• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1<?xml version="1.0" encoding="UTF-8" ?>
2<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
3<html xmlns="http://www.w3.org/1999/xhtml" lang="en">
4<head>
5  <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
6  <link rel="stylesheet" href="resources/doc.css" charset="UTF-8" type="text/css" />
7  <link rel="shortcut icon" href="resources/report.gif" type="image/gif" />
8  <title>JaCoCo - Class Ids</title>
9</head>
10<body>
11
12<div class="breadcrumb">
13  <a href="../index.html" class="el_report">JaCoCo</a> &gt;
14  <a href="index.html" class="el_group">Documentation</a> &gt;
15  <span class="el_source">Class Ids</span>
16</div>
17<div id="content">
18
19<h1>Class Ids</h1>
20
21<p>
22  As JaCoCo's class identifiers are sometimes causing confusion this chapter
23  answers the concepts and common issues with class ids in FAQ style format.
24</p>
25
26<h3>What are class ids and how are they created?</h3>
27<p>
28  Class ids are 64-bit integer values, for example
29  <code>0x638e104737889183</code> in hex notation. Their calculation is
30  considered an implementation detail of JaCoCo. Currently ids are created with
31  a CRC64 checksum of the raw class file.
32</p>
33
34<h3>What are class ids used for?</h3>
35<p>
36  Class ids are used to unambiguously identify Java classes. At runtime execution
37  data is sampled for every loaded class and typically stored to
38  <code>*.exec</code> files. At analysis time &mdash; for example for report
39  generation &mdash; the class ids are used to relate analyzed classes with the
40  execution data.
41</p>
42
43<h3>What are the advantages of JaCoCo class ids?</h3>
44<p>
45  The concept of class ids allows distinguishing different versions of classes,
46  for example when multiple versions of an application are deployed to an
47  application server or different versions of libraries are included.
48</p>
49<p>
50  Also class ids are the prerequisite for JaCoCo's minimal runtime-overhead and
51  small <code>*.exec</code> files even for very large applications under test.
52</p>
53
54<h3>What is the disadvantage of JaCoCo class ids?</h3>
55<p>
56  The fact that class ids identify a specific version of a class causes problems
57  in setups where different classes are used at runtime and at analysis time.
58</p>
59
60<h3>What happens if different classes are used at runtime and at analysis time?</h3>
61<p>
62  In this case execution data cannot be related to the analyzed classes. As a
63  consequence such classes are reported with 0% coverage.
64</p>
65
66<h3>How can I detect that I have a problem with class ids?</h3>
67<p>
68  The typical symptom of class id mismatch is classes not shown as covered
69  although they have been executed during the test. This situation can be easily
70  detected e.g. in the HTML report: Open the <i>Sessions</i> page with the link
71  on the top-right corner. You see a list of all classes where execution data
72  has been collected for. Find the class in questions and check whether the
73  entry has a link to the corresponding coverage report page. If the entry is
74  not linked this means there is a class id mismatch between the class used at
75  runtime and the class provided to create the report.
76</p>
77
78<h3>What can cause different class ids?</h3>
79<p>
80  Class ids are identical for the exact same class file only (byte-by-byte).
81  There is a couple of reasons why you might get different class files. First
82  compiling Java source files will result in different class files if you use
83  a different tool chain:
84</p>
85<ul>
86  <li>Different compiler vendor (e.g. Eclipse vs. Oracle JDK)</li>
87  <li>Different compiler versions</li>
88  <li>Different compiler settings (e.g. debug vs. non-debug) </li>
89</ul>
90<p>
91  Also post-processing class files (obfuscation, AspectJ, etc.) will typically
92  change the class files. JaCoCo will work well if you simply use the same class
93  files for runtime as well as for analysis. So the tool chain to create these
94  class files does not matter.
95</p>
96<p>
97  Even if the class files on the file system are the same there is possible that
98  classes seen by the JaCoCo runtime agent are different anyways. This typically
99  happens when another Java agent is configured <i>before</i> the JaCoCo agent
100  or special class loaders pre-process the class files. Typical candidates are:
101</p>
102<ul>
103  <li>Mocking frameworks</li>
104  <li>Application servers</li>
105  <li>Persistence frameworks</li>
106</ul>
107
108<h3>What workarounds exist to deal with runtime-modified classes?</h3>
109<p>
110  If classes get modified at runtime in your setup there are some workarounds to
111  make JaCoCo work anyways:
112</p>
113<ul>
114  <li>If you use another Java agent make sure the <a href="agent.html">JaCoCo
115      agent</a> is specified at first in the command line. This way the JaCoCo
116      agent should see the original class files.</li>
117  <li>Specify the <code>classdumpdir</code> option of the
118      <a href="agent.html">JaCoCo agent</a> and use the dumped classes at report
119      generation. Note that only loaded classes will be dumped, i.e. classes not
120      executed at all will not show-up in your report as not covered.</li>
121  <li>Use <a href="offline.html">offline instrumentation</a> before you run your
122      tests. This way classes get instrumented by JaCoCo before any runtime
123      modification can take place. Note that in this case the report has to be
124      generated with the <i>original</i> classes, not with instrumented ones.</li>
125</ul>
126
127<h3>Why can't JaCoCo simply use the class name to identify classes?</h3>
128<p>
129  To understand why JaCoCo can't rely on class names we need to have a look at
130  the way how JaCoCo measures code coverage.
131</p>
132<p>
133  JaCoCo tracks execution with so called <i>probes</i>. Probes are additional
134  byte code instructions inserted in the original class file which will note
135  when they are executed and report this to the JaCoCo runtime. This process is
136  called <i>instrumentation</i>. To keep the runtime overhead minimal, only a
137  few probes are inserted at "strategic" places. These probe positions are
138  determined by <a href="flow.html">analyzing the control flow</a> of all
139  methods of a class. As a result every instrumented class produces a list of
140  <code>n</code> boolean flags indicating whether the probe has been executed or
141  not. A JaCoCo <code>*.exec</code> file simply stores a boolean array per
142  class id.
143</p>
144<p>
145  At analysis time, for example for report generation, the <code>*.exec</code>
146  file is used to get information about probe execution status. But as probes
147  are stored in a plain boolean array there is no information like corresponding
148  methods or lines. To retrieve this information we need the original class
149  files and perform the exact same control flow analysis than at instrumentation
150  time. Because this is a deterministic process we get the same probe positions.
151  With this information we can now interfere the execution status of every
152  single instruction and branch of a method. Using the debug information
153  embedded in the class files we can also calculate line coverage.
154</p>
155<p>
156  If we would use just slightly different classes at analysis time than at
157  runtime &mdash; e.g. different method ordering or additional branches &mdash;
158  we would end-up with different probes. For example the probe at index
159  <code>i</code> would be in method <code>a()</code> and not in method
160  <b>b()</b>. Obviously this will create random coverage results.
161</p>
162
163<h3>Why do I get an error when I try to analyze multiple versions of the same
164    class with a group?</h3>
165<p>
166  JaCoCo always analyzes a set of class as a group. The group is used to
167  aggregate data for source files and packages (both can contain multiple
168  classes). Within the reporting API classes are identified by their fully
169  qualified name (e.g. to create stable file names in the HTML reports).
170  Therefore it is not possible to include two different classes with the same
171  name within a group. Anyhow it is possible to analyze different versions of
172  class files in separate groups, for example the <a href="ant.html#report">Ant
173  report task</a> can be configured with multiple groups.
174</p>
175
176</div>
177<div class="footer">
178  <span class="right"><a href="@jacoco.home.url@">JaCoCo</a> @qualified.bundle.version@</span>
179  <a href="license.html">Copyright</a> &copy; @copyright.years@ Mountainminds GmbH &amp; Co. KG and Contributors
180</div>
181
182</body>
183</html>
184