termux-packages/packages/edbrowse/pcre2.patch

From 1d9a2e966ddb5242d79e4b6e238124c4be81f6dd Mon Sep 17 00:00:00 2001
From: Karl Dahlke <eklhad@comcast.net>
Date: Thu, 30 Sep 2021 20:21:39 -0400
Subject: [PATCH] Upgrade from pcre1 to pcre2. Patch courtesy of Sylvain
 BERTRAND.

I only had to tweak it in a couple places.
---
 src/buffers.c | 74 ++++++++++++++++++++++++++++++++-------------------
 src/main.c    |  5 ----
 src/makefile  |  2 +-
 5 files changed, 54 insertions(+), 37 deletions(-)

diff --git a/src/buffers.c b/src/buffers.c
index 4324b4a6..b18a7674 100644
--- a/src/buffers.c
+++ b/src/buffers.c
@@ -11,7 +11,8 @@
 
 /* If this include file is missing, you need the pcre package,
  * and the pcre-devel package. */
-#include <pcre.h>
+#define PCRE2_CODE_UNIT_WIDTH 8
+#include <pcre2.h>
 static bool pcre_utf8_error_stop = false;
 
 #include <readline/readline.h>
@@ -3147,22 +3148,23 @@ regexpCheck(const char *line, bool isleft, bool ebmuck,
 
 /* regexp variables */
 static int re_count;
-static int re_vector[11 * 3];
-static pcre *re_cc;		/* compiled */
+static PCRE2_SIZE *re_vector;
+static pcre2_match_data *match_data;
+static pcre2_code *re_cc;	/* compiled */
 static bool re_utf8 = true;
 
 static void regexpCompile(const char *re, bool ci)
 {
 	static signed char try8 = 0;	/* 1 is utf8 on, -1 is utf8 off */
-	const char *re_error;
-	int re_offset;
+	int re_error;
+	PCRE2_SIZE re_offset;
 	int re_opt;
 
 top:
 /* Do we need PCRE_NO_AUTO_CAPTURE? */
 	re_opt = 0;
 	if (ci)
-		re_opt |= PCRE_CASELESS;
+		re_opt |= PCRE2_CASELESS;
 
 	if (re_utf8) {
 		if (cons_utf8 && !cw->binMode && try8 >= 0) {
@@ -3174,23 +3176,26 @@ static void regexpCompile(const char *re, bool ci)
 				}
 			}
 			try8 = 1;
-			re_opt |= PCRE_UTF8;
+			re_opt |= PCRE2_UTF;
 		}
 	}
 
-	re_cc = pcre_compile(re, re_opt, &re_error, &re_offset, 0);
-	if (!re_cc && try8 > 0 && strstr(re_error, "PCRE_UTF8 support")) {
+	re_cc = pcre2_compile((uchar*)re, PCRE2_ZERO_TERMINATED, re_opt, &re_error, &re_offset, 0);
+	if (!re_cc && try8 > 0 && re_error == PCRE2_ERROR_UTF_IS_DISABLED) {
 		i_puts(MSG_PcreUtf8);
 		try8 = -1;
 		goto top;
 	}
-	if (!re_cc && try8 > 0 && strstr(re_error, "invalid UTF-8 string")) {
+	if (!re_cc && try8 > 0 && (PCRE2_ERROR_UTF32_ERR2 <= re_error && re_error <= PCRE2_ERROR_UTF8_ERR1)) {
 		i_puts(MSG_BadUtf8String);
 	}
 
 	if (!re_cc)
-		setError(MSG_RexpError, re_error);
-}				/* regexpCompile */
+		setError(MSG_RexpError, "ERROR");
+	else
+// re_cc and match_data rise and fall together.
+		match_data = pcre2_match_data_create_from_pattern(re_cc, NULL);
+}
 
 /* Get the start or end of a range.
  * Pass the line containing the address. */
@@ -3246,7 +3251,8 @@ const char **split)
 			char *subject;
 			ln += incr;
 			if (!searchWrap && (ln == 0 || ln > cw->dol)) {
-				pcre_free(re_cc);
+				pcre2_match_data_free(match_data);
+				pcre2_code_free(re_cc);
 				setError(MSG_NotFound);
 				return false;
 			}
@@ -3256,26 +3262,31 @@ const char **split)
 				ln = cw->dol;
 			subject = (char *)fetchLine(ln, 1);
 			re_count =
-			    pcre_exec(re_cc, 0, subject,
+			    pcre2_match(re_cc, (uchar*)subject,
 				      pstLength((pst) subject) - 1, 0, 0,
-				      re_vector, 33);
+				      match_data, NULL);
+//  {uchar snork[300]; pcre2_get_error_message(re_count, snork, 300); puts(snork); }
+			re_vector = pcre2_get_ovector_pointer(match_data);
 			free(subject);
 // An error in evaluation is treated like text not found.
 // This usually happens because this particular line has bad binary, not utf8.
 			if (re_count < -1 && pcre_utf8_error_stop) {
-				pcre_free(re_cc);
+				pcre2_match_data_free(match_data);
+				pcre2_code_free(re_cc);
 				setError(MSG_RexpError2, ln);
 				return (globSub = false);
 			}
 			if (re_count >= 0)
 				break;
 			if (ln == cw->dot) {
-				pcre_free(re_cc);
+				pcre2_match_data_free(match_data);
+				pcre2_code_free(re_cc);
 				setError(MSG_NotFound);
 				return false;
 			}
 		}		/* loop over lines */
-		pcre_free(re_cc);
+		pcre2_match_data_free(match_data);
+		pcre2_code_free(re_cc);
 /* and ln is the line that matches */
 	}
 	/* Now add or subtract from this number */
@@ -3363,11 +3374,14 @@ static bool doGlobal(const char *line)
 	for (i = startRange; i <= endRange; ++i) {
 		char *subject = (char *)fetchLine(i, 1);
 		re_count =
-		    pcre_exec(re_cc, 0, subject, pstLength((pst) subject) - 1,
-			      0, 0, re_vector, 33);
+		    pcre2_match(re_cc, (uchar*)subject, pstLength((pst) subject) - 1,
+			      0, 0, match_data, NULL);
+		re_vector = pcre2_get_ovector_pointer(match_data);
+
 		free(subject);
 		if (re_count < -1 && pcre_utf8_error_stop) {
-			pcre_free(re_cc);
+  			pcre2_match_data_free(match_data);
+			pcre2_code_free(re_cc);
 			setError(MSG_RexpError2, i);
 			return false;
 		}
@@ -3377,7 +3391,8 @@ static bool doGlobal(const char *line)
 			cw->map[i].gflag = true;
 		}
 	}			/* loop over line */
-	pcre_free(re_cc);
+	pcre2_match_data_free(match_data);
+	pcre2_code_free(re_cc);
 
 	if (!gcnt) {
 		setError((cmd == 'v') + MSG_NoMatchG);
@@ -3477,7 +3492,8 @@ replaceText(const char *line, int len, const char *rhs,
 	while (true) {
 /* find the next match */
 		re_count =
-		    pcre_exec(re_cc, 0, line, len, offset, 0, re_vector, 33);
+		    pcre2_match(re_cc, (uchar*)line, len, offset, 0, match_data, NULL);
+		re_vector = pcre2_get_ovector_pointer(match_data);
 		if (re_count < -1 &&
 		    (pcre_utf8_error_stop || startRange == endRange)) {
 			setError(MSG_RexpError2, ln);
@@ -4092,8 +4108,10 @@ static int substituteText(const char *line)
 		breakLineResult = 0;
 	}			/* loop over lines in the range */
 
-	if (re_cc)
-		pcre_free(re_cc);
+	if (re_cc) {
+		pcre2_match_data_free(match_data);
+		pcre2_code_free(re_cc);
+	}
 
 	if (intFlag) {
 		setError(MSG_Interrupted);
@@ -4113,8 +4131,10 @@ static int substituteText(const char *line)
 	return true;
 
 abort:
-	if (re_cc)
-		pcre_free(re_cc);
+	if (re_cc) {
+		pcre2_match_data_free(match_data);
+		pcre2_code_free(re_cc);
+	}
 	nzFree(replaceString);
 /* we may have just freed the result of a breakline command */
 	breakLineResult = 0;
diff --git a/src/main.c b/src/main.c
index 862e5dc3..838b0ce0 100644
--- a/src/main.c
+++ b/src/main.c
@@ -6,7 +6,6 @@
 #include "eb.h"
 
 #include <pthread.h>
-#include <pcre.h>
 #include <signal.h>
 
 /* Define the globals that are declared in eb.h. */
@@ -577,10 +576,6 @@ int main(int argc, char **argv)
 	ttySaveSettings();
 	initializeReadline();
 
-/* Let's everybody use my malloc and free routines */
-	pcre_malloc = allocMem;
-	pcre_free = nzFree;
-
 	loadReplacements();
 
 	if (argc && stringEqual(argv[0], "-c")) {
diff --git a/src/makefile b/src/makefile
index e30866e9..82076114 100644
--- a/src/makefile
+++ b/src/makefile
@@ -11,7 +11,7 @@ CFLAGS += $(PLATFORM_CFLAGS)
 PERL !=		which perl
 TIDY_LIBS !=	pkg-config --libs tidy
 CURL_LIBS !=	pkg-config --libs libcurl
-PCRE_LIBS !=	pkg-config --libs libpcre
+PCRE_LIBS !=	pkg-config --libs libpcre2-8
 EDBR_ODBC_OBJS != ./make-helper.sh --odbc-objs "${BUILD_EDBR_ODBC}"
 EDBR_ODBC_LIBS != ./make-helper.sh --odbc-libs "${BUILD_EDBR_ODBC}"
 #  Set EBDEMIN to a nonempty string to support dynamic js deminimization