8 Logging used to be done using Android's [android.util.Log]
9 (http://developer.android.com/reference/android/util/Log.html).
11 A wrapper on that is now available: org.chromium.base.Log. It is designed to
12 write logs as belonging to logical groups going beyond single classes, and to
13 make it easy to switch logging on or off for individual groups.
18 private static final String TAG = "YourModuleTag";
20 Log.i(TAG, "Logged INFO message.");
21 Log.d(TAG, "Some DEBUG info: %s", data);
27 I/cr_YourModuleTag: ( 999): Logged INFO message
28 D/cr_YourModuleTag: ( 999): [MyClass.java:42] Some DEBUG info: data.toString
31 Here, **TAG** will be a feature or package name, "MediaRemote" or "NFC" for
32 example. In most cases, the class name is not needed. It will be prepended by
33 the "cr_" prefix to make obvious which logs are coming from Chrome.
35 ### Verbose and Debug logs have special handling ###
37 * `Log.v` and `Log.d` Calls made using `org.chromium.base.Log` are stripped
38 out of production binaries using Proguard. There is no way to get those logs
41 * The file name and line number will be prepended to the log message.
42 For higher priority logs, those are not added for performance concerns.
44 ### An exception trace is printed when the exception is the last parameter ###
46 As with `java.util.Log`, putting a throwable as last parameter will dump the
47 corresponding stack trace:
50 Log.i(TAG, "An error happened: %s", e)
54 I/cr_YourModuleTag: ( 999): An error happened: This is the exception's message
55 I/cr_YourModuleTag: ( 999): java.lang.Exception: This is the exception's message
56 I/cr_YourModuleTag: ( 999): at foo.bar.MyClass.test(MyClass.java:42)
57 I/cr_YourModuleTag: ( 999): ...
60 Having the exception as last parameter doesn't prevent it from being used for
63 ## Logging Best Practices
65 ### Rule #1: Never log PII (Personal Identification Information):
67 This is a huge concern, because other applications can access the log and
68 extract a lot of data from your own by doing so. Even if JellyBean restricted
69 this, people are going to run your application on rooted devices and allow some
70 apps to access it. Also anyone with USB access to the device can use ADB to get
71 the full logcat and get the same data right now.
73 If you really need to print something , print a series of Xs instead
74 (e.g. "XXXXXX"), or print a truncated hash of the PII instead. Truncation is
75 required to make it harder for an attacker to recover the full data through
76 rainbow tables and similar methods.
78 Similarly, avoid dumping API keys, cookies, etc...
80 ### Rule #2: Do not build debug logs in production code:
82 The log methods are removed in release builds using Proguard. Because log
83 messages might not be written, the cost of creating them should also be avoided.
84 This can be done using three complementary ways:
86 #### Use string formatting instead of concatenations
90 Log.d(TAG, "I " + preference + " writing logs.");
93 Log.d(TAG, "I %s writing logs.", preference);
96 Proguard removes the method call itself, but doesn't do anything about the
97 arguments. The method's arguments will still be computed and provided as
98 input. The first call above will always lead to the creation of a
99 `StringBuilder` and a few concatenations, while the second just passes the
100 arguments and won't need that.
102 #### Guard expensive calls
104 Sometimes the values to log aren't readily available and need to be computed
105 specially. This should be avoided when logging is disabled.
108 static private final boolean DEBUG = false; // debug toggle.
111 Log.i(TAG, createThatExpensiveLogMessage(activity))
115 Because the variable is a `static final` that can be evaluated at compile
116 time, the Java compiler will optimize out all guarded calls from the
117 generated `.class` file. Changing it however requires editing each of the
118 files for which debug should be enabled and recompiling.
120 #### Annotate debug functions with the `@RemovableInRelease` annotation.
122 That annotation tells Proguard to assume that a given function has no side
123 effects, and is called only for its returned value. If this value is unused,
124 the call will be removed. If the function is not called at all, it will also
125 be removed. Since Proguard is already used to strip debug and verbose calls
126 out of release builds, this annotation allows it to have a deeper action by
127 removing also function calls used to generate the log call's arguments.
130 /* If that function is only used in Log.d calls, proguard should
131 * completely remove it from the release builds. */
133 private static String getSomeDebugLogString(Thing[] things) {
134 StringBuilder sb = new StringBuilder(
135 "Reporting " + thing.length + " things: ");
136 for (Thing thing : things) {
137 sb.append('\n').append(thing.id).append(' ').append(report.foo);
139 return sb.toString();
144 Log.d(TAG, getSomeDebugLogString(things)); /* The line is removed in
149 Again, this is useful only if the input to that function are variables
150 already available in the scope. The idea is to move computations,
151 concatenations, etc. to a place where that can be removed when not needed,
152 without invading the main function's logic. It can then have a similar
153 effect as guarding with a static final property that would be enabled in
154 Debug and disabled in Release.
156 ### Rule #3: Favor small log messages
158 This is still related to the global fixed-sized kernel buffer used to keep all
159 logs. Try to make your log information as terse as possible. This reduces the
160 risk of pushing interesting log data out of the buffer when something really
161 nasty happens. It's really better to have a single-line log message, than
162 several ones. I.e. don't use:
165 Log.GROUP.d(TAG, "field1 = %s", value1);
166 Log.GROUP.d(TAG, "field2 = %s", value2);
167 Log.GROUP.d(TAG, "field3 = %s", value3);
170 Instead, write this as:
173 Log.d(TAG, "field1 = %s, field2 = %s, field3 = %s", value1, value2, value3);
176 That doesn't seem to be much different if you count overall character counts,
177 but each independent log entry also implies a small, but non-trivial header, in
178 the kernel log buffer. And since every byte count, you can also try something
182 Log.d(TAG, "fields [%s,%s,%s]", value1, value2, value3);
187 Logcat allows filtering by specifying tags and the associated level:
190 adb logcat [TAG_EXPR:LEVEL]...
191 adb logcat cr_YourModuleTag:D *:S
194 This shows only logs having a level higher or equal to DEBUG for
195 `cr_YourModuleTag`, and SILENT (nothing is logged at this level or higher, so it
196 silences the tags) for everything else. You can persist a filter by setting an
197 environment variable:
200 export ANDROID_LOG_TAGS="cr_YourModuleTag:D *:S"
203 For more, see the [related page on developer.android.com]
204 (http://developer.android.com/tools/debugging/debugging-log.html#filteringOutput)