r5116 | eht16 | 2010-08-05 22:13:47 +0100 (Thu, 05 Aug 2010) | 3 lines
[geany-mirror.git] / scintilla / LexPython.cxx
blobb8921865c2ad8905b1072ec564145247d7fbbbbd
1 // Scintilla source code edit control
2 /** @file LexPython.cxx
3 ** Lexer for Python.
4 **/
5 // Copyright 1998-2002 by Neil Hodgson <neilh@scintilla.org>
6 // The License.txt file describes the conditions under which this software may be distributed.
8 #include <stdlib.h>
9 #include <string.h>
10 #include <ctype.h>
11 #include <stdio.h>
12 #include <stdarg.h>
14 #include "Platform.h"
16 #include "PropSet.h"
17 #include "Accessor.h"
18 #include "StyleContext.h"
19 #include "KeyWords.h"
20 #include "Scintilla.h"
21 #include "SciLexer.h"
23 #ifdef SCI_NAMESPACE
24 using namespace Scintilla;
25 #endif
27 /* kwCDef, kwCTypeName only used for Cython */
28 enum kwType { kwOther, kwClass, kwDef, kwImport, kwCDef, kwCTypeName };
30 static const int indicatorWhitespace = 1;
32 static bool IsPyComment(Accessor &styler, int pos, int len) {
33 return len > 0 && styler[pos] == '#';
36 enum literalsAllowed { litNone=0, litU=1, litB=2};
38 static bool IsPyStringTypeChar(int ch, literalsAllowed allowed) {
39 return
40 ((allowed & litB) && (ch == 'b' || ch == 'B')) ||
41 ((allowed & litU) && (ch == 'u' || ch == 'U'));
44 static bool IsPyStringStart(int ch, int chNext, int chNext2, literalsAllowed allowed) {
45 if (ch == '\'' || ch == '"')
46 return true;
47 if (IsPyStringTypeChar(ch, allowed)) {
48 if (chNext == '"' || chNext == '\'')
49 return true;
50 if ((chNext == 'r' || chNext == 'R') && (chNext2 == '"' || chNext2 == '\''))
51 return true;
53 if ((ch == 'r' || ch == 'R') && (chNext == '"' || chNext == '\''))
54 return true;
56 return false;
59 /* Return the state to use for the string starting at i; *nextIndex will be set to the first index following the quote(s) */
60 static int GetPyStringState(Accessor &styler, int i, unsigned int *nextIndex, literalsAllowed allowed) {
61 char ch = styler.SafeGetCharAt(i);
62 char chNext = styler.SafeGetCharAt(i + 1);
64 // Advance beyond r, u, or ur prefix (or r, b, or br in Python 3.0), but bail if there are any unexpected chars
65 if (ch == 'r' || ch == 'R') {
66 i++;
67 ch = styler.SafeGetCharAt(i);
68 chNext = styler.SafeGetCharAt(i + 1);
69 } else if (IsPyStringTypeChar(ch, allowed)) {
70 if (chNext == 'r' || chNext == 'R')
71 i += 2;
72 else
73 i += 1;
74 ch = styler.SafeGetCharAt(i);
75 chNext = styler.SafeGetCharAt(i + 1);
78 if (ch != '"' && ch != '\'') {
79 *nextIndex = i + 1;
80 return SCE_P_DEFAULT;
83 if (ch == chNext && ch == styler.SafeGetCharAt(i + 2)) {
84 *nextIndex = i + 3;
86 if (ch == '"')
87 return SCE_P_TRIPLEDOUBLE;
88 else
89 return SCE_P_TRIPLE;
90 } else {
91 *nextIndex = i + 1;
93 if (ch == '"')
94 return SCE_P_STRING;
95 else
96 return SCE_P_CHARACTER;
100 static inline bool IsAWordChar(int ch) {
101 return (ch < 0x80) && (isalnum(ch) || ch == '.' || ch == '_');
104 static inline bool IsAWordStart(int ch) {
105 return (ch < 0x80) && (isalnum(ch) || ch == '_');
108 static void ColourisePyDoc(unsigned int startPos, int length, int initStyle,
109 WordList *keywordlists[], Accessor &styler) {
111 int endPos = startPos + length;
113 // Backtrack to previous line in case need to fix its tab whinging
114 int lineCurrent = styler.GetLine(startPos);
115 if (startPos > 0) {
116 if (lineCurrent > 0) {
117 lineCurrent--;
118 // Look for backslash-continued lines
119 while (lineCurrent > 0) {
120 int eolPos = styler.LineStart(lineCurrent) - 1;
121 int eolStyle = styler.StyleAt(eolPos);
122 if (eolStyle == SCE_P_STRING
123 || eolStyle == SCE_P_CHARACTER
124 || eolStyle == SCE_P_STRINGEOL) {
125 lineCurrent -= 1;
126 } else {
127 break;
130 startPos = styler.LineStart(lineCurrent);
132 initStyle = startPos == 0 ? SCE_P_DEFAULT : styler.StyleAt(startPos - 1);
135 WordList &keywords = *keywordlists[0];
136 WordList &keywords2 = *keywordlists[1];
138 // property tab.timmy.whinge.level
139 // For Python code, checks whether indenting is consistent.
140 // The default, 0 turns off indentation checking,
141 // 1 checks whether each line is potentially inconsistent with the previous line,
142 // 2 checks whether any space characters occur before a tab character in the indentation,
143 // 3 checks whether any spaces are in the indentation, and
144 // 4 checks for any tab characters in the indentation.
145 // 1 is a good level to use.
146 const int whingeLevel = styler.GetPropertyInt("tab.timmy.whinge.level");
148 // property lexer.python.literals.binary
149 // Set to 0 to not recognise Python 3 binary and octal literals: 0b1011 0o712.
150 bool base2or8Literals = styler.GetPropertyInt("lexer.python.literals.binary", 1) != 0;
152 // property lexer.python.strings.u
153 // Set to 0 to not recognise Python Unicode literals u"x" as used before Python 3.
154 literalsAllowed allowedLiterals = (styler.GetPropertyInt("lexer.python.strings.u", 1)) ? litU : litNone;
156 // property lexer.python.strings.b
157 // Set to 0 to not recognise Python 3 bytes literals b"x".
158 if (styler.GetPropertyInt("lexer.python.strings.b", 1))
159 allowedLiterals = static_cast<literalsAllowed>(allowedLiterals | litB);
161 // property lexer.python.strings.over.newline
162 // Set to 1 to allow strings to span newline characters.
163 bool stringsOverNewline = styler.GetPropertyInt("lexer.python.strings.over.newline") != 0;
165 initStyle = initStyle & 31;
166 if (initStyle == SCE_P_STRINGEOL) {
167 initStyle = SCE_P_DEFAULT;
170 kwType kwLast = kwOther;
171 int spaceFlags = 0;
172 styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
173 bool base_n_number = false;
175 StyleContext sc(startPos, endPos - startPos, initStyle, styler);
177 bool indentGood = true;
178 int startIndicator = sc.currentPos;
179 bool inContinuedString = false;
181 for (; sc.More(); sc.Forward()) {
183 if (sc.atLineStart) {
184 styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
185 indentGood = true;
186 if (whingeLevel == 1) {
187 indentGood = (spaceFlags & wsInconsistent) == 0;
188 } else if (whingeLevel == 2) {
189 indentGood = (spaceFlags & wsSpaceTab) == 0;
190 } else if (whingeLevel == 3) {
191 indentGood = (spaceFlags & wsSpace) == 0;
192 } else if (whingeLevel == 4) {
193 indentGood = (spaceFlags & wsTab) == 0;
195 if (!indentGood) {
196 styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 0);
197 startIndicator = sc.currentPos;
201 if (sc.atLineEnd) {
202 if ((sc.state == SCE_P_DEFAULT) ||
203 (sc.state == SCE_P_TRIPLE) ||
204 (sc.state == SCE_P_TRIPLEDOUBLE)) {
205 // Perform colourisation of white space and triple quoted strings at end of each line to allow
206 // tab marking to work inside white space and triple quoted strings
207 sc.SetState(sc.state);
209 lineCurrent++;
210 if ((sc.state == SCE_P_STRING) || (sc.state == SCE_P_CHARACTER)) {
211 if (inContinuedString || stringsOverNewline) {
212 inContinuedString = false;
213 } else {
214 sc.ChangeState(SCE_P_STRINGEOL);
215 sc.ForwardSetState(SCE_P_DEFAULT);
218 if (!sc.More())
219 break;
222 bool needEOLCheck = false;
224 // Check for a state end
225 if (sc.state == SCE_P_OPERATOR) {
226 kwLast = kwOther;
227 sc.SetState(SCE_P_DEFAULT);
228 } else if (sc.state == SCE_P_NUMBER) {
229 if (!IsAWordChar(sc.ch) &&
230 !(!base_n_number && ((sc.ch == '+' || sc.ch == '-') && (sc.chPrev == 'e' || sc.chPrev == 'E')))) {
231 sc.SetState(SCE_P_DEFAULT);
233 } else if (sc.state == SCE_P_IDENTIFIER) {
234 if ((sc.ch == '.') || (!IsAWordChar(sc.ch))) {
235 char s[100];
236 sc.GetCurrent(s, sizeof(s));
237 int style = SCE_P_IDENTIFIER;
238 if ((kwLast == kwImport) && (strcmp(s, "as") == 0)) {
239 style = SCE_P_WORD;
240 } else if (keywords.InList(s)) {
241 style = SCE_P_WORD;
242 } else if (kwLast == kwClass) {
243 style = SCE_P_CLASSNAME;
244 } else if (kwLast == kwDef) {
245 style = SCE_P_DEFNAME;
246 } else if (kwLast == kwCDef) {
247 int pos = sc.currentPos;
248 unsigned char ch = styler.SafeGetCharAt(pos, '\0');
249 while (ch != '\0') {
250 if (ch == '(') {
251 style = SCE_P_DEFNAME;
252 break;
253 } else if (ch == ':') {
254 style = SCE_P_CLASSNAME;
255 break;
256 } else if (ch == ' ' || ch == '\t' || ch == '\n' || ch == '\r') {
257 pos++;
258 ch = styler.SafeGetCharAt(pos, '\0');
259 } else {
260 break;
263 } else if (keywords2.InList(s)) {
264 style = SCE_P_WORD2;
266 sc.ChangeState(style);
267 sc.SetState(SCE_P_DEFAULT);
268 if (style == SCE_P_WORD) {
269 if (0 == strcmp(s, "class"))
270 kwLast = kwClass;
271 else if (0 == strcmp(s, "def"))
272 kwLast = kwDef;
273 else if (0 == strcmp(s, "import"))
274 kwLast = kwImport;
275 else if (0 == strcmp(s, "cdef"))
276 kwLast = kwCDef;
277 else if (0 == strcmp(s, "cimport"))
278 kwLast = kwImport;
279 else if (kwLast != kwCDef)
280 kwLast = kwOther;
281 } else if (kwLast != kwCDef) {
282 kwLast = kwOther;
285 } else if ((sc.state == SCE_P_COMMENTLINE) || (sc.state == SCE_P_COMMENTBLOCK)) {
286 if (sc.ch == '\r' || sc.ch == '\n') {
287 sc.SetState(SCE_P_DEFAULT);
289 } else if (sc.state == SCE_P_DECORATOR) {
290 if (!IsAWordChar(sc.ch)) {
291 sc.SetState(SCE_P_DEFAULT);
293 } else if ((sc.state == SCE_P_STRING) || (sc.state == SCE_P_CHARACTER)) {
294 if (sc.ch == '\\') {
295 if ((sc.chNext == '\r') && (sc.GetRelative(2) == '\n')) {
296 sc.Forward();
298 if (sc.chNext == '\n' || sc.chNext == '\r') {
299 inContinuedString = true;
300 } else {
301 // Don't roll over the newline.
302 sc.Forward();
304 } else if ((sc.state == SCE_P_STRING) && (sc.ch == '\"')) {
305 sc.ForwardSetState(SCE_P_DEFAULT);
306 needEOLCheck = true;
307 } else if ((sc.state == SCE_P_CHARACTER) && (sc.ch == '\'')) {
308 sc.ForwardSetState(SCE_P_DEFAULT);
309 needEOLCheck = true;
311 } else if (sc.state == SCE_P_TRIPLE) {
312 if (sc.ch == '\\') {
313 sc.Forward();
314 } else if (sc.Match("\'\'\'")) {
315 sc.Forward();
316 sc.Forward();
317 sc.ForwardSetState(SCE_P_DEFAULT);
318 needEOLCheck = true;
320 } else if (sc.state == SCE_P_TRIPLEDOUBLE) {
321 if (sc.ch == '\\') {
322 sc.Forward();
323 } else if (sc.Match("\"\"\"")) {
324 sc.Forward();
325 sc.Forward();
326 sc.ForwardSetState(SCE_P_DEFAULT);
327 needEOLCheck = true;
331 if (!indentGood && !IsASpaceOrTab(sc.ch)) {
332 styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 1);
333 startIndicator = sc.currentPos;
334 indentGood = true;
337 // One cdef line, clear kwLast only at end of line
338 if (kwLast == kwCDef && sc.atLineEnd) {
339 kwLast = kwOther;
342 // State exit code may have moved on to end of line
343 if (needEOLCheck && sc.atLineEnd) {
344 lineCurrent++;
345 styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
346 if (!sc.More())
347 break;
350 // Check for a new state starting character
351 if (sc.state == SCE_P_DEFAULT) {
352 if (IsADigit(sc.ch) || (sc.ch == '.' && IsADigit(sc.chNext))) {
353 if (sc.ch == '0' && (sc.chNext == 'x' || sc.chNext == 'X')) {
354 base_n_number = true;
355 sc.SetState(SCE_P_NUMBER);
356 } else if (sc.ch == '0' &&
357 (sc.chNext == 'o' || sc.chNext == 'O' || sc.chNext == 'b' || sc.chNext == 'B')) {
358 if (base2or8Literals) {
359 base_n_number = true;
360 sc.SetState(SCE_P_NUMBER);
361 } else {
362 sc.SetState(SCE_P_NUMBER);
363 sc.ForwardSetState(SCE_P_IDENTIFIER);
365 } else {
366 base_n_number = false;
367 sc.SetState(SCE_P_NUMBER);
369 } else if ((isascii(sc.ch) && isoperator(static_cast<char>(sc.ch))) || sc.ch == '`') {
370 sc.SetState(SCE_P_OPERATOR);
371 } else if (sc.ch == '#') {
372 sc.SetState(sc.chNext == '#' ? SCE_P_COMMENTBLOCK : SCE_P_COMMENTLINE);
373 } else if (sc.ch == '@') {
374 sc.SetState(SCE_P_DECORATOR);
375 } else if (IsPyStringStart(sc.ch, sc.chNext, sc.GetRelative(2), allowedLiterals)) {
376 unsigned int nextIndex = 0;
377 sc.SetState(GetPyStringState(styler, sc.currentPos, &nextIndex, allowedLiterals));
378 while (nextIndex > (sc.currentPos + 1) && sc.More()) {
379 sc.Forward();
381 } else if (IsAWordStart(sc.ch)) {
382 sc.SetState(SCE_P_IDENTIFIER);
386 styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 0);
387 sc.Complete();
390 static bool IsCommentLine(int line, Accessor &styler) {
391 int pos = styler.LineStart(line);
392 int eol_pos = styler.LineStart(line + 1) - 1;
393 for (int i = pos; i < eol_pos; i++) {
394 char ch = styler[i];
395 if (ch == '#')
396 return true;
397 else if (ch != ' ' && ch != '\t')
398 return false;
400 return false;
403 static bool IsQuoteLine(int line, Accessor &styler) {
404 int style = styler.StyleAt(styler.LineStart(line)) & 31;
405 return ((style == SCE_P_TRIPLE) || (style == SCE_P_TRIPLEDOUBLE));
409 static void FoldPyDoc(unsigned int startPos, int length, int /*initStyle - unused*/,
410 WordList *[], Accessor &styler) {
411 const int maxPos = startPos + length;
412 const int maxLines = styler.GetLine(maxPos - 1); // Requested last line
413 const int docLines = styler.GetLine(styler.Length() - 1); // Available last line
415 // property fold.comment.python
416 // This option enables folding multi-line comments when using the Python lexer.
417 const bool foldComment = styler.GetPropertyInt("fold.comment.python") != 0;
419 // property fold.quotes.python
420 // This option enables folding multi-line quoted strings when using the Python lexer.
421 const bool foldQuotes = styler.GetPropertyInt("fold.quotes.python") != 0;
423 const bool foldCompact = styler.GetPropertyInt("fold.compact") != 0;
425 // Backtrack to previous non-blank line so we can determine indent level
426 // for any white space lines (needed esp. within triple quoted strings)
427 // and so we can fix any preceding fold level (which is why we go back
428 // at least one line in all cases)
429 int spaceFlags = 0;
430 int lineCurrent = styler.GetLine(startPos);
431 int indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL);
432 while (lineCurrent > 0) {
433 lineCurrent--;
434 indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL);
435 if (!(indentCurrent & SC_FOLDLEVELWHITEFLAG) &&
436 (!IsCommentLine(lineCurrent, styler)) &&
437 (!IsQuoteLine(lineCurrent, styler)))
438 break;
440 int indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK;
442 // Set up initial loop state
443 startPos = styler.LineStart(lineCurrent);
444 int prev_state = SCE_P_DEFAULT & 31;
445 if (lineCurrent >= 1)
446 prev_state = styler.StyleAt(startPos - 1) & 31;
447 int prevQuote = foldQuotes && ((prev_state == SCE_P_TRIPLE) || (prev_state == SCE_P_TRIPLEDOUBLE));
448 int prevComment = 0;
449 if (lineCurrent >= 1)
450 prevComment = foldComment && IsCommentLine(lineCurrent - 1, styler);
452 // Process all characters to end of requested range or end of any triple quote
453 // or comment that hangs over the end of the range. Cap processing in all cases
454 // to end of document (in case of unclosed quote or comment at end).
455 while ((lineCurrent <= docLines) && ((lineCurrent <= maxLines) || prevQuote || prevComment)) {
457 // Gather info
458 int lev = indentCurrent;
459 int lineNext = lineCurrent + 1;
460 int indentNext = indentCurrent;
461 int quote = false;
462 if (lineNext <= docLines) {
463 // Information about next line is only available if not at end of document
464 indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL);
465 int style = styler.StyleAt(styler.LineStart(lineNext)) & 31;
466 quote = foldQuotes && ((style == SCE_P_TRIPLE) || (style == SCE_P_TRIPLEDOUBLE));
468 const int quote_start = (quote && !prevQuote);
469 const int quote_continue = (quote && prevQuote);
470 const int comment = foldComment && IsCommentLine(lineCurrent, styler);
471 const int comment_start = (comment && !prevComment && (lineNext <= docLines) &&
472 IsCommentLine(lineNext, styler) && (lev > SC_FOLDLEVELBASE));
473 const int comment_continue = (comment && prevComment);
474 if ((!quote || !prevQuote) && !comment)
475 indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK;
476 if (quote)
477 indentNext = indentCurrentLevel;
478 if (indentNext & SC_FOLDLEVELWHITEFLAG)
479 indentNext = SC_FOLDLEVELWHITEFLAG | indentCurrentLevel;
481 if (quote_start) {
482 // Place fold point at start of triple quoted string
483 lev |= SC_FOLDLEVELHEADERFLAG;
484 } else if (quote_continue || prevQuote) {
485 // Add level to rest of lines in the string
486 lev = lev + 1;
487 } else if (comment_start) {
488 // Place fold point at start of a block of comments
489 lev |= SC_FOLDLEVELHEADERFLAG;
490 } else if (comment_continue) {
491 // Add level to rest of lines in the block
492 lev = lev + 1;
495 // Skip past any blank lines for next indent level info; we skip also
496 // comments (all comments, not just those starting in column 0)
497 // which effectively folds them into surrounding code rather
498 // than screwing up folding.
500 while (!quote &&
501 (lineNext < docLines) &&
502 ((indentNext & SC_FOLDLEVELWHITEFLAG) ||
503 (lineNext <= docLines && IsCommentLine(lineNext, styler)))) {
505 lineNext++;
506 indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL);
509 const int levelAfterComments = indentNext & SC_FOLDLEVELNUMBERMASK;
510 const int levelBeforeComments = Platform::Maximum(indentCurrentLevel,levelAfterComments);
512 // Now set all the indent levels on the lines we skipped
513 // Do this from end to start. Once we encounter one line
514 // which is indented more than the line after the end of
515 // the comment-block, use the level of the block before
517 int skipLine = lineNext;
518 int skipLevel = levelAfterComments;
520 while (--skipLine > lineCurrent) {
521 int skipLineIndent = styler.IndentAmount(skipLine, &spaceFlags, NULL);
523 if (foldCompact) {
524 if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments)
525 skipLevel = levelBeforeComments;
527 int whiteFlag = skipLineIndent & SC_FOLDLEVELWHITEFLAG;
529 styler.SetLevel(skipLine, skipLevel | whiteFlag);
530 } else {
531 if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments &&
532 !(skipLineIndent & SC_FOLDLEVELWHITEFLAG) &&
533 !IsCommentLine(skipLine, styler))
534 skipLevel = levelBeforeComments;
536 styler.SetLevel(skipLine, skipLevel);
540 // Set fold header on non-quote/non-comment line
541 if (!quote && !comment && !(indentCurrent & SC_FOLDLEVELWHITEFLAG)) {
542 if ((indentCurrent & SC_FOLDLEVELNUMBERMASK) < (indentNext & SC_FOLDLEVELNUMBERMASK))
543 lev |= SC_FOLDLEVELHEADERFLAG;
546 // Keep track of triple quote and block comment state of previous line
547 prevQuote = quote;
548 prevComment = comment_start || comment_continue;
550 // Set fold level for this line and move to next line
551 styler.SetLevel(lineCurrent, lev);
552 indentCurrent = indentNext;
553 lineCurrent = lineNext;
556 // NOTE: Cannot set level of last line here because indentCurrent doesn't have
557 // header flag set; the loop above is crafted to take care of this case!
558 //styler.SetLevel(lineCurrent, indentCurrent);
561 static const char *const pythonWordListDesc[] = {
562 "Keywords",
563 "Highlighted identifiers",
567 LexerModule lmPython(SCLEX_PYTHON, ColourisePyDoc, "python", FoldPyDoc,
568 pythonWordListDesc);