This is the Metrics Team Java Style Guide, which not only is concerned with coding style but which also contains recommendations for setting up Java projects in Metrics.
== Dependencies ==
Ideally the Java JDK alone should do. If that is not sufficient, mature libraries available in current Debian stable are fine.
A 'mature' library is a well tested and widely used library, e.g. Apache commons-lang or logback. In addition, we follow these [wiki:org/operations/Guidelines guidelines].
Additional dependencies that are not part of the release and not necessary for preparing a release are less restricted. For example, for coding style conformance we use checkstyle version > 6, which is not part of Debian stable, as the Debian stable version cannot support the functionality needed here.
== Build Environment ==
Metrics Java projects use Ant as opposed to Maven or similar tools.
Ant gives us tighter control over dependencies and is very flexible to serve all development needs.
Metrics' java projects also should make use of [https://gitweb.torproject.org/metrics-base.git metrics-base] in order to avoid multiplying ant task implementations (see [https://gitweb.torproject.org/collector.git/tree CollecTor's git repository] for a usage example).
=== Unit Testing ===
Currently, testing is facilitated by JUnit.
Not every project has a test coverage of 90% or more (yet).
In order to achieve this goal, patches should aim at 100% test coverage.
== Code Metrics ==
In order to provide high quality code we use various code metrics.
In the future, we will consider to add other metrics to this list.
Ticket #19731 discusses these topics in detail.
=== Coverage and Complexity ===
Test coverage is measured by Cobertura. The Cobertura results also contain code complexity metrics.
We currently have a higher priority for test coverage, but eventually we will also take code complexity into account for improvement.
The cobertura check is always set (and should be adapted) to ensure that coverage doesn't decrease.
=== Conformance to the Style Guide ===
The coding style is checked by checkstyle. Currently, not all projects confirm to the style guide (completely).
That's why patches should not introduce new checkstyle complaints.
== File Structure of Java Projects ==
This is the common file structure for Metrics Team's java projects.
It is adapted from [https://maven.apache.org/guides/introduction/introduction-to-the-standard-directory-layout.html Maven's standard structure]:
<tdalign="left">classes instrumented by cobertura</td></tr>
<tr>
<tdalign="left"><tt>generated/javadoc</tt></td>
<tdalign="left">javadoc</td></tr>
</table>
}}}
The top-level should hold license, readme, and build.xml.
== Logging ==
Metrics' Java projects use SLF4J throughout.
Use the efficient style for parametrized logging statements as explained [http://logback.qos.ch/manual/architecture.html#parametrized here].
Usage example:
Bad:
~~`log.warn("Number " + x + " is not prime.")`
Good:
`log.warn("Number {} is not prime.", x)`
The different log levels provided by SLF4J should be used as follows:
* trace: Shouldn't be used and neither enabled.
* debug: Is the finest level. It should be used for detailed messages to debug a problem. The debug log level is usually turned off and is only enabled when debugging a problem.
* info: This is the log level for normal operation and is turned on in production. Info level messages should indicate when major operations during a run begin or end. The messages should be informative and readable for the general operator.
* warn: Warnings inform about problems that can be overcome by the application or that at least don't hinder basic operation. The message should indicate if the operator needs to interfere and state what ought to be done, e.g. "provide more disk space" etc.
* error: Errors are problems the application cannot recover from alone, e.g., "no space left", or problems that were not anticipated. The unexpected errors should have ask the operator to post the error log message to metrics-team or the bug tracker.
Log messages are written to loggers. The usual logger is identified by the simple class name, e.g., `LoggerFactory.getLogger(Main.class)`. For special purposes other loggers could be used, i.e., `LoggerFactory.getLogger("statistics")` for performance measurements.
== Coding Style
This style guide is derived from [https://google.github.io/styleguide/javaguide.html Google's Java Style Guide], which is licensed under [https://github.com/google/styleguide CC-By 3.0 License]. The exact git revision used as basis can be found [https://github.com/google/styleguide/commit/7969290bacb1965d09677a79d523b4871c9d039c here].
<p>This document serves as the <strong>complete</strong> definition of Metrics-Team's coding standards for
source code in the Java™ Programming Language. A Java source file is described as being <em>in
Metrics-Team's Style</em> if and only if it adheres to the rules herein.</p><p>Like other programming style guides, the issues covered span not only aesthetic issues of
formatting, but other types of conventions or coding standards as well. However, this document
focuses primarily on the <strong>hard-and-fast rules</strong> that we follow universally, and
avoids giving <em>advice</em> that isn't clearly enforceable (whether by human or tool).
<p>In this document, unless otherwise clarified:</p><ol><li>The term <em>class</em> is used inclusively to mean an "ordinary" class, enum class,
interface or annotation type (<code class="prettyprint lang-java">@interface</code>).</li><li>The term <em>comment</em> always refers to <em>implementation</em> comments. We do not
use the phrase "documentation comments", instead using the common term "Javadoc."</li></ol><p>Other "terminology notes" will appear occasionally throughout the document.</p><aname="s1.2-guide-notes"/>
<p>Source files are encoded in <strong>UTF-8</strong>.</p><aname="s2.3-special-characters"/>
<h3>2.3 Special characters <ahref="#s2.3-special-characters"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>Aside from the line terminator sequence, the <strong>ASCII horizontal space
character</strong> (<strong>0x20</strong>) is the only whitespace character that appears
anywhere in a source file. This implies that:</p><ol><li>All other whitespace characters in string and character literals are escaped.</li><li>Tab characters are <strong>not</strong> used for indentation.</li></ol><aname="s2.3.2-special-escape-sequences"/>
<h4>2.3.2 Special escape sequences <ahref="#s2.3.2-special-escape-sequences"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>For any character that has a special escape sequence
(<codeclass="prettyprint lang-java">\b</code>,
<codeclass="prettyprint lang-java">\t</code>,
<codeclass="prettyprint lang-java">\n</code>,
<codeclass="prettyprint lang-java">\f</code>,
<codeclass="prettyprint lang-java">\r</code>,
<codeclass="prettyprint lang-java">\"</code>,
<codeclass="prettyprint lang-java">\'</code> and
<codeclass="prettyprint lang-java">\\</code>), that sequence
is used rather than the corresponding octal
(e.g. <codeclass="badcode">\012</code>) or Unicode
<p>For the remaining non-ASCII characters, either the actual Unicode character
(e.g. <codeclass="prettyprint lang-java">∞</code>) or the equivalent Unicode escape
(e.g. <codeclass="prettyprint lang-java">\u221e</code>) is used, depending only on which
makes the code <strong>easier to read and understand</strong>.</p><pclass="tip"><strong>Tip:</strong> In the Unicode escape case, and occasionally even when actual
Unicode characters are used, an explanatory comment can be very helpful.</p><p>Examples:</p><table><tr><th>Example</th><th>Discussion</th></tr><tr><td><codeclass="prettyprint lang-java">String unitAbbrev = "μs";</code></td><td>Best: perfectly clear even without a comment.</td></tr><tr><td><codeclass="prettyprint lang-java">String unitAbbrev = "\u03bcs"; // "μs"</code></td><td>Allowed, but there's no reason to do this.</td></tr><tr><td><codeclass="prettyprint lang-java">String unitAbbrev = "\u03bcs";
// Greek letter mu, "s"</code></td><td>Allowed, but awkward and prone to mistakes.</td></tr><tr><td><codeclass="badcode">String unitAbbrev = "\u03bcs";</code></td><td>Poor: the reader has no idea what this is.</td></tr><tr><td><codeclass="prettyprint lang-java">return '\ufeff' + content;
// byte order mark</code></td><td>Good: use escapes for non-printable characters, and comment if necessary.</td></tr></table><pclass="tip"><strong>Tip:</strong> Never make your code less readable simply out of fear that
some programs might not handle non-ASCII characters properly. If that should happen, those
programs are <strong>broken</strong> and they must be <strong>fixed</strong>.</p><aname="filestructure"/><aname="s3-source-file-structure"/>
<div><p>A source file consists of, <strong>in order</strong>:</p><ol><li>License or copyright information, if present</li><li>Package statement</li><li>Import statements</li><li>Exactly one top-level class</li></ol></div><p><strong>Exactly one blank line</strong> separates each section that is present.</p><aname="s3.1-copyright-statement"/>
<h3>3.1 License or copyright information, if present <ahref="#s3.1-copyright-statement"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>If license or copyright information belongs in a file, it belongs here.</p><aname="s3.2-package-statement"/>
<h4>3.3.1 No wildcard imports <ahref="#s3.3.1-wildcard-imports"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p><strong>Wildcard imports</strong>, static or otherwise, <strong>are not used</strong>.</p><aname="s3.3.2-import-line-wrapping"/>
<h4>3.3.2 No line-wrapping <ahref="#s3.3.2-import-line-wrapping"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Import statements are <strong>not line-wrapped</strong>. The column limit (Section 4.4,
<ahref="#s4.4-column-limit">Column limit: 80</a>) does not apply to import
<h4>3.3.3 Ordering and spacing <ahref="#s3.3.3-import-ordering-and-spacing"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Import statements are divided into the following groups, in this order, with each group
separated by a single blank line:</p><ol><li>All static imports in a single group</li><li><code>org.torproject</code> imports
(only if this source file is in the <code>org.torproject</code> package
space)</li><li>Third-party imports, one group per top-level package, in ASCII sort order
<code>sun</code></li></ul></li><li><code>java</code> imports</li><li><code>javax</code> imports</li></ol><p>Within a group there are no blank lines, and the imported names appear in ASCII sort
order. (<strong>Note:</strong> this is not the same as the import <em>statements</em> being in
ASCII sort order; the presence of semicolons warps the result.)</p><aname="s3.4-class-declaration"/>
<h3>3.4 Class declaration <ahref="#s3.4-class-declaration"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<h4>3.4.1 Exactly one top-level class declaration <ahref="#s3.4.1-one-top-level-class"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Each top-level class resides in a source file of its own.</p><aname="s3.4.2-class-member-ordering"/>
<h4>3.4.2 Class member ordering <ahref="#s3.4.2-class-member-ordering"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>All attributes declaration should be in alphabetical order at the beginning of a class. When looking at a java file in a simple text editor it's easy
to jump to the very beginning, look at the declaration and then jump back to the method declaration. The declaration might be far away from its first use,
but easier to locate.</p>
<p>
Method declarations should appear in <em>depth first</em> order by usage. Example:
<h5>3.4.2.1 Overloads: never split <ahref="#s3.4.2.1-overloads-never-split"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h5>
<p>When a class has multiple constructors, or multiple methods with the same name, these appear
sequentially, with no intervening members.</p><aname="s4-formatting"/>
<h4>4.1.1 Braces are used where optional <ahref="#s4.1.1-braces-always-used"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Braces are used with
<codeclass="prettyprint lang-java">if</code>,
<codeclass="prettyprint lang-java">else</code>,
<codeclass="prettyprint lang-java">for</code>,
<codeclass="prettyprint lang-java">do</code> and
<codeclass="prettyprint lang-java">while</code> statements, even when the
body is empty or contains only a single statement.</p><aname="s4.1.2-blocks-k-r-style"/>
<h4>4.1.2 Nonempty blocks: K & R style <ahref="#s4.1.2-blocks-k-r-style"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
for <em>nonempty</em> blocks and block-like constructs:</p><ul><li>No line break before the opening brace.</li><li>Line break after the opening brace.</li><li>Line break before the closing brace.</li><li>Line break after the closing brace <em>if</em> that brace terminates a statement or the body
of a method, constructor or <em>named</em> class. For example, there is <em>no</em> line break
after the brace if it is followed by <codeclass="prettyprint lang-java">else</code> or a
<h4>4.1.3 Empty blocks: may be concise <ahref="#s4.1.3-braces-empty-blocks"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>An empty block or block-like construct <em>may</em> be closed immediately after it is
opened, with no characters or line break in between
(<codeclass="prettyprint lang-java">{}</code>), <strong>unless</strong> it is part of a
<em>multi-block statement</em> (one that directly contains multiple blocks:
<codeclass="prettyprint lang-java">if/else-if/else</code> or
<p>Each time a new block or block-like construct is opened, the indent increases by two
spaces. When the block ends, the indent returns to the previous indent level. The indent level
applies to both code and comments throughout the block. (See the example in Section 4.1.2,
<ahref="#s4.1.2-blocks-k-r-style">Nonempty blocks: K & R Style</a>.)</p><aname="s4.3-one-statement-per-line"/>
<h3>4.3 One statement per line <ahref="#s4.3-one-statement-per-line"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>Each statement is followed by a line-break.</p><aname="columnlimit"/><aname="s4.4-column-limit"/>
</p><p><strong>Exceptions:</strong></p><ol><li>Lines where obeying the column limit is not possible (for example, a long URL in Javadoc,
or a long JSNI method reference).</li><li><codeclass="prettyprint lang-java">package</code> and
<codeclass="prettyprint lang-java">import</code> statements (see Sections
3.2 <ahref="#s3.2-package-statement">Package statement</a> and
3.3 <ahref="#s3.3-import-statements">Import statements</a>).</li><li>Command lines in a comment that may be cut-and-pasted into a shell.</li></ol><aname="s4.5-line-wrapping"/>
<pclass="terminology"><strong>Terminology Note:</strong> When code that might otherwise legally
occupy a single line is divided into multiple lines, typically to avoid overflowing the column
limit, this activity is called
<em>line-wrapping</em>.</p><p>There is no comprehensive, deterministic formula showing <em>exactly</em> how to line-wrap in
every situation. Very often there are several valid ways to line-wrap the same piece of code.</p><pclass="tip"><strong>Tip:</strong> Extracting a method or local variable may solve the problem
without the need to line-wrap.</p><aname="s4.5.1-line-wrapping-where-to-break"/>
<h4>4.5.1 Where to break <ahref="#s4.5.1-line-wrapping-where-to-break"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>The prime directive of line-wrapping is: prefer to break at a
<strong>higher syntactic level</strong>. Also:</p><ol><li>When a line is broken at a <em>non-assignment</em> operator the break comes <em>before</em>
the symbol. (Note that this is not the same practice used in Google style for other languages,
such as C++ and JavaScript.)
<ul><li>This also applies to the following "operator-like" symbols: the dot separator
(<codeclass="prettyprint lang-java">.</code>), the ampersand in type bounds
(<codeclass="prettyprint lang-java"><T extends Foo & Bar></code>), and the pipe in
catch blocks
(<codeclass="prettyprint lang-java">catch (FooException | BarException e)</code>).</li></ul></li><li>When a line is broken at an <em>assignment</em> operator the break typically comes
<em>after</em> the symbol, but either way is acceptable.
<ul><li>This also applies to the "assignment-operator-like" colon in an enhanced
<codeclass="prettyprint lang-java">for</code> ("foreach") statement.</li></ul></li><li>A method or constructor name stays attached to the open parenthesis
(<codeclass="prettyprint lang-java">(</code>) that follows it.</li><li>A comma (<codeclass="prettyprint lang-java">,</code>) stays attached to the token that
<h4>4.5.2 Indent continuation lines at least +4 spaces <ahref="#s4.5.2-line-wrapping-indent"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>When line-wrapping, each line after the first (each <em>continuation line</em>) is indented
at least +4 from the original line.</p><p>When there are multiple continuation lines, indentation may be varied beyond +4 as
desired. In general, two continuation lines use the same indentation level if and only if they
begin with syntactically parallel elements.</p><p>Section 4.6.3 on <ahref="#s4.6.3-horizontal-alignment">Horizontal alignment</a> addresses
the discouraged practice of using a variable number of spaces to align certain tokens with
<ul><li><spanclass="exception"><strong>Exception:</strong> A blank line between two consecutive
fields (having no other code between them) is optional. Such blank lines are used as needed to
create <em>logical groupings</em> of fields.</span></li></ul></li><li>Within method bodies, as needed to create <em>logical groupings</em> of statements.</li><li><em>Optionally</em> before the first member or after the last member of the class (neither
encouraged nor discouraged).</li><li>As required by other sections of this document (such as Section 3.3,
<ahref="#s3.3-import-statements">Import statements</a>).</li></ol><p><em>Multiple</em> consecutive blank lines are permitted, but never required (or encouraged).</p><aname="s4.6.2-horizontal-whitespace"/>
<p>Beyond where required by the language or other style rules, and apart from literals, comments and
Javadoc, a single ASCII space also appears in the following places <strong>only</strong>.</p><ol><li>Separating any reserved word, such as
<codeclass="prettyprint lang-java">if</code>,
<codeclass="prettyprint lang-java">for</code> or
<codeclass="prettyprint lang-java">catch</code>, from an open parenthesis
(<codeclass="prettyprint lang-java">(</code>)
that follows it on that line</li><li>Separating any reserved word, such as
<codeclass="prettyprint lang-java">else</code> or
<codeclass="prettyprint lang-java">catch</code>, from a closing curly brace
(<codeclass="prettyprint lang-java">}</code>) that precedes it on that line</li><li>Before any open curly brace
(<codeclass="prettyprint lang-java">{</code>), with two exceptions:
<ul><li><codeclass="prettyprint lang-java">@SomeAnnotation({a, b})</code> (no space is used)</li><li><codeclass="prettyprint lang-java">String[][] x = {{"foo"}};</code> (no space is required
between <codeclass="prettyprint lang-java">{{</code>, by item 8 below)</li></ul></li><li>On both sides of any binary or ternary operator. This also applies to the following
"operator-like" symbols:
<ul><li>the ampersand in a conjunctive type bound:
<codeclass="prettyprint lang-java"><T extends Foo & Bar></code></li><li>the pipe for a catch block that handles multiple exceptions:
<codeclass="prettyprint lang-java">catch (FooException | BarException e)</code></li><li>the colon (<codeclass="prettyprint lang-java">:</code>) in an enhanced
<codeclass="prettyprint lang-java">for</code> ("foreach") statement</li></ul></li><li>After <codeclass="prettyprint lang-java">,:;</code> or the closing parenthesis
(<codeclass="prettyprint lang-java">)</code>) of a cast</li><li>On both sides of the double slash (<codeclass="prettyprint lang-java">//</code>) that
begins an end-of-line comment. Here, multiple spaces are allowed, but not required.</li><li>Between the type and variable of a declaration:
<codeclass="prettyprint lang-java">List<String> list</code></li><li><em>Optional</em> just inside both braces of an array initializer
<ul><li><codeclass="prettyprint lang-java">new int[] {5, 6}</code> and
<codeclass="prettyprint lang-java">new int[] { 5, 6 }</code> are both valid</li></ul></li></ol><pclass="note"><strong>Note:</strong> This rule never requires or forbids additional space at the
start or end of a line, only <em>interior</em> space.</p><aname="s4.6.3-horizontal-alignment"/>
<h4>4.6.3 Horizontal alignment: never required <ahref="#s4.6.3-horizontal-alignment"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<pclass="terminology"><strong>Terminology Note:</strong><em>Horizontal alignment</em> is the
practice of adding a variable number of additional spaces in your code with the goal of making
certain tokens appear directly below certain other tokens on previous lines.</p><p>This practice is permitted, but is <strong>never required</strong> by Google Style. It is not
even required to <em>maintain</em> horizontal alignment in places where it was already used.</p><p>Here is an example without alignment, then using alignment:</p><preclass="prettyprint lang-java">
private int x; // this is fine
private Color color; // this too
private int x; // permitted, but future edits
private Color color; // may leave it unaligned
</pre><pclass="tip"><strong>Tip:</strong> Alignment can aid readability, but it creates problems for
future maintenance. Consider a future change that needs to touch just one line. This change may
leave the formerly-pleasing formatting mangled, and that is <strong>allowed</strong>. More often
it prompts the coder (perhaps you) to adjust whitespace on nearby lines as well, possibly
triggering a cascading series of reformattings. That one-line change now has a "blast radius."
This can at worst result in pointless busywork, but at best it still corrupts version history
information, slows down reviewers and exacerbates merge conflicts.</p><aname="parentheses"/><aname="s4.7-grouping-parentheses"/>
<h3>4.8 Specific constructs <ahref="#s4.8-specific-constructs"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>After each comma that follows an enum constant, a line-break is optional.</p><p>An enum class with no methods and no documentation on its constants may optionally be formatted
as if it were an array initializer (see Section 4.8.3.1 on
</pre><p>Since enum classes <em>are classes</em>, all other rules for formatting classes apply.</p><aname="localvariables"/><aname="s4.8.2-variable-declarations"/>
<h5>4.8.2.1 One variable per declaration <ahref="#s4.8.2.1-variables-per-declaration"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h5>
<p>Every variable declaration (field or local) declares only one variable: declarations such as
<codeclass="badcode">int a, b;</code> are not used.</p><aname="s4.8.2.2-variables-limited-scope"/>
<h5>4.8.2.2 Declared when needed, initialized as soon as
<h5>4.8.3.1 Array initializers: can be "block-like" <ahref="#s4.8.3.1-array-initializers"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h5>
<p>Any array initializer may <em>optionally</em> be formatted as if it were a "block-like
construct." For example, the following are all valid (<strong>not</strong> an exhaustive
list):</p><preclass="prettyprint lang-java">
new int[] { new int[] {
0, 1, 2, 3 0,
} 1,
2,
new int[] { 3,
0, 1, }
2, 3
} new int[]
{0, 1, 2, 3}
</pre><aname="s4.8.3.2-array-declarations"/>
<h5>4.8.3.2 No C-style array declarations <ahref="#s4.8.3.2-array-declarations"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h5>
<p>The square brackets form a part of the <em>type</em>, not the variable:
<codeclass="prettyprint lang-java">String[] args</code>, not
<p>As with any other block, the contents of a switch block are indented +2.</p><p>After a switch label, a newline appears, and the indentation level is increased +2, exactly as
if a block were being opened. The following switch label returns to the previous indentation
level, as if a block had been closed.</p><aname="fallthrough"/><aname="s4.8.4.2-switch-fall-through"/>
<codeclass="prettyprint lang-java">return</code> or thrown exception), or is marked with a comment
to indicate that execution will or <em>might</em> continue into the next statement group. Any
comment that communicates the idea of fall-through is sufficient (typically
<codeclass="prettyprint lang-java">// fall through</code>). This special comment is not required in
the last statement group of the switch block. Example:</p><preclass="prettyprint lang-java">
switch (input) {
case 1:
case 2:
prepareOneOrTwo();
// fall through
case 3:
handleOneTwoOrThree();
break;
default:
handleLargeNumber(input);
}
</pre><aname="s4.8.4.3-switch-default"/>
<h5>4.8.4.3 The default case is present <ahref="#s4.8.4.3-switch-default"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h5>
<p>Each switch statement includes a <codeclass="prettyprint lang-java">default</code> statement
group, even if it contains no code.</p><aname="annotations"/><aname="s4.8.5-annotations"/>
<p>Block comments are indented at the same level as the surrounding code. They may be in
<codeclass="prettyprint lang-java">/* ... */</code> style or
<codeclass="prettyprint lang-java">// ...</code> style. For multi-line
<codeclass="prettyprint lang-java">/* ... */</code> comments, subsequent lines must start with
<code>*</code> aligned with the <code>*</code> on the previous line.</p><preclass="prettyprint lang-java">
/*
* This is // And so /* Or you can
* okay. // is this. * even do this. */
*/
</pre><p>Comments are not enclosed in boxes drawn with asterisks or other characters.</p><pclass="tip"><strong>Tip:</strong> When writing multi-line comments, use the
<codeclass="prettyprint lang-java">/* ... */</code> style if you want automatic code formatters to
re-wrap the lines when necessary (paragraph-style). Most formatters don't re-wrap lines in
<h3>5.1 Rules common to all identifiers <ahref="#s5.1-identifier-names"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>Identifiers use only ASCII letters and digits, and in two cases noted below, underscores. Thus
each valid identifier name is matched by the regular expression <code>\w+</code> .</p><p> In Google Style special prefixes or
suffixes, like those seen in the examples <codeclass="badcode">name_</code>,
<codeclass="badcode">mName</code>, <codeclass="badcode">s_name</code> and
<codeclass="badcode">kName</code>, are <strong>not</strong> used.</p><aname="s5.2-specific-identifier-names"/>
<h3>5.2 Rules by identifier type <ahref="#s5.2-specific-identifier-names"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<h4>5.2.2 Class names <ahref="#s5.2.2-class-names"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Class names are written in <ahref="#s5.3-camel-case">UpperCamelCase</a>.</p><p>Class names are typically nouns or noun phrases. For example,
<codeclass="prettyprint lang-java">Character</code> or
<codeclass="prettyprint lang-java">ImmutableList</code>. Interface names may also be nouns or
noun phrases (for example, <codeclass="prettyprint lang-java">List</code>), but may sometimes be
adjectives or adjective phrases instead (for example,
<codeclass="prettyprint lang-java">Readable</code>).</p><p>There are no specific rules or even well-established conventions for naming annotation types.</p><p><em>Test</em> classes are named starting with the name of the class they are testing, and ending
with <codeclass="prettyprint lang-java">Test</code>. For example,
<codeclass="prettyprint lang-java">HashTest</code> or
<p>Constant names use <codeclass="prettyprint lang-java">CONSTANT_CASE</code>: all uppercase
letters, with words separated by underscores. But what <em>is</em> a constant, exactly?</p><p>Every constant is a static final field, but not all static final fields are constants. Before
choosing constant case, consider whether the field really <em>feels like</em> a constant. For
example, if any of that instance's observable state can change, it is almost certainly not a
constant. Merely <em>intending</em> to never mutate the object is generally not
static final ImmutableList<String> NAMES = ImmutableList.of("Ed", "Ann");
static final Joiner COMMA_JOINER = Joiner.on(','); // because Joiner is immutable
static final SomeMutableType[] EMPTY_ARRAY = {};
enum SomeEnum { ENUM_CONSTANT }
// Not constants
static String nonFinal = "non-final";
final String nonStatic = "non-static";
static final Set<String> mutableCollection = new HashSet<String>();
static final ImmutableSet<SomeMutableType> mutableElements = ImmutableSet.of(mutable);
static final Logger logger = Logger.getLogger(MyClass.getName());
static final String[] nonEmptyArray = {"these", "can", "change"};
</pre><p>These names are typically nouns or noun phrases.</p><aname="s5.2.5-non-constant-field-names"/>
<h4>5.2.5 Non-constant field names <ahref="#s5.2.5-non-constant-field-names"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Non-constant field names (static or otherwise) are written
in <ahref="#s5.3-camel-case">lowerCamelCase</a>.</p><p>These names are typically nouns or noun phrases. For example,
<codeclass="prettyprint lang-java">computedValues</code> or
<p>Parameter names are written in <ahref="#s5.3-camel-case">lowerCamelCase</a>.</p><p>One-character parameter names should be avoided.</p><aname="s5.2.7-local-variable-names"/>
<h4>5.2.7 Local variable names <ahref="#s5.2.7-local-variable-names"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Local variable names are written in <ahref="#s5.3-camel-case">lowerCamelCase</a>, and can be
abbreviated more liberally than other types of names.</p><p>However, one-character names should be avoided, except for temporary and looping variables.</p><p>Even when final and immutable, local variables are not considered to be constants, and should not
be styled as constants.</p><aname="s5.2.8-type-variable-names"/>
<h4>5.2.8 Type variable names <ahref="#s5.2.8-type-variable-names"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>Each type variable is named in one of two styles:</p><ul><li>A single capital letter, optionally followed by a single numeral (such as
<p>Sometimes there is more than one reasonable way to convert an English phrase into camel case,
such as when acronyms or unusual constructs like "IPv6" or "iOS" are present. To improve
predictability, Google Style specifies the following (nearly) deterministic scheme.</p><p>Beginning with the prose form of the name:</p><ol><li>Convert the phrase to plain ASCII and remove any apostrophes. For example, "Müller's
algorithm" might become "Muellers algorithm".</li><li>Divide this result into words, splitting on spaces and any remaining punctuation (typically
hyphens).
<ul><li><em>Recommended:</em> if any word already has a conventional camel-case appearance in common
usage, split this into its constituent parts (e.g., "AdWords" becomes "ad words"). Note
that a word such as "iOS" is not really in camel case <em>per se</em>; it defies <em>any</em>
convention, so this recommendation does not apply.</li></ul></li><li>Now lowercase <em>everything</em> (including acronyms), then uppercase only the first
character of:
<ul><li>... each word, to yield <em>upper camel case</em>, or</li><li>... each word except the first, to yield <em>lower camel case</em></li></ul></li><li>Finally, join all the words into a single identifier.</li></ol><p>Note that the casing of the original words is almost entirely disregarded. Examples:</p><table><tr><th>Prose form</th><th>Correct</th><th>Incorrect</th></tr><tr><td>"XML HTTP request"</td><td><codeclass="prettyprint lang-java">XmlHttpRequest</code></td><td><codeclass="badcode">XMLHTTPRequest</code></td></tr><tr><td>"new customer ID"</td><td><codeclass="prettyprint lang-java">newCustomerId</code></td><td><codeclass="badcode">newCustomerID</code></td></tr><tr><td>"inner stopwatch"</td><td><codeclass="prettyprint lang-java">innerStopwatch</code></td><td><codeclass="badcode">innerStopWatch</code></td></tr><tr><td>"supports IPv6 on iOS?"</td><td><codeclass="prettyprint lang-java">supportsIpv6OnIos</code></td><td><codeclass="badcode">supportsIPv6OnIOS</code></td></tr><tr><td>"YouTube importer"</td><td><codeclass="prettyprint lang-java">YouTubeImporter</code><br/><codeclass="prettyprint lang-java">YoutubeImporter</code>*</td><td/></tr></table><p>*Acceptable, but not recommended.</p><pclass="note"><strong>Note:</strong> Some words are ambiguously hyphenated in the English
language: for example "nonempty" and "non-empty" are both correct, so the method names
<codeclass="prettyprint lang-java">checkNonempty</code> and
<codeclass="prettyprint lang-java">checkNonEmpty</code> are likewise both correct.</p><aname="s6-programming-practices"/>
<h3>6.2 Caught exceptions: not ignored <ahref="#s6.2-caught-exceptions"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>Except as noted below, it is very rarely correct to do nothing in response to a caught
exception. (Typical responses are to log it, or if it is considered "impossible", rethrow it as an
<codeclass="prettyprint lang-java">AssertionError</code>.)</p><p>When it truly is appropriate to take no action whatsoever in a catch block, the reason this is
justified is explained in a comment.</p><preclass="prettyprint lang-java">
try {
int i = Integer.parseInt(response);
return handleNumericResponse(i);
} catch (NumberFormatException ok) {
// it's not numeric; that's fine, just continue
}
return handleTextResponse(response);
</pre><pclass="exception"><strong>Exception:</strong> In tests, a caught exception may be ignored
without comment <em>if</em> it is named <codeclass="prettyprint lang-java">expected</code>. The
following is a very common idiom for ensuring that the method under test <em>does</em> throw an
exception of the expected type, so a comment is unnecessary here.</p><preclass="prettyprint lang-java">
try {
emptyStack.pop();
fail();
} catch (NoSuchElementException expected) {
}
</pre><aname="s6.3-static-members"/>
<h3>6.3 Static members: qualified using class <ahref="#s6.3-static-members"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>When a reference to a static class member must be qualified, it is qualified with that class's
name, not with a reference or expression of that class's type.</p><preclass="prettyprint lang-java">
Foo aFoo = ...;
Foo.aStaticMethod(); // good
<spanclass="badcode">aFoo.aStaticMethod();</span> // bad
<spanclass="badcode">somethingThatYieldsAFoo().aStaticMethod();</span> // very bad
<h3>6.4 Finalizers: not used <ahref="#s6.4-finalizers"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>It is <strong>extremely rare</strong> to override <codeclass="prettyprint lang-java">Object.finalize</code>.</p><pclass="tip"><strong>Tip:</strong> Don't do it. If you absolutely must, first read and understand
<h4>7.1.1 General form <ahref="#s7.1.1-javadoc-multi-line"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h4>
<p>The <em>basic</em> formatting of Javadoc blocks is as seen in this example:</p><preclass="prettyprint lang-java">
/**
* Multiple lines of Javadoc text are written here,
* wrapped normally...
*/
public int method(String p1) { ... }
</pre><p>... or in this single-line example:</p><preclass="prettyprint lang-java">
/** An especially short bit of Javadoc. */
</pre><p>The basic form is always acceptable. The single-line form may be substituted when there are no
at-clauses present, and the entirety of the Javadoc block (including comment markers) can fit on a
single line.</p><aname="s7.1.2-javadoc-paragraphs"/>
<p>Any of the standard "at-clauses" that are used appear in the order <code>@param</code>,
<code>@return</code>, <code>@throws</code>, <code>@deprecated</code>, and these four types never
appear with an empty description. When an at-clause doesn't fit on a single line, continuation lines
are indented four (or more) spaces from the position of the <code>@</code>.
</p><aname="s7.2-summary-fragment"/>
<h3>7.2 The summary fragment <ahref="#s7.2-summary-fragment"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>The Javadoc for each class and member begins with a brief <strong>summary fragment</strong>. This
fragment is very important: it is the only part of the text that appears in certain contexts such as
class and method indexes.</p><p>This is a fragment—a noun phrase or verb phrase, not a complete sentence. It does
<strong>not</strong> begin with <code class="badcode">A {@code Foo} is a...</code>, or
<codeclass="badcode">This method returns...</code>, nor does it form a complete imperative sentence
like <codeclass="badcode">Save the record.</code>. However, the fragment is capitalized and
punctuated as if it were a complete sentence.</p><pclass="tip"><strong>Tip:</strong> A common mistake is to write simple Javadoc in the form
<codeclass="badcode">/** @return the customer ID */</code>. This is incorrect, and should be
changed to <codeclass="prettyprint lang-java">/** Returns the customer ID. */</code>.</p><aname="s7.3.3-javadoc-optional"/><aname="s7.3-javadoc-where-required"/>
<h3>7.3 Where Javadoc is used <ahref="#s7.3-javadoc-where-required"><imgheight="21"width="21"src="/projects/tor/raw-attachment/wiki/org/teams/MetricsTeam/MetricsJavaStyleGuide/javaguidelink.png"/></a></h3>
<p>At the <em>minimum</em>, Javadoc is present for every
<codeclass="prettyprint lang-java">public</code> class, and every
<codeclass="prettyprint lang-java">public</code> or
<codeclass="prettyprint lang-java">protected</code> member of such a class, with a few exceptions
noted below.</p><p>Other classes and members still have Javadoc <em>as needed</em>. Whenever an implementation
comment would be used to define the overall purpose or behavior of a class, method or field, that
comment is written as Javadoc instead. (It's more uniform, and more tool-friendly.)</p><aname="s7.3.1-javadoc-exception-self-explanatory"/>
<p>Javadoc is optional for "simple, obvious" methods like
<codeclass="prettyprint lang-java">getFoo</code>, in cases where there <em>really and truly</em> is
nothing else worthwhile to say but "Returns the foo".</p><pclass="note"><strong>Important:</strong> it is not appropriate to cite this exception to justify
omitting relevant information that a typical reader might need to know. For example, for a method
named <codeclass="prettyprint lang-java">getCanonicalName</code>, don't omit its documentation
(with the rationale that it would say only
<codeclass="badcode">/** Returns the canonical name. */</code>) if a typical reader may have no idea
what the term "canonical name" means!</p><aname="s7.3.2-javadoc-exception-overrides"/>
