patch 9.1.0645: regex: wrong match when searching multi-byte char case-insensitivev9.1.0645

Problem:  regex: wrong match when searching multi-byte char
          case-insensitive (diffsetter)
Solution: Apply proper case-folding for characters and search-string

This patch does the following 4 things:

1) When the regexp engine compares two utf-8 codepoints case
   insensitive it may match an adjacent character, because it assumes
   it can step over as many bytes as the pattern contains.

   This however is not necessarily true because of case-folding, a
   multi-byte UTF-8 character can be considered equal to some
   single-byte value.

   Let's consider the pattern 'ſ' and the string 's'. When comparing and
   ignoring case, the single character 's' matches, and since it matches
   Vim will try to step over the match (by the amount of bytes of the
   pattern), assuming that since it matches, the length of both strings is
   the same.

   However in that case, it should only step over the single byte value
   's' by 1 byte and try to start matching after it again. So for the
   backtracking engine we need to ensure:
   * we try to match the correct length for the pattern and the text
   * in case of a match, we step over it correctly

   There is one tricky thing for the backtracing engine. We also need to
   calculate correctly the number of bytes to compare the 2 different
   utf-8 strings s1 and s2. So we will count the number of characters in
   s1 that the byte len specified. Then we count the number of bytes to
   step over the same number of characters in string s2 and then we can
   correctly compare the 2 utf-8 strings.

2) A similar thing can happen for the NFA engine, when skipping to the
   next character to test for a match. We are skipping over the regstart
   pointer, however we do not consider the case that because of
   case-folding we may need to adjust the number of bytes to skip over.
   So this needs to be adjusted in find_match_text() as well.

3) A related issue turned out, when prog->match_text is actually empty.
   In that case we should try to find the next match and skip this
   condition.

4) When comparing characters using collections, we must also apply case
   folding to each character in the collection and not just to the
   current character from the search string.  This doesn't apply to the
   NFA engine, because internally it converts collections to branches
   [abc] -> a\|b\|c

fixes: #14294
closes: #14756

Signed-off-by: Christian Brabandt <cb@256bit.org>

9 files changed, 109 insertions, 15 deletions

diff --git a/runtime/doc/version9.txt b/runtime/doc/version9.txt
index e8e9194dc5..73e54a1cf3 100644
--- a/runtime/doc/version9.txt
+++ b/runtime/doc/version9.txt
@@ -1,4 +1,4 @@
-*version9.txt*  For Vim version 9.1.  Last change: 2024 Jul 28
+*version9.txt*  For Vim version 9.1.  Last change: 2024 Jul 30
 
 
 		  VIM REFERENCE MANUAL    by Bram Moolenaar
@@ -41590,6 +41590,8 @@ Changed~
   behaviour/inconsistency (see |d-special| and |cw|).
 - allow to specify additional attributes in the completion menu (allows to
   mark deprecated attributes from LSP server) |complete-items|
+- the regex engines match correctly case-insensitive multi-byte characters
+  (and apply proper case folding)
 
 							*added-9.2*
 Added ~
diff --git a/src/mbyte.c b/src/mbyte.c
index a68ba7be3d..d8c47acdd1 100644
--- a/src/mbyte.c
+++ b/src/mbyte.c
@@ -3801,6 +3801,15 @@ utf_strnicmp(
  * two characters otherwise.
  */
     int
+mb_strnicmp2(char_u *s1, char_u *s2, size_t n1, size_t n2)
+{
+    if (n1 == n2 || !enc_utf8)
+	return mb_strnicmp(s1, s2, n1);
+    else
+	return utf_strnicmp(s1, s2, n1, n2);
+}
+
+    int
 mb_strnicmp(char_u *s1, char_u *s2, size_t nn)
 {
     int		i, l;
diff --git a/src/proto/mbyte.pro b/src/proto/mbyte.pro
index c57c94c8ae..bb976e3bf8 100644
--- a/src/proto/mbyte.pro
+++ b/src/proto/mbyte.pro
@@ -48,6 +48,7 @@ int utf_islower(int a);
 int utf_tolower(int a);
 int utf_isupper(int a);
 int mb_strnicmp(char_u *s1, char_u *s2, size_t nn);
+int mb_strnicmp2(char_u *s1, char_u *s2, size_t n1, size_t n2);
 void show_utf8(void);
 int latin_head_off(char_u *base, char_u *p);
 int dbcs_screen_head_off(char_u *base, char_u *p);
diff --git a/src/regexp.c b/src/regexp.c
index ff201d9ffe..a1b080e7d9 100644
--- a/src/regexp.c
+++ b/src/regexp.c
@@ -1729,7 +1729,9 @@ mb_decompose(int c, int *c1, int *c2, int *c3)
 /*
  * Compare two strings, ignore case if rex.reg_ic set.
  * Return 0 if strings match, non-zero otherwise.
- * Correct the length "*n" when composing characters are ignored.
+ * Correct the length "*n" when composing characters are ignored
+ * or for utf8 when both utf codepoints are considered equal because of
+ * case-folding but have different length (e.g. 's' and 'ſ')
  */
     static int
 cstrncmp(char_u *s1, char_u *s2, int *n)
@@ -1738,6 +1740,29 @@ cstrncmp(char_u *s1, char_u *s2, int *n)
 
     if (!rex.reg_ic)
 	result = STRNCMP(s1, s2, *n);
+    else if (enc_utf8)
+    {
+	char_u *p = s1;
+	size_t n2 = 0;
+	int n1 = *n;
+	// count the number of characters for byte-length of s1
+	while (n1 > 0 && *p != NUL)
+	{
+	    n1 -= mb_ptr2len(s1);
+	    MB_PTR_ADV(p);
+	    n2++;
+	}
+	// count the number of bytes to advance the same number of chars for s2
+	p = s2;
+	while (n2-- > 0 && *p != NUL)
+	    MB_PTR_ADV(p);
+
+	n2 = p - s2;
+
+	result = MB_STRNICMP2(s1, s2, *n, n2);
+	if (result == 0 && (int)n2 < *n)
+	    *n = n2;
+    }
     else
 	result = MB_STRNICMP(s1, s2, *n);
 
@@ -1787,7 +1812,7 @@ cstrncmp(char_u *s1, char_u *s2, int *n)
 cstrchr(char_u *s, int c)
 {
     char_u	*p;
-    int		cc;
+    int		cc, lc;
 
     if (!rex.reg_ic || (!enc_utf8 && mb_char2len(c) > 1))
 	return vim_strchr(s, c);
@@ -1796,26 +1821,35 @@ cstrchr(char_u *s, int c)
     // faster (esp. when using MS Visual C++!).
     // For UTF-8 need to use folded case.
     if (enc_utf8 && c > 0x80)
+    {
 	cc = utf_fold(c);
+	lc = cc;
+    }
     else
-	 if (MB_ISUPPER(c))
-	cc = MB_TOLOWER(c);
-    else if (MB_ISLOWER(c))
-	cc = MB_TOUPPER(c);
-    else
-	return vim_strchr(s, c);
+	if (MB_ISUPPER(c))
+	{
+	    cc = MB_TOLOWER(c);
+	    lc = cc;
+	}
+	else if (MB_ISLOWER(c))
+	{
+	    cc = MB_TOUPPER(c);
+	    lc = c;
+	}
+	else
+	    return vim_strchr(s, c);
 
     if (has_mbyte)
     {
 	for (p = s; *p != NUL; p += (*mb_ptr2len)(p))
 	{
-	    if (enc_utf8 && c > 0x80)
+	    int uc = utf_ptr2char(p);
+	    if (enc_utf8 && (c > 0x80 || uc > 0x80))
 	    {
-		int uc = utf_ptr2char(p);
-
 		// Do not match an illegal byte.  E.g. 0xff matches 0xc3 0xbf,
 		// not 0xff.
-		if ((uc < 0x80 || uc != *p) && utf_fold(uc) == cc)
+		// compare with lower case of the character
+		if ((uc < 0x80 || uc != *p) && utf_fold(uc) == lc)
 		    return p;
 	    }
 	    else if (*p == c || *p == cc)
diff --git a/src/regexp_bt.c b/src/regexp_bt.c
index 5452dda0f6..16dac730de 100644
--- a/src/regexp_bt.c
+++ b/src/regexp_bt.c
@@ -3823,6 +3823,14 @@ regmatch(
 			}
 		    }
 		}
+		else if (enc_utf8)
+		{
+		    if (cstrncmp(opnd, rex.input, &len) != 0)
+		    {
+			status = RA_NOMATCH;
+			break;
+		    }
+		}
 		else
 		    for (i = 0; i < len; ++i)
 			if (opnd[i] != rex.input[i])
diff --git a/src/regexp_nfa.c b/src/regexp_nfa.c
index 4f07a21d5d..6db4134628 100644
--- a/src/regexp_nfa.c
+++ b/src/regexp_nfa.c
@@ -5666,7 +5666,12 @@ find_match_text(colnr_T *startcol, int regstart, char_u *match_text)
     for (;;)
     {
 	match = TRUE;
-	len2 = MB_CHAR2LEN(regstart); // skip regstart
+	// skip regstart
+	len2 = MB_CHAR2LEN(regstart);
+	if (enc_utf8 && len2 > 1 && MB_CHAR2LEN(PTR2CHAR(rex.line + col)) != len2)
+	    // because of case-folding of the previously matched text, we may need
+	    // to skip fewer bytes than mb_char2len(regstart)
+	    len2 = mb_char2len(utf_fold(regstart));
 	for (len1 = 0; match_text[len1] != NUL; len1 += MB_CHAR2LEN(c1))
 	{
 	    c1 = PTR2CHAR(match_text + len1);
@@ -7502,7 +7507,7 @@ nfa_regexec_both(
 
 	// If match_text is set it contains the full text that must match.
 	// Nothing else to try. Doesn't handle combining chars well.
-	if (prog->match_text != NULL && !rex.reg_icombine)
+	if (prog->match_text != NULL && *prog->match_text != NUL && !rex.reg_icombine)
 	{
 	    retval = find_match_text(&col, prog->regstart, prog->match_text);
 	    if (REG_MULTI)
diff --git a/src/testdir/test_regexp_utf8.vim b/src/testdir/test_regexp_utf8.vim
index bc705441e7..51c0984adb 100644
--- a/src/testdir/test_regexp_utf8.vim
+++ b/src/testdir/test_regexp_utf8.vim
@@ -587,4 +587,36 @@ func Test_combining_chars_in_collection()
   bw!
 endfunc
 
+func Test_search_multibyte_match_ascii()
+  new
+  " Match single 'ſ' and 's'
+  call setline(1,  'das abc heraus abc ſich abc ſind')
+  for i in range(0, 2)
+    exe "set re="..i
+    let ic_match = matchbufline('%', '\c\%u17f', 1, '$')->mapnew({idx, val -> val.text})
+    let noic_match = matchbufline('%', '\C\%u17f', 1, '$')->mapnew({idx, val -> val.text})
+    call assert_equal(['s', 's', 'ſ','ſ'], ic_match, "Ignorecase Regex-engine: " .. &re)
+    call assert_equal(['ſ','ſ'], noic_match, "No-Ignorecase Regex-engine: " .. &re)
+  endfor
+  " Match several 'ſſ' and 'ss'
+  call setline(1,  'das abc herauss abc ſſich abc ſind')
+  for i in range(0, 2)
+    exe "set re="..i
+    let ic_match = matchbufline('%', '\c\%u17f\%u17f', 1, '$')->mapnew({idx, val -> val.text})
+    let noic_match = matchbufline('%', '\C\%u17f\%u17f', 1, '$')->mapnew({idx, val -> val.text})
+    let ic_match2 = matchbufline('%', '\c\%u17f\+', 1, '$')->mapnew({idx, val -> val.text})
+    let noic_match2 = matchbufline('%', '\C\%u17f\+', 1, '$')->mapnew({idx, val -> val.text})
+    let ic_match3 = matchbufline('%', '\c[\u17f]\+', 1, '$')->mapnew({idx, val -> val.text})
+    let noic_match3 = matchbufline('%', '\C[\u17f]\+', 1, '$')->mapnew({idx, val -> val.text})
+
+    call assert_equal(['ss', 'ſſ'], ic_match, "Ignorecase Regex-engine: " .. &re)
+    call assert_equal(['ſſ'], noic_match, "No-Ignorecase Regex-engine: " .. &re)
+    call assert_equal(['s', 'ss', 'ſſ', 'ſ'], ic_match2, "Ignorecase Regex-engine: " .. &re)
+    call assert_equal(['ſſ','ſ'], noic_match2, "No-Ignorecase Regex-engine: " .. &re)
+    call assert_equal(['s', 'ss', 'ſſ', 'ſ'], ic_match3, "Ignorecase Collection Regex-engine: " .. &re)
+    call assert_equal(['ſſ','ſ'], noic_match3, "No-Ignorecase Collection Regex-engine: " .. &re)
+  endfor
+  bw!
+endfunc
+
 " vim: shiftwidth=2 sts=2 expandtab
diff --git a/src/version.c b/src/version.c
index a2662861e3..f19e9415e9 100644
--- a/src/version.c
+++ b/src/version.c
@@ -705,6 +705,8 @@ static char *(features[]) =
 static int included_patches[] =
 {   /* Add new patch number below this line */
 /**/
+    645,
+/**/
     644,
 /**/
     643,
diff --git a/src/vim.h b/src/vim.h
index c022f2e7f4..9c1434cc63 100644
--- a/src/vim.h
+++ b/src/vim.h
@@ -1769,6 +1769,7 @@ void *vim_memset(void *, int, size_t);
 
 # define MB_STRICMP(d, s)	mb_strnicmp((char_u *)(d), (char_u *)(s), (int)MAXCOL)
 # define MB_STRNICMP(d, s, n)	mb_strnicmp((char_u *)(d), (char_u *)(s), (int)(n))
+# define MB_STRNICMP2(d, s, n1, n2)	mb_strnicmp2((char_u *)(d), (char_u *)(s), (n1), (n2))
 
 #define STRCAT(d, s)	    strcat((char *)(d), (char *)(s))
 #define STRNCAT(d, s, n)    strncat((char *)(d), (char *)(s), (size_t)(n))