From 4f94b49a318c1130e1c9fe82eca2ad4a894604f0 Mon Sep 17 00:00:00 2001
From: Teodor Sigaev <teodor@sigaev.ru>
Date: Wed, 7 Dec 2005 13:12:54 +0000
Subject: [PATCH] Improve word parser.  - allow ~ in filenames  - -8.2.1 now is
 '-' and '8.2.1' instead of '-8.2' '.' '3'  - '.text' now is not a file

---
 contrib/tsearch2/wordparser/parser.c | 48 +++++++++++++++++++++++++++-
 contrib/tsearch2/wordparser/parser.h |  4 +++
 2 files changed, 51 insertions(+), 1 deletion(-)

diff --git a/contrib/tsearch2/wordparser/parser.c b/contrib/tsearch2/wordparser/parser.c
index deccdb284a..37f020ef68 100644
--- a/contrib/tsearch2/wordparser/parser.c
+++ b/contrib/tsearch2/wordparser/parser.c
@@ -243,6 +243,15 @@ SpecialHyphen(TParser * prs)
 	prs->state->poschar -= prs->state->lencharlexeme;
 }
 
+static void
+SpecialVerVersion(TParser * prs)
+{
+	prs->state->posbyte -= prs->state->lenbytelexeme;
+	prs->state->poschar -= prs->state->lencharlexeme;
+	prs->state->lenbytelexeme = 0;
+	prs->state->lencharlexeme = 0;
+}
+
 static int
 p_isstophost(TParser * prs)
 {
@@ -326,8 +335,9 @@ static TParserStateActionItem actionTPS_Base[] = {
 	{p_iseqC, '-', A_PUSH, TPS_InSignedIntFirst, 0, NULL},
 	{p_iseqC, '+', A_PUSH, TPS_InSignedIntFirst, 0, NULL},
 	{p_iseqC, '&', A_PUSH, TPS_InHTMLEntityFirst, 0, NULL},
+	{p_iseqC, '~', A_PUSH, TPS_InFileTwiddle, 0, NULL},
 	{p_iseqC, '/', A_PUSH, TPS_InFileFirst, 0, NULL},
-	{p_iseqC, '.', A_PUSH, TPS_InPathFirst, 0, NULL},
+	{p_iseqC, '.', A_PUSH, TPS_InPathFirstFirst, 0, NULL},
 	{NULL, 0, A_NEXT, TPS_InSpace, 0, NULL}
 };
 
@@ -429,11 +439,25 @@ static TParserStateActionItem actionTPS_InDecimalFirst[] = {
 static TParserStateActionItem actionTPS_InDecimal[] = {
 	{p_isEOF, 0, A_BINGO, TPS_Base, DECIMAL, NULL},
 	{p_isdigit, 0, A_NEXT, TPS_InDecimal, 0, NULL},
+	{p_iseqC, '.', A_PUSH, TPS_InVerVersion, 0, NULL},
 	{p_iseqC, 'e', A_PUSH, TPS_InMantissaFirst, 0, NULL},
 	{p_iseqC, 'E', A_PUSH, TPS_InMantissaFirst, 0, NULL},
 	{NULL, 0, A_BINGO, TPS_Base, DECIMAL, NULL}
 };
 
+static TParserStateActionItem actionTPS_InVerVersion[] = {
+	{p_isEOF, 0, A_POP, TPS_Null, 0, NULL},
+	{p_isdigit, 0, A_RERUN, TPS_InSVerVersion, 0, SpecialVerVersion},
+	{NULL, 0, A_POP, TPS_Null, 0, NULL}
+};
+
+static TParserStateActionItem actionTPS_InSVerVersion[] = {
+	{p_isEOF, 0, A_POP, TPS_Null, 0, NULL},
+	{p_isdigit, 0, A_BINGO, TPS_InUnsignedInt, SPACE, NULL},
+	{NULL, 0, A_NEXT, TPS_Null, 0, NULL}
+};
+
+
 static TParserStateActionItem actionTPS_InVersionFirst[] = {
 	{p_isEOF, 0, A_POP, TPS_Null, 0, NULL},
 	{p_isdigit, 0, A_CLEAR, TPS_InVersion, 0, NULL},
@@ -537,6 +561,7 @@ static TParserStateActionItem actionTPS_InTag[] = {
 	{p_iseqC, '&', A_NEXT, TPS_Null, 0, NULL},
 	{p_iseqC, '?', A_NEXT, TPS_Null, 0, NULL},
 	{p_iseqC, '%', A_NEXT, TPS_Null, 0, NULL},
+	{p_iseqC, '~', A_NEXT, TPS_Null, 0, NULL},
 	{p_isspace, 0, A_NEXT, TPS_Null, 0, SpecialTags},
 	{NULL, 0, A_POP, TPS_Null, 0, NULL}
 };
@@ -676,6 +701,16 @@ static TParserStateActionItem actionTPS_InFileFirst[] = {
 	{p_iseqC, '.', A_NEXT, TPS_InPathFirst, 0, NULL},
 	{p_iseqC, '_', A_NEXT, TPS_InFile, 0, NULL},
 	{p_iseqC, '?', A_PUSH, TPS_InURIFirst, 0, NULL},
+	{p_iseqC, '~', A_PUSH, TPS_InFileTwiddle, 0, NULL},
+	{NULL, 0, A_POP, TPS_Null, 0, NULL}
+};
+
+static TParserStateActionItem actionTPS_InFileTwiddle[] = {
+	{p_isEOF, 0, A_POP, TPS_Null, 0, NULL},
+	{p_islatin, 0, A_NEXT, TPS_InFile, 0, NULL},
+	{p_isdigit, 0, A_NEXT, TPS_InFile, 0, NULL},
+	{p_iseqC, '_', A_NEXT, TPS_InFile, 0, NULL},
+	{p_iseqC, '/', A_NEXT, TPS_InFileFirst, 0, NULL},
 	{NULL, 0, A_POP, TPS_Null, 0, NULL}
 };
 
@@ -689,6 +724,13 @@ static TParserStateActionItem actionTPS_InPathFirst[] = {
 	{NULL, 0, A_POP, TPS_Null, 0, NULL}
 };
 
+static TParserStateActionItem actionTPS_InPathFirstFirst[] = {
+	{p_isEOF, 0, A_POP, TPS_Null, 0, NULL},
+	{p_iseqC, '.', A_NEXT, TPS_InPathSecond, 0, NULL},
+	{p_iseqC, '/', A_NEXT, TPS_InFileFirst, 0, NULL},
+	{NULL, 0, A_POP, TPS_Null, 0, NULL}
+};
+
 static TParserStateActionItem actionTPS_InPathSecond[] = {
 	{p_isEOF, 0, A_BINGO|A_CLEAR, TPS_Base, FILEPATH, NULL},
 	{p_iseqC, '/', A_NEXT|A_PUSH, TPS_InFileFirst, 0, NULL},
@@ -920,6 +962,8 @@ static const TParserStateAction Actions[] = {
 	{TPS_InUDecimal, actionTPS_InUDecimal},
 	{TPS_InDecimalFirst, actionTPS_InDecimalFirst},
 	{TPS_InDecimal, actionTPS_InDecimal},
+	{TPS_InVerVersion, actionTPS_InVerVersion},
+	{TPS_InSVerVersion, actionTPS_InSVerVersion},
 	{TPS_InVersionFirst, actionTPS_InVersionFirst},
 	{TPS_InVersion, actionTPS_InVersion},
 	{TPS_InMantissaFirst, actionTPS_InMantissaFirst},
@@ -953,7 +997,9 @@ static const TParserStateAction Actions[] = {
 	{TPS_InHost, actionTPS_InHost},
 	{TPS_InEmail, actionTPS_InEmail},
 	{TPS_InFileFirst, actionTPS_InFileFirst},
+	{TPS_InFileTwiddle, actionTPS_InFileTwiddle},
 	{TPS_InPathFirst, actionTPS_InPathFirst},
+	{TPS_InPathFirstFirst, actionTPS_InPathFirstFirst},
 	{TPS_InPathSecond, actionTPS_InPathSecond},
 	{TPS_InFile, actionTPS_InFile},
 	{TPS_InFileNext, actionTPS_InFileNext},
diff --git a/contrib/tsearch2/wordparser/parser.h b/contrib/tsearch2/wordparser/parser.h
index 9cdd141efd..3f7962feea 100644
--- a/contrib/tsearch2/wordparser/parser.h
+++ b/contrib/tsearch2/wordparser/parser.h
@@ -19,6 +19,8 @@ typedef enum
 	TPS_InUDecimal,
 	TPS_InDecimalFirst,
 	TPS_InDecimal,
+	TPS_InVerVersion,
+	TPS_InSVerVersion,
 	TPS_InVersionFirst,
 	TPS_InVersion,
 	TPS_InMantissaFirst,
@@ -52,7 +54,9 @@ typedef enum
 	TPS_InHost,
 	TPS_InEmail,
 	TPS_InFileFirst,
+	TPS_InFileTwiddle,
 	TPS_InPathFirst,
+	TPS_InPathFirstFirst,
 	TPS_InPathSecond,
 	TPS_InFile,
 	TPS_InFileNext,
-- 
GitLab