<!doctype html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=iso-8859-1">
<meta http-equiv="content-style-type" content="text/css">
<link rel="stylesheet" type="text/css" href="style.css">
<title>Optimizations</title>
<script type="text/javascript" language="JavaScript">
<!--
if (window.self==window.top)
window.top.location.replace("../index.html#"+window.location.pathname+window.location.hash);
else {
var hash="#"+window.location.pathname.replace(window.top.location.pathname.replace("index.html", ""), "");
if (window.top.location.hash!=hash)
window.top.location.hash=hash;
}
//-->
</script>
</head>
<body>
<h2>Optimizations</h2>
The optimization step of ProGuard can be switched off with the
<a href="usage.html#dontoptimize"><code>-dontoptimize</code></a> option. For
more fine-grained control over individual optimizations, experts can use the
<a href="usage.html#optimizations"><code>-optimizations</code></a> option,
with a filter based on the optimization names listed below. The filter works
like any <a href="usage.html#filters">filter</a> in ProGuard.
<p>
The following wildcards are supported:
<table cellspacing="10">
<tr><td valign="top"><code><b>?</b></code></td>
<td>matches any single character in an optimization name.</td></tr>
<tr><td valign="top"><code><b>*</b></code></td>
<td>matches any part of an optimization name.</td></tr>
</table>
An optimization that is preceded by an exclamation mark '<b>!</b>' is
<i>excluded</i> from further attempts to match with <i>subsequent</i>
optimization names in the filter. Make sure to specify filters correctly,
since they are not checked for potential typos.
<p>
For example,
"<code>code/simplification/variable,code/simplification/arithmetic</code>"
only performs the two specified peephole optimizations.
<p>
For example, "<code>!method/propagation/*</code>" performs all optimizations,
except the ones that propagate values between methods.
<p>
For example,
"<code>!code/simplification/advanced,code/simplification/*</code>" only
performs all peephole optimizations.
<p>
Some optimizations necessarily imply other optimizations. These are then
indicated. Note that the list is likely to change over time, as optimizations
are added and reorganized.
<p>
<dl>
<dt><code><b>class/marking/final</b></code></dt>
<dd>Marks classes as final, whenever possible.</dd>
<dt><code><b>class/merging/vertical</b></code></dt>
<dd>Merges classes vertically in the class hierarchy, whenever possible.</dd>
<dt><code><b>class/merging/horizontal</b></code></dt>
<dd>Merges classes horizontally in the class hierarchy, whenever possible.</dd>
<dt><div>(⇒ <code>code/removal/advanced</code>)</div>
<code><b>field/removal/writeonly</b></code></dt>
<dd>Removes write-only fields.</dd>
<dt><code><b>field/marking/private</b></code></dt>
<dd>Marks fields as private, whenever possible.</dd>
<dt><div>(⇒ <code>code/simplification/advanced</code>)</div>
<code><b>field/propagation/value</b></code></dt>
<dd>Propagates the values of fields across methods.</dd>
<dt><code><b>method/marking/private</b></code></dt>
<dd>Marks methods as private, whenever possible (<i>devirtualization</i>).</dd>
<dt><div>(⇒ <code>code/removal/advanced</code>)</div>
<code><b>method/marking/static</b></code></dt>
<dd>Marks methods as static, whenever possible (<i>devirtualization</i>).</dd>
<dt><code><b>method/marking/final</b></code></dt>
<dd>Marks methods as final, whenever possible.</dd>
<dt><div>(⇒ <code>code/removal/advanced</code>)</div>
<code><b>method/removal/parameter</b></code></dt>
<dd>Removes unused method parameters.</dd>
<dt><div>(⇒ <code>code/simplification/advanced</code>)</div>
<code><b>method/propagation/parameter</b></code></dt>
<dd>Propagates the values of method parameters from method invocations to
the invoked methods.</dd>
<dt><div>(⇒ <code>code/simplification/advanced</code>)</div>
<code><b>method/propagation/returnvalue</b></code></dt>
<dd>Propagates the values of method return values from methods to their
invocations.</dd>
<dt><code><b>method/inlining/short</b></code></dt>
<dd>Inlines short methods.</dd>
<dt><code><b>method/inlining/unique</b></code></dt>
<dd>Inlines methods that are only called once.</dd>
<dt><code><b>method/inlining/tailrecursion</b></code></dt>
<dd>Simplifies tail recursion calls, whenever possible.</dd>
<dt><code><b>code/merging</b></code></dt>
<dd>Merges identical blocks of code by modifying branch targets.</dd>
<dt><code><b>code/simplification/variable</b></code></dt>
<dd>Performs peephole optimizations for variable loading and storing.</dd>
<dt><code><b>code/simplification/arithmetic</b></code></dt>
<dd>Performs peephole optimizations for arithmetic instructions.</dd>
<dt><code><b>code/simplification/cast</b></code></dt>
<dd>Performs peephole optimizations for casting operations.</dd>
<dt><code><b>code/simplification/field</b></code></dt>
<dd>Performs peephole optimizations for field loading and storing.</dd>
<dt><div>(⇒ <code>code/removal/simple</code>)</div>
<code><b>code/simplification/branch</b></code></dt>
<dd>Performs peephole optimizations for branch instructions.</dd>
<dt><code><b>code/simplification/string</b></code></dt>
<dd>Performs peephole optimizations for constant strings.</dd>
<dt><div>(<i>best used with</i> <code>code/removal/advanced</code>)</div>
<code><b>code/simplification/advanced</b></code></dt>
<dd>Simplifies code based on control flow analysis and data flow
analysis.</dd>
<dt><div>(⇒ <code>code/removal/exception</code>)</div>
<code><b>code/removal/advanced</b></code></dt>
<dd>Removes dead code based on control flow analysis and data flow
analysis.</dd>
<dt><div>(⇒ <code>code/removal/exception</code>)</div>
<code><b>code/removal/simple</b></code></dt>
<dd>Removes dead code based on a simple control flow analysis.</dd>
<dt><code><b>code/removal/variable</b></code></dt>
<dd>Removes unused variables from the local variable frame.</dd>
<dt><code><b>code/removal/exception</b></code></dt>
<dd>Removes exceptions with empty try blocks.</dd>
<dt><code><b>code/allocation/variable</b></code></dt>
<dd>Optimizes variable allocation on the local variable frame.</dd>
</dl>
<p>
ProGuard also provides some unofficial settings to control optimizations, that
may disappear in future versions. These are Java system properties, which
can be set as JVM arguments (with <code>-D.....)</code>:
<dl>
<dt><code><b>maximum.inlined.code.length</b></code> (default = 8 bytes)</dt>
<dd>Specifies the maximum code length (expressed in bytes) of short methods
that are eligible to be inlined. Inlining methods that are too long may
unnecessarily inflate the code size.</dd>
<dt><code><b>maximum.resulting.code.length</b></code> (default = 8000 bytes
for JSE, 2000 bytes for JME)</dt>
<dd>Specifies the maximum resulting code length (expressed in bytes) allowed
when inlining methods. Many Java virtual machines do not apply just-in-time
compilation to methods that are too long, so it's important not to let them
grow too large.</dd>
<dt><code><b>optimize.conservatively</b></code> (default = unset)</dt>
<dd>Allows input code with ordinary instructions intentionally throwing
<code>NullPointerException</code>,
<code>ArrayIndexOutOfBoundsException</code>, or
<code>ClassCastException</code>, without any other useful purposes. By
default, ProGuard may just discard such seemingly useless instructions,
resulting in better optimization of most common code.</dd>
</dl>
<hr />
<noscript><div><a target="_top" href="../index.html" class="button">Show menu</a></div></noscript>
<address>
Copyright © 2002-2013
<a target="other" href="http://www.lafortune.eu/">Eric Lafortune</a>.
</address>
</body>
</html>