From e251c6463913bcd7c9c40b4dc168010f6de68b5a Mon Sep 17 00:00:00 2001
From: Thomas Voss <mail@thomasvoss.com>
Date: Tue, 5 Nov 2024 21:09:53 +0100
Subject: Check for whitespace after regex using islbrk()

---
 src/main.c | 11 +++++------
 src/util.c | 14 ++++++++++++++
 src/util.h |  3 +++
 src/work.c | 13 -------------
 4 files changed, 22 insertions(+), 19 deletions(-)

diff --git a/src/main.c b/src/main.c
index 89e3007..120a88e 100644
--- a/src/main.c
+++ b/src/main.c
@@ -330,14 +330,13 @@ pattern_comp(u8view_t pat)
 				reopts |= PCRE2_UCP;
 
 			for (;;) {
-				if (ucsgnext(&g, &pat) == 0)
+				if (ucsgnext(&g, &pat) == 0 || islbrk(g))
 					break;
-				if (g.len != 1)
-					goto bad_flag;
-				if (uprop_is_pat_ws(*g.p))
+				int w = ucsnext(&ch, &g);
+				if (uprop_is_pat_ws(ch))
 					break;
 
-				switch (*g.p) {
+				switch (ch) {
 				case 'i': reopts |=  PCRE2_CASELESS; break;
 				case 'I': reopts &= ~PCRE2_CASELESS; break;
 				case 'l': reopts |=  PCRE2_LITERAL;  break;
@@ -345,7 +344,7 @@ pattern_comp(u8view_t pat)
 				case 'u': reopts |=  PCRE2_UCP;      break;
 				case 'U': reopts &= ~PCRE2_UCP;      break;
 				default:
-				bad_flag:
+					VSHFT(&g, -w);
 					cerr(EXIT_FATAL, "Unknown regex flag %s%.*s%s",
 					     lquot, SV_PRI_ARGS(g), rquot);
 				}
diff --git a/src/util.c b/src/util.c
index b5b1f8a..a946119 100644
--- a/src/util.c
+++ b/src/util.c
@@ -4,6 +4,8 @@
 #include <stdlib.h>
 
 #include <errors.h>
+#include <macros.h>
+#include <mbstring.h>
 #include <pcre2.h>
 
 #include "exitcodes.h"
@@ -47,4 +49,16 @@ getenv_posnum(const char *ev, int fallback)
 			return (int)n;
 	}
 	return fallback;
+}
+
+bool
+islbrk(u8view_t g)
+{
+	return ucseq(g, U8("\n"))
+	    || ucseq(g, U8("\v"))
+	    || ucseq(g, U8("\f"))
+	    || ucseq(g, U8("\r\n"))
+	    || ucseq(g, U8("\x85"))
+	    || ucseq(g, U8("\u2028"))
+	    || ucseq(g, U8("\u2029"));
 }
\ No newline at end of file
diff --git a/src/util.h b/src/util.h
index 705d0ee..ab3e6d6 100644
--- a/src/util.h
+++ b/src/util.h
@@ -1,7 +1,10 @@
 #ifndef GRAB_UTIL_H
 #define GRAB_UTIL_H
 
+#include <mbstring.h>
+
 int getenv_posnum(const char *ev, int fallback);
 void pcre2_bitch_and_die(int ec, const char *fmt, ...);
+bool islbrk(u8view_t g);
 
 #endif /* !GRAB_UTIL_H */
\ No newline at end of file
diff --git a/src/work.c b/src/work.c
index c8a3e21..5bc8b24 100644
--- a/src/work.c
+++ b/src/work.c
@@ -41,7 +41,6 @@ typedef struct {
 
 static void compute_pos(const char8_t *p, pos_state_t *ps);
 static bool has_lbrk_p(u8view_t sv);
-static bool islbrk(u8view_t g);
 static int svposcmp(const void *a, const void *b);
 static void write_match_to_buffer(u8view_t sv, u8view_t *hl);
 
@@ -482,18 +481,6 @@ has_lbrk_p(u8view_t sv)
 	return false;
 }
 
-bool
-islbrk(u8view_t g)
-{
-	return ucseq(g, U8("\n"))
-	    || ucseq(g, U8("\v"))
-	    || ucseq(g, U8("\f"))
-	    || ucseq(g, U8("\r\n"))
-	    || ucseq(g, U8("\x85"))
-	    || ucseq(g, U8("\u2028"))
-	    || ucseq(g, U8("\u2029"));
-}
-
 int
 svposcmp(const void *a_, const void *b_)
 {
-- 
cgit v1.2.3