// Scintilla source code edit control
/** @file LexPython.cxx
 ** Lexer for Python.
 **/
// Copyright 1998-2002 by Neil Hodgson <neilh@scintilla.org>
// The License.txt file describes the conditions under which this software may be distributed.

#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <stdarg.h>
#include <assert.h>
#include <ctype.h>

#include <string>
#include <vector>
#include <map>
#include <algorithm>

#include "ILexer.h"
#include "Scintilla.h"
#include "SciLexer.h"

#include "StringCopy.h"
#include "WordList.h"
#include "LexAccessor.h"
#include "Accessor.h"
#include "StyleContext.h"
#include "CharacterSet.h"
#include "CharacterCategory.h"
#include "LexerModule.h"
#include "OptionSet.h"
#include "SubStyles.h"
#include "DefaultLexer.h"

using namespace Scintilla;

namespace {
// Use an unnamed namespace to protect the functions and classes from name conflicts

/* Notes on f-strings: f-strings are strings prefixed with f (e.g. f'') that may
   have arbitrary expressions in {}.  The tokens in the expressions are lexed as if
   they were outside of any string.  Expressions may contain { and } characters as
   long as there is a closing } for every {, may be 2+ lines in a triple quoted
   string, and may have a formatting specifier following a ! or :, but both !
   and : are valid inside of a bracketed expression and != is a valid
   expression token even outside of a bracketed expression.

   When in an f-string expression, the lexer keeps track of the state value of
   the f-string and the nesting count for the expression (# of [, (, { seen - # of
   }, ), ] seen).  f-strings may be nested (e.g. f'{ a + f"{1+2}"') so a stack of
   states and nesting counts is kept.  If a f-string expression continues beyond
   the end of a line, this stack is saved in a std::map that maps a line number to
   the stack at the end of that line.  std::vector is used for the stack.

   The PEP for f-strings is at https://www.python.org/dev/peps/pep-0498/
*/
struct SingleFStringExpState {
	int state;
	int nestingCount;
};

/* kwCDef, kwCTypeName only used for Cython */
enum kwType { kwOther, kwClass, kwDef, kwImport, kwCDef, kwCTypeName, kwCPDef };

enum literalsAllowed { litNone = 0, litU = 1, litB = 2, litF = 4 };

const int indicatorWhitespace = 1;

bool IsPyComment(Accessor &styler, Sci_Position pos, Sci_Position len) {
	return len > 0 && styler[pos] == '#';
}

bool IsPyStringTypeChar(int ch, literalsAllowed allowed) {
	return
		((allowed & litB) && (ch == 'b' || ch == 'B')) ||
		((allowed & litU) && (ch == 'u' || ch == 'U')) ||
		((allowed & litF) && (ch == 'f' || ch == 'F'));
}

bool IsPyStringStart(int ch, int chNext, int chNext2, literalsAllowed allowed) {
	if (ch == '\'' || ch == '"')
		return true;
	if (IsPyStringTypeChar(ch, allowed)) {
		if (chNext == '"' || chNext == '\'')
			return true;
		if ((chNext == 'r' || chNext == 'R') && (chNext2 == '"' || chNext2 == '\''))
			return true;
	}
	if ((ch == 'r' || ch == 'R') && (chNext == '"' || chNext == '\''))
		return true;

	return false;
}

bool IsPyFStringState(int st) {
	return ((st == SCE_P_FCHARACTER) || (st == SCE_P_FSTRING) ||
		(st == SCE_P_FTRIPLE) || (st == SCE_P_FTRIPLEDOUBLE));
}

bool IsPySingleQuoteStringState(int st) {
	return ((st == SCE_P_CHARACTER) || (st == SCE_P_STRING) ||
		(st == SCE_P_FCHARACTER) || (st == SCE_P_FSTRING));
}

bool IsPyTripleQuoteStringState(int st) {
	return ((st == SCE_P_TRIPLE) || (st == SCE_P_TRIPLEDOUBLE) ||
		(st == SCE_P_FTRIPLE) || (st == SCE_P_FTRIPLEDOUBLE));
}

char GetPyStringQuoteChar(int st) {
	if ((st == SCE_P_CHARACTER) || (st == SCE_P_FCHARACTER) ||
			(st == SCE_P_TRIPLE) || (st == SCE_P_FTRIPLE))
		return '\'';
	if ((st == SCE_P_STRING) || (st == SCE_P_FSTRING) ||
			(st == SCE_P_TRIPLEDOUBLE) || (st == SCE_P_FTRIPLEDOUBLE))
		return '"';

	return '\0';
}

void PushStateToStack(int state, std::vector<SingleFStringExpState> &stack, SingleFStringExpState *&currentFStringExp) {
	SingleFStringExpState single = {state, 0};
	stack.push_back(single);

	currentFStringExp = &stack.back();
}

int PopFromStateStack(std::vector<SingleFStringExpState> &stack, SingleFStringExpState *&currentFStringExp) {
	int state = 0;

	if (!stack.empty()) {
		state = stack.back().state;
		stack.pop_back();
	}

	if (stack.empty()) {
		currentFStringExp = NULL;
	} else {
		currentFStringExp = &stack.back();
	}

	return state;
}

/* Return the state to use for the string starting at i; *nextIndex will be set to the first index following the quote(s) */
int GetPyStringState(Accessor &styler, Sci_Position i, Sci_PositionU *nextIndex, literalsAllowed allowed) {
	char ch = styler.SafeGetCharAt(i);
	char chNext = styler.SafeGetCharAt(i + 1);
	const int firstIsF = (ch == 'f' || ch == 'F');

	// 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
	if (ch == 'r' || ch == 'R') {
		i++;
		ch = styler.SafeGetCharAt(i);
		chNext = styler.SafeGetCharAt(i + 1);
	} else if (IsPyStringTypeChar(ch, allowed)) {
		if (chNext == 'r' || chNext == 'R')
			i += 2;
		else
			i += 1;
		ch = styler.SafeGetCharAt(i);
		chNext = styler.SafeGetCharAt(i + 1);
	}

	if (ch != '"' && ch != '\'') {
		*nextIndex = i + 1;
		return SCE_P_DEFAULT;
	}

	if (ch == chNext && ch == styler.SafeGetCharAt(i + 2)) {
		*nextIndex = i + 3;

		if (ch == '"')
			return (firstIsF ? SCE_P_FTRIPLEDOUBLE : SCE_P_TRIPLEDOUBLE);
		else
			return (firstIsF ? SCE_P_FTRIPLE : SCE_P_TRIPLE);
	} else {
		*nextIndex = i + 1;

		if (ch == '"')
			return (firstIsF ? SCE_P_FSTRING : SCE_P_STRING);
		else
			return (firstIsF ? SCE_P_FCHARACTER : SCE_P_CHARACTER);
	}
}

inline bool IsAWordChar(int ch, bool unicodeIdentifiers) {
	if (ch < 0x80)
		return (isalnum(ch) || ch == '.' || ch == '_');

	if (!unicodeIdentifiers)
		return false;

	// Python uses the XID_Continue set from unicode data
	return IsXidContinue(ch);
}

inline bool IsAWordStart(int ch, bool unicodeIdentifiers) {
	if (ch < 0x80)
		return (isalpha(ch) || ch == '_');

	if (!unicodeIdentifiers)
		return false;

	// Python uses the XID_Start set from unicode data
	return IsXidStart(ch);
}

static bool IsFirstNonWhitespace(Sci_Position pos, Accessor &styler) {
	Sci_Position line = styler.GetLine(pos);
	Sci_Position start_pos = styler.LineStart(line);
	for (Sci_Position i = start_pos; i < pos; i++) {
		const char ch = styler[i];
		if (!(ch == ' ' || ch == '\t'))
			return false;
	}
	return true;
}

// Options used for LexerPython
struct OptionsPython {
	int whingeLevel;
	bool base2or8Literals;
	bool stringsU;
	bool stringsB;
	bool stringsF;
	bool stringsOverNewline;
	bool keywords2NoSubIdentifiers;
	bool fold;
	bool foldQuotes;
	bool foldCompact;
	bool unicodeIdentifiers;

	OptionsPython() {
		whingeLevel = 0;
		base2or8Literals = true;
		stringsU = true;
		stringsB = true;
		stringsF = true;
		stringsOverNewline = false;
		keywords2NoSubIdentifiers = false;
		fold = false;
		foldQuotes = false;
		foldCompact = false;
		unicodeIdentifiers = true;
	}

	literalsAllowed AllowedLiterals() const {
		literalsAllowed allowedLiterals = stringsU ? litU : litNone;
		if (stringsB)
			allowedLiterals = static_cast<literalsAllowed>(allowedLiterals | litB);
		if (stringsF)
			allowedLiterals = static_cast<literalsAllowed>(allowedLiterals | litF);
		return allowedLiterals;
	}
};

static const char *const pythonWordListDesc[] = {
	"Keywords",
	"Highlighted identifiers",
	0
};

struct OptionSetPython : public OptionSet<OptionsPython> {
	OptionSetPython() {
		DefineProperty("tab.timmy.whinge.level", &OptionsPython::whingeLevel,
			       "For Python code, checks whether indenting is consistent. "
			       "The default, 0 turns off indentation checking, "
			       "1 checks whether each line is potentially inconsistent with the previous line, "
			       "2 checks whether any space characters occur before a tab character in the indentation, "
			       "3 checks whether any spaces are in the indentation, and "
			       "4 checks for any tab characters in the indentation. "
			       "1 is a good level to use.");

		DefineProperty("lexer.python.literals.binary", &OptionsPython::base2or8Literals,
			       "Set to 0 to not recognise Python 3 binary and octal literals: 0b1011 0o712.");

		DefineProperty("lexer.python.strings.u", &OptionsPython::stringsU,
			       "Set to 0 to not recognise Python Unicode literals u\"x\" as used before Python 3.");

		DefineProperty("lexer.python.strings.b", &OptionsPython::stringsB,
			       "Set to 0 to not recognise Python 3 bytes literals b\"x\".");

		DefineProperty("lexer.python.strings.f", &OptionsPython::stringsF,
			       "Set to 0 to not recognise Python 3.6 f-string literals f\"var={var}\".");

		DefineProperty("lexer.python.strings.over.newline", &OptionsPython::stringsOverNewline,
			       "Set to 1 to allow strings to span newline characters.");

		DefineProperty("lexer.python.keywords2.no.sub.identifiers", &OptionsPython::keywords2NoSubIdentifiers,
			       "When enabled, it will not style keywords2 items that are used as a sub-identifier. "
			       "Example: when set, will not highlight \"foo.open\" when \"open\" is a keywords2 item.");

		DefineProperty("fold", &OptionsPython::fold);

		DefineProperty("fold.quotes.python", &OptionsPython::foldQuotes,
			       "This option enables folding multi-line quoted strings when using the Python lexer.");

		DefineProperty("fold.compact", &OptionsPython::foldCompact);

		DefineProperty("lexer.python.unicode.identifiers", &OptionsPython::unicodeIdentifiers,
			       "Set to 0 to not recognise Python 3 unicode identifiers.");

		DefineWordListSets(pythonWordListDesc);
	}
};

const char styleSubable[] = { SCE_P_IDENTIFIER, 0 };

LexicalClass lexicalClasses[] = {
	// Lexer Python SCLEX_PYTHON SCE_P_:
	0, "SCE_P_DEFAULT", "default", "White space",
	1, "SCE_P_COMMENTLINE", "comment line", "Comment",
	2, "SCE_P_NUMBER", "literal numeric", "Number",
	3, "SCE_P_STRING", "literal string", "String",
	4, "SCE_P_CHARACTER", "literal string", "Single quoted string",
	5, "SCE_P_WORD", "keyword", "Keyword",
	6, "SCE_P_TRIPLE", "literal string", "Triple quotes",
	7, "SCE_P_TRIPLEDOUBLE", "literal string", "Triple double quotes",
	8, "SCE_P_CLASSNAME", "identifier", "Class name definition",
	9, "SCE_P_DEFNAME", "identifier", "Function or method name definition",
	10, "SCE_P_OPERATOR", "operator", "Operators",
	11, "SCE_P_IDENTIFIER", "identifier", "Identifiers",
	12, "SCE_P_COMMENTBLOCK", "comment", "Comment-blocks",
	13, "SCE_P_STRINGEOL", "error literal string", "End of line where string is not closed",
	14, "SCE_P_WORD2", "identifier", "Highlighted identifiers",
	15, "SCE_P_DECORATOR", "preprocessor", "Decorators",
	16, "SCE_P_FSTRING", "literal string interpolated", "F-String",
	17, "SCE_P_FCHARACTER", "literal string interpolated", "Single quoted f-string",
	18, "SCE_P_FTRIPLE", "literal string interpolated", "Triple quoted f-string",
	19, "SCE_P_FTRIPLEDOUBLE", "literal string interpolated", "Triple double quoted f-string",
};

}

class LexerPython : public DefaultLexer {
	WordList keywords;
	WordList keywords2;
	OptionsPython options;
	OptionSetPython osPython;
	enum { ssIdentifier };
	SubStyles subStyles;
	std::map<Sci_Position, std::vector<SingleFStringExpState> > ftripleStateAtEol;
public:
	explicit LexerPython() :
		DefaultLexer(lexicalClasses, ELEMENTS(lexicalClasses)),
		subStyles(styleSubable, 0x80, 0x40, 0) {
	}
	~LexerPython() override {
	}
	void SCI_METHOD Release() override {
		delete this;
	}
	int SCI_METHOD Version() const override {
		return lvSubStyles;
	}
	const char *SCI_METHOD PropertyNames() override {
		return osPython.PropertyNames();
	}
	int SCI_METHOD PropertyType(const char *name) override {
		return osPython.PropertyType(name);
	}
	const char *SCI_METHOD DescribeProperty(const char *name) override {
		return osPython.DescribeProperty(name);
	}
	Sci_Position SCI_METHOD PropertySet(const char *key, const char *val) override;
	const char *SCI_METHOD DescribeWordListSets() override {
		return osPython.DescribeWordListSets();
	}
	Sci_Position SCI_METHOD WordListSet(int n, const char *wl) override;
	void SCI_METHOD Lex(Sci_PositionU startPos, Sci_Position length, int initStyle, IDocument *pAccess) override;
	void SCI_METHOD Fold(Sci_PositionU startPos, Sci_Position length, int initStyle, IDocument *pAccess) override;

	void *SCI_METHOD PrivateCall(int, void *) override {
		return 0;
	}

	int SCI_METHOD LineEndTypesSupported() override {
		return SC_LINE_END_TYPE_UNICODE;
	}

	int SCI_METHOD AllocateSubStyles(int styleBase, int numberStyles) override {
		return subStyles.Allocate(styleBase, numberStyles);
	}
	int SCI_METHOD SubStylesStart(int styleBase) override {
		return subStyles.Start(styleBase);
	}
	int SCI_METHOD SubStylesLength(int styleBase) override {
		return subStyles.Length(styleBase);
	}
	int SCI_METHOD StyleFromSubStyle(int subStyle) override {
		const int styleBase = subStyles.BaseStyle(subStyle);
		return styleBase;
	}
	int SCI_METHOD PrimaryStyleFromStyle(int style) override {
		return style;
	}
	void SCI_METHOD FreeSubStyles() override {
		subStyles.Free();
	}
	void SCI_METHOD SetIdentifiers(int style, const char *identifiers) override {
		subStyles.SetIdentifiers(style, identifiers);
	}
	int SCI_METHOD DistanceToSecondaryStyles() override {
		return 0;
	}
	const char *SCI_METHOD GetSubStyleBases() override {
		return styleSubable;
	}

	static ILexer *LexerFactoryPython() {
		return new LexerPython();
	}

private:
	void ProcessLineEnd(StyleContext &sc, std::vector<SingleFStringExpState> &fstringStateStack, SingleFStringExpState *&currentFStringExp, bool &inContinuedString);
};

Sci_Position SCI_METHOD LexerPython::PropertySet(const char *key, const char *val) {
	if (osPython.PropertySet(&options, key, val)) {
		return 0;
	}
	return -1;
}

Sci_Position SCI_METHOD LexerPython::WordListSet(int n, const char *wl) {
	WordList *wordListN = 0;
	switch (n) {
	case 0:
		wordListN = &keywords;
		break;
	case 1:
		wordListN = &keywords2;
		break;
	}
	Sci_Position firstModification = -1;
	if (wordListN) {
		WordList wlNew;
		wlNew.Set(wl);
		if (*wordListN != wlNew) {
			wordListN->Set(wl);
			firstModification = 0;
		}
	}
	return firstModification;
}

void LexerPython::ProcessLineEnd(StyleContext &sc, std::vector<SingleFStringExpState> &fstringStateStack, SingleFStringExpState *&currentFStringExp, bool &inContinuedString) {
	long deepestSingleStateIndex = -1;
	unsigned long i;

	// Find the deepest single quote state because that string will end; no \ continuation in f-string
	for (i = 0; i < fstringStateStack.size(); i++) {
		if (IsPySingleQuoteStringState(fstringStateStack[i].state)) {
			deepestSingleStateIndex = i;
			break;
		}
	}

	if (deepestSingleStateIndex != -1) {
		sc.SetState(fstringStateStack[deepestSingleStateIndex].state);
		while (fstringStateStack.size() > static_cast<unsigned long>(deepestSingleStateIndex)) {
			PopFromStateStack(fstringStateStack, currentFStringExp);
		}
	}
	if (!fstringStateStack.empty()) {
		std::pair<Sci_Position, std::vector<SingleFStringExpState> > val;
		val.first = sc.currentLine;
		val.second = fstringStateStack;

		ftripleStateAtEol.insert(val);
	}

	if ((sc.state == SCE_P_DEFAULT)
			|| IsPyTripleQuoteStringState(sc.state)) {
		// Perform colourisation of white space and triple quoted strings at end of each line to allow
		// tab marking to work inside white space and triple quoted strings
		sc.SetState(sc.state);
	}
	if (IsPySingleQuoteStringState(sc.state)) {
		if (inContinuedString || options.stringsOverNewline) {
			inContinuedString = false;
		} else {
			sc.ChangeState(SCE_P_STRINGEOL);
			sc.ForwardSetState(SCE_P_DEFAULT);
		}
	}
}

void SCI_METHOD LexerPython::Lex(Sci_PositionU startPos, Sci_Position length, int initStyle, IDocument *pAccess) {
	Accessor styler(pAccess, NULL);

	// Track whether in f-string expression; vector is used for a stack to
	// handle nested f-strings such as f"""{f'''{f"{f'{1}'}"}'''}"""
	std::vector<SingleFStringExpState> fstringStateStack;
	SingleFStringExpState *currentFStringExp = NULL;

	const Sci_Position endPos = startPos + length;

	// Backtrack to previous line in case need to fix its tab whinging
	Sci_Position lineCurrent = styler.GetLine(startPos);
	if (startPos > 0) {
		if (lineCurrent > 0) {
			lineCurrent--;
			// Look for backslash-continued lines
			while (lineCurrent > 0) {
				Sci_Position eolPos = styler.LineStart(lineCurrent) - 1;
				const int eolStyle = styler.StyleAt(eolPos);
				if (eolStyle == SCE_P_STRING
						|| eolStyle == SCE_P_CHARACTER
						|| eolStyle == SCE_P_STRINGEOL) {
					lineCurrent -= 1;
				} else {
					break;
				}
			}
			startPos = styler.LineStart(lineCurrent);
		}
		initStyle = startPos == 0 ? SCE_P_DEFAULT : styler.StyleAt(startPos - 1);
	}

	const literalsAllowed allowedLiterals = options.AllowedLiterals();

	initStyle = initStyle & 31;
	if (initStyle == SCE_P_STRINGEOL) {
		initStyle = SCE_P_DEFAULT;
	}

	// Set up fstate stack from last line and remove any subsequent ftriple at eol states
	std::map<Sci_Position, std::vector<SingleFStringExpState> >::iterator it;
	it = ftripleStateAtEol.find(lineCurrent - 1);
	if (it != ftripleStateAtEol.end() && !it->second.empty()) {
		fstringStateStack = it->second;
		currentFStringExp = &fstringStateStack.back();
	}
	it = ftripleStateAtEol.lower_bound(lineCurrent);
	if (it != ftripleStateAtEol.end()) {
		ftripleStateAtEol.erase(it, ftripleStateAtEol.end());
	}

	kwType kwLast = kwOther;
	int spaceFlags = 0;
	styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
	bool base_n_number = false;

	const WordClassifier &classifierIdentifiers = subStyles.Classifier(SCE_P_IDENTIFIER);

	StyleContext sc(startPos, endPos - startPos, initStyle, styler);

	bool indentGood = true;
	Sci_Position startIndicator = sc.currentPos;
	bool inContinuedString = false;

	for (; sc.More(); sc.Forward()) {

		if (sc.atLineStart) {
			styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
			indentGood = true;
			if (options.whingeLevel == 1) {
				indentGood = (spaceFlags & wsInconsistent) == 0;
			} else if (options.whingeLevel == 2) {
				indentGood = (spaceFlags & wsSpaceTab) == 0;
			} else if (options.whingeLevel == 3) {
				indentGood = (spaceFlags & wsSpace) == 0;
			} else if (options.whingeLevel == 4) {
				indentGood = (spaceFlags & wsTab) == 0;
			}
			if (!indentGood) {
				styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 0);
				startIndicator = sc.currentPos;
			}
		}

		if (sc.atLineEnd) {
			ProcessLineEnd(sc, fstringStateStack, currentFStringExp, inContinuedString);
			lineCurrent++;
			if (!sc.More())
				break;
		}

		bool needEOLCheck = false;


		if (sc.state == SCE_P_OPERATOR) {
			kwLast = kwOther;
			sc.SetState(SCE_P_DEFAULT);
		} else if (sc.state == SCE_P_NUMBER) {
			if (!IsAWordChar(sc.ch, false) &&
					!(!base_n_number && ((sc.ch == '+' || sc.ch == '-') && (sc.chPrev == 'e' || sc.chPrev == 'E')))) {
				sc.SetState(SCE_P_DEFAULT);
			}
		} else if (sc.state == SCE_P_IDENTIFIER) {
			if ((sc.ch == '.') || (!IsAWordChar(sc.ch, options.unicodeIdentifiers))) {
				char s[100];
				sc.GetCurrent(s, sizeof(s));
				int style = SCE_P_IDENTIFIER;
				if ((kwLast == kwImport) && (strcmp(s, "as") == 0)) {
					style = SCE_P_WORD;
				} else if (keywords.InList(s)) {
					style = SCE_P_WORD;
				} else if (kwLast == kwClass) {
					style = SCE_P_CLASSNAME;
				} else if (kwLast == kwDef) {
					style = SCE_P_DEFNAME;
				} else if (kwLast == kwCDef || kwLast == kwCPDef) {
					Sci_Position pos = sc.currentPos;
					unsigned char ch = styler.SafeGetCharAt(pos, '\0');
					while (ch != '\0') {
						if (ch == '(') {
							style = SCE_P_DEFNAME;
							break;
						} else if (ch == ':') {
							style = SCE_P_CLASSNAME;
							break;
						} else if (ch == ' ' || ch == '\t' || ch == '\n' || ch == '\r') {
							pos++;
							ch = styler.SafeGetCharAt(pos, '\0');
						} else {
							break;
						}
					}
				} else if (keywords2.InList(s)) {
					if (options.keywords2NoSubIdentifiers) {
						// We don't want to highlight keywords2
						// that are used as a sub-identifier,
						// i.e. not open in "foo.open".
						Sci_Position pos = styler.GetStartSegment() - 1;
						if (pos < 0 || (styler.SafeGetCharAt(pos, '\0') != '.'))
							style = SCE_P_WORD2;
					} else {
						style = SCE_P_WORD2;
					}
				} else {
					int subStyle = classifierIdentifiers.ValueFor(s);
					if (subStyle >= 0) {
						style = subStyle;
					}
				}
				sc.ChangeState(style);
				sc.SetState(SCE_P_DEFAULT);
				if (style == SCE_P_WORD) {
					if (0 == strcmp(s, "class"))
						kwLast = kwClass;
					else if (0 == strcmp(s, "def"))
						kwLast = kwDef;
					else if (0 == strcmp(s, "import"))
						kwLast = kwImport;
					else if (0 == strcmp(s, "cdef"))
						kwLast = kwCDef;
					else if (0 == strcmp(s, "cpdef"))
						kwLast = kwCPDef;
					else if (0 == strcmp(s, "cimport"))
						kwLast = kwImport;
					else if (kwLast != kwCDef && kwLast != kwCPDef)
						kwLast = kwOther;
				} else if (kwLast != kwCDef && kwLast != kwCPDef) {
					kwLast = kwOther;
				}
			}
		} else if ((sc.state == SCE_P_COMMENTLINE) || (sc.state == SCE_P_COMMENTBLOCK)) {
			if (sc.ch == '\r' || sc.ch == '\n') {
				sc.SetState(SCE_P_DEFAULT);
			}
		} else if (sc.state == SCE_P_DECORATOR) {
			if (!IsAWordStart(sc.ch, options.unicodeIdentifiers)) {
				sc.SetState(SCE_P_DEFAULT);
			}
		} else if (IsPySingleQuoteStringState(sc.state)) {
			if (sc.ch == '\\') {
				if ((sc.chNext == '\r') && (sc.GetRelative(2) == '\n')) {
					sc.Forward();
				}
				if (sc.chNext == '\n' || sc.chNext == '\r') {
					inContinuedString = true;
				} else {
					// Don't roll over the newline.
					sc.Forward();
				}
			} else if (sc.ch == GetPyStringQuoteChar(sc.state)) {
				sc.ForwardSetState(SCE_P_DEFAULT);
				needEOLCheck = true;
			}
		} else if ((sc.state == SCE_P_TRIPLE) || (sc.state == SCE_P_FTRIPLE)) {
			if (sc.ch == '\\') {
				sc.Forward();
			} else if (sc.Match(R"(''')")) {
				sc.Forward();
				sc.Forward();
				sc.ForwardSetState(SCE_P_DEFAULT);
				needEOLCheck = true;
			}
		} else if ((sc.state == SCE_P_TRIPLEDOUBLE) || (sc.state == SCE_P_FTRIPLEDOUBLE)) {
			if (sc.ch == '\\') {
				sc.Forward();
			} else if (sc.Match(R"(""")")) {
				sc.Forward();
				sc.Forward();
				sc.ForwardSetState(SCE_P_DEFAULT);
				needEOLCheck = true;
			}
		}

		// Note if used and not if else because string states also match
		// some of the above clauses
		if (IsPyFStringState(sc.state) && sc.ch == '{') {
			if (sc.chNext == '{') {
				sc.Forward();
			} else {
				PushStateToStack(sc.state, fstringStateStack, currentFStringExp);
				sc.ForwardSetState(SCE_P_DEFAULT);
			}
			needEOLCheck = true;
		}

		// If in an f-string expression, check for the ending quote(s)
		// and end f-string to handle syntactically incorrect cases like
		// f'{' and f"""{"""
		if (!fstringStateStack.empty() && (sc.ch == '\'' || sc.ch == '"')) {
			long matching_stack_i = -1;
			for (unsigned long stack_i = 0; stack_i < fstringStateStack.size() && matching_stack_i == -1; stack_i++) {
				const int stack_state = fstringStateStack[stack_i].state;
				const char quote = GetPyStringQuoteChar(stack_state);
				if (sc.ch == quote) {
					if (IsPySingleQuoteStringState(stack_state)) {
						matching_stack_i = stack_i;
					} else if (quote == '"' ? sc.Match(R"(""")") : sc.Match("'''")) {
						matching_stack_i = stack_i;
					}
				}
			}

			if (matching_stack_i != -1) {
				sc.SetState(fstringStateStack[matching_stack_i].state);
				if (IsPyTripleQuoteStringState(fstringStateStack[matching_stack_i].state)) {
					sc.Forward();
					sc.Forward();
				}
				sc.ForwardSetState(SCE_P_DEFAULT);
				needEOLCheck = true;

				while (fstringStateStack.size() > static_cast<unsigned long>(matching_stack_i)) {
					PopFromStateStack(fstringStateStack, currentFStringExp);
				}
			}
		}
		// End of code to find the end of a state

		if (!indentGood && !IsASpaceOrTab(sc.ch)) {
			styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 1);
			startIndicator = sc.currentPos;
			indentGood = true;
		}

		// One cdef or cpdef line, clear kwLast only at end of line
		if ((kwLast == kwCDef || kwLast == kwCPDef) && sc.atLineEnd) {
			kwLast = kwOther;
		}

		// State exit code may have moved on to end of line
		if (needEOLCheck && sc.atLineEnd) {
			ProcessLineEnd(sc, fstringStateStack, currentFStringExp, inContinuedString);
			lineCurrent++;
			styler.IndentAmount(lineCurrent, &spaceFlags, IsPyComment);
			if (!sc.More())
				break;
		}

		// If in f-string expression, check for }, :, ! to resume f-string state or update nesting count
		if (currentFStringExp != NULL && !IsPySingleQuoteStringState(sc.state) && !IsPyTripleQuoteStringState(sc.state)) {
			if (currentFStringExp->nestingCount == 0 && (sc.ch == '}' || sc.ch == ':' || (sc.ch == '!' && sc.chNext != '='))) {
				sc.SetState(PopFromStateStack(fstringStateStack, currentFStringExp));
			} else {
				if (sc.ch == '{' || sc.ch == '[' || sc.ch == '(') {
					currentFStringExp->nestingCount++;
				} else if (sc.ch == '}' || sc.ch == ']' || sc.ch == ')') {
					currentFStringExp->nestingCount--;
				}
			}
		}

		// Check for a new state starting character
		if (sc.state == SCE_P_DEFAULT) {
			if (IsADigit(sc.ch) || (sc.ch == '.' && IsADigit(sc.chNext))) {
				if (sc.ch == '0' && (sc.chNext == 'x' || sc.chNext == 'X')) {
					base_n_number = true;
					sc.SetState(SCE_P_NUMBER);
				} else if (sc.ch == '0' &&
						(sc.chNext == 'o' || sc.chNext == 'O' || sc.chNext == 'b' || sc.chNext == 'B')) {
					if (options.base2or8Literals) {
						base_n_number = true;
						sc.SetState(SCE_P_NUMBER);
					} else {
						sc.SetState(SCE_P_NUMBER);
						sc.ForwardSetState(SCE_P_IDENTIFIER);
					}
				} else {
					base_n_number = false;
					sc.SetState(SCE_P_NUMBER);
				}
			} else if ((IsASCII(sc.ch) && isoperator(static_cast<char>(sc.ch))) || sc.ch == '`') {
				sc.SetState(SCE_P_OPERATOR);
			} else if (sc.ch == '#') {
				sc.SetState(sc.chNext == '#' ? SCE_P_COMMENTBLOCK : SCE_P_COMMENTLINE);
			} else if (sc.ch == '@') {
				if (IsFirstNonWhitespace(sc.currentPos, styler))
					sc.SetState(SCE_P_DECORATOR);
				else
					sc.SetState(SCE_P_OPERATOR);
			} else if (IsPyStringStart(sc.ch, sc.chNext, sc.GetRelative(2), allowedLiterals)) {
				Sci_PositionU nextIndex = 0;
				sc.SetState(GetPyStringState(styler, sc.currentPos, &nextIndex, allowedLiterals));
				while (nextIndex > (sc.currentPos + 1) && sc.More()) {
					sc.Forward();
				}
			} else if (IsAWordStart(sc.ch, options.unicodeIdentifiers)) {
				sc.SetState(SCE_P_IDENTIFIER);
			}
		}
	}
	styler.IndicatorFill(startIndicator, sc.currentPos, indicatorWhitespace, 0);
	sc.Complete();
}

static bool IsCommentLine(Sci_Position line, Accessor &styler) {
	Sci_Position pos = styler.LineStart(line);
	const Sci_Position eol_pos = styler.LineStart(line + 1) - 1;
	for (Sci_Position i = pos; i < eol_pos; i++) {
		const char ch = styler[i];
		if (ch == '#')
			return true;
		else if (ch != ' ' && ch != '\t')
			return false;
	}
	return false;
}

static bool IsQuoteLine(Sci_Position line, const Accessor &styler) {
	const int style = styler.StyleAt(styler.LineStart(line)) & 31;
	return IsPyTripleQuoteStringState(style);
}


void SCI_METHOD LexerPython::Fold(Sci_PositionU startPos, Sci_Position length, int /*initStyle - unused*/, IDocument *pAccess) {
	if (!options.fold)
		return;

	Accessor styler(pAccess, NULL);

	const Sci_Position maxPos = startPos + length;
	const Sci_Position maxLines = (maxPos == styler.Length()) ? styler.GetLine(maxPos) : styler.GetLine(maxPos - 1);	// Requested last line
	const Sci_Position docLines = styler.GetLine(styler.Length());	// Available last line

	// Backtrack to previous non-blank line so we can determine indent level
	// for any white space lines (needed esp. within triple quoted strings)
	// and so we can fix any preceding fold level (which is why we go back
	// at least one line in all cases)
	int spaceFlags = 0;
	Sci_Position lineCurrent = styler.GetLine(startPos);
	int indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL);
	while (lineCurrent > 0) {
		lineCurrent--;
		indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL);
		if (!(indentCurrent & SC_FOLDLEVELWHITEFLAG) &&
				(!IsCommentLine(lineCurrent, styler)) &&
				(!IsQuoteLine(lineCurrent, styler)))
			break;
	}
	int indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK;

	// Set up initial loop state
	startPos = styler.LineStart(lineCurrent);
	int prev_state = SCE_P_DEFAULT & 31;
	if (lineCurrent >= 1)
		prev_state = styler.StyleAt(startPos - 1) & 31;
	int prevQuote = options.foldQuotes && IsPyTripleQuoteStringState(prev_state);

	// Process all characters to end of requested range or end of any triple quote
	//that hangs over the end of the range.  Cap processing in all cases
	// to end of document (in case of unclosed quote at end).
	while ((lineCurrent <= docLines) && ((lineCurrent <= maxLines) || prevQuote)) {

		// Gather info
		int lev = indentCurrent;
		Sci_Position lineNext = lineCurrent + 1;
		int indentNext = indentCurrent;
		int quote = false;
		if (lineNext <= docLines) {
			// Information about next line is only available if not at end of document
			indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL);
			Sci_Position lookAtPos = (styler.LineStart(lineNext) == styler.Length()) ? styler.Length() - 1 : styler.LineStart(lineNext);
			const int style = styler.StyleAt(lookAtPos) & 31;
			quote = options.foldQuotes && IsPyTripleQuoteStringState(style);
		}
		const int quote_start = (quote && !prevQuote);
		const int quote_continue = (quote && prevQuote);
		if (!quote || !prevQuote)
			indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK;
		if (quote)
			indentNext = indentCurrentLevel;
		if (indentNext & SC_FOLDLEVELWHITEFLAG)
			indentNext = SC_FOLDLEVELWHITEFLAG | indentCurrentLevel;

		if (quote_start) {
			// Place fold point at start of triple quoted string
			lev |= SC_FOLDLEVELHEADERFLAG;
		} else if (quote_continue || prevQuote) {
			// Add level to rest of lines in the string
			lev = lev + 1;
		}

		// Skip past any blank lines for next indent level info; we skip also
		// comments (all comments, not just those starting in column 0)
		// which effectively folds them into surrounding code rather
		// than screwing up folding.  If comments end file, use the min
		// comment indent as the level after

		int minCommentLevel = indentCurrentLevel;
		while (!quote &&
				(lineNext < docLines) &&
				((indentNext & SC_FOLDLEVELWHITEFLAG) ||
				 (lineNext <= docLines && IsCommentLine(lineNext, styler)))) {

			if (IsCommentLine(lineNext, styler) && indentNext < minCommentLevel) {
				minCommentLevel = indentNext;
			}

			lineNext++;
			indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL);
		}

		const int levelAfterComments = ((lineNext < docLines) ? indentNext & SC_FOLDLEVELNUMBERMASK : minCommentLevel);
		const int levelBeforeComments = std::max(indentCurrentLevel, levelAfterComments);

		// Now set all the indent levels on the lines we skipped
		// Do this from end to start.  Once we encounter one line
		// which is indented more than the line after the end of
		// the comment-block, use the level of the block before

		Sci_Position skipLine = lineNext;
		int skipLevel = levelAfterComments;

		while (--skipLine > lineCurrent) {
			const int skipLineIndent = styler.IndentAmount(skipLine, &spaceFlags, NULL);

			if (options.foldCompact) {
				if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments)
					skipLevel = levelBeforeComments;

				int whiteFlag = skipLineIndent & SC_FOLDLEVELWHITEFLAG;

				styler.SetLevel(skipLine, skipLevel | whiteFlag);
			} else {
				if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments &&
						!(skipLineIndent & SC_FOLDLEVELWHITEFLAG) &&
						!IsCommentLine(skipLine, styler))
					skipLevel = levelBeforeComments;

				styler.SetLevel(skipLine, skipLevel);
			}
		}

		// Set fold header on non-quote line
		if (!quote && !(indentCurrent & SC_FOLDLEVELWHITEFLAG)) {
			if ((indentCurrent & SC_FOLDLEVELNUMBERMASK) < (indentNext & SC_FOLDLEVELNUMBERMASK))
				lev |= SC_FOLDLEVELHEADERFLAG;
		}

		// Keep track of triple quote state of previous line
		prevQuote = quote;

		// Set fold level for this line and move to next line
		styler.SetLevel(lineCurrent, options.foldCompact ? lev : lev & ~SC_FOLDLEVELWHITEFLAG);
		indentCurrent = indentNext;
		lineCurrent = lineNext;
	}

	// NOTE: Cannot set level of last line here because indentCurrent doesn't have
	// header flag set; the loop above is crafted to take care of this case!
	//styler.SetLevel(lineCurrent, indentCurrent);
}

LexerModule lmPython(SCLEX_PYTHON, LexerPython::LexerFactoryPython, "python",
		     pythonWordListDesc);