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.
7
8 #include <stdlib.h>
9 #include <string.h>
10 #include <stdio.h>
11 #include <stdarg.h>
12 #include <assert.h>
13 #include <ctype.h>
14
15 #include <string>
16 #include <vector>
17 #include <map>
18 #include <algorithm>
19
20 #include "ILexer.h"
21 #include "Scintilla.h"
22 #include "SciLexer.h"
23
24 #include "StringCopy.h"
25 #include "WordList.h"
26 #include "LexAccessor.h"
27 #include "Accessor.h"
28 #include "StyleContext.h"
29 #include "CharacterSet.h"
30 #include "CharacterCategory.h"
31 #include "LexerModule.h"
32 #include "OptionSet.h"
33 #include "SubStyles.h"
34 #include "DefaultLexer.h"
35
36 using namespace Scintilla;
37
38 namespace {
39 // Use an unnamed namespace to protect the functions and classes from name conflicts
40
41 /* Notes on f-strings: f-strings are strings prefixed with f (e.g. f'') that may
42 have arbitrary expressions in {}. The tokens in the expressions are lexed as if
43 they were outside of any string. Expressions may contain { and } characters as
44 long as there is a closing } for every {, may be 2+ lines in a triple quoted
45 string, and may have a formatting specifier following a ! or :, but both !
46 and : are valid inside of a bracketed expression and != is a valid
47 expression token even outside of a bracketed expression.
48
49 When in an f-string expression, the lexer keeps track of the state value of
50 the f-string and the nesting count for the expression (# of [, (, { seen - # of
51 }, ), ] seen). f-strings may be nested (e.g. f'{ a + f"{1+2}"') so a stack of
52 states and nesting counts is kept. If a f-string expression continues beyond
53 the end of a line, this stack is saved in a std::map that maps a line number to
54 the stack at the end of that line. std::vector is used for the stack.
55
56 The PEP for f-strings is at https://www.python.org/dev/peps/pep-0498/
57 */
58 struct SingleFStringExpState {
59 int state;
60 int nestingCount;
61 };
62
63 /* kwCDef, kwCTypeName only used for Cython */
64 enum kwType { kwOther, kwClass, kwDef, kwImport, kwCDef, kwCTypeName, kwCPDef };
65
66 enum literalsAllowed { litNone = 0, litU = 1, litB = 2, litF = 4 };
67
68 const int indicatorWhitespace = 1;
69
IsPyComment(Accessor & styler,Sci_Position pos,Sci_Position len)70 bool IsPyComment(Accessor &styler, Sci_Position pos, Sci_Position len) {
71 return len > 0 && styler[pos] == '#';
72 }
73
IsPyStringTypeChar(int ch,literalsAllowed allowed)74 bool IsPyStringTypeChar(int ch, literalsAllowed allowed) {
75 return
76 ((allowed & litB) && (ch == 'b' || ch == 'B')) ||
77 ((allowed & litU) && (ch == 'u' || ch == 'U')) ||
78 ((allowed & litF) && (ch == 'f' || ch == 'F'));
79 }
80
IsPyStringStart(int ch,int chNext,int chNext2,literalsAllowed allowed)81 bool IsPyStringStart(int ch, int chNext, int chNext2, literalsAllowed allowed) {
82 if (ch == '\'' || ch == '"')
83 return true;
84 if (IsPyStringTypeChar(ch, allowed)) {
85 if (chNext == '"' || chNext == '\'')
86 return true;
87 if ((chNext == 'r' || chNext == 'R') && (chNext2 == '"' || chNext2 == '\''))
88 return true;
89 }
90 if ((ch == 'r' || ch == 'R') && (chNext == '"' || chNext == '\''))
91 return true;
92
93 return false;
94 }
95
IsPyFStringState(int st)96 bool IsPyFStringState(int st) {
97 return ((st == SCE_P_FCHARACTER) || (st == SCE_P_FSTRING) ||
98 (st == SCE_P_FTRIPLE) || (st == SCE_P_FTRIPLEDOUBLE));
99 }
100
IsPySingleQuoteStringState(int st)101 bool IsPySingleQuoteStringState(int st) {
102 return ((st == SCE_P_CHARACTER) || (st == SCE_P_STRING) ||
103 (st == SCE_P_FCHARACTER) || (st == SCE_P_FSTRING));
104 }
105
IsPyTripleQuoteStringState(int st)106 bool IsPyTripleQuoteStringState(int st) {
107 return ((st == SCE_P_TRIPLE) || (st == SCE_P_TRIPLEDOUBLE) ||
108 (st == SCE_P_FTRIPLE) || (st == SCE_P_FTRIPLEDOUBLE));
109 }
110
GetPyStringQuoteChar(int st)111 char GetPyStringQuoteChar(int st) {
112 if ((st == SCE_P_CHARACTER) || (st == SCE_P_FCHARACTER) ||
113 (st == SCE_P_TRIPLE) || (st == SCE_P_FTRIPLE))
114 return '\'';
115 if ((st == SCE_P_STRING) || (st == SCE_P_FSTRING) ||
116 (st == SCE_P_TRIPLEDOUBLE) || (st == SCE_P_FTRIPLEDOUBLE))
117 return '"';
118
119 return '\0';
120 }
121
PushStateToStack(int state,std::vector<SingleFStringExpState> & stack,SingleFStringExpState * & currentFStringExp)122 void PushStateToStack(int state, std::vector<SingleFStringExpState> &stack, SingleFStringExpState *¤tFStringExp) {
123 SingleFStringExpState single = {state, 0};
124 stack.push_back(single);
125
126 currentFStringExp = &stack.back();
127 }
128
PopFromStateStack(std::vector<SingleFStringExpState> & stack,SingleFStringExpState * & currentFStringExp)129 int PopFromStateStack(std::vector<SingleFStringExpState> &stack, SingleFStringExpState *¤tFStringExp) {
130 int state = 0;
131
132 if (!stack.empty()) {
133 state = stack.back().state;
134 stack.pop_back();
135 }
136
137 if (stack.empty()) {
138 currentFStringExp = NULL;
139 } else {
140 currentFStringExp = &stack.back();
141 }
142
143 return state;
144 }
145
146 /* Return the state to use for the string starting at i; *nextIndex will be set to the first index following the quote(s) */
GetPyStringState(Accessor & styler,Sci_Position i,Sci_PositionU * nextIndex,literalsAllowed allowed)147 int GetPyStringState(Accessor &styler, Sci_Position i, Sci_PositionU *nextIndex, literalsAllowed allowed) {
148 char ch = styler.SafeGetCharAt(i);
149 char chNext = styler.SafeGetCharAt(i + 1);
150 const int firstIsF = (ch == 'f' || ch == 'F');
151
152 // Advance beyond r, u, or ur prefix (or r, b, or br in Python 2.7+ and r, f, or fr in Python 3.6+), but bail if there are any unexpected chars
153 if (ch == 'r' || ch == 'R') {
154 i++;
155 ch = styler.SafeGetCharAt(i);
156 chNext = styler.SafeGetCharAt(i + 1);
157 } else if (IsPyStringTypeChar(ch, allowed)) {
158 if (chNext == 'r' || chNext == 'R')
159 i += 2;
160 else
161 i += 1;
162 ch = styler.SafeGetCharAt(i);
163 chNext = styler.SafeGetCharAt(i + 1);
164 }
165
166 if (ch != '"' && ch != '\'') {
167 *nextIndex = i + 1;
168 return SCE_P_DEFAULT;
169 }
170
171 if (ch == chNext && ch == styler.SafeGetCharAt(i + 2)) {
172 *nextIndex = i + 3;
173
174 if (ch == '"')
175 return (firstIsF ? SCE_P_FTRIPLEDOUBLE : SCE_P_TRIPLEDOUBLE);
176 else
177 return (firstIsF ? SCE_P_FTRIPLE : SCE_P_TRIPLE);
178 } else {
179 *nextIndex = i + 1;
180
181 if (ch == '"')
182 return (firstIsF ? SCE_P_FSTRING : SCE_P_STRING);
183 else
184 return (firstIsF ? SCE_P_FCHARACTER : SCE_P_CHARACTER);
185 }
186 }
187
IsAWordChar(int ch,bool unicodeIdentifiers)188 inline bool IsAWordChar(int ch, bool unicodeIdentifiers) {
189 if (ch < 0x80)
190 return (isalnum(ch) || ch == '.' || ch == '_');
191
192 if (!unicodeIdentifiers)
193 return false;
194
195 // Python uses the XID_Continue set from unicode data
196 return IsXidContinue(ch);
197 }
198
IsAWordStart(int ch,bool unicodeIdentifiers)199 inline bool IsAWordStart(int ch, bool unicodeIdentifiers) {
200 if (ch < 0x80)
201 return (isalpha(ch) || ch == '_');
202
203 if (!unicodeIdentifiers)
204 return false;
205
206 // Python uses the XID_Start set from unicode data
207 return IsXidStart(ch);
208 }
209
IsFirstNonWhitespace(Sci_Position pos,Accessor & styler)210 static bool IsFirstNonWhitespace(Sci_Position pos, Accessor &styler) {
211 Sci_Position line = styler.GetLine(pos);
212 Sci_Position start_pos = styler.LineStart(line);
213 for (Sci_Position i = start_pos; i < pos; i++) {
214 const char ch = styler[i];
215 if (!(ch == ' ' || ch == '\t'))
216 return false;
217 }
218 return true;
219 }
220
221 // Options used for LexerPython
222 struct OptionsPython {
223 int whingeLevel;
224 bool base2or8Literals;
225 bool stringsU;
226 bool stringsB;
227 bool stringsF;
228 bool stringsOverNewline;
229 bool keywords2NoSubIdentifiers;
230 bool fold;
231 bool foldQuotes;
232 bool foldCompact;
233 bool unicodeIdentifiers;
234
OptionsPython__anond1800df60111::OptionsPython235 OptionsPython() {
236 whingeLevel = 0;
237 base2or8Literals = true;
238 stringsU = true;
239 stringsB = true;
240 stringsF = true;
241 stringsOverNewline = false;
242 keywords2NoSubIdentifiers = false;
243 fold = false;
244 foldQuotes = false;
245 foldCompact = false;
246 unicodeIdentifiers = true;
247 }
248
AllowedLiterals__anond1800df60111::OptionsPython249 literalsAllowed AllowedLiterals() const {
250 literalsAllowed allowedLiterals = stringsU ? litU : litNone;
251 if (stringsB)
252 allowedLiterals = static_cast<literalsAllowed>(allowedLiterals | litB);
253 if (stringsF)
254 allowedLiterals = static_cast<literalsAllowed>(allowedLiterals | litF);
255 return allowedLiterals;
256 }
257 };
258
259 static const char *const pythonWordListDesc[] = {
260 "Keywords",
261 "Highlighted identifiers",
262 0
263 };
264
265 struct OptionSetPython : public OptionSet<OptionsPython> {
OptionSetPython__anond1800df60111::OptionSetPython266 OptionSetPython() {
267 DefineProperty("tab.timmy.whinge.level", &OptionsPython::whingeLevel,
268 "For Python code, checks whether indenting is consistent. "
269 "The default, 0 turns off indentation checking, "
270 "1 checks whether each line is potentially inconsistent with the previous line, "
271 "2 checks whether any space characters occur before a tab character in the indentation, "
272 "3 checks whether any spaces are in the indentation, and "
273 "4 checks for any tab characters in the indentation. "
274 "1 is a good level to use.");
275
276 DefineProperty("lexer.python.literals.binary", &OptionsPython::base2or8Literals,
277 "Set to 0 to not recognise Python 3 binary and octal literals: 0b1011 0o712.");
278
279 DefineProperty("lexer.python.strings.u", &OptionsPython::stringsU,
280 "Set to 0 to not recognise Python Unicode literals u\"x\" as used before Python 3.");
281
282 DefineProperty("lexer.python.strings.b", &OptionsPython::stringsB,
283 "Set to 0 to not recognise Python 3 bytes literals b\"x\".");
284
285 DefineProperty("lexer.python.strings.f", &OptionsPython::stringsF,
286 "Set to 0 to not recognise Python 3.6 f-string literals f\"var={var}\".");
287
288 DefineProperty("lexer.python.strings.over.newline", &OptionsPython::stringsOverNewline,
289 "Set to 1 to allow strings to span newline characters.");
290
291 DefineProperty("lexer.python.keywords2.no.sub.identifiers", &OptionsPython::keywords2NoSubIdentifiers,
292 "When enabled, it will not style keywords2 items that are used as a sub-identifier. "
293 "Example: when set, will not highlight \"foo.open\" when \"open\" is a keywords2 item.");
294
295 DefineProperty("fold", &OptionsPython::fold);
296
297 DefineProperty("fold.quotes.python", &OptionsPython::foldQuotes,
298 "This option enables folding multi-line quoted strings when using the Python lexer.");
299
300 DefineProperty("fold.compact", &OptionsPython::foldCompact);
301
302 DefineProperty("lexer.python.unicode.identifiers", &OptionsPython::unicodeIdentifiers,
303 "Set to 0 to not recognise Python 3 unicode identifiers.");
304
305 DefineWordListSets(pythonWordListDesc);
306 }
307 };
308
309 const char styleSubable[] = { SCE_P_IDENTIFIER, 0 };
310
311 LexicalClass lexicalClasses[] = {
312 // Lexer Python SCLEX_PYTHON SCE_P_:
313 0, "SCE_P_DEFAULT", "default", "White space",
314 1, "SCE_P_COMMENTLINE", "comment line", "Comment",
315 2, "SCE_P_NUMBER", "literal numeric", "Number",
316 3, "SCE_P_STRING", "literal string", "String",
317 4, "SCE_P_CHARACTER", "literal string", "Single quoted string",
318 5, "SCE_P_WORD", "keyword", "Keyword",
319 6, "SCE_P_TRIPLE", "literal string", "Triple quotes",
320 7, "SCE_P_TRIPLEDOUBLE", "literal string", "Triple double quotes",
321 8, "SCE_P_CLASSNAME", "identifier", "Class name definition",
322 9, "SCE_P_DEFNAME", "identifier", "Function or method name definition",
323 10, "SCE_P_OPERATOR", "operator", "Operators",
324 11, "SCE_P_IDENTIFIER", "identifier", "Identifiers",
325 12, "SCE_P_COMMENTBLOCK", "comment", "Comment-blocks",
326 13, "SCE_P_STRINGEOL", "error literal string", "End of line where string is not closed",
327 14, "SCE_P_WORD2", "identifier", "Highlighted identifiers",
328 15, "SCE_P_DECORATOR", "preprocessor", "Decorators",
329 16, "SCE_P_FSTRING", "literal string interpolated", "F-String",
330 17, "SCE_P_FCHARACTER", "literal string interpolated", "Single quoted f-string",
331 18, "SCE_P_FTRIPLE", "literal string interpolated", "Triple quoted f-string",
332 19, "SCE_P_FTRIPLEDOUBLE", "literal string interpolated", "Triple double quoted f-string",
333 };
334
335 }
336
337 class LexerPython : public DefaultLexer {
338 WordList keywords;
339 WordList keywords2;
340 OptionsPython options;
341 OptionSetPython osPython;
342 enum { ssIdentifier };
343 SubStyles subStyles;
344 std::map<Sci_Position, std::vector<SingleFStringExpState> > ftripleStateAtEol;
345 public:
LexerPython()346 explicit LexerPython() :
347 DefaultLexer("python", SCLEX_PYTHON, lexicalClasses, ELEMENTS(lexicalClasses)),
348 subStyles(styleSubable, 0x80, 0x40, 0) {
349 }
~LexerPython()350 ~LexerPython() override {
351 }
Release()352 void SCI_METHOD Release() override {
353 delete this;
354 }
Version() const355 int SCI_METHOD Version() const override {
356 return lvRelease5;
357 }
PropertyNames()358 const char *SCI_METHOD PropertyNames() override {
359 return osPython.PropertyNames();
360 }
PropertyType(const char * name)361 int SCI_METHOD PropertyType(const char *name) override {
362 return osPython.PropertyType(name);
363 }
DescribeProperty(const char * name)364 const char *SCI_METHOD DescribeProperty(const char *name) override {
365 return osPython.DescribeProperty(name);
366 }
367 Sci_Position SCI_METHOD PropertySet(const char *key, const char *val) override;
PropertyGet(const char * key)368 const char * SCI_METHOD PropertyGet(const char *key) override {
369 return osPython.PropertyGet(key);
370 }
DescribeWordListSets()371 const char *SCI_METHOD DescribeWordListSets() override {
372 return osPython.DescribeWordListSets();
373 }
374 Sci_Position SCI_METHOD WordListSet(int n, const char *wl) override;
375 void SCI_METHOD Lex(Sci_PositionU startPos, Sci_Position length, int initStyle, IDocument *pAccess) override;
376 void SCI_METHOD Fold(Sci_PositionU startPos, Sci_Position length, int initStyle, IDocument *pAccess) override;
377
PrivateCall(int,void *)378 void *SCI_METHOD PrivateCall(int, void *) override {
379 return 0;
380 }
381
LineEndTypesSupported()382 int SCI_METHOD LineEndTypesSupported() override {
383 return SC_LINE_END_TYPE_UNICODE;
384 }
385
AllocateSubStyles(int styleBase,int numberStyles)386 int SCI_METHOD AllocateSubStyles(int styleBase, int numberStyles) override {
387 return subStyles.Allocate(styleBase, numberStyles);
388 }
SubStylesStart(int styleBase)389 int SCI_METHOD SubStylesStart(int styleBase) override {
390 return subStyles.Start(styleBase);
391 }
SubStylesLength(int styleBase)392 int SCI_METHOD SubStylesLength(int styleBase) override {
393 return subStyles.Length(styleBase);
394 }
StyleFromSubStyle(int subStyle)395 int SCI_METHOD StyleFromSubStyle(int subStyle) override {
396 const int styleBase = subStyles.BaseStyle(subStyle);
397 return styleBase;
398 }
PrimaryStyleFromStyle(int style)399 int SCI_METHOD PrimaryStyleFromStyle(int style) override {
400 return style;
401 }
FreeSubStyles()402 void SCI_METHOD FreeSubStyles() override {
403 subStyles.Free();
404 }
SetIdentifiers(int style,const char * identifiers)405 void SCI_METHOD SetIdentifiers(int style, const char *identifiers) override {
406 subStyles.SetIdentifiers(style, identifiers);
407 }
DistanceToSecondaryStyles()408 int SCI_METHOD DistanceToSecondaryStyles() override {
409 return 0;
410 }
GetSubStyleBases()411 const char *SCI_METHOD GetSubStyleBases() override {
412 return styleSubable;
413 }
414
LexerFactoryPython()415 static ILexer5 *LexerFactoryPython() {
416 return new LexerPython();
417 }
418
419 private:
420 void ProcessLineEnd(StyleContext &sc, std::vector<SingleFStringExpState> &fstringStateStack, SingleFStringExpState *¤tFStringExp, bool &inContinuedString);
421 };
422
PropertySet(const char * key,const char * val)423 Sci_Position SCI_METHOD LexerPython::PropertySet(const char *key, const char *val) {
424 if (osPython.PropertySet(&options, key, val)) {
425 return 0;
426 }
427 return -1;
428 }
429
WordListSet(int n,const char * wl)430 Sci_Position SCI_METHOD LexerPython::WordListSet(int n, const char *wl) {
431 WordList *wordListN = 0;
432 switch (n) {
433 case 0:
434 wordListN = &keywords;
435 break;
436 case 1:
437 wordListN = &keywords2;
438 break;
439 }
440 Sci_Position firstModification = -1;
441 if (wordListN) {
442 WordList wlNew;
443 wlNew.Set(wl);
444 if (*wordListN != wlNew) {
445 wordListN->Set(wl);
446 firstModification = 0;
447 }
448 }
449 return firstModification;
450 }
451
ProcessLineEnd(StyleContext & sc,std::vector<SingleFStringExpState> & fstringStateStack,SingleFStringExpState * & currentFStringExp,bool & inContinuedString)452 void LexerPython::ProcessLineEnd(StyleContext &sc, std::vector<SingleFStringExpState> &fstringStateStack, SingleFStringExpState *¤tFStringExp, bool &inContinuedString) {
453 long deepestSingleStateIndex = -1;
454 unsigned long i;
455
456 // Find the deepest single quote state because that string will end; no \ continuation in f-string
457 for (i = 0; i < fstringStateStack.size(); i++) {
458 if (IsPySingleQuoteStringState(fstringStateStack[i].state)) {
459 deepestSingleStateIndex = i;
460 break;
461 }
462 }
463
464 if (deepestSingleStateIndex != -1) {
465 sc.SetState(fstringStateStack[deepestSingleStateIndex].state);
466 while (fstringStateStack.size() > static_cast<unsigned long>(deepestSingleStateIndex)) {
467 PopFromStateStack(fstringStateStack, currentFStringExp);
468 }
469 }
470 if (!fstringStateStack.empty()) {
471 std::pair<Sci_Position, std::vector<SingleFStringExpState> > val;
472 val.first = sc.currentLine;
473 val.second = fstringStateStack;
474
475 ftripleStateAtEol.insert(val);
476 }
477
478 if ((sc.state == SCE_P_DEFAULT)
479 || IsPyTripleQuoteStringState(sc.state)) {
480 // Perform colourisation of white space and triple quoted strings at end of each line to allow
481 // tab marking to work inside white space and triple quoted strings
482 sc.SetState(sc.state);
483 }
484 if (IsPySingleQuoteStringState(sc.state)) {
485 if (inContinuedString || options.stringsOverNewline) {
486 inContinuedString = false;
487 } else {
488 sc.ChangeState(SCE_P_STRINGEOL);
489 sc.ForwardSetState(SCE_P_DEFAULT);
490 }
491 }
492 }
493
Lex(Sci_PositionU startPos,Sci_Position length,int initStyle,IDocument * pAccess)494 void SCI_METHOD LexerPython::Lex(Sci_PositionU startPos, Sci_Position length, int initStyle, IDocument *pAccess) {
495 Accessor styler(pAccess, NULL);
496
497 // Track whether in f-string expression; vector is used for a stack to
498 // handle nested f-strings such as f"""{f'''{f"{f'{1}'}"}'''}"""
499 std::vector<SingleFStringExpState> fstringStateStack;
500 SingleFStringExpState *currentFStringExp = NULL;
501
502 const Sci_Position endPos = startPos + length;
503
504 // Backtrack to previous line in case need to fix its tab whinging
505 Sci_Position lineCurrent = styler.GetLine(startPos);
506 if (startPos > 0) {
507 if (lineCurrent > 0) {
508 lineCurrent--;
509 // Look for backslash-continued lines
510 while (lineCurrent > 0) {
511 Sci_Position eolPos = styler.LineStart(lineCurrent) - 1;
512 const int eolStyle = styler.StyleAt(eolPos);
513 if (eolStyle == SCE_P_STRING
514 || eolStyle == SCE_P_CHARACTER
515 || eolStyle == SCE_P_STRINGEOL) {
516 lineCurrent -= 1;
517 } else {
518 break;
519 }
520 }
521 startPos = styler.LineStart(lineCurrent);
522 }
523 initStyle = startPos == 0 ? SCE_P_DEFAULT : styler.StyleAt(startPos - 1);
524 }
525
526 const literalsAllowed allowedLiterals = options.AllowedLiterals();
527
528 initStyle = initStyle & 31;
529 if (initStyle == SCE_P_STRINGEOL) {
530 initStyle = SCE_P_DEFAULT;
531 }
532
533 // Set up fstate stack from last line and remove any subsequent ftriple at eol states
534 std::map<Sci_Position, std::vector<SingleFStringExpState> >::iterator it;
535 it = ftripleStateAtEol.find(lineCurrent - 1);
536 if (it != ftripleStateAtEol.end() && !it->second.empty()) {
537 fstringStateStack = it->second;
538 currentFStringExp = &fstringStateStack.back();
539 }
540 it = ftripleStateAtEol.lower_bound(lineCurrent);
541 if (it != ftripleStateAtEol.end()) {
542 ftripleStateAtEol.erase(it, ftripleStateAtEol.end());
543 }
544
545 kwType kwLast = kwOther;
546 int spaceFlags = 0;
547 styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
548 bool base_n_number = false;
549
550 const WordClassifier &classifierIdentifiers = subStyles.Classifier(SCE_P_IDENTIFIER);
551
552 StyleContext sc(startPos, endPos - startPos, initStyle, styler);
553
554 bool indentGood = true;
555 Sci_Position startIndicator = sc.currentPos;
556 bool inContinuedString = false;
557
558 for (; sc.More(); sc.Forward()) {
559
560 if (sc.atLineStart) {
561 styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
562 indentGood = true;
563 if (options.whingeLevel == 1) {
564 indentGood = (spaceFlags & wsInconsistent) == 0;
565 } else if (options.whingeLevel == 2) {
566 indentGood = (spaceFlags & wsSpaceTab) == 0;
567 } else if (options.whingeLevel == 3) {
568 indentGood = (spaceFlags & wsSpace) == 0;
569 } else if (options.whingeLevel == 4) {
570 indentGood = (spaceFlags & wsTab) == 0;
571 }
572 if (!indentGood) {
573 styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 0);
574 startIndicator = sc.currentPos;
575 }
576 }
577
578 if (sc.atLineEnd) {
579 ProcessLineEnd(sc, fstringStateStack, currentFStringExp, inContinuedString);
580 lineCurrent++;
581 if (!sc.More())
582 break;
583 }
584
585 bool needEOLCheck = false;
586
587
588 if (sc.state == SCE_P_OPERATOR) {
589 kwLast = kwOther;
590 sc.SetState(SCE_P_DEFAULT);
591 } else if (sc.state == SCE_P_NUMBER) {
592 if (!IsAWordChar(sc.ch, false) &&
593 !(!base_n_number && ((sc.ch == '+' || sc.ch == '-') && (sc.chPrev == 'e' || sc.chPrev == 'E')))) {
594 sc.SetState(SCE_P_DEFAULT);
595 }
596 } else if (sc.state == SCE_P_IDENTIFIER) {
597 if ((sc.ch == '.') || (!IsAWordChar(sc.ch, options.unicodeIdentifiers))) {
598 char s[100];
599 sc.GetCurrent(s, sizeof(s));
600 int style = SCE_P_IDENTIFIER;
601 if ((kwLast == kwImport) && (strcmp(s, "as") == 0)) {
602 style = SCE_P_WORD;
603 } else if (keywords.InList(s)) {
604 style = SCE_P_WORD;
605 } else if (kwLast == kwClass) {
606 style = SCE_P_CLASSNAME;
607 } else if (kwLast == kwDef) {
608 style = SCE_P_DEFNAME;
609 } else if (kwLast == kwCDef || kwLast == kwCPDef) {
610 Sci_Position pos = sc.currentPos;
611 unsigned char ch = styler.SafeGetCharAt(pos, '\0');
612 while (ch != '\0') {
613 if (ch == '(') {
614 style = SCE_P_DEFNAME;
615 break;
616 } else if (ch == ':') {
617 style = SCE_P_CLASSNAME;
618 break;
619 } else if (ch == ' ' || ch == '\t' || ch == '\n' || ch == '\r') {
620 pos++;
621 ch = styler.SafeGetCharAt(pos, '\0');
622 } else {
623 break;
624 }
625 }
626 } else if (keywords2.InList(s)) {
627 if (options.keywords2NoSubIdentifiers) {
628 // We don't want to highlight keywords2
629 // that are used as a sub-identifier,
630 // i.e. not open in "foo.open".
631 Sci_Position pos = styler.GetStartSegment() - 1;
632 if (pos < 0 || (styler.SafeGetCharAt(pos, '\0') != '.'))
633 style = SCE_P_WORD2;
634 } else {
635 style = SCE_P_WORD2;
636 }
637 } else {
638 int subStyle = classifierIdentifiers.ValueFor(s);
639 if (subStyle >= 0) {
640 style = subStyle;
641 }
642 }
643 sc.ChangeState(style);
644 sc.SetState(SCE_P_DEFAULT);
645 if (style == SCE_P_WORD) {
646 if (0 == strcmp(s, "class"))
647 kwLast = kwClass;
648 else if (0 == strcmp(s, "def"))
649 kwLast = kwDef;
650 else if (0 == strcmp(s, "import"))
651 kwLast = kwImport;
652 else if (0 == strcmp(s, "cdef"))
653 kwLast = kwCDef;
654 else if (0 == strcmp(s, "cpdef"))
655 kwLast = kwCPDef;
656 else if (0 == strcmp(s, "cimport"))
657 kwLast = kwImport;
658 else if (kwLast != kwCDef && kwLast != kwCPDef)
659 kwLast = kwOther;
660 } else if (kwLast != kwCDef && kwLast != kwCPDef) {
661 kwLast = kwOther;
662 }
663 }
664 } else if ((sc.state == SCE_P_COMMENTLINE) || (sc.state == SCE_P_COMMENTBLOCK)) {
665 if (sc.ch == '\r' || sc.ch == '\n') {
666 sc.SetState(SCE_P_DEFAULT);
667 }
668 } else if (sc.state == SCE_P_DECORATOR) {
669 if (!IsAWordStart(sc.ch, options.unicodeIdentifiers)) {
670 sc.SetState(SCE_P_DEFAULT);
671 }
672 } else if (IsPySingleQuoteStringState(sc.state)) {
673 if (sc.ch == '\\') {
674 if ((sc.chNext == '\r') && (sc.GetRelative(2) == '\n')) {
675 sc.Forward();
676 }
677 if (sc.chNext == '\n' || sc.chNext == '\r') {
678 inContinuedString = true;
679 } else {
680 // Don't roll over the newline.
681 sc.Forward();
682 }
683 } else if (sc.ch == GetPyStringQuoteChar(sc.state)) {
684 sc.ForwardSetState(SCE_P_DEFAULT);
685 needEOLCheck = true;
686 }
687 } else if ((sc.state == SCE_P_TRIPLE) || (sc.state == SCE_P_FTRIPLE)) {
688 if (sc.ch == '\\') {
689 sc.Forward();
690 } else if (sc.Match(R"(''')")) {
691 sc.Forward();
692 sc.Forward();
693 sc.ForwardSetState(SCE_P_DEFAULT);
694 needEOLCheck = true;
695 }
696 } else if ((sc.state == SCE_P_TRIPLEDOUBLE) || (sc.state == SCE_P_FTRIPLEDOUBLE)) {
697 if (sc.ch == '\\') {
698 sc.Forward();
699 } else if (sc.Match(R"(""")")) {
700 sc.Forward();
701 sc.Forward();
702 sc.ForwardSetState(SCE_P_DEFAULT);
703 needEOLCheck = true;
704 }
705 }
706
707 // Note if used and not if else because string states also match
708 // some of the above clauses
709 if (IsPyFStringState(sc.state) && sc.ch == '{') {
710 if (sc.chNext == '{') {
711 sc.Forward();
712 } else {
713 PushStateToStack(sc.state, fstringStateStack, currentFStringExp);
714 sc.ForwardSetState(SCE_P_DEFAULT);
715 }
716 needEOLCheck = true;
717 }
718
719 // If in an f-string expression, check for the ending quote(s)
720 // and end f-string to handle syntactically incorrect cases like
721 // f'{' and f"""{"""
722 if (!fstringStateStack.empty() && (sc.ch == '\'' || sc.ch == '"')) {
723 long matching_stack_i = -1;
724 for (unsigned long stack_i = 0; stack_i < fstringStateStack.size() && matching_stack_i == -1; stack_i++) {
725 const int stack_state = fstringStateStack[stack_i].state;
726 const char quote = GetPyStringQuoteChar(stack_state);
727 if (sc.ch == quote) {
728 if (IsPySingleQuoteStringState(stack_state)) {
729 matching_stack_i = stack_i;
730 } else if (quote == '"' ? sc.Match(R"(""")") : sc.Match("'''")) {
731 matching_stack_i = stack_i;
732 }
733 }
734 }
735
736 if (matching_stack_i != -1) {
737 sc.SetState(fstringStateStack[matching_stack_i].state);
738 if (IsPyTripleQuoteStringState(fstringStateStack[matching_stack_i].state)) {
739 sc.Forward();
740 sc.Forward();
741 }
742 sc.ForwardSetState(SCE_P_DEFAULT);
743 needEOLCheck = true;
744
745 while (fstringStateStack.size() > static_cast<unsigned long>(matching_stack_i)) {
746 PopFromStateStack(fstringStateStack, currentFStringExp);
747 }
748 }
749 }
750 // End of code to find the end of a state
751
752 if (!indentGood && !IsASpaceOrTab(sc.ch)) {
753 styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 1);
754 startIndicator = sc.currentPos;
755 indentGood = true;
756 }
757
758 // One cdef or cpdef line, clear kwLast only at end of line
759 if ((kwLast == kwCDef || kwLast == kwCPDef) && sc.atLineEnd) {
760 kwLast = kwOther;
761 }
762
763 // State exit code may have moved on to end of line
764 if (needEOLCheck && sc.atLineEnd) {
765 ProcessLineEnd(sc, fstringStateStack, currentFStringExp, inContinuedString);
766 lineCurrent++;
767 styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
768 if (!sc.More())
769 break;
770 }
771
772 // If in f-string expression, check for }, :, ! to resume f-string state or update nesting count
773 if (currentFStringExp != NULL && !IsPySingleQuoteStringState(sc.state) && !IsPyTripleQuoteStringState(sc.state)) {
774 if (currentFStringExp->nestingCount == 0 && (sc.ch == '}' || sc.ch == ':' || (sc.ch == '!' && sc.chNext != '='))) {
775 sc.SetState(PopFromStateStack(fstringStateStack, currentFStringExp));
776 } else {
777 if (sc.ch == '{' || sc.ch == '[' || sc.ch == '(') {
778 currentFStringExp->nestingCount++;
779 } else if (sc.ch == '}' || sc.ch == ']' || sc.ch == ')') {
780 currentFStringExp->nestingCount--;
781 }
782 }
783 }
784
785 // Check for a new state starting character
786 if (sc.state == SCE_P_DEFAULT) {
787 if (IsADigit(sc.ch) || (sc.ch == '.' && IsADigit(sc.chNext))) {
788 if (sc.ch == '0' && (sc.chNext == 'x' || sc.chNext == 'X')) {
789 base_n_number = true;
790 sc.SetState(SCE_P_NUMBER);
791 } else if (sc.ch == '0' &&
792 (sc.chNext == 'o' || sc.chNext == 'O' || sc.chNext == 'b' || sc.chNext == 'B')) {
793 if (options.base2or8Literals) {
794 base_n_number = true;
795 sc.SetState(SCE_P_NUMBER);
796 } else {
797 sc.SetState(SCE_P_NUMBER);
798 sc.ForwardSetState(SCE_P_IDENTIFIER);
799 }
800 } else {
801 base_n_number = false;
802 sc.SetState(SCE_P_NUMBER);
803 }
804 } else if ((IsASCII(sc.ch) && isoperator(static_cast<char>(sc.ch))) || sc.ch == '`') {
805 sc.SetState(SCE_P_OPERATOR);
806 } else if (sc.ch == '#') {
807 sc.SetState(sc.chNext == '#' ? SCE_P_COMMENTBLOCK : SCE_P_COMMENTLINE);
808 } else if (sc.ch == '@') {
809 if (IsFirstNonWhitespace(sc.currentPos, styler))
810 sc.SetState(SCE_P_DECORATOR);
811 else
812 sc.SetState(SCE_P_OPERATOR);
813 } else if (IsPyStringStart(sc.ch, sc.chNext, sc.GetRelative(2), allowedLiterals)) {
814 Sci_PositionU nextIndex = 0;
815 sc.SetState(GetPyStringState(styler, sc.currentPos, &nextIndex, allowedLiterals));
816 while (nextIndex > (sc.currentPos + 1) && sc.More()) {
817 sc.Forward();
818 }
819 } else if (IsAWordStart(sc.ch, options.unicodeIdentifiers)) {
820 sc.SetState(SCE_P_IDENTIFIER);
821 }
822 }
823 }
824 styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 0);
825 sc.Complete();
826 }
827
IsCommentLine(Sci_Position line,Accessor & styler)828 static bool IsCommentLine(Sci_Position line, Accessor &styler) {
829 Sci_Position pos = styler.LineStart(line);
830 const Sci_Position eol_pos = styler.LineStart(line + 1) - 1;
831 for (Sci_Position i = pos; i < eol_pos; i++) {
832 const char ch = styler[i];
833 if (ch == '#')
834 return true;
835 else if (ch != ' ' && ch != '\t')
836 return false;
837 }
838 return false;
839 }
840
IsQuoteLine(Sci_Position line,const Accessor & styler)841 static bool IsQuoteLine(Sci_Position line, const Accessor &styler) {
842 const int style = styler.StyleAt(styler.LineStart(line)) & 31;
843 return IsPyTripleQuoteStringState(style);
844 }
845
846
Fold(Sci_PositionU startPos,Sci_Position length,int,IDocument * pAccess)847 void SCI_METHOD LexerPython::Fold(Sci_PositionU startPos, Sci_Position length, int /*initStyle - unused*/, IDocument *pAccess) {
848 if (!options.fold)
849 return;
850
851 Accessor styler(pAccess, NULL);
852
853 const Sci_Position maxPos = startPos + length;
854 const Sci_Position maxLines = (maxPos == styler.Length()) ? styler.GetLine(maxPos) : styler.GetLine(maxPos - 1); // Requested last line
855 const Sci_Position docLines = styler.GetLine(styler.Length()); // Available last line
856
857 // Backtrack to previous non-blank line so we can determine indent level
858 // for any white space lines (needed esp. within triple quoted strings)
859 // and so we can fix any preceding fold level (which is why we go back
860 // at least one line in all cases)
861 int spaceFlags = 0;
862 Sci_Position lineCurrent = styler.GetLine(startPos);
863 int indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL);
864 while (lineCurrent > 0) {
865 lineCurrent--;
866 indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL);
867 if (!(indentCurrent & SC_FOLDLEVELWHITEFLAG) &&
868 (!IsCommentLine(lineCurrent, styler)) &&
869 (!IsQuoteLine(lineCurrent, styler)))
870 break;
871 }
872 int indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK;
873
874 // Set up initial loop state
875 startPos = styler.LineStart(lineCurrent);
876 int prev_state = SCE_P_DEFAULT & 31;
877 if (lineCurrent >= 1)
878 prev_state = styler.StyleAt(startPos - 1) & 31;
879 int prevQuote = options.foldQuotes && IsPyTripleQuoteStringState(prev_state);
880
881 // Process all characters to end of requested range or end of any triple quote
882 //that hangs over the end of the range. Cap processing in all cases
883 // to end of document (in case of unclosed quote at end).
884 while ((lineCurrent <= docLines) && ((lineCurrent <= maxLines) || prevQuote)) {
885
886 // Gather info
887 int lev = indentCurrent;
888 Sci_Position lineNext = lineCurrent + 1;
889 int indentNext = indentCurrent;
890 int quote = false;
891 if (lineNext <= docLines) {
892 // Information about next line is only available if not at end of document
893 indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL);
894 Sci_Position lookAtPos = (styler.LineStart(lineNext) == styler.Length()) ? styler.Length() - 1 : styler.LineStart(lineNext);
895 const int style = styler.StyleAt(lookAtPos) & 31;
896 quote = options.foldQuotes && IsPyTripleQuoteStringState(style);
897 }
898 const int quote_start = (quote && !prevQuote);
899 const int quote_continue = (quote && prevQuote);
900 if (!quote || !prevQuote)
901 indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK;
902 if (quote)
903 indentNext = indentCurrentLevel;
904 if (indentNext & SC_FOLDLEVELWHITEFLAG)
905 indentNext = SC_FOLDLEVELWHITEFLAG | indentCurrentLevel;
906
907 if (quote_start) {
908 // Place fold point at start of triple quoted string
909 lev |= SC_FOLDLEVELHEADERFLAG;
910 } else if (quote_continue || prevQuote) {
911 // Add level to rest of lines in the string
912 lev = lev + 1;
913 }
914
915 // Skip past any blank lines for next indent level info; we skip also
916 // comments (all comments, not just those starting in column 0)
917 // which effectively folds them into surrounding code rather
918 // than screwing up folding. If comments end file, use the min
919 // comment indent as the level after
920
921 int minCommentLevel = indentCurrentLevel;
922 while (!quote &&
923 (lineNext < docLines) &&
924 ((indentNext & SC_FOLDLEVELWHITEFLAG) ||
925 (lineNext <= docLines && IsCommentLine(lineNext, styler)))) {
926
927 if (IsCommentLine(lineNext, styler) && indentNext < minCommentLevel) {
928 minCommentLevel = indentNext;
929 }
930
931 lineNext++;
932 indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL);
933 }
934
935 const int levelAfterComments = ((lineNext < docLines) ? indentNext & SC_FOLDLEVELNUMBERMASK : minCommentLevel);
936 const int levelBeforeComments = std::max(indentCurrentLevel, levelAfterComments);
937
938 // Now set all the indent levels on the lines we skipped
939 // Do this from end to start. Once we encounter one line
940 // which is indented more than the line after the end of
941 // the comment-block, use the level of the block before
942
943 Sci_Position skipLine = lineNext;
944 int skipLevel = levelAfterComments;
945
946 while (--skipLine > lineCurrent) {
947 const int skipLineIndent = styler.IndentAmount(skipLine, &spaceFlags, NULL);
948
949 if (options.foldCompact) {
950 if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments)
951 skipLevel = levelBeforeComments;
952
953 int whiteFlag = skipLineIndent & SC_FOLDLEVELWHITEFLAG;
954
955 styler.SetLevel(skipLine, skipLevel | whiteFlag);
956 } else {
957 if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments &&
958 !(skipLineIndent & SC_FOLDLEVELWHITEFLAG) &&
959 !IsCommentLine(skipLine, styler))
960 skipLevel = levelBeforeComments;
961
962 styler.SetLevel(skipLine, skipLevel);
963 }
964 }
965
966 // Set fold header on non-quote line
967 if (!quote && !(indentCurrent & SC_FOLDLEVELWHITEFLAG)) {
968 if ((indentCurrent & SC_FOLDLEVELNUMBERMASK) < (indentNext & SC_FOLDLEVELNUMBERMASK))
969 lev |= SC_FOLDLEVELHEADERFLAG;
970 }
971
972 // Keep track of triple quote state of previous line
973 prevQuote = quote;
974
975 // Set fold level for this line and move to next line
976 styler.SetLevel(lineCurrent, options.foldCompact ? lev : lev & ~SC_FOLDLEVELWHITEFLAG);
977 indentCurrent = indentNext;
978 lineCurrent = lineNext;
979 }
980
981 // NOTE: Cannot set level of last line here because indentCurrent doesn't have
982 // header flag set; the loop above is crafted to take care of this case!
983 //styler.SetLevel(lineCurrent, indentCurrent);
984 }
985
986 LexerModule lmPython(SCLEX_PYTHON, LexerPython::LexerFactoryPython, "python",
987 pythonWordListDesc);
988