Merge pull request #20943 from zeertzjq/vim-8.2.4688

vim-patch:8.2.{4688,4693,4978},9.0.0053: regexp fixes

runtime/doc/pattern.txt

@@ -372,7 +372,7 @@ Vim includes two regexp engines:
 1. An old, backtracking engine that supports everything.
 2. A new, NFA engine that works much faster on some patterns, possibly slower
    on some patterns.
-
+								 *E1281*
 Vim will automatically select the right engine for you.  However, if you run
 into a problem or want to specifically select one engine or the other, you can
 prepend one of the following to the pattern:

src/nvim/regexp.c

@@ -97,20 +97,24 @@ static int toggle_Magic(int x)
 
 #define MAX_LIMIT       (32767L << 16L)
 
-static char_u e_missingbracket[] = N_("E769: Missing ] after %s[");
-static char_u e_reverse_range[] = N_("E944: Reverse range in character class");
-static char_u e_large_class[] = N_("E945: Range too large in character class");
-static char_u e_unmatchedpp[] = N_("E53: Unmatched %s%%(");
-static char_u e_unmatchedp[] = N_("E54: Unmatched %s(");
-static char_u e_unmatchedpar[] = N_("E55: Unmatched %s)");
-static char_u e_z_not_allowed[] = N_("E66: \\z( not allowed here");
-static char_u e_z1_not_allowed[] = N_("E67: \\z1 - \\z9 not allowed here");
-static char_u e_missing_sb[] = N_("E69: Missing ] after %s%%[");
-static char_u e_empty_sb[] = N_("E70: Empty %s%%[]");
-static char_u e_recursive[] = N_("E956: Cannot use pattern recursively");
-static char_u e_regexp_number_after_dot_pos_search[]
+static char e_missingbracket[] = N_("E769: Missing ] after %s[");
+static char e_reverse_range[] = N_("E944: Reverse range in character class");
+static char e_large_class[] = N_("E945: Range too large in character class");
+static char e_unmatchedpp[] = N_("E53: Unmatched %s%%(");
+static char e_unmatchedp[] = N_("E54: Unmatched %s(");
+static char e_unmatchedpar[] = N_("E55: Unmatched %s)");
+static char e_z_not_allowed[] = N_("E66: \\z( not allowed here");
+static char e_z1_not_allowed[] = N_("E67: \\z1 - \\z9 not allowed here");
+static char e_missing_sb[] = N_("E69: Missing ] after %s%%[");
+static char e_empty_sb[] = N_("E70: Empty %s%%[]");
+static char e_recursive[] = N_("E956: Cannot use pattern recursively");
+static char e_regexp_number_after_dot_pos_search_chr[]
   = N_("E1204: No Number allowed after .: '\\%%%c'");
-static char_u e_substitute_nesting_too_deep[] = N_("E1290: substitute nesting too deep");
+static char e_nfa_regexp_missing_value_in_chr[]
+  = N_("E1273: (NFA regexp) missing value in '\\%%%c'");
+static char e_atom_engine_must_be_at_start_of_pattern[]
+  = N_("E1281: Atom '\\%%#=%c' must be at the start of the pattern");
+static char e_substitute_nesting_too_deep[] = N_("E1290: substitute nesting too deep");
 
 #define NOT_MULTI       0
 #define MULTI_ONE       1

src/nvim/regexp_bt.c

@@ -1971,6 +1971,11 @@ static char_u *regatom(int *flagp)
       break;
 
     case '#':
+      if (regparse[0] == '=' && regparse[1] >= 48 && regparse[1] <= 50) {
+        // misplaced \%#=1
+        semsg(_(e_atom_engine_must_be_at_start_of_pattern), regparse[1]);
+        return FAIL;
+      }
       ret = regnode(CURSOR);
       break;
 
@@ -2091,6 +2096,7 @@ static char_u *regatom(int *flagp)
         uint32_t n = 0;
         int cmp;
         bool cur = false;
+        bool got_digit = false;
 
         cmp = c;
         if (cmp == '<' || cmp == '>') {
@@ -2101,6 +2107,7 @@ static char_u *regatom(int *flagp)
           c = getchr();
         }
         while (ascii_isdigit(c)) {
+          got_digit = true;
           n = n * 10 + (uint32_t)(c - '0');
           c = getchr();
         }
@@ -2115,9 +2122,9 @@ static char_u *regatom(int *flagp)
             *regcode++ = (char_u)cmp;
           }
           break;
-        } else if (c == 'l' || c == 'c' || c == 'v') {
+        } else if ((c == 'l' || c == 'c' || c == 'v') && (cur || got_digit)) {
           if (cur && n) {
-            semsg(_(e_regexp_number_after_dot_pos_search), no_Magic(c));
+            semsg(_(e_regexp_number_after_dot_pos_search_chr), no_Magic(c));
             rc_did_emsg = true;
             return NULL;
           }

src/nvim/regexp_nfa.c

@@ -2094,6 +2094,12 @@ static int nfa_regatom(void)
       break;
 
     case '#':
+      if (regparse[0] == '=' && regparse[1] >= 48
+          && regparse[1] <= 50) {
+        // misplaced \%#=1
+        semsg(_(e_atom_engine_must_be_at_start_of_pattern), regparse[1]);
+        return FAIL;
+      }
       EMIT(NFA_CURSOR);
       break;
 
@@ -2141,6 +2147,7 @@ static int nfa_regatom(void)
       int64_t n = 0;
       const int cmp = c;
       bool cur = false;
+      bool got_digit = false;
 
       if (c == '<' || c == '>') {
         c = getchr();
@@ -2151,7 +2158,7 @@ static int nfa_regatom(void)
       }
       while (ascii_isdigit(c)) {
         if (cur) {
-          semsg(_(e_regexp_number_after_dot_pos_search), no_Magic(c));
+          semsg(_(e_regexp_number_after_dot_pos_search_chr), no_Magic(c));
           return FAIL;
         }
         if (n > (INT32_MAX - (c - '0')) / 10) {
@@ -2161,10 +2168,15 @@ static int nfa_regatom(void)
         }
         n = n * 10 + (c - '0');
         c = getchr();
+        got_digit = true;
       }
       if (c == 'l' || c == 'c' || c == 'v') {
         int32_t limit = INT32_MAX;
 
+        if (!cur && !got_digit) {
+          semsg(_(e_nfa_regexp_missing_value_in_chr), no_Magic(c));
+          return FAIL;
+        }
         if (c == 'l') {
           if (cur) {
             n = curwin->w_cursor.lnum;

src/nvim/testdir/test_regexp_latin.vim

@@ -105,6 +105,31 @@ func Test_multi_failure()
   set re=0
 endfunc
 
+func Test_column_success_failure()
+  new
+  call setline(1, 'xbar')
+
+  set re=1
+  %s/\%>0v./A/
+  call assert_equal('Abar', getline(1))
+  call assert_fails('/\%v', 'E71:')
+  call assert_fails('/\%>v', 'E71:')
+  call assert_fails('/\%c', 'E71:')
+  call assert_fails('/\%<c', 'E71:')
+  call assert_fails('/\%l', 'E71:')
+  set re=2
+  %s/\%>0v./B/
+  call assert_equal('Bbar', getline(1))
+  call assert_fails('/\%v', 'E1273:')
+  call assert_fails('/\%>v', 'E1273:')
+  call assert_fails('/\%c', 'E1273:')
+  call assert_fails('/\%<c', 'E1273:')
+  call assert_fails('/\%l', 'E1273:')
+
+  set re=0
+  bwipe!
+endfunc
+
 func Test_recursive_addstate()
   throw 'skipped: TODO: '
   " This will call addstate() recursively until it runs into the limit.
@@ -1033,6 +1058,28 @@ func Test_using_invalid_visual_position()
   bwipe!
 endfunc
 
+func Test_using_two_engines_pattern()
+  new
+  call setline(1, ['foobar=0', 'foobar=1', 'foobar=2'])
+  " \%#= at the end of the pattern
+  for i in range(0, 2)
+    for j in range(0, 2)
+      exe "set re=" .. i
+      call cursor(j + 1, 7)
+      call assert_fails("%s/foobar\\%#=" .. j, 'E1281:')
+    endfor
+  endfor
+  set re=0
+
+  " \%#= at the start of the pattern
+  for i in range(0, 2)
+    call cursor(i + 1, 7)
+    exe ":%s/\\%#=" .. i .. "foobar=" .. i .. "/xx"
+  endfor
+  call assert_equal(['xx', 'xx', 'xx'], getline(1, '$'))
+  bwipe!
+endfunc
+
 func Test_recursive_substitute_expr()
   new
   func Repl()