to customize it for your
needs. The default configuration file only covers general cases, so you most likely have to edit
it for your own needs. See the following section about Configuring ProGuard for information on
customizing the ProGuard configuration file.
To enable ProGuard so that it runs as part of an Ant or Eclipse build, set the
proguard.config
property in the <project_root>/default.properties
file. The path can be an absolute path or a path relative to the project's root.
If you left the proguard.cfg
file in its default location (the project's root directory),
you can specify its location like this:
proguard.config=proguard.cfg
You can also move the the file to anywhere you want, and specify the absolute path to it:
proguard.config=/path/to/proguard.cfg
When you build your application in release mode, either by running ant release
or
by using the Export Wizard in Eclipse, the build system automatically checks to see if
the proguard.config
property is set. If it is, ProGuard automatically processes
the application's bytecode before packaging everything into an .apk
file. Building in debug mode
does not invoke ProGuard, because it makes debugging more cumbersome.
ProGuard outputs the following files after it runs:
dump.txt
.apk
filemapping.txt
seeds.txt
usage.txt
.apk
These files are located in the following directories:
<project_root>/bin/proguard
if you are using Ant.<project_root>/proguard
if you are using Eclipse.Caution: Every time you run a build in release mode, these files are overwritten with the latest files generated by ProGuard. Save a copy of them each time you release your application in order to de-obfuscate bug reports from your release builds. For more information on why saving these files is important, see Debugging considerations for published applications.
For some situations, the default configurations in the proguard.cfg
file will
suffice. However, many situations are hard for ProGuard to analyze correctly and it might remove code
that it thinks is not used, but your application actually needs. Some examples include:
AndroidManifest.xml
fileThe default proguard.cfg
file tries to cover general cases, but you might
encounter exceptions such as ClassNotFoundException
, which happens when ProGuard
strips away an entire class that your application calls.
You can fix errors when ProGuard strips away your code by adding a -keep
line in
the proguard.cfg
file. For example:
-keep public class <MyClass>
There are many options and considerations when using the -keep
option, so it is
highly recommended that you read the ProGuard
Manual for more information about customizing your configuration file. The Overview of Keep options and
Examples section
are particularly helpful. The Troubleshooting section of the
ProGuard Manual outlines other common problems you might encounter when your code gets stripped
away.
When your obfuscated code outputs a stack trace, the method names are obfuscated, which makes
debugging hard, if not impossible. Fortunately, whenever ProGuard runs, it outputs a
<project_root>/bin/proguard/mapping.txt
file, which shows you the original
class, method, and field names mapped to their obfuscated names.
The retrace.bat
script on Windows or the retrace.sh
script on Linux
or Mac OS X can convert an obfuscated stack trace to a readable one. It is located in the
<sdk_root>/tools/proguard/
directory. The syntax for executing the
retrace
tool is:
retrace.bat|retrace.sh [-verbose] mapping.txt [<stacktrace_file>]
For example:
retrace.bat -verbose mapping.txt obfuscated_trace.txt
If you do not specify a value for <stacktrace_file>, the retrace
tool reads
from standard input.
Save the mapping.txt
file for every release that you publish to your users.
By retaining a copy of the mapping.txt
file for each release build,
you ensure that you can debug a problem if a user encounters a bug and submits an obfuscated stack trace.
A project's mapping.txt
file is overwritten every time you do a release build, so you must be
careful about saving the versions that you need.
For example, say you publish an application and continue developing new features of
the application for a new version. You then do a release build using ProGuard soon after. The
build overwrites the previous mapping.txt
file. A user submits a bug report
containing a stack trace from the application that is currently published. You no longer have a way
of debugging the user's stack trace, because the mapping.txt
file associated with the version
on the user's device is gone. There are other situations where your mapping.txt
file can be overwritten, so
ensure that you save a copy for every release that you anticipate you have to debug.
How you save the mapping.txt
file is your decision. For example, you can rename them to
include a version or build number, or you can version control them along with your source
code.