From c66fb7cf6ea5f6c2895d0489d5f25f3005a52246 Mon Sep 17 00:00:00 2001
From: vdimir <vdimir@clickhouse.com>
Date: Thu, 10 Mar 2022 13:31:45 +0000
Subject: [PATCH 1/2] Attempt to fix replaceRegexpAll

---
 src/Functions/ReplaceRegexpImpl.h             |  9 +++++--
 ...gexp_all_empty_match_alternative.reference | 15 ++++++++++++
 ...ace_regexp_all_empty_match_alternative.sql | 24 ++++++++++++++++---
 3 files changed, 43 insertions(+), 5 deletions(-)

diff --git a/src/Functions/ReplaceRegexpImpl.h b/src/Functions/ReplaceRegexpImpl.h
index 5d2549239c86..feacfd507524 100644
--- a/src/Functions/ReplaceRegexpImpl.h
+++ b/src/Functions/ReplaceRegexpImpl.h
@@ -137,8 +137,13 @@ struct ReplaceRegexpImpl
 
                 if (replace_one)
                     can_finish_current_string = true;
-                else if (match.length() == 0)
-                    ++match_pos; /// Step one character to avoid infinite loop.
+
+                if (match.length() == 0)
+                {
+                    /// Step one character to avoid infinite loop
+                    if (++match_pos >= static_cast<size_t>(input.length()))
+                        can_finish_current_string = true;
+                }
             }
             else
                 can_finish_current_string = true;
diff --git a/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.reference b/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.reference
index e8183f05f5db..da7b788b1575 100644
--- a/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.reference
+++ b/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.reference
@@ -1,3 +1,18 @@
 1
 1
 1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
diff --git a/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.sql b/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.sql
index 6725fa041140..ebbc6ce97e0f 100644
--- a/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.sql
+++ b/tests/queries/0_stateless/02151_replace_regexp_all_empty_match_alternative.sql
@@ -1,3 +1,21 @@
-select replaceRegexpAll(',,1,,', '^[,]*|[,]*$', '') x;
-select replaceRegexpAll(',,1', '^[,]*|[,]*$', '') x;
-select replaceRegexpAll('1,,', '^[,]*|[,]*$', '') x;
+SELECT replaceRegexpAll(',,1,,', '^[,]*|[,]*$', '');
+SELECT replaceRegexpAll(',,1', '^[,]*|[,]*$', '');
+SELECT replaceRegexpAll('1,,', '^[,]*|[,]*$', '');
+
+SELECT replaceRegexpAll(materialize(',,1,,'), '^[,]*|[,]*$', '');
+SELECT replaceRegexpAll(materialize(',,1'), '^[,]*|[,]*$', '');
+SELECT replaceRegexpAll(materialize('1,,'), '^[,]*|[,]*$', '');
+
+SELECT replaceRegexpAll('a', 'z*', '') == 'a';
+SELECT replaceRegexpAll('aa', 'z*', '') == 'aa';
+SELECT replaceRegexpAll('aaq', 'z*', '') == 'aaq';
+SELECT replaceRegexpAll('aazq', 'z*', '') == 'aaq';
+SELECT replaceRegexpAll('aazzq', 'z*', '') == 'aaq';
+SELECT replaceRegexpAll('aazzqa', 'z*', '') == 'aaqa';
+
+SELECT replaceRegexpAll(materialize('a'), 'z*', '') == 'a';
+SELECT replaceRegexpAll(materialize('aa'), 'z*', '') == 'aa';
+SELECT replaceRegexpAll(materialize('aaq'), 'z*', '') == 'aaq';
+SELECT replaceRegexpAll(materialize('aazq'), 'z*', '') == 'aaq';
+SELECT replaceRegexpAll(materialize('aazzq'), 'z*', '') == 'aaq';
+SELECT replaceRegexpAll(materialize('aazzqa'), 'z*', '') == 'aaqa';

From e4e058d667d4f9956df000703a2d54eff7b09afc Mon Sep 17 00:00:00 2001
From: Alexey Milovidov <milovidov@yandex-team.ru>
Date: Sat, 12 Mar 2022 01:35:37 +0300
Subject: [PATCH 2/2] Update ReplaceRegexpImpl.h

---
 src/Functions/ReplaceRegexpImpl.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/Functions/ReplaceRegexpImpl.h b/src/Functions/ReplaceRegexpImpl.h
index feacfd507524..ad0ce08d9051 100644
--- a/src/Functions/ReplaceRegexpImpl.h
+++ b/src/Functions/ReplaceRegexpImpl.h
@@ -141,7 +141,8 @@ struct ReplaceRegexpImpl
                 if (match.length() == 0)
                 {
                     /// Step one character to avoid infinite loop
-                    if (++match_pos >= static_cast<size_t>(input.length()))
+                    ++match_pos;
+                    if (match_pos >= static_cast<size_t>(input.length()))
                         can_finish_current_string = true;
                 }
             }