Remove experimental regex syntax, see PR #722.

The regex code to find words in current document is actually much slow than
plain find. the removal makes it possible to use external regex engine.

locale/de/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\tEnd of a line\n\
 *? or +?\tNon-greedy matching of quantifiers ""?"" and ""+""\n\
 (\tStart of a region\n\
 )\tEnd of a region\n\
-\\n\tRefers to a region when replacing (n is 1-9)\n\n\
-Experimental Syntax:\n\
-\\h\tStart of a word, according to IsWordStartAt()\n\
-\\H\tEnd of a word, according to IsWordEndAt()\n\
-\\i\tMatches 1 or more characters to the end of a word\n\
-\\i?\tMatches 0 or more characters to the end of a word"
+\\n\tRefers to a region when replacing (n is 1-9)"
 
 	IDS_WILDCARDHELP		"Wildcard Search\n\n\
 *\tMatches zero or more characters.\n\

locale/fr-FR/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\tfind e ligne\n\
 *? or +?\tNon-greedy matching of quantifiers ""?"" and ""+""\n\
 (\tDébut d'une région\n\
 )\tFin  d'une région\n\
-\\n\tSe réfère à une région lors du remplacement (avec n étant compris entre 1 et 9)\n\n\
-Syntaxe expérimentale :\n\
-\\h\tDébut d'un mot, en accord avec IsWordStartAt()\n\
-\\H\tFin d'un mot, en accord avec IsWordEndAt()\n\
-\\i\tTrouvera 1 caractère ou plus jusqu'à la fin du mot\n\
-\\i?\tTrouvera 0 caractère ou plus jusqu'à la fin du mot"
+\\n\tSe réfère à une région lors du remplacement (avec n étant compris entre 1 et 9)"
 
 	IDS_WILDCARDHELP		"Jocker de recherche\n\n\
 *\tTrouvera zéro ou plus de caractères.\n\

locale/it/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\tFine di una linea\n\
 *? o +?\tCorrispondenza inaccurata di quantificatori ""?"" e ""+""\n\
 (\tInizio di un'area\n\
 )\tFine di un'area\n\
-\\n\tRiferimento a un'area quando sostituita (n è 1-9)\n\n\
-Sintassi sperimentale:\n\
-\\h\tInizio di una parola, secondo IsWordStartAt()\n\
-\\H\tFine di una parola, secondo IsWordEndAt()\n\
-\\i\tCorrisponde a 1 o più caratteri alla fine di una \t\tparola\n\
-\\i?\tCorrisponde a 0 o più caratteri alla fine di una \t\tparola"
+\\n\tRiferimento a un'area quando sostituita (n è 1-9)"
 
 	IDS_WILDCARDHELP		"Ricerca con caratteri Jolly\n\n\
 *\tCorrisponde a zero o più caratteri.\n\

locale/ja/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\t行の末尾\n\
 *? か +?\t上記記号の ""?"" と ""+"" を最長合致ではなくする\n\
 (\t範囲指定の括弧開始\n\
 )\t範囲指定の閉じ括弧\n\
-\\n\t置換時に対応括弧内の文字列を呼出 (n は 1-9)\n\n\
-試験段階の構文:\n\
-\\h\tIsWordStartAt() による単語の開始\n\
-\\H\tIsWordEndAt() による単語の終わり\n\
-\\i\t単語の終わりまで1文字以上に合致\n\
-\\i?\t同 0文字以上に合致"
+\\n\t置換時に対応括弧内の文字列を呼出 (n は 1-9)"
 
 	IDS_WILDCARDHELP		"ワイルドカード検索\n\n\
 *\t0文字以上の文字列に合致\n\

locale/ko/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\t줄의 끝\n\
 *? or +?\t""?"" 및 ""+""의 게으른 수량자\n\
 (\t그룹의 시작\n\
 )\t그룹의 끝\n\
-\\n\t교체시 영역을 나타냄 (n은 1-9)\n\n\
-시험적 구문:\n\
-\\h\tIsWordStartAt()에 따른 단어 시작\n\
-\\H\tIsWordEndAt()에 따른 단어의 끝\n\
-\\i\t단어의 끝까지 1개 이상의 문자를 찾습니다\n\
-\\i?\t단어의 끝까지 0개 이상의 문자를 찾습니다"
+\\n\t교체시 영역을 나타냄 (n은 1-9)"
 
 	IDS_WILDCARDHELP        "와일드카드 검색\n\n\
 *\t0개 이상의 문자와 일치합니다.\n\

locale/pt-BR/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\tEnd of a line\n\
 *? or +?\tNon-greedy matching of quantifiers ""?"" and ""+""\n\
 (\tStart of a region\n\
 )\tEnd of a region\n\
-\\n\tRefers to a region when replacing (n is 1-9)\n\n\
-Experimental Syntax:\n\
-\\h\tStart of a word, according to IsWordStartAt()\n\
-\\H\tEnd of a word, according to IsWordEndAt()\n\
-\\i\tMatches 1 or more characters to the end of a word\n\
-\\i?\tMatches 0 or more characters to the end of a word"
+\\n\tRefers to a region when replacing (n is 1-9)"
 
 	IDS_WILDCARDHELP		"Wildcard Search\n\n\
 *\tMatches zero or more characters.\n\

locale/zh-Hans/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\t行结束\n\
 *? 或 +?\t对限量词 ""?"" 和 ""+"" 做非贪婪匹配\n\
 (\t区域开始\n\
 )\t区域结束\n\
-\\n\t替换时引用的区域(n 为 1-9)\n\n\
-实验性语法:\n\
-\\h\t单词开始，根据 IsWordStartAt()\n\
-\\H\t单词结束，根据 IsWordEndAt()\n\
-\\i\t匹配 1 个或多个字符到单词结束\n\
-\\i?\t匹配 0 个或多个字符到单词结束"
+\\n\t替换时引用的区域(n 为 1-9)"
 
 	IDS_WILDCARDHELP		"通配符搜索\n\n\
 *\t匹配 0 个或多个字符\n\

locale/zh-Hant/Notepad2.rc

@@ -1764,12 +1764,7 @@ $\t行結束。\n\
 *? 或 +?\t對限量詞 ""?"" 和 ""+"" 做非貪婪比對。\n\
 (\t區域開始。\n\
 )\t區域結束。\n\
-\\n\t取代時參考的區域(n 為 1-9)。\n\n\
-實驗性語法:\n\
-\\h\t單詞開始，根據 IsWordStartAt()。\n\
-\\H\t單詞結束，根據 IsWordEndAt()。\n\
-\\i\t符合 1 個或多個字元到單詞結束。\n\
-\\i?\t符合 0 個或多個字元到單詞結束。"
+\\n\t取代時參考的區域(n 為 1-9)。"
 
 	IDS_WILDCARDHELP		"通配符搜尋\n\n\
 *\t符合 0 個或多個字元。\n\

scintilla/include/Scintilla.h

@@ -478,6 +478,7 @@ typedef sptr_t (*SciFnDirectStatus)(sptr_t ptr, unsigned int iMessage, uptr_t wP
 #define SCFIND_NONE 0x0
 #define SCFIND_WHOLEWORD 0x2
 #define SCFIND_MATCHCASE 0x4
+#define SCFIND_MATCH_TO_WORD_END 0x8
 #define SCFIND_WORDSTART 0x00100000
 #define SCFIND_REGEXP 0x00200000
 #define SCFIND_POSIX 0x00400000

scintilla/include/Scintilla.iface

@@ -1259,6 +1259,7 @@ enu FindOption=SCFIND_
 val SCFIND_NONE=0x0
 val SCFIND_WHOLEWORD=0x2
 val SCFIND_MATCHCASE=0x4
+val SCFIND_MATCH_TO_WORD_END=0x8
 val SCFIND_WORDSTART=0x00100000
 val SCFIND_REGEXP=0x00200000
 val SCFIND_POSIX=0x00400000

scintilla/include/ScintillaTypes.h

@@ -271,6 +271,7 @@ enum class FindOption {
 	None = 0x0,
 	WholeWord = 0x2,
 	MatchCase = 0x4,
+	MatchToWordEnd = 0x8,
 	WordStart = 0x00100000,
 	RegExp = 0x00200000,
 	Posix = 0x00400000,

scintilla/src/Document.cxx

@@ -2951,16 +2951,12 @@ class BuiltinRegex final : public RegexSearchBase {
 
 	const char *SubstituteByPosition(Document *doc, const char *text, Sci::Position *length) override;
 
-#ifdef NO_CXX11_REGEX
 	void ClearCache() noexcept override {
 		search.ClearCache();
 	}
-#endif
 
 private:
-#ifdef NO_CXX11_REGEX
 	RESearch search;
-#endif
 	std::string substituted;
 };
 
@@ -3020,25 +3016,9 @@ class DocumentIndexer final : public CharacterIndexer {
 			return '\0';
 	}
 
-	bool IsWordStartAt(Sci::Position pos) const noexcept override {
-		return pdoc->IsWordStartAt(pos);
-	}
-
-	bool IsWordEndAt(Sci::Position pos) const noexcept override {
-		return pdoc->IsWordEndAt(pos);
-	}
-
 	Sci::Position MovePositionOutsideChar(Sci::Position pos, Sci::Position moveDir) const noexcept override {
 		return pdoc->MovePositionOutsideChar(pos, moveDir, true);
 	}
-
-	Sci::Position NextPosition(Sci::Position pos, int moveDir) const noexcept override {
-		return pdoc->NextPosition(pos, moveDir);
-	}
-
-	Sci::Position ExtendWordSelect(Sci::Position pos, int delta) const noexcept override {
-		return pdoc->ExtendWordSelect(pos, delta, true);
-	}
 };
 
 #ifndef NO_CXX11_REGEX
@@ -3365,11 +3345,13 @@ Sci::Position Cxx11RegexFindText(const Document *doc, Sci::Position minPos, Sci:
 #endif // NO_CXX11_REGEX
 
 Sci::Position BuiltinRegex::FindText(const Document *doc, Sci::Position minPos, Sci::Position maxPos, const char *s,
-	bool caseSensitive, [[maybe_unused]] FindOption flags, Sci::Position *length) {
+	bool caseSensitive, FindOption flags, Sci::Position *length) {
 
 #ifndef NO_CXX11_REGEX
-	return Cxx11RegexFindText(doc, minPos, maxPos, s, caseSensitive, length, search);
-#else
+	if (FlagSet(flags, FindOption::Cxx11RegEx)) {
+		return Cxx11RegexFindText(doc, minPos, maxPos, s, caseSensitive, length, search);
+	}
+#endif
 
 	const RESearchRange resr(doc, minPos, maxPos);
 
@@ -3442,7 +3424,6 @@ Sci::Position BuiltinRegex::FindText(const Document *doc, Sci::Position minPos,
 	}
 	*length = lenRet;
 	return pos;
-#endif // NO_CXX11_REGEX
 }
 
 const char *BuiltinRegex::SubstituteByPosition(Document *doc, const char *text, Sci::Position *length) {

scintilla/src/Editor.cxx

@@ -4181,8 +4181,12 @@ Sci::Position Editor::FindTextFull(
 			static_cast<FindOption>(wParam),
 			&lengthFound);
 		if (pos >= 0) {
+			Sci::Position endPos = pos + lengthFound;
+			if (wParam & static_cast<int>(FindOption::MatchToWordEnd)) {
+				endPos = pdoc->ExtendWordSelect(endPos, 1, true);
+			}
 			ft->chrgText.cpMin = pos;
-			ft->chrgText.cpMax = pos + lengthFound;
+			ft->chrgText.cpMax = endPos;
 		}
 		return pos;
 	} catch (const RegexError &) {

scintilla/src/RESearch.cxx

@@ -2,7 +2,7 @@
 /** @file RESearch.cxx
  ** Regular expression search library.
  **/
-#if !defined(SCI_OWNREGEX) && defined(NO_CXX11_REGEX)
+#ifndef SCI_OWNREGEX
 
 /*
  * regex - Regular expression pattern matching and replacement
@@ -238,12 +238,6 @@ using namespace Scintilla::Internal;
 #define CLQ     12 /* 0 to 1 closure */
 #define LCLO    13 /* lazy closure */
 
-// experimental
-#define EXP_MATCH_WORD_START	14
-#define EXP_MATCH_WORD_END		15
-#define EXP_MATCH_TO_WORD_END		16
-#define EXP_MATCH_TO_WORD_END_OPT	17
-
 #define END     0
 
 /*
@@ -640,11 +634,6 @@ const char *RESearch::DoCompile(const char *pattern, Sci::Position length, bool
 				break;
 			}
 
-			if (*p == '?' && *lp == EXP_MATCH_TO_WORD_END) {
-				*lp = EXP_MATCH_TO_WORD_END_OPT;
-				break;
-			}
-
 			if (*p == '+') {
 				for (sp = mp; lp < sp; lp++) {
 					*mp++ = *lp;
@@ -675,17 +664,6 @@ const char *RESearch::DoCompile(const char *pattern, Sci::Position length, bool
 					return badpat("Null pattern inside \\<\\>");
 				*mp++ = EOW;
 				break;
-			case 'h':
-				*mp++ = EXP_MATCH_WORD_START;
-				break;
-			case 'H':
-				if (*sp == EXP_MATCH_WORD_START)
-					return badpat("Null pattern inside \\h\\H");
-				*mp++ = EXP_MATCH_WORD_END;
-				break;
-			case 'i':
-				*mp++ = EXP_MATCH_TO_WORD_END;
-				break;
 			case '1':
 			case '2':
 			case '3':
@@ -841,11 +819,10 @@ int RESearch::Execute(const CharacterIndexer &ci, Sci::Position lp, Sci::Positio
 	}
 	default:			/* regular matching all the way. */
 		while (lp < endp) {
-			Sci::Position offset = 1;
-			ep = PMatch(ci, lp, endp, ap, 1, &offset);
+			ep = PMatch(ci, lp, endp, ap);
 			if (ep != NOTFOUND)
 				break;
-			lp += offset;
+			lp++;
 		}
 		break;
 	case END:			/* munged automaton. fail always */
@@ -898,7 +875,7 @@ int RESearch::Execute(const CharacterIndexer &ci, Sci::Position lp, Sci::Positio
 #define CHRSKIP 3	/* [CLO] CHR chr END      */
 #define CCLSKIP 34	/* [CLO] CCL 32 bytes END */
 
-Sci::Position RESearch::PMatch(const CharacterIndexer &ci, Sci::Position lp, Sci::Position endp, char *ap, int moveDir, Sci::Position *offset) {
+Sci::Position RESearch::PMatch(const CharacterIndexer &ci, Sci::Position lp, Sci::Position endp, char *ap) {
 	uint8_t op;
 
 	while ((op = *ap++) != END) {
@@ -943,39 +920,6 @@ Sci::Position RESearch::PMatch(const CharacterIndexer &ci, Sci::Position lp, Sci
 			if (lp == bol || !iswordc(ci.CharAt(lp - 1)) || iswordc(ci.CharAt(lp)))
 				return NOTFOUND;
 			break;
-		case EXP_MATCH_WORD_START:
-			if (!ci.IsWordStartAt(lp)) {
-				if (offset) {
-					Sci::Position e = ci.MovePositionOutsideChar(lp, moveDir);
-					e = (e == lp) ? ci.NextPosition(lp, moveDir) - lp : e - lp;
-					*offset = (e == 0) ? moveDir : e;
-				}
-				return NOTFOUND;
-			}
-			break;
-		case EXP_MATCH_WORD_END:
-			if (lp == bol || !ci.IsWordEndAt(lp)) {
-				if (offset) {
-					Sci::Position e = ci.MovePositionOutsideChar(lp, moveDir);
-					e = (e == lp) ? ci.NextPosition(lp, moveDir) - lp : e - lp;
-					*offset = (e == 0) ? moveDir : e;
-				}
-				return NOTFOUND;
-			}
-			break;
-		case EXP_MATCH_TO_WORD_END:
-		case EXP_MATCH_TO_WORD_END_OPT: {
-			Sci::Position e = ci.ExtendWordSelect(lp, moveDir);
-			if ((e == lp && op != EXP_MATCH_TO_WORD_END_OPT) || !ci.IsWordEndAt(e)) {
-				if (offset) {
-					e = (e == lp) ? ci.NextPosition(lp, moveDir) - lp : e - lp;
-					*offset = (e == 0) ? moveDir : e;
-				}
-				return NOTFOUND;
-			}
-			lp = e;
-		}
-		break;
 		case REF: {
 			const int n = static_cast<uint8_t>(*ap++);
 			Sci::Position bp = bopat[n];	/* beginning of subpat... */
@@ -1029,15 +973,14 @@ Sci::Position RESearch::PMatch(const CharacterIndexer &ci, Sci::Position lp, Sci
 			Sci::Position llp = lp;	/* lazy lp for LCLO       */
 			Sci::Position e = NOTFOUND; /* extra pointer for CLO  */
 			while (llp >= are) {
-				Sci::Position qoff = -1;
-				const Sci::Position q = PMatch(ci, llp, endp, ap, -1, &qoff);
+				const Sci::Position q = PMatch(ci, llp, endp, ap);
 				if (q != NOTFOUND) {
 					e = q;
 					lp = llp;
 					if (op != LCLO) return e;
 				}
 				if (*ap == END) return e;
-				llp += qoff;
+				--llp;
 			}
 			if (*ap == EOT)
 				PMatch(ci, lp, endp, ap);

scintilla/src/RESearch.h

@@ -6,18 +6,14 @@
 // Based on the work of Ozan S. Yigit.
 // This file is in the public domain.
 #pragma once
-#if !defined(SCI_OWNREGEX) && defined(NO_CXX11_REGEX)
+#ifndef SCI_OWNREGEX
 
 namespace Scintilla::Internal {
 
 class CharacterIndexer {
 public:
 	virtual char CharAt(Sci::Position index) const noexcept = 0;
-	virtual bool IsWordStartAt(Sci::Position pos) const noexcept = 0;
-	virtual bool IsWordEndAt(Sci::Position pos) const noexcept = 0;
 	virtual Sci::Position MovePositionOutsideChar(Sci::Position pos, Sci::Position moveDir) const noexcept = 0;
-	virtual Sci::Position NextPosition(Sci::Position pos, int moveDir) const noexcept = 0;
-	virtual Sci::Position ExtendWordSelect(Sci::Position pos, int delta) const noexcept = 0;
 };
 
 class RESearch {
@@ -49,7 +45,7 @@ class RESearch {
 	int GetBackslashExpression(const char *pattern, int &incr) noexcept;
 
 	const char *DoCompile(const char *pattern, Sci::Position length, bool caseSensitive, bool posix) noexcept;
-	Sci::Position PMatch(const CharacterIndexer &ci, Sci::Position lp, Sci::Position endp, char *ap, int moveDir = 1, Sci::Position *offset = nullptr);
+	Sci::Position PMatch(const CharacterIndexer &ci, Sci::Position lp, Sci::Position endp, char *ap);
 
 	Sci::Position bol;
 	Sci::Position tagstk[MAXTAG];  /* subpat tag stack */