1983d4c8aSAndrew Rist /**************************************************************
2cdf0e10cSrcweir *
3983d4c8aSAndrew Rist * Licensed to the Apache Software Foundation (ASF) under one
4983d4c8aSAndrew Rist * or more contributor license agreements. See the NOTICE file
5983d4c8aSAndrew Rist * distributed with this work for additional information
6983d4c8aSAndrew Rist * regarding copyright ownership. The ASF licenses this file
7983d4c8aSAndrew Rist * to you under the Apache License, Version 2.0 (the
8983d4c8aSAndrew Rist * "License"); you may not use this file except in compliance
9983d4c8aSAndrew Rist * with the License. You may obtain a copy of the License at
10983d4c8aSAndrew Rist *
11983d4c8aSAndrew Rist * http://www.apache.org/licenses/LICENSE-2.0
12983d4c8aSAndrew Rist *
13983d4c8aSAndrew Rist * Unless required by applicable law or agreed to in writing,
14983d4c8aSAndrew Rist * software distributed under the License is distributed on an
15983d4c8aSAndrew Rist * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16983d4c8aSAndrew Rist * KIND, either express or implied. See the License for the
17983d4c8aSAndrew Rist * specific language governing permissions and limitations
18983d4c8aSAndrew Rist * under the License.
19983d4c8aSAndrew Rist *
20983d4c8aSAndrew Rist *************************************************************/
21983d4c8aSAndrew Rist
22983d4c8aSAndrew Rist
23cdf0e10cSrcweir
24cdf0e10cSrcweir
25cdf0e10cSrcweir #ifndef TX3_WTRATREE_HXX
26cdf0e10cSrcweir #define TX3_WTRATREE_HXX
27cdf0e10cSrcweir
28cdf0e10cSrcweir // USED
29cdf0e10cSrcweir // Base Classes
30cdf0e10cSrcweir // Components
31cdf0e10cSrcweir // Parameters
32cdf0e10cSrcweir #include <tools/string.hxx>
33cdf0e10cSrcweir
34cdf0e10cSrcweir const INT16 C_NR_OF_WTT_RESULTS = 5;
35cdf0e10cSrcweir const INT16 C_NR_OF_POSSIBLE_CHARS = 256;
36cdf0e10cSrcweir
37cdf0e10cSrcweir
38cdf0e10cSrcweir typedef unsigned char u_char;
39cdf0e10cSrcweir typedef const char * constr;
40cdf0e10cSrcweir
41cdf0e10cSrcweir
42cdf0e10cSrcweir class WTT_Node;
43cdf0e10cSrcweir
44cdf0e10cSrcweir
45cdf0e10cSrcweir /** @task
46cdf0e10cSrcweir This class implements the functionality, that class WordTransformer
47cdf0e10cSrcweir offers.
48*a893be29SPedro Giffuni WordTransformer is dependent of this class, but NOT the other way!
49cdf0e10cSrcweir **/
50cdf0e10cSrcweir class WordTransTree
51cdf0e10cSrcweir {
52cdf0e10cSrcweir public:
53cdf0e10cSrcweir enum E_Result
54cdf0e10cSrcweir {
55cdf0e10cSrcweir OK = 0,
56cdf0e10cSrcweir HOTKEY_LOST,
57cdf0e10cSrcweir OUTPUT_OVERFLOW
58cdf0e10cSrcweir };
59cdf0e10cSrcweir
60cdf0e10cSrcweir
61cdf0e10cSrcweir // LIFECYCLE
62cdf0e10cSrcweir WordTransTree(
63cdf0e10cSrcweir CharSet i_nWorkingCharSet = RTL_TEXTENCODING_MS_1252);
64cdf0e10cSrcweir void SetCharSet(
65cdf0e10cSrcweir CharSet i_nWorkingCharSet);
66cdf0e10cSrcweir ~WordTransTree();
67cdf0e10cSrcweir
68cdf0e10cSrcweir void AddWordPair(
69cdf0e10cSrcweir const ByteString & i_sOldString,
70cdf0e10cSrcweir const ByteString & i_sReplaceString );
71cdf0e10cSrcweir
72cdf0e10cSrcweir // OPERATIONS
73cdf0e10cSrcweir void InitTransformation(
74cdf0e10cSrcweir const char * i_sInput, /// [!=0], a range of i_nInputLength must be valid memory for read.
75cdf0e10cSrcweir UINT32 i_nInputLength,
76cdf0e10cSrcweir UINT32 i_nOutputMaxLength = STRING_MAXLEN - 12 );
77cdf0e10cSrcweir E_Result TransformNextToken();
78cdf0e10cSrcweir
79cdf0e10cSrcweir // INQUIRY
80cdf0e10cSrcweir sal_Bool TextEndReached() const;
81cdf0e10cSrcweir const char * Output() const;
82cdf0e10cSrcweir
83cdf0e10cSrcweir // These 3 functions are valid between two calls of
84cdf0e10cSrcweir // TransformNextToken():
85cdf0e10cSrcweir E_Result CurResult() const;
86cdf0e10cSrcweir ByteString CurReplacedString() const;
87cdf0e10cSrcweir ByteString CurReplacingString() const;
88cdf0e10cSrcweir char CurHotkey() const;
89cdf0e10cSrcweir
90cdf0e10cSrcweir private:
91cdf0e10cSrcweir // SERVICE FUNCTONS
92cdf0e10cSrcweir UINT8 CalculateBranch(
93cdf0e10cSrcweir u_char i_cInputChar ) const;
94cdf0e10cSrcweir
95cdf0e10cSrcweir void Handle_Hotkey();
96cdf0e10cSrcweir void Handle_TokenToKeep();
97cdf0e10cSrcweir void Handle_TokenToTransform();
98cdf0e10cSrcweir
99cdf0e10cSrcweir // DATA
100cdf0e10cSrcweir // Fixed data
101cdf0e10cSrcweir const u_char * sInput;
102cdf0e10cSrcweir UINT32 nInputLength;
103cdf0e10cSrcweir const u_char * pInputEnd;
104cdf0e10cSrcweir
105cdf0e10cSrcweir u_char * sOutput; // DYN
106cdf0e10cSrcweir UINT32 nOutputMaxLength;
107cdf0e10cSrcweir
108cdf0e10cSrcweir WTT_Node * dpParsingTreeTop; // DYN
109cdf0e10cSrcweir WTT_Node * pUnknownAlpha;
110cdf0e10cSrcweir u_char cChar2Branch[C_NR_OF_POSSIBLE_CHARS];
111cdf0e10cSrcweir u_char c_AE, c_OE, c_UE, c_ae, c_oe, c_ue;
112cdf0e10cSrcweir
113cdf0e10cSrcweir // Working data
114cdf0e10cSrcweir const u_char * pInputCurTokenStart;
115cdf0e10cSrcweir const u_char * pInputPosition;
116cdf0e10cSrcweir u_char * pOutputPosition;
117cdf0e10cSrcweir WTT_Node * pCurParseNode;
118cdf0e10cSrcweir
119cdf0e10cSrcweir // Data which are valid only after a completed call to TransformNextToken()
120cdf0e10cSrcweir E_Result eCurResult;
121*a893be29SPedro Giffuni u_char cCurHotkey; // Letter which is used as hotkey
122cdf0e10cSrcweir u_char cCurHotkeySign; // Letter which is used to assign hotkey ('~'or '&') .
123cdf0e10cSrcweir };
124cdf0e10cSrcweir
125cdf0e10cSrcweir
126cdf0e10cSrcweir
127cdf0e10cSrcweir
128cdf0e10cSrcweir
129cdf0e10cSrcweir
130cdf0e10cSrcweir
131cdf0e10cSrcweir inline sal_Bool
TextEndReached() const132cdf0e10cSrcweir WordTransTree::TextEndReached() const
133cdf0e10cSrcweir { return pInputPosition == pInputEnd; }
134cdf0e10cSrcweir inline const char *
Output() const135cdf0e10cSrcweir WordTransTree::Output() const
136cdf0e10cSrcweir { return TextEndReached() ? (constr) sOutput : ""; }
137cdf0e10cSrcweir inline WordTransTree::E_Result
CurResult() const138cdf0e10cSrcweir WordTransTree::CurResult() const
139cdf0e10cSrcweir { return eCurResult; }
140cdf0e10cSrcweir inline ByteString
CurReplacedString() const141cdf0e10cSrcweir WordTransTree::CurReplacedString() const
142cdf0e10cSrcweir { return ByteString((constr) pInputCurTokenStart,pInputPosition-pInputCurTokenStart); }
143cdf0e10cSrcweir inline char
CurHotkey() const144cdf0e10cSrcweir WordTransTree::CurHotkey() const
145cdf0e10cSrcweir { return cCurHotkey; }
146cdf0e10cSrcweir inline UINT8
CalculateBranch(u_char i_cInputChar) const147cdf0e10cSrcweir WordTransTree::CalculateBranch(u_char i_cInputChar) const
148cdf0e10cSrcweir { return cChar2Branch[i_cInputChar]; }
149cdf0e10cSrcweir
150cdf0e10cSrcweir
151cdf0e10cSrcweir
152cdf0e10cSrcweir #endif
153cdf0e10cSrcweir
154cdf0e10cSrcweir
155cdf0e10cSrcweir
156