vim-patch:9.1.0296: regexp: engines do not handle case-folding well (#28259)

Problem:  Regex engines do not handle case-folding well
Solution: Correctly calculate byte length of characters to skip

When the regexp engine compares two utf-8 codepoints case insensitively
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' so 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

The same 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.

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.

fixes: vim/vim#14294
closes: vim/vim#14433

7a27c108e0

Co-authored-by: Christian Brabandt <cb@256bit.org>

runtime/doc/dev_vimpatch.txt

@@ -204,6 +204,7 @@ information.
   mb_ptr2char                                          utf_ptr2char
   mb_head_off                                          utf_head_off
   mb_tail_off                                          utf_cp_bounds
+  mb_strnicmp2                                         utf_strnicmp
   mb_lefthalve                                        grid_lefthalve
   mb_fix_col                                           grid_fix_col
   utf_off2cells                                       grid_off2cells

src/nvim/mbyte.c

@@ -1387,7 +1387,7 @@ bool mb_isalpha(int a)
   return mb_islower(a) || mb_isupper(a);
 }
 
-static int utf_strnicmp(const char *s1, const char *s2, size_t n1, size_t n2)
+int utf_strnicmp(const char *s1, const char *s2, size_t n1, size_t n2)
 {
   int c1, c2;
   char buffer[6];

src/nvim/regexp.c

@@ -1627,7 +1627,9 @@ static void 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 *s1, char *s2, int *n)
 {
   int result;
@@ -1635,8 +1637,11 @@ static int cstrncmp(char *s1, char *s2, int *n)
   if (!rex.reg_ic) {
     result = strncmp(s1, s2, (size_t)(*n));
   } else {
-    assert(*n >= 0);
-    result = mb_strnicmp(s1, s2, (size_t)(*n));
+    int l2 = utfc_ptr2len(s2);
+    result = utf_strnicmp(s1, s2, (size_t)(*n), (size_t)l2);
+    if (result == 0 && l2 < *n) {
+      *n = l2;
+    }
   }
 
   // if it failed and it's utf8 and we want to combineignore:
@@ -6490,11 +6495,9 @@ static bool regmatch(uint8_t *scan, const proftime_T *tm, int *timed_out)
               }
             }
           } else {
-            for (i = 0; i < len; i++) {
-              if (opnd[i] != rex.input[i]) {
-                status = RA_NOMATCH;
-                break;
-              }
+            if (cstrncmp((char *)opnd, (char *)rex.input, &len) != 0) {
+              status = RA_NOMATCH;
+              break;
             }
           }
           rex.input += len;
@@ -13845,23 +13848,26 @@ static int skip_to_start(int c, colnr_T *colp)
 // Returns zero for no match, 1 for a match.
 static int find_match_text(colnr_T *startcol, int regstart, uint8_t *match_text)
 {
-#define PTR2LEN(x) utf_ptr2len(x)
-
   colnr_T col = *startcol;
-  int regstart_len = PTR2LEN((char *)rex.line + col);
+  const int regstart_len = utf_char2len(regstart);
 
   while (true) {
     bool match = true;
     uint8_t *s1 = match_text;
-    uint8_t *s2 = rex.line + col + regstart_len;  // skip regstart
+    // skip regstart
+    uint8_t *s2 = rex.line + col + regstart_len;
+    if (regstart_len > 1
+        && utf_char2len(utf_ptr2char((char *)rex.line + col)) != regstart_len) {
+      // because of case-folding of the previously matched text, we may need
+      // to skip fewer bytes than utf_char2len(regstart)
+      s2 = rex.line + col + utf_char2len(utf_fold(regstart));
+    }
     while (*s1) {
-      int c1_len = PTR2LEN((char *)s1);
+      int c1_len = utf_ptr2len((char *)s1);
       int c1 = utf_ptr2char((char *)s1);
-      int c2_len = PTR2LEN((char *)s2);
+      int c2_len = utf_ptr2len((char *)s2);
       int c2 = utf_ptr2char((char *)s2);
-
-      if ((c1 != c2 && (!rex.reg_ic || utf_fold(c1) != utf_fold(c2)))
-          || c1_len != c2_len) {
+      if (c1 != c2 && (!rex.reg_ic || utf_fold(c1) != utf_fold(c2))) {
         match = false;
         break;
       }
@@ -13894,8 +13900,6 @@ static int find_match_text(colnr_T *startcol, int regstart, uint8_t *match_text)
 
   *startcol = col;
   return 0L;
-
-#undef PTR2LEN
 }
 
 static int nfa_did_time_out(void)
@@ -15527,7 +15531,7 @@ static int nfa_regexec_both(uint8_t *line, colnr_T startcol, proftime_T *tm, int
 
     // 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) {
         rex.reg_mmatch->rmm_matchcol = col;

test/old/testdir/test_regexp_utf8.vim

@@ -611,4 +611,32 @@ 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})
+
+    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)
+  endfor
+  bw!
+endfunc
+
 " vim: shiftwidth=2 sts=2 expandtab