1/*
2 * Copyright (C) 2010 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 *      http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17#define LOG_TAG "Pattern"
18
19#include <stdlib.h>
20
21#include "JNIHelp.h"
22#include "JniConstants.h"
23#include "ScopedJavaUnicodeString.h"
24#include "jni.h"
25#include "unicode/parseerr.h"
26#include "unicode/regex.h"
27
28// ICU documentation: http://icu-project.org/apiref/icu4c/classRegexPattern.html
29
30static const char* regexDetailMessage(UErrorCode status) {
31    // These human-readable error messages were culled from "utypes.h", and then slightly tuned
32    // to make more sense in context.
33    // If we don't have a special-case, we'll just return the textual name of
34    // the enum value (such as U_REGEX_RULE_SYNTAX), which is better than nothing.
35    switch (status) {
36    case U_REGEX_INTERNAL_ERROR: return "An internal error was detected";
37    case U_REGEX_RULE_SYNTAX: return "Syntax error in regexp pattern";
38    case U_REGEX_INVALID_STATE: return "Matcher in invalid state for requested operation";
39    case U_REGEX_BAD_ESCAPE_SEQUENCE: return "Unrecognized backslash escape sequence in pattern";
40    case U_REGEX_PROPERTY_SYNTAX: return "Incorrect Unicode property";
41    case U_REGEX_UNIMPLEMENTED: return "Use of unimplemented feature";
42    case U_REGEX_MISMATCHED_PAREN: return "Incorrectly nested parentheses in regexp pattern";
43    case U_REGEX_NUMBER_TOO_BIG: return "Decimal number is too large";
44    case U_REGEX_BAD_INTERVAL: return "Error in {min,max} interval";
45    case U_REGEX_MAX_LT_MIN: return "In {min,max}, max is less than min";
46    case U_REGEX_INVALID_BACK_REF: return "Back-reference to a non-existent capture group";
47    case U_REGEX_INVALID_FLAG: return "Invalid value for match mode flags";
48    case U_REGEX_LOOK_BEHIND_LIMIT: return "Look-behind pattern matches must have a bounded maximum length";
49    case U_REGEX_SET_CONTAINS_STRING: return "Regular expressions cannot have UnicodeSets containing strings";
50    case U_REGEX_OCTAL_TOO_BIG: return "Octal character constants must be <= 0377.";
51    case U_REGEX_MISSING_CLOSE_BRACKET: return "Missing closing bracket in character class";
52    case U_REGEX_INVALID_RANGE: return "In a character range [x-y], x is greater than y";
53    case U_REGEX_STACK_OVERFLOW: return "Regular expression backtrack stack overflow";
54    case U_REGEX_TIME_OUT: return "Maximum allowed match time exceeded";
55    case U_REGEX_STOPPED_BY_CALLER: return "Matching operation aborted by user callback function";
56    default:
57        return u_errorName(status);
58    }
59}
60
61static void throwPatternSyntaxException(JNIEnv* env, UErrorCode status, jstring pattern, UParseError error) {
62    static jmethodID method = env->GetMethodID(JniConstants::patternSyntaxExceptionClass,
63            "<init>", "(Ljava/lang/String;Ljava/lang/String;I)V");
64    jstring message = env->NewStringUTF(regexDetailMessage(status));
65    jclass exceptionClass = JniConstants::patternSyntaxExceptionClass;
66    jobject exception = env->NewObject(exceptionClass, method, message, pattern, error.offset);
67    env->Throw(reinterpret_cast<jthrowable>(exception));
68}
69
70static void Pattern_free(void* addr) {
71    delete reinterpret_cast<icu::RegexPattern*>(addr);
72}
73
74static jlong Pattern_getNativeFinalizer(JNIEnv*, jclass) {
75    return reinterpret_cast<jlong>(&Pattern_free);
76}
77
78// Return a guess of the amount of native memory to be deallocated by a typical call to
79// Pattern_free().
80static jint Pattern_nativeSize(JNIEnv*, jclass) {
81    return 500;  // Very rough guess based on a quick look at the implementation.
82}
83
84static jlong Pattern_compileImpl(JNIEnv* env, jclass, jstring javaRegex, jint flags) {
85    flags |= UREGEX_ERROR_ON_UNKNOWN_ESCAPES;
86
87    UErrorCode status = U_ZERO_ERROR;
88    UParseError error;
89    error.offset = -1;
90
91    ScopedJavaUnicodeString regex(env, javaRegex);
92    if (!regex.valid()) {
93        return 0;
94    }
95    icu::UnicodeString& regexString(regex.unicodeString());
96    icu::RegexPattern* result = icu::RegexPattern::compile(regexString, flags, error, status);
97    if (!U_SUCCESS(status)) {
98        throwPatternSyntaxException(env, status, javaRegex, error);
99    }
100    return static_cast<jlong>(reinterpret_cast<uintptr_t>(result));
101}
102
103static JNINativeMethod gMethods[] = {
104    NATIVE_METHOD(Pattern, compileImpl, "(Ljava/lang/String;I)J"),
105    NATIVE_METHOD(Pattern, getNativeFinalizer, "()J"),
106    NATIVE_METHOD(Pattern, nativeSize, "()I"),
107};
108
109void register_java_util_regex_Pattern(JNIEnv* env) {
110    jniRegisterNativeMethods(env, "java/util/regex/Pattern", gMethods, NELEM(gMethods));
111}
112