1 // Scintilla source code edit control
2 /** @file LexCaml.cxx
3 ** Lexer for Objective Caml.
4 **/
5 // Copyright 2005 by Robert Roessler <robertr@rftp.com>
6 // The License.txt file describes the conditions under which this software may be distributed.
7 /* Release History
8 20050204 Initial release.
9 20050205 Quick compiler standards/"cleanliness" adjustment.
10 20050206 Added cast for IsLeadByte().
11 20050209 Changes to "external" build support.
12 20050306 Fix for 1st-char-in-doc "corner" case.
13 20050502 Fix for [harmless] one-past-the-end coloring.
14 20050515 Refined numeric token recognition logic.
15 20051125 Added 2nd "optional" keywords class.
16 20051129 Support "magic" (read-only) comments for RCaml.
17 20051204 Swtich to using StyleContext infrastructure.
18 */
19
20 #include <stdlib.h>
21 #include <string.h>
22 #include <ctype.h>
23 #include <stdio.h>
24 #include <stdarg.h>
25
26 #include "Platform.h"
27
28 #include "PropSet.h"
29 #include "Accessor.h"
30 #include "StyleContext.h"
31 #include "KeyWords.h"
32 #include "Scintilla.h"
33 #include "SciLexer.h"
34
35 // Since the Microsoft __iscsym[f] funcs are not ANSI...
iscaml(int c)36 inline int iscaml(int c) {return isalnum(c) || c == '_';}
iscamlf(int c)37 inline int iscamlf(int c) {return isalpha(c) || c == '_';}
iscamld(int c)38 inline int iscamld(int c) {return isdigit(c) || c == '_';}
39
40 static const int baseT[24] = {
41 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A - L */
42 0, 0, 8, 0, 0, 0, 0, 0, 0, 0, 0,16 /* M - X */
43 };
44
45 #ifdef BUILD_AS_EXTERNAL_LEXER
46 /*
47 (actually seems to work!)
48 */
49 #include "WindowAccessor.h"
50 #include "ExternalLexer.h"
51
52 #if PLAT_WIN
53 #include <windows.h>
54 #endif
55
56 static void ColouriseCamlDoc(
57 unsigned int startPos, int length,
58 int initStyle,
59 WordList *keywordlists[],
60 Accessor &styler);
61
62 static void FoldCamlDoc(
63 unsigned int startPos, int length,
64 int initStyle,
65 WordList *keywordlists[],
66 Accessor &styler);
67
68 static void InternalLexOrFold(int lexOrFold, unsigned int startPos, int length,
69 int initStyle, char *words[], WindowID window, char *props);
70
71 static const char* LexerName = "caml";
72
73 #ifdef TRACE
DebugPrintf(const char * format,...)74 void Platform::DebugPrintf(const char *format, ...) {
75 char buffer[2000];
76 va_list pArguments;
77 va_start(pArguments, format);
78 vsprintf(buffer,format,pArguments);
79 va_end(pArguments);
80 Platform::DebugDisplay(buffer);
81 }
82 #else
DebugPrintf(const char *,...)83 void Platform::DebugPrintf(const char *, ...) {
84 }
85 #endif
86
IsDBCSLeadByte(int codePage,char ch)87 bool Platform::IsDBCSLeadByte(int codePage, char ch) {
88 return ::IsDBCSLeadByteEx(codePage, ch) != 0;
89 }
90
SendScintilla(WindowID w,unsigned int msg,unsigned long wParam,long lParam)91 long Platform::SendScintilla(WindowID w, unsigned int msg, unsigned long wParam, long lParam) {
92 return ::SendMessage(reinterpret_cast<HWND>(w), msg, wParam, lParam);
93 }
94
SendScintillaPointer(WindowID w,unsigned int msg,unsigned long wParam,void * lParam)95 long Platform::SendScintillaPointer(WindowID w, unsigned int msg, unsigned long wParam, void *lParam) {
96 return ::SendMessage(reinterpret_cast<HWND>(w), msg, wParam,
97 reinterpret_cast<LPARAM>(lParam));
98 }
99
Fold(unsigned int lexer,unsigned int startPos,int length,int initStyle,char * words[],WindowID window,char * props)100 void EXT_LEXER_DECL Fold(unsigned int lexer, unsigned int startPos, int length,
101 int initStyle, char *words[], WindowID window, char *props)
102 {
103 // below useless evaluation(s) to supress "not used" warnings
104 lexer;
105 // build expected data structures and do the Fold
106 InternalLexOrFold(1, startPos, length, initStyle, words, window, props);
107
108 }
109
GetLexerCount()110 int EXT_LEXER_DECL GetLexerCount()
111 {
112 return 1; // just us [Objective] Caml lexers here!
113 }
114
GetLexerName(unsigned int Index,char * name,int buflength)115 void EXT_LEXER_DECL GetLexerName(unsigned int Index, char *name, int buflength)
116 {
117 // below useless evaluation(s) to supress "not used" warnings
118 Index;
119 // return as much of our lexer name as will fit (what's up with Index?)
120 if (buflength > 0) {
121 buflength--;
122 int n = strlen(LexerName);
123 if (n > buflength)
124 n = buflength;
125 memcpy(name, LexerName, n), name[n] = '\0';
126 }
127 }
128
Lex(unsigned int lexer,unsigned int startPos,int length,int initStyle,char * words[],WindowID window,char * props)129 void EXT_LEXER_DECL Lex(unsigned int lexer, unsigned int startPos, int length,
130 int initStyle, char *words[], WindowID window, char *props)
131 {
132 // below useless evaluation(s) to supress "not used" warnings
133 lexer;
134 // build expected data structures and do the Lex
135 InternalLexOrFold(0, startPos, length, initStyle, words, window, props);
136 }
137
InternalLexOrFold(int foldOrLex,unsigned int startPos,int length,int initStyle,char * words[],WindowID window,char * props)138 static void InternalLexOrFold(int foldOrLex, unsigned int startPos, int length,
139 int initStyle, char *words[], WindowID window, char *props)
140 {
141 // create and initialize a WindowAccessor (including contained PropSet)
142 PropSet ps;
143 ps.SetMultiple(props);
144 WindowAccessor wa(window, ps);
145 // create and initialize WordList(s)
146 int nWL = 0;
147 for (; words[nWL]; nWL++) ; // count # of WordList PTRs needed
148 WordList** wl = new WordList* [nWL + 1];// alloc WordList PTRs
149 int i = 0;
150 for (; i < nWL; i++) {
151 wl[i] = new WordList(); // (works or THROWS bad_alloc EXCEPTION)
152 wl[i]->Set(words[i]);
153 }
154 wl[i] = 0;
155 // call our "internal" folder/lexer (... then do Flush!)
156 if (foldOrLex)
157 FoldCamlDoc(startPos, length, initStyle, wl, wa);
158 else
159 ColouriseCamlDoc(startPos, length, initStyle, wl, wa);
160 wa.Flush();
161 // clean up before leaving
162 for (i = nWL - 1; i >= 0; i--)
163 delete wl[i];
164 delete [] wl;
165 }
166
167 static
168 #endif /* BUILD_AS_EXTERNAL_LEXER */
169
ColouriseCamlDoc(unsigned int startPos,int length,int initStyle,WordList * keywordlists[],Accessor & styler)170 void ColouriseCamlDoc(
171 unsigned int startPos, int length,
172 int initStyle,
173 WordList *keywordlists[],
174 Accessor &styler)
175 {
176 // initialize styler
177 StyleContext sc(startPos, length, initStyle, styler);
178 // set up [initial] state info (terminating states that shouldn't "bleed")
179 int nesting = 0;
180 if (sc.state < SCE_CAML_STRING)
181 sc.state = SCE_CAML_DEFAULT;
182 if (sc.state >= SCE_CAML_COMMENT)
183 nesting = (sc.state & 0x0f) - SCE_CAML_COMMENT;
184
185 int chBase = 0, chToken = 0, chLit = 0;
186 WordList& keywords = *keywordlists[0];
187 WordList& keywords2 = *keywordlists[1];
188 WordList& keywords3 = *keywordlists[2];
189 const int useMagic = styler.GetPropertyInt("lexer.caml.magic", 0);
190
191 // foreach char in range...
192 while (sc.More()) {
193 // set up [per-char] state info
194 int state2 = -1; // (ASSUME no state change)
195 int chColor = sc.currentPos - 1;// (ASSUME standard coloring range)
196 bool advance = true; // (ASSUME scanner "eats" 1 char)
197
198 // step state machine
199 switch (sc.state & 0x0f) {
200 case SCE_CAML_DEFAULT:
201 chToken = sc.currentPos; // save [possible] token start (JIC)
202 // it's wide open; what do we have?
203 if (iscamlf(sc.ch))
204 state2 = SCE_CAML_IDENTIFIER;
205 else if (sc.Match('`') && iscamlf(sc.chNext))
206 state2 = SCE_CAML_TAGNAME;
207 else if (sc.Match('#') && isdigit(sc.chNext))
208 state2 = SCE_CAML_LINENUM;
209 else if (isdigit(sc.ch)) {
210 state2 = SCE_CAML_NUMBER, chBase = 10;
211 if (sc.Match('0') && strchr("bBoOxX", sc.chNext))
212 chBase = baseT[tolower(sc.chNext) - 'a'], sc.Forward();
213 } else if (sc.Match('\'')) /* (char literal?) */
214 state2 = SCE_CAML_CHAR, chLit = 0;
215 else if (sc.Match('\"'))
216 state2 = SCE_CAML_STRING;
217 else if (sc.Match('(', '*'))
218 state2 = SCE_CAML_COMMENT,
219 sc.ch = ' ', // (make SURE "(*)" isn't seen as a closed comment)
220 sc.Forward();
221 else if (strchr("!?~" /* Caml "prefix-symbol" */
222 "=<>@^|&+-*/$%" /* Caml "infix-symbol" */
223 "()[]{};,:.#", sc.ch)) /* Caml "bracket" or ;,:.# */
224 state2 = SCE_CAML_OPERATOR;
225 break;
226
227 case SCE_CAML_IDENTIFIER:
228 // [try to] interpret as [additional] identifier char
229 if (!(iscaml(sc.ch) || sc.Match('\''))) {
230 const int n = sc.currentPos - chToken;
231 if (n < 24) {
232 // length is believable as keyword, [re-]construct token
233 char t[24];
234 for (int i = -n; i < 0; i++)
235 t[n + i] = static_cast<char>(sc.GetRelative(i));
236 t[n] = '\0';
237 // special-case "_" token as KEYWORD
238 if ((n == 1 && sc.chPrev == '_') || keywords.InList(t))
239 sc.ChangeState(SCE_CAML_KEYWORD);
240 else if (keywords2.InList(t))
241 sc.ChangeState(SCE_CAML_KEYWORD2);
242 else if (keywords3.InList(t))
243 sc.ChangeState(SCE_CAML_KEYWORD3);
244 }
245 state2 = SCE_CAML_DEFAULT, advance = false;
246 }
247 break;
248
249 case SCE_CAML_TAGNAME:
250 // [try to] interpret as [additional] tagname char
251 if (!(iscaml(sc.ch) || sc.Match('\'')))
252 state2 = SCE_CAML_DEFAULT, advance = false;
253 break;
254
255 /*case SCE_CAML_KEYWORD:
256 case SCE_CAML_KEYWORD2:
257 case SCE_CAML_KEYWORD3:
258 // [try to] interpret as [additional] keyword char
259 if (!iscaml(ch))
260 state2 = SCE_CAML_DEFAULT, advance = false;
261 break;*/
262
263 case SCE_CAML_LINENUM:
264 // [try to] interpret as [additional] linenum directive char
265 if (!isdigit(sc.ch))
266 state2 = SCE_CAML_DEFAULT, advance = false;
267 break;
268
269 case SCE_CAML_OPERATOR: {
270 // [try to] interpret as [additional] operator char
271 const char* o = 0;
272 if (iscaml(sc.ch) || isspace(sc.ch) /* ident or whitespace */
273 || (o = strchr(")]};,\'\"`#", sc.ch),o)/* "termination" chars */
274 || !strchr("!$%&*+-./:<=>?@^|~", sc.ch)/* "operator" chars */) {
275 // check for INCLUSIVE termination
276 if (o && strchr(")]};,", sc.ch)) {
277 if ((sc.Match(')') && sc.chPrev == '(')
278 || (sc.Match(']') && sc.chPrev == '['))
279 // special-case "()" and "[]" tokens as KEYWORDS
280 sc.ChangeState(SCE_CAML_KEYWORD);
281 chColor++;
282 } else
283 advance = false;
284 state2 = SCE_CAML_DEFAULT;
285 }
286 break;
287 }
288
289 case SCE_CAML_NUMBER:
290 // [try to] interpret as [additional] numeric literal char
291 // N.B. - improperly accepts "extra" digits in base 2 or 8 literals
292 if (iscamld(sc.ch) || IsADigit(sc.ch, chBase))
293 break;
294 // how about an integer suffix?
295 if ((sc.Match('l') || sc.Match('L') || sc.Match('n'))
296 && (iscamld(sc.chPrev) || IsADigit(sc.chPrev, chBase)))
297 break;
298 // or a floating-point literal?
299 if (chBase == 10) {
300 // with a decimal point?
301 if (sc.Match('.') && iscamld(sc.chPrev))
302 break;
303 // with an exponent? (I)
304 if ((sc.Match('e') || sc.Match('E'))
305 && (iscamld(sc.chPrev) || sc.chPrev == '.'))
306 break;
307 // with an exponent? (II)
308 if ((sc.Match('+') || sc.Match('-'))
309 && (sc.chPrev == 'e' || sc.chPrev == 'E'))
310 break;
311 }
312 // it looks like we have run out of number
313 state2 = SCE_CAML_DEFAULT, advance = false;
314 break;
315
316 case SCE_CAML_CHAR:
317 // [try to] interpret as [additional] char literal char
318 if (sc.Match('\\')) {
319 chLit = 1; // (definitely IS a char literal)
320 if (sc.chPrev == '\\')
321 sc.ch = ' '; // (so termination test isn't fooled)
322 // should we be terminating - one way or another?
323 } else if ((sc.Match('\'') && sc.chPrev != '\\') || sc.atLineEnd) {
324 state2 = SCE_CAML_DEFAULT;
325 if (sc.Match('\''))
326 chColor++;
327 else
328 sc.ChangeState(SCE_CAML_IDENTIFIER);
329 // ... maybe a char literal, maybe not
330 } else if (chLit < 1 && sc.currentPos - chToken >= 2)
331 sc.ChangeState(SCE_CAML_IDENTIFIER), advance = false;
332 break;
333
334 case SCE_CAML_STRING:
335 // [try to] interpret as [additional] string literal char
336 if (sc.Match('\\') && sc.chPrev == '\\')
337 sc.ch = ' '; // (so '\\' doesn't cause us trouble)
338 else if (sc.Match('\"') && sc.chPrev != '\\')
339 state2 = SCE_CAML_DEFAULT, chColor++;
340 break;
341
342 case SCE_CAML_COMMENT:
343 case SCE_CAML_COMMENT1:
344 case SCE_CAML_COMMENT2:
345 case SCE_CAML_COMMENT3:
346 // we're IN a comment - does this start a NESTED comment?
347 if (sc.Match('(', '*'))
348 state2 = sc.state + 1, chToken = sc.currentPos,
349 sc.ch = ' ', // (make SURE "(*)" isn't seen as a closed comment)
350 sc.Forward(), nesting++;
351 // [try to] interpret as [additional] comment char
352 else if (sc.Match(')') && sc.chPrev == '*') {
353 if (nesting)
354 state2 = (sc.state & 0x0f) - 1, chToken = 0, nesting--;
355 else
356 state2 = SCE_CAML_DEFAULT;
357 chColor++;
358 // enable "magic" (read-only) comment AS REQUIRED
359 } else if (useMagic && sc.currentPos - chToken == 4
360 && sc.Match('c') && sc.chPrev == 'r' && sc.GetRelative(-2) == '@')
361 sc.state |= 0x10; // (switch to read-only comment style)
362 break;
363 }
364
365 // handle state change and char coloring as required
366 if (state2 >= 0)
367 styler.ColourTo(chColor, sc.state), sc.ChangeState(state2);
368 // move to next char UNLESS re-scanning current char
369 if (advance)
370 sc.Forward();
371 }
372
373 // do any required terminal char coloring (JIC)
374 sc.Complete();
375 }
376
377 #ifdef BUILD_AS_EXTERNAL_LEXER
378 static
379 #endif /* BUILD_AS_EXTERNAL_LEXER */
FoldCamlDoc(unsigned int startPos,int length,int initStyle,WordList * keywordlists[],Accessor & styler)380 void FoldCamlDoc(
381 unsigned int startPos, int length,
382 int initStyle,
383 WordList *keywordlists[],
384 Accessor &styler)
385 {
386 // below useless evaluation(s) to supress "not used" warnings
387 startPos || length || initStyle || keywordlists[0] || styler.Length();
388 }
389
390 static const char * const camlWordListDesc[] = {
391 "Keywords", // primary Objective Caml keywords
392 "Keywords2", // "optional" keywords (typically from Pervasives)
393 "Keywords3", // "optional" keywords (typically typenames)
394 0
395 };
396
397 #ifndef BUILD_AS_EXTERNAL_LEXER
398 LexerModule lmCaml(SCLEX_CAML, ColouriseCamlDoc, "caml", FoldCamlDoc, camlWordListDesc);
399 #endif /* BUILD_AS_EXTERNAL_LEXER */
400