1*01aa44aaSAndrew Rist /**************************************************************
2cdf0e10cSrcweir  *
3*01aa44aaSAndrew Rist  * Licensed to the Apache Software Foundation (ASF) under one
4*01aa44aaSAndrew Rist  * or more contributor license agreements.  See the NOTICE file
5*01aa44aaSAndrew Rist  * distributed with this work for additional information
6*01aa44aaSAndrew Rist  * regarding copyright ownership.  The ASF licenses this file
7*01aa44aaSAndrew Rist  * to you under the Apache License, Version 2.0 (the
8*01aa44aaSAndrew Rist  * "License"); you may not use this file except in compliance
9*01aa44aaSAndrew Rist  * with the License.  You may obtain a copy of the License at
10*01aa44aaSAndrew Rist  *
11*01aa44aaSAndrew Rist  *   http://www.apache.org/licenses/LICENSE-2.0
12*01aa44aaSAndrew Rist  *
13*01aa44aaSAndrew Rist  * Unless required by applicable law or agreed to in writing,
14*01aa44aaSAndrew Rist  * software distributed under the License is distributed on an
15*01aa44aaSAndrew Rist  * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16*01aa44aaSAndrew Rist  * KIND, either express or implied.  See the License for the
17*01aa44aaSAndrew Rist  * specific language governing permissions and limitations
18*01aa44aaSAndrew Rist  * under the License.
19*01aa44aaSAndrew Rist  *
20*01aa44aaSAndrew Rist  *************************************************************/
21*01aa44aaSAndrew Rist 
22*01aa44aaSAndrew Rist 
23cdf0e10cSrcweir #ifndef _SVTOOLS_SYNTAXHIGHLIGHT_HXX
24cdf0e10cSrcweir #define _SVTOOLS_SYNTAXHIGHLIGHT_HXX
25cdf0e10cSrcweir 
26cdf0e10cSrcweir #include <list>
27cdf0e10cSrcweir 
28cdf0e10cSrcweir #include <vos/macros.hxx>
29cdf0e10cSrcweir #include <vos/mutex.hxx>
30cdf0e10cSrcweir 
31cdf0e10cSrcweir #include <vcl/svapp.hxx>
32cdf0e10cSrcweir 
33cdf0e10cSrcweir #include <tools/stream.hxx>
34cdf0e10cSrcweir #include <tools/shl.hxx>
35cdf0e10cSrcweir 
36cdf0e10cSrcweir #include <svl/brdcst.hxx>
37cdf0e10cSrcweir #include <svtools/svtdllapi.h>
38cdf0e10cSrcweir 
39cdf0e10cSrcweir 
40cdf0e10cSrcweir // for the bsearch
41cdf0e10cSrcweir #ifdef WNT
42cdf0e10cSrcweir #define CDECL _cdecl
43cdf0e10cSrcweir #endif
44cdf0e10cSrcweir #if defined(UNX) || defined(OS2)
45cdf0e10cSrcweir #define CDECL
46cdf0e10cSrcweir #endif
47cdf0e10cSrcweir #ifdef UNX
48cdf0e10cSrcweir #include <sys/resource.h>
49cdf0e10cSrcweir #endif
50cdf0e10cSrcweir 
51cdf0e10cSrcweir #include <stdio.h>
52cdf0e10cSrcweir 
53cdf0e10cSrcweir #include <tools/string.hxx>
54cdf0e10cSrcweir #include <tools/gen.hxx>
55cdf0e10cSrcweir 
56cdf0e10cSrcweir 
57cdf0e10cSrcweir // Token-Typen TT_...
58cdf0e10cSrcweir enum TokenTypes
59cdf0e10cSrcweir {
60cdf0e10cSrcweir 	TT_UNKNOWN,
61cdf0e10cSrcweir 	TT_IDENTIFIER,
62cdf0e10cSrcweir 	TT_WHITESPACE,
63cdf0e10cSrcweir 	TT_NUMBER,
64cdf0e10cSrcweir 	TT_STRING,
65cdf0e10cSrcweir 	TT_EOL,
66cdf0e10cSrcweir 	TT_COMMENT,
67cdf0e10cSrcweir 	TT_ERROR,
68cdf0e10cSrcweir 	TT_OPERATOR,
69cdf0e10cSrcweir 	TT_KEYWORDS,
70cdf0e10cSrcweir 	TT_PARAMETER
71cdf0e10cSrcweir };
72cdf0e10cSrcweir 
73cdf0e10cSrcweir struct HighlightPortion { sal_uInt16 nBegin; sal_uInt16 nEnd; TokenTypes tokenType; };
74cdf0e10cSrcweir 
75cdf0e10cSrcweir 
76cdf0e10cSrcweir typedef std::vector<HighlightPortion> HighlightPortions;
77cdf0e10cSrcweir 
78cdf0e10cSrcweir /////////////////////////////////////////////////////////////////////////
79cdf0e10cSrcweir // Hilfsklasse zur Untersuchung von JavaScript-Modulen, zunaechst zum
80cdf0e10cSrcweir // Heraussuchen der Funktionen, spaeter auch zum Syntax-Highlighting verwenden
81cdf0e10cSrcweir 
82cdf0e10cSrcweir //	Flags fuer Zeichen-Eigenschaften
83cdf0e10cSrcweir #define CHAR_START_IDENTIFIER	0x0001
84cdf0e10cSrcweir #define CHAR_IN_IDENTIFIER		0x0002
85cdf0e10cSrcweir #define CHAR_START_NUMBER		0x0004
86cdf0e10cSrcweir #define CHAR_IN_NUMBER			0x0008
87cdf0e10cSrcweir #define CHAR_IN_HEX_NUMBER		0x0010
88cdf0e10cSrcweir #define CHAR_IN_OCT_NUMBER		0x0020
89cdf0e10cSrcweir #define CHAR_START_STRING		0x0040
90cdf0e10cSrcweir #define CHAR_OPERATOR			0x0080
91cdf0e10cSrcweir #define CHAR_SPACE				0x0100
92cdf0e10cSrcweir #define CHAR_EOL				0x0200
93cdf0e10cSrcweir 
94cdf0e10cSrcweir #define CHAR_EOF				0x00
95cdf0e10cSrcweir 
96cdf0e10cSrcweir 
97cdf0e10cSrcweir // Sprachmodus des HighLighters (spaeter eventuell feiner
98cdf0e10cSrcweir // differenzieren mit Keyword-Liste, C-Kommentar-Flag)
99cdf0e10cSrcweir enum HighlighterLanguage
100cdf0e10cSrcweir {
101cdf0e10cSrcweir 	HIGHLIGHT_BASIC,
102cdf0e10cSrcweir 	HIGHLIGHT_SQL
103cdf0e10cSrcweir };
104cdf0e10cSrcweir 
105cdf0e10cSrcweir class SimpleTokenizer_Impl
106cdf0e10cSrcweir {
107cdf0e10cSrcweir 	HighlighterLanguage aLanguage;
108cdf0e10cSrcweir 	// Zeichen-Info-Tabelle
109cdf0e10cSrcweir 	sal_uInt16 aCharTypeTab[256];
110cdf0e10cSrcweir 
111cdf0e10cSrcweir 	const sal_Unicode* mpStringBegin;
112cdf0e10cSrcweir 	const sal_Unicode* mpActualPos;
113cdf0e10cSrcweir 
114cdf0e10cSrcweir 	// Zeile und Spalte
115cdf0e10cSrcweir 	sal_uInt32 nLine;
116cdf0e10cSrcweir 	sal_uInt32 nCol;
117cdf0e10cSrcweir 
peekChar(void)118cdf0e10cSrcweir 	sal_Unicode peekChar( void )	{ return *mpActualPos; }
getChar(void)119cdf0e10cSrcweir 	sal_Unicode getChar( void )		{ nCol++; return *mpActualPos++; }
120cdf0e10cSrcweir 
121cdf0e10cSrcweir 	// Hilfsfunktion: Zeichen-Flag Testen
122cdf0e10cSrcweir 	sal_Bool testCharFlags( sal_Unicode c, sal_uInt16 nTestFlags );
123cdf0e10cSrcweir 
124cdf0e10cSrcweir 	// Neues Token holen, Leerstring == nix mehr da
125cdf0e10cSrcweir 	sal_Bool getNextToken( /*out*/TokenTypes& reType,
126cdf0e10cSrcweir 		/*out*/const sal_Unicode*& rpStartPos, /*out*/const sal_Unicode*& rpEndPos );
127cdf0e10cSrcweir 
128cdf0e10cSrcweir 	String getTokStr( /*out*/const sal_Unicode* pStartPos, /*out*/const sal_Unicode* pEndPos );
129cdf0e10cSrcweir 
130cdf0e10cSrcweir #ifdef DBG_UTIL
131cdf0e10cSrcweir 	// TEST: Token ausgeben
132cdf0e10cSrcweir 	String getFullTokenStr( /*out*/TokenTypes eType,
133cdf0e10cSrcweir 		/*out*/const sal_Unicode* pStartPos, /*out*/const sal_Unicode* pEndPos );
134cdf0e10cSrcweir #endif
135cdf0e10cSrcweir 
136cdf0e10cSrcweir 	const char** ppListKeyWords;
137cdf0e10cSrcweir 	sal_uInt16 nKeyWordCount;
138cdf0e10cSrcweir 
139cdf0e10cSrcweir public:
140cdf0e10cSrcweir 	SimpleTokenizer_Impl( HighlighterLanguage aLang = HIGHLIGHT_BASIC );
141cdf0e10cSrcweir 	~SimpleTokenizer_Impl( void );
142cdf0e10cSrcweir 
143cdf0e10cSrcweir 	sal_uInt16 parseLine( sal_uInt32 nLine, const String* aSource );
144cdf0e10cSrcweir 	void getHighlightPortions( sal_uInt32 nParseLine, const String& rLine,
145cdf0e10cSrcweir 													/*out*/HighlightPortions& portions );
146cdf0e10cSrcweir 	void setKeyWords( const char** ppKeyWords, sal_uInt16 nCount );
147cdf0e10cSrcweir };
148cdf0e10cSrcweir 
149cdf0e10cSrcweir 
150cdf0e10cSrcweir //*** SyntaxHighlighter-Klasse ***
151cdf0e10cSrcweir // Konzept: Der Highlighter wird ueber alle Aenderungen im Source
152cdf0e10cSrcweir // informiert (notifyChange) und liefert dem Aufrufer jeweils die
153cdf0e10cSrcweir // Information zurueck, welcher Zeilen-Bereich des Source-Codes
154cdf0e10cSrcweir // aufgrund dieser Aenderung neu gehighlighted werden muss.
155cdf0e10cSrcweir // Dazu merkt sich Highlighter intern fuer jede Zeile, ob dort
156cdf0e10cSrcweir // C-Kommentare beginnen oder enden.
157cdf0e10cSrcweir class SVT_DLLPUBLIC SyntaxHighlighter
158cdf0e10cSrcweir {
159cdf0e10cSrcweir 	HighlighterLanguage eLanguage;
160cdf0e10cSrcweir 	SimpleTokenizer_Impl* m_pSimpleTokenizer;
161cdf0e10cSrcweir 	char* m_pKeyWords;
162cdf0e10cSrcweir 	sal_uInt16 m_nKeyWordCount;
163cdf0e10cSrcweir 
164cdf0e10cSrcweir //	void initializeKeyWords( HighlighterLanguage eLanguage );
165cdf0e10cSrcweir 
166cdf0e10cSrcweir public:
167cdf0e10cSrcweir 	SyntaxHighlighter( void );
168cdf0e10cSrcweir 	~SyntaxHighlighter( void );
169cdf0e10cSrcweir 
170cdf0e10cSrcweir 	// HighLighter (neu) initialisieren, die Zeilen-Tabelle wird
171cdf0e10cSrcweir 	// dabei komplett geloescht, d.h. im Abschluss wird von einem
172cdf0e10cSrcweir 	// leeren Source ausgegangen. In notifyChange() kann dann
173cdf0e10cSrcweir 	// nur Zeile 0 angegeben werden.
174cdf0e10cSrcweir 	void initialize( HighlighterLanguage eLanguage_ );
175cdf0e10cSrcweir 
176cdf0e10cSrcweir 	const Range notifyChange( sal_uInt32 nLine, sal_Int32 nLineCountDifference,
177cdf0e10cSrcweir 								const String* pChangedLines, sal_uInt32 nArrayLength);
178cdf0e10cSrcweir 
179cdf0e10cSrcweir 	void getHighlightPortions( sal_uInt32 nLine, const String& rLine,
180cdf0e10cSrcweir 											HighlightPortions& pPortions );
181cdf0e10cSrcweir 
GetLanguage()182cdf0e10cSrcweir 	HighlighterLanguage GetLanguage() { return eLanguage;}
183cdf0e10cSrcweir };
184cdf0e10cSrcweir #endif
185