vim-patch:8.2.2278: falling back to old regexp engine can some patterns

Problem:    Falling back to old regexp engine can some patterns.
Solution:   Do not fall back once [[:lower:]] or [[:upper:]] is used.
            (Christian Brabandt, closes vim/vim#7572)
https://github.com/vim/vim/commit/66c50c565321d4d49d8d5620912e5e8fe4825644

1 files changed, 19 insertions, 4 deletions

diff --git a/src/nvim/regexp_nfa.c b/src/nvim/regexp_nfa.c
index 73308c6870..5047e0db03 100644
--- a/src/nvim/regexp_nfa.c
+++ b/src/nvim/regexp_nfa.c
@@ -328,6 +328,11 @@ static int *post_start;   ///< holds the postfix form of r.e.
 static int *post_end;
 static int *post_ptr;
 
+// Set when the pattern should use the NFA engine.
+// E.g. [[:upper:]] only allows 8bit characters for BT engine,
+// while NFA engine handles multibyte characters correctly.
+static bool wants_nfa;
+
 static int nstate;  ///< Number of states in the NFA. Also used when executing.
 static int istate;  ///< Index in the state vector, used in alloc_state()
 
@@ -377,6 +382,7 @@ nfa_regcomp_start (
   post_start = (int *)xmalloc(postfix_size);
   post_ptr = post_start;
   post_end = post_start + nstate_max;
+  wants_nfa = false;
   rex.nfa_has_zend = false;
   rex.nfa_has_backref = false;
 
@@ -1618,6 +1624,7 @@ collection:
               EMIT(NFA_CLASS_GRAPH);
               break;
             case CLASS_LOWER:
+              wants_nfa = true;
               EMIT(NFA_CLASS_LOWER);
               break;
             case CLASS_PRINT:
@@ -1630,6 +1637,7 @@ collection:
               EMIT(NFA_CLASS_SPACE);
               break;
             case CLASS_UPPER:
+              wants_nfa = true;
               EMIT(NFA_CLASS_UPPER);
               break;
             case CLASS_XDIGIT:
@@ -1998,10 +2006,17 @@ static int nfa_regpiece(void)
       return OK;
     }
 
-    // The engine is very inefficient (uses too many states) when the maximum
-    // is much larger than the minimum and when the maximum is large. Bail out
-    // if we can use the other engine.
-    if ((nfa_re_flags & RE_AUTO) && (maxval > 500 || maxval > minval + 200)) {
+    // The engine is very inefficient (uses too many states) when the
+    // maximum is much larger than the minimum and when the maximum is
+    // large.  However, when maxval is MAX_LIMIT, it is okay, as this
+    // will emit NFA_STAR.
+    // Bail out if we can use the other engine, but only, when the
+    // pattern does not need the NFA engine like (e.g. [[:upper:]]\{2,\}
+    // does not work with with characters > 8 bit with the BT engine)
+    if ((nfa_re_flags & RE_AUTO)
+        && (maxval > 500 || maxval > minval + 200)
+        && (maxval != MAX_LIMIT && minval < 200)
+        && !wants_nfa) {
       return FAIL;
     }