author | Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com> |
Fri, 14 May 2010 17:40:32 +0300 | |
changeset 24 | 71313a964664 |
parent 0 | 1fb32624e06b |
permissions | -rw-r--r-- |
0
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
1 |
/* |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
2 |
********************************************************************** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
3 |
* Copyright (c) 2000-2005, International Business Machines |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
4 |
* Corporation and others. All Rights Reserved. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
5 |
********************************************************************** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
6 |
* Date Name Description |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
7 |
* 02/04/00 aliu Creation. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
8 |
********************************************************************** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
9 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
10 |
#ifndef SYMTABLE_H |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
11 |
#define SYMTABLE_H |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
12 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
13 |
#include "unicode/utypes.h" |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
14 |
#include "unicode/uobject.h" |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
15 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
16 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
17 |
* \file |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
18 |
* \brief C++ API: An interface that defines both lookup protocol and parsing of |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
19 |
* symbolic names. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
20 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
21 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
22 |
U_NAMESPACE_BEGIN |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
23 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
24 |
class ParsePosition; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
25 |
class UnicodeFunctor; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
26 |
class UnicodeSet; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
27 |
class UnicodeString; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
28 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
29 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
30 |
* An interface that defines both lookup protocol and parsing of |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
31 |
* symbolic names. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
32 |
* |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
33 |
* <p>A symbol table maintains two kinds of mappings. The first is |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
34 |
* between symbolic names and their values. For example, if the |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
35 |
* variable with the name "start" is set to the value "alpha" |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
36 |
* (perhaps, though not necessarily, through an expression such as |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
37 |
* "$start=alpha"), then the call lookup("start") will return the |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
38 |
* char[] array ['a', 'l', 'p', 'h', 'a']. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
39 |
* |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
40 |
* <p>The second kind of mapping is between character values and |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
41 |
* UnicodeMatcher objects. This is used by RuleBasedTransliterator, |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
42 |
* which uses characters in the private use area to represent objects |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
43 |
* such as UnicodeSets. If U+E015 is mapped to the UnicodeSet [a-z], |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
44 |
* then lookupMatcher(0xE015) will return the UnicodeSet [a-z]. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
45 |
* |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
46 |
* <p>Finally, a symbol table defines parsing behavior for symbolic |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
47 |
* names. All symbolic names start with the SYMBOL_REF character. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
48 |
* When a parser encounters this character, it calls parseReference() |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
49 |
* with the position immediately following the SYMBOL_REF. The symbol |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
50 |
* table parses the name, if there is one, and returns it. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
51 |
* |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
52 |
* @stable ICU 2.8 |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
53 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
54 |
class U_COMMON_API SymbolTable /* not : public UObject because this is an interface/mixin class */ { |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
55 |
public: |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
56 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
57 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
58 |
* The character preceding a symbol reference name. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
59 |
* @stable ICU 2.8 |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
60 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
61 |
enum { SYMBOL_REF = 0x0024 /*$*/ }; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
62 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
63 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
64 |
* Destructor. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
65 |
* @stable ICU 2.8 |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
66 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
67 |
virtual ~SymbolTable(); |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
68 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
69 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
70 |
* Lookup the characters associated with this string and return it. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
71 |
* Return <tt>NULL</tt> if no such name exists. The resultant |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
72 |
* string may have length zero. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
73 |
* @param s the symbolic name to lookup |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
74 |
* @return a string containing the name's value, or <tt>NULL</tt> if |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
75 |
* there is no mapping for s. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
76 |
* @stable ICU 2.8 |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
77 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
78 |
virtual const UnicodeString* lookup(const UnicodeString& s) const = 0; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
79 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
80 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
81 |
* Lookup the UnicodeMatcher associated with the given character, and |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
82 |
* return it. Return <tt>NULL</tt> if not found. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
83 |
* @param ch a 32-bit code point from 0 to 0x10FFFF inclusive. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
84 |
* @return the UnicodeMatcher object represented by the given |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
85 |
* character, or NULL if there is no mapping for ch. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
86 |
* @stable ICU 2.8 |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
87 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
88 |
virtual const UnicodeFunctor* lookupMatcher(UChar32 ch) const = 0; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
89 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
90 |
/** |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
91 |
* Parse a symbol reference name from the given string, starting |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
92 |
* at the given position. If no valid symbol reference name is |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
93 |
* found, return the empty string and leave pos unchanged. That is, if the |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
94 |
* character at pos cannot start a name, or if pos is at or after |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
95 |
* text.length(), then return an empty string. This indicates an |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
96 |
* isolated SYMBOL_REF character. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
97 |
* @param text the text to parse for the name |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
98 |
* @param pos on entry, the index of the first character to parse. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
99 |
* This is the character following the SYMBOL_REF character. On |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
100 |
* exit, the index after the last parsed character. If the parse |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
101 |
* failed, pos is unchanged on exit. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
102 |
* @param limit the index after the last character to be parsed. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
103 |
* @return the parsed name, or an empty string if there is no |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
104 |
* valid symbolic name at the given position. |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
105 |
* @stable ICU 2.8 |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
106 |
*/ |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
107 |
virtual UnicodeString parseReference(const UnicodeString& text, |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
108 |
ParsePosition& pos, int32_t limit) const = 0; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
109 |
}; |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
110 |
U_NAMESPACE_END |
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
111 |
|
1fb32624e06b
Revision: 201003
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
112 |
#endif |