Merge pull request #61853 from ClickHouse/cherrypick/23.8/cd3edf3f5b4…

…e716caf1cea9e34fa8e2984d767f6

Cherry pick #61749 to 23.8: Fix crash in `multiSearchAllPositionsCaseInsensitiveUTF8` for incorrect UTF-8

src/Common/Volnitsky.h

@@ -191,7 +191,8 @@ namespace VolnitskyTraits
                         if (length_l != length_r)
                             return false;
 
-                        assert(length_l >= 2 && length_r >= 2);
+                        if (length_l < 2 || length_r < 2)
+                            return false;  /// Some part of the given ngram contains an invalid UTF-8 sequence.
 
                         chars.c0 = seq_l[seq_ngram_offset];
                         chars.c1 = seq_l[seq_ngram_offset + 1];
@@ -253,7 +254,9 @@ namespace VolnitskyTraits
                     if (size_l != size_u)
                         return false;
 
-                    assert(size_l >= 1 && size_u >= 1);
+                    if (size_l == 0 || size_u == 0)
+                        return false;  /// Some part of the given ngram contains an invalid UTF-8 sequence.
+
                     chars.c1 = seq_l[0];
                     putNGramBase(n, offset);
 
@@ -276,7 +279,8 @@ namespace VolnitskyTraits
                     if (size_l != size_u)
                         return false;
 
-                    assert(size_l > seq_ngram_offset && size_u > seq_ngram_offset);
+                    if (size_l <= seq_ngram_offset || size_u <= seq_ngram_offset)
+                        return false;  /// Some part of the given ngram contains an invalid UTF-8 sequence.
 
                     chars.c0 = seq_l[seq_ngram_offset];
                     putNGramBase(n, offset);
@@ -302,10 +306,8 @@ namespace VolnitskyTraits
                     if (size_first_l != size_first_u || size_second_l != size_second_u)
                         return false;
 
-                    assert(size_first_l > seq_ngram_offset);
-                    assert(size_first_u > seq_ngram_offset);
-                    assert(size_second_l > 0);
-                    assert(size_second_u > 0);
+                    if (size_first_l <= seq_ngram_offset || size_first_u <= seq_ngram_offset || size_second_l == 0 || size_second_u == 0)
+                        return false;
 
                     auto c0l = first_l_seq[seq_ngram_offset];
                     auto c0u = first_u_seq[seq_ngram_offset];
@@ -399,7 +401,7 @@ class VolnitskyBase
         if (fallback || fallback_searcher.force_fallback)
             return;
 
-        hash = std::unique_ptr<VolnitskyTraits::Offset[]>(new VolnitskyTraits::Offset[VolnitskyTraits::hash_size]{});
+        hash = std::make_unique<VolnitskyTraits::Offset[]>(VolnitskyTraits::hash_size);
 
         auto callback = [this](const VolnitskyTraits::Ngram ngram, const int offset) { return this->putNGramBase(ngram, offset); };
         /// ssize_t is used here because unsigned can't be used with condition like `i >= 0`, unsigned always >= 0

tests/queries/0_stateless/02364_multiSearch_function_family.reference

@@ -12872,3 +12872,4 @@
 1
 1
 1
+1

tests/queries/0_stateless/02364_multiSearch_function_family.sql

@@ -223,6 +223,8 @@ select [2] = multiSearchAllPositions(materialize('abab'), materialize(['ba']));
 select [1] = multiSearchAllPositionsCaseInsensitive(materialize('aBaB'), materialize(['abab']));
 select [3] = multiSearchAllPositionsUTF8(materialize('ab€ab'), materialize(['€']));
 select [3] = multiSearchAllPositionsCaseInsensitiveUTF8(materialize('ab€AB'), materialize(['€ab']));
+-- checks the correct handling of broken utf-8 sequence
+select [0] = multiSearchAllPositionsCaseInsensitiveUTF8(materialize(''), materialize(['a\x90\x90\x90\x90\x90\x90']));
 
 select 1 = multiSearchAny(materialize('abcdefgh'), ['b']);
 select 1 = multiSearchAny(materialize('abcdefgh'), ['bc']);