Slide 1
Slide 1 text
Regexp in Android and Java
Keishin Yokomaku @ Drivemode, Inc.
potatotips #22
Slide 2
Slide 2 text
@KeithYokoma
• Keishin Yokomaku at Drivemode, Inc.
• Work
• Android apps
• Android Training and its publication
• Like
• Bicycle, Photography, Tumblr and Motorsport
Slide 3
Slide 3 text
Pattern.compile()
Slide 4
Slide 4 text
Android is not Java
• Android and Java have different implementation respectively on regexp API.
• Same regexp, different result.
• Write once, not run anywhere (ˑ ‿ ˑ)
Slide 5
Slide 5 text
What’s the matter?
Slide 6
Slide 6 text
Test env vs. runtime
• If your tests are running on JVM(e.g. with Robolectric)…
• Some patterns pass the test, but won’t work at runtime.
• Some patterns work at runtime, but won’t pass the test.
Slide 7
Slide 7 text
ʉ\ʊ(π)ʊ/ʉ
Slide 8
Slide 8 text
Differences in detail
Slide 9
Slide 9 text
Supported flags
• Java
• All flags defined at Pattern are supported.
• Android
• Only CASE_INSENSITIVE, COMMENTS, DOTALL, LITERAL,
MULTILINE, UNICODE_CASE, UNIX_LINES are supported.
• If any other flags are set, RuntimeException will be thrown.
Slide 10
Slide 10 text
Android Pattern
public class Pattern {
private Pattern(String pattern, int flags) throws PatternSyntaxException {
if ((flags & CANON_EQ) != 0) {
throw new UnsupportedOperationException(“CANON_EQ flag not supported”);
}
int supportedFlags = CASE_INSENSITIVE | COMMENTS | DOTALL |
LITERAL | MULTILINE | UNICODE_CASE | UNIX_LINES;
if ((flags & ~supportedFlags) != 0) {
throw new IllegalArgumentException(“Unsupported flags: “ + (flags & ~supportedFlags));
}
this.pattern = pattern;
this.flags = flags;
compile();
}
}
Slide 11
Slide 11 text
Android Pattern
public class Pattern {
private Pattern(String pattern, int flags) throws PatternSyntaxException {
if ((flags & CANON_EQ) != 0) {
throw new UnsupportedOperationException(“CANON_EQ flag not supported”);
}
int supportedFlags = CASE_INSENSITIVE | COMMENTS | DOTALL |
LITERAL | MULTILINE | UNICODE_CASE | UNIX_LINES;
if ((flags & ~supportedFlags) != 0) {
throw new IllegalArgumentException(“Unsupported flags: “ + (flags & ~supportedFlags));
}
this.pattern = pattern;
this.flags = flags;
compile();
}
}
Slide 12
Slide 12 text
Android Pattern
public class Pattern {
private Pattern(String pattern, int flags) throws PatternSyntaxException {
if ((flags & CANON_EQ) != 0) {
throw new UnsupportedOperationException(“CANON_EQ flag not supported”);
}
int supportedFlags = CASE_INSENSITIVE | COMMENTS | DOTALL |
LITERAL | MULTILINE | UNICODE_CASE | UNIX_LINES;
if ((flags & ~supportedFlags) != 0) {
throw new IllegalArgumentException(“Unsupported flags: “ + (flags & ~supportedFlags));
}
this.pattern = pattern;
this.flags = flags;
compile();
}
}
Slide 13
Slide 13 text
Java Pattern
public class Pattern {
private void compile() {
if (has(CANON_EQ) && !has(LITERAL)) {
normalize();
} else {
normalizedPattern = pattern;
}
patternLength = normalizedPattern.length();
// Copy pattern to int array for convenience
// Use double zero to terminate pattern
temp = new int[patternLength + 2];
hasSupplementary = false;
int c, count = 0;
// Convert all chars into code points
for (int x = 0; x < patternLength; x += Character.charCount(c)) {
c = normalizedPattern.codePointAt(x);
if (isSupplementary(c)) {
hasSupplementary = true;
}
temp[count++] = c;
}
patternLength = count; // patternLength now in code points
// ……
}
}
Slide 14
Slide 14 text
Character class
• Java
• Matches only single byte characters
• Android
• Matches both single byte and multi byte characters.
• Details here: http://bit.ly/1R73wkM
Slide 15
Slide 15 text
Regular expression engines
• Java
• java.util.regex Engine
• Conform Unicode Technical Standard #18 Level1 and Release
2.1”Canonical Equivalents”.
• Android
• ICU(International Components for Unicode) Engine
• Conform Unicode Technical Standard #18 Level 1 and Default Word
Boundaries and Name Properties from Level2
Slide 16
Slide 16 text
Canonical Equivalents
• Canonically equivalent code point sequences are assumed to have the
same appearance and meaning when printed or displayed.
• e.g. “ü” and “u¨” are canonically equivalent
Slide 17
Slide 17 text
Android is not Java
ʉ\ʊ(π)ʊ/ʉ
Slide 18
Slide 18 text
Regexp in Android and Java
Keishin Yokomaku @ Drivemode, Inc.
potatotips #22