/src/icu/icu4c/source/common/uniquecharstr.h

Source (jump to first uncovered line)
// © 2020 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html

// uniquecharstr.h
// created: 2020sep01 Frank Yung-Fong Tang

#ifndef __UNIQUECHARSTR_H__
#define __UNIQUECHARSTR_H__

#include "charstr.h"
#include "uassert.h"
#include "uhash.h"
#include "cmemory.h"

U_NAMESPACE_BEGIN

/**
 * Stores NUL-terminated strings with duplicate elimination.
 * Checks for unique UTF-16 string pointers and converts to invariant characters.
 *
 * Intended to be stack-allocated. Add strings, get a unique number for each,
 * freeze the object, get a char * pointer for each string,
 * call orphanCharStrings() to capture the string storage, and let this object go out of scope.
 */
class UniqueCharStrings {
public:
    UniqueCharStrings(UErrorCode &errorCode) : strings(nullptr) {
        // Note: We hash on string contents but store stable char16_t * pointers.
        // If the strings are stored in resource bundles which should be built with
        // duplicate elimination, then we should be able to hash on just the pointer values.
        uhash_init(&map, uhash_hashUChars, uhash_compareUChars, uhash_compareLong, &errorCode);
        if (U_FAILURE(errorCode)) { return; }
        strings = new CharString();
        if (strings == nullptr) {
            errorCode = U_MEMORY_ALLOCATION_ERROR;
        }
    }
    ~UniqueCharStrings() {
        uhash_close(&map);
        delete strings;
    }

    /** Returns/orphans the CharString that contains all strings. */
    CharString *orphanCharStrings() {
        CharString *result = strings;
        strings = nullptr;
        return result;
    }

    /**
     * Adds a NUL-terminated string and returns a unique number for it.
     * The string must not change, nor move around in memory,
     * while this UniqueCharStrings is in use.
     *
     * Best used with string data in a stable storage, such as strings returned
     * by resource bundle functions.
     */
    int32_t add(const char16_t*p, UErrorCode &errorCode) {
        if (U_FAILURE(errorCode)) { return -1; }
        if (isFrozen) {
            errorCode = U_NO_WRITE_PERMISSION;
            return -1;
        }
        // The string points into the resource bundle.
        int32_t oldIndex = uhash_geti(&map, p);
        if (oldIndex != 0) {  // found duplicate
            return oldIndex;
        }
        // Explicit NUL terminator for the previous string.
        // The strings object is also terminated with one implicit NUL.
        strings->append(0, errorCode);
        int32_t newIndex = strings->length();
        strings->appendInvariantChars(p, u_strlen(p), errorCode);
        uhash_puti(&map, const_cast<char16_t *>(p), newIndex, &errorCode);
        return newIndex;
    }

    /**
     * Adds a unicode string by value and returns a unique number for it.
     */
    int32_t addByValue(UnicodeString s, UErrorCode &errorCode) {
        if (U_FAILURE(errorCode)) { return -1; }
        if (isFrozen) {
            errorCode = U_NO_WRITE_PERMISSION;
            return -1;
        }
        int32_t oldIndex = uhash_geti(&map, s.getTerminatedBuffer());
        if (oldIndex != 0) {  // found duplicate
            return oldIndex;
        }
        // We need to store the string content of the UnicodeString.
        UnicodeString *key = keyStore.create(s);
        if (key == nullptr) {
            errorCode = U_MEMORY_ALLOCATION_ERROR;
            return -1;
        }
        return add(key->getTerminatedBuffer(), errorCode);
    }

    void freeze() { isFrozen = true; }

    /**
     * Returns a string pointer for its unique number, if this object is frozen.
     * Otherwise nullptr.
     */
    const char *get(int32_t i) const {
        U_ASSERT(isFrozen);
        return isFrozen && i > 0 ? strings->data() + i : nullptr;
    }

private:
    UHashtable map;
    CharString *strings;
    MemoryPool<UnicodeString> keyStore;
    bool isFrozen = false;
};

U_NAMESPACE_END

#endif  // __UNIQUECHARSTR_H__

Line	Count	Source (jump to first uncovered line)
1		// © 2020 and later: Unicode, Inc. and others.
2		// License & terms of use: http://www.unicode.org/copyright.html
3
4		// uniquecharstr.h
5		// created: 2020sep01 Frank Yung-Fong Tang
6
7		#ifndef __UNIQUECHARSTR_H__
8		#define __UNIQUECHARSTR_H__
9
10		#include "charstr.h"
11		#include "uassert.h"
12		#include "uhash.h"
13		#include "cmemory.h"
14
15		U_NAMESPACE_BEGIN
16
17		/**
18		* Stores NUL-terminated strings with duplicate elimination.
19		* Checks for unique UTF-16 string pointers and converts to invariant characters.
20		*
21		* Intended to be stack-allocated. Add strings, get a unique number for each,
22		* freeze the object, get a char * pointer for each string,
23		* call orphanCharStrings() to capture the string storage, and let this object go out of scope.
24		*/
25		class UniqueCharStrings {
26		public:
27	0	UniqueCharStrings(UErrorCode &errorCode) : strings(nullptr) {
28		// Note: We hash on string contents but store stable char16_t * pointers.
29		// If the strings are stored in resource bundles which should be built with
30		// duplicate elimination, then we should be able to hash on just the pointer values.
31	0	uhash_init(&map, uhash_hashUChars, uhash_compareUChars, uhash_compareLong, &errorCode);
32	0	if (U_FAILURE(errorCode)) { return; }
33	0	strings = new CharString();
34	0	if (strings == nullptr) {
35	0	errorCode = U_MEMORY_ALLOCATION_ERROR;
36	0	}
37	0	}
38	0	~UniqueCharStrings() {
39	0	uhash_close(&map);
40	0	delete strings;
41	0	}
42
43		/** Returns/orphans the CharString that contains all strings. */
44	0	CharString *orphanCharStrings() {
45	0	CharString *result = strings;
46	0	strings = nullptr;
47	0	return result;
48	0	}
49
50		/**
51		* Adds a NUL-terminated string and returns a unique number for it.
52		* The string must not change, nor move around in memory,
53		* while this UniqueCharStrings is in use.
54		*
55		* Best used with string data in a stable storage, such as strings returned
56		* by resource bundle functions.
57		*/
58	0	int32_t add(const char16_t*p, UErrorCode &errorCode) {
59	0	if (U_FAILURE(errorCode)) { return -1; }
60	0	if (isFrozen) {
61	0	errorCode = U_NO_WRITE_PERMISSION;
62	0	return -1;
63	0	}
64		// The string points into the resource bundle.
65	0	int32_t oldIndex = uhash_geti(&map, p);
66	0	if (oldIndex != 0) { // found duplicate
67	0	return oldIndex;
68	0	}
69		// Explicit NUL terminator for the previous string.
70		// The strings object is also terminated with one implicit NUL.
71	0	strings->append(0, errorCode);
72	0	int32_t newIndex = strings->length();
73	0	strings->appendInvariantChars(p, u_strlen(p), errorCode);
74	0	uhash_puti(&map, const_cast<char16_t *>(p), newIndex, &errorCode);
75	0	return newIndex;
76	0	}
77
78		/**
79		* Adds a unicode string by value and returns a unique number for it.
80		*/
81	0	int32_t addByValue(UnicodeString s, UErrorCode &errorCode) {
82	0	if (U_FAILURE(errorCode)) { return -1; }
83	0	if (isFrozen) {
84	0	errorCode = U_NO_WRITE_PERMISSION;
85	0	return -1;
86	0	}
87	0	int32_t oldIndex = uhash_geti(&map, s.getTerminatedBuffer());
88	0	if (oldIndex != 0) { // found duplicate
89	0	return oldIndex;
90	0	}
91		// We need to store the string content of the UnicodeString.
92	0	UnicodeString *key = keyStore.create(s);
93	0	if (key == nullptr) {
94	0	errorCode = U_MEMORY_ALLOCATION_ERROR;
95	0	return -1;
96	0	}
97	0	return add(key->getTerminatedBuffer(), errorCode);
98	0	}
99
100	0	void freeze() { isFrozen = true; }
101
102		/**
103		* Returns a string pointer for its unique number, if this object is frozen.
104		* Otherwise nullptr.
105		*/
106	0	const char *get(int32_t i) const {
107	0	U_ASSERT(isFrozen);
108	0	return isFrozen && i > 0 ? strings->data() + i : nullptr;
109	0	}
110
111		private:
112		UHashtable map;
113		CharString *strings;
114		MemoryPool<UnicodeString> keyStore;
115		bool isFrozen = false;
116		};
117
118		U_NAMESPACE_END
119
120		#endif // __UNIQUECHARSTR_H__

Coverage Report

Created: 2025-06-13 06:34