MDEV-6027 RLIKE: "." no longer matching new line

Added a new system variable:
default_regex_flags='DOTALL,DUPNAMES,EXTENDED,EXTRA,MULTILINE,UNGREEDY'

10 files changed, 287 insertions, 1 deletions

diff --git a/mysql-test/r/func_regexp_pcre.result b/mysql-test/r/func_regexp_pcre.result
index ebbdedc6999..b777af767cb 100644
--- a/mysql-test/r/func_regexp_pcre.result
+++ b/mysql-test/r/func_regexp_pcre.result
@@ -754,3 +754,88 @@ DROP TABLE t1;
 SELECT REGEXP_SUBSTR('See https://mariadb.org/en/foundation/ for details', 'https?://[^/]*');
 REGEXP_SUBSTR('See https://mariadb.org/en/foundation/ for details', 'https?://[^/]*')
 https://mariadb.org
+#
+# MDEV-6027 RLIKE: "." no longer matching new line
+#
+SELECT  'cat and\ndog'  RLIKE 'cat.*dog';
+'cat and\ndog'  RLIKE 'cat.*dog'
+0
+SELECT  'cat and\r\ndog'  RLIKE 'cat.*dog';
+'cat and\r\ndog'  RLIKE 'cat.*dog'
+0
+SELECT 'a\nb' RLIKE 'a.b';
+'a\nb' RLIKE 'a.b'
+0
+SELECT 'a\nb' RLIKE '(?-s)a.b';
+'a\nb' RLIKE '(?-s)a.b'
+0
+SET default_regex_flags='DOTALL';
+SELECT @@default_regex_flags;
+@@default_regex_flags
+DOTALL
+SELECT  'cat and\ndog'  RLIKE 'cat.*dog';
+'cat and\ndog'  RLIKE 'cat.*dog'
+1
+SELECT  'cat and\r\ndog'  RLIKE 'cat.*dog';
+'cat and\r\ndog'  RLIKE 'cat.*dog'
+1
+SELECT 'a\nb' RLIKE 'a.b';
+'a\nb' RLIKE 'a.b'
+1
+SELECT 'a\nb' RLIKE '(?-s)a.b';
+'a\nb' RLIKE '(?-s)a.b'
+0
+SET default_regex_flags=DEFAULT;
+SELECT REGEXP_SUBSTR('Monday Mon','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$');
+ERROR 42000: Got error 'two named subpatterns have the same name at offset 29' from regexp
+SET default_regex_flags='DUPNAMES';
+SELECT REGEXP_SUBSTR('Monday Mon','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$');
+REGEXP_SUBSTR('Monday Mon','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$')
+Monday Mon
+SELECT REGEXP_SUBSTR('Tuesday Tue','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$');
+REGEXP_SUBSTR('Tuesday Tue','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$')
+Tuesday Tue
+SET default_regex_flags=DEFAULT;
+SELECT 'AB' RLIKE 'A B';
+'AB' RLIKE 'A B'
+0
+SELECT 'AB' RLIKE 'A# this is a comment\nB';
+'AB' RLIKE 'A# this is a comment\nB'
+0
+SET default_regex_flags='EXTENDED';
+SELECT 'AB' RLIKE 'A B';
+'AB' RLIKE 'A B'
+1
+SELECT 'AB' RLIKE 'A# this is a comment\nB';
+'AB' RLIKE 'A# this is a comment\nB'
+1
+SET default_regex_flags=DEFAULT;
+SELECT 'Aq' RLIKE 'A\\q';
+'Aq' RLIKE 'A\\q'
+1
+SET default_regex_flags='EXTRA';
+SELECT 'Aq' RLIKE 'A\\q';
+ERROR 42000: Got error 'unrecognized character follows \ at offset 2' from regexp
+SET default_regex_flags=DEFAULT;
+SELECT 'a\nb\nc' RLIKE '^b$';
+'a\nb\nc' RLIKE '^b$'
+0
+SET default_regex_flags='MULTILINE';
+SELECT 'a\nb\nc' RLIKE '^b$';
+'a\nb\nc' RLIKE '^b$'
+1
+SET default_regex_flags=DEFAULT;
+SELECT REGEXP_SUBSTR('abc','.+');
+REGEXP_SUBSTR('abc','.+')
+abc
+SELECT REGEXP_REPLACE('abc','^(.*)(.*)$','\\1/\\2');
+REGEXP_REPLACE('abc','^(.*)(.*)$','\\1/\\2')
+abc/
+SET default_regex_flags='UNGREEDY';
+SELECT REGEXP_SUBSTR('abc','.+');
+REGEXP_SUBSTR('abc','.+')
+a
+SELECT REGEXP_REPLACE('abc','^(.*)(.*)$','\\1/\\2');
+REGEXP_REPLACE('abc','^(.*)(.*)$','\\1/\\2')
+/abc
+SET default_regex_flags=DEFAULT;
diff --git a/mysql-test/r/mysqld--help.result b/mysql-test/r/mysqld--help.result
index 5a468cdf236..f8ce0e35904 100644
--- a/mysql-test/r/mysqld--help.result
+++ b/mysql-test/r/mysqld--help.result
@@ -136,6 +136,10 @@ The following options may be given as the first argument:
  --deadlock-timeout-short=# 
  Short timeout for the two-step deadlock detection (in
  microseconds)
+ --default-regex-flags=name 
+ Default flags for the regex library. Syntax:
+ default-regex-flags='[flag[,flag[,flag...]]]'. See the
+ manual for the complete list of valid flags
  --default-storage-engine=name 
  The default storage engine for new tables
  --default-time-zone=name 
@@ -1079,6 +1083,7 @@ deadlock-search-depth-long 15
 deadlock-search-depth-short 4
 deadlock-timeout-long 50000000
 deadlock-timeout-short 10000
+default-regex-flags 
 default-storage-engine myisam
 default-time-zone (No default value)
 default-week-format 0
diff --git a/mysql-test/suite/sys_vars/r/default_regex_flags_basic.result b/mysql-test/suite/sys_vars/r/default_regex_flags_basic.result
new file mode 100644
index 00000000000..84c55f3d2b6
--- /dev/null
+++ b/mysql-test/suite/sys_vars/r/default_regex_flags_basic.result
@@ -0,0 +1,57 @@
+SET default_regex_flags='';
+SELECT @@default_regex_flags;
+@@default_regex_flags
+
+SET default_regex_flags=DEFAULT;
+SELECT @@default_regex_flags;
+@@default_regex_flags
+
+SET default_regex_flags=NULL;
+ERROR 42000: Variable 'default_regex_flags' can't be set to the value of 'NULL'
+SELECT @@default_regex_flags;
+@@default_regex_flags
+
+SET default_regex_flags='UNKNOWN';
+ERROR 42000: Variable 'default_regex_flags' can't be set to the value of 'UNKNOWN'
+SET default_regex_flags=123;
+ERROR 42000: Variable 'default_regex_flags' can't be set to the value of '123'
+SET default_regex_flags=123.0;
+ERROR 42000: Incorrect argument type to variable 'default_regex_flags'
+SET default_regex_flags=123e0;
+ERROR 42000: Incorrect argument type to variable 'default_regex_flags'
+SET default_regex_flags='DOTALL';
+SELECT @@default_regex_flags;
+@@default_regex_flags
+DOTALL
+SET default_regex_flags=NULL;
+ERROR 42000: Variable 'default_regex_flags' can't be set to the value of 'NULL'
+SELECT @@default_regex_flags;
+@@default_regex_flags
+DOTALL
+SET @@default_regex_flags=63;
+SELECT @@default_regex_flags;
+@@default_regex_flags
+DOTALL,DUPNAMES,EXTENDED,EXTRA,MULTILINE,UNGREEDY
+SET @@default_regex_flags='DOTALL';
+SELECT @@default_regex_flags;
+@@default_regex_flags
+DOTALL
+SET @@default_regex_flags='DOTALL,DUPNAMES,EXTENDED,EXTRA,MULTILINE,UNGREEDY';
+SELECT @@default_regex_flags;
+@@default_regex_flags
+DOTALL,DUPNAMES,EXTENDED,EXTRA,MULTILINE,UNGREEDY
+SET @@default_regex_flags=DEFAULT;
+SET @@global.default_regex_flags='MULTILINE';
+SELECT @@session.default_regex_flags;
+@@session.default_regex_flags
+
+# connection con1
+SELECT @@session.default_regex_flags;
+@@session.default_regex_flags
+MULTILINE
+# connection default
+SELECT @@session.default_regex_flags;
+@@session.default_regex_flags
+
+SET @@global.default_regex_flags=DEFAULT;
+SET default_regex_flags=DEFAULT;
diff --git a/mysql-test/suite/sys_vars/t/default_regex_flags_basic.test b/mysql-test/suite/sys_vars/t/default_regex_flags_basic.test
new file mode 100644
index 00000000000..1e826293949
--- /dev/null
+++ b/mysql-test/suite/sys_vars/t/default_regex_flags_basic.test
@@ -0,0 +1,43 @@
+
+SET default_regex_flags='';
+SELECT @@default_regex_flags;
+SET default_regex_flags=DEFAULT;
+SELECT @@default_regex_flags;
+--error ER_WRONG_VALUE_FOR_VAR
+SET default_regex_flags=NULL;
+SELECT @@default_regex_flags;
+--error ER_WRONG_VALUE_FOR_VAR
+SET default_regex_flags='UNKNOWN';
+--error ER_WRONG_VALUE_FOR_VAR
+SET default_regex_flags=123;
+--error ER_WRONG_TYPE_FOR_VAR
+SET default_regex_flags=123.0;
+--error ER_WRONG_TYPE_FOR_VAR
+SET default_regex_flags=123e0;
+SET default_regex_flags='DOTALL';
+SELECT @@default_regex_flags;
+--error ER_WRONG_VALUE_FOR_VAR
+SET default_regex_flags=NULL;
+SELECT @@default_regex_flags;
+
+SET @@default_regex_flags=63;
+SELECT @@default_regex_flags;
+SET @@default_regex_flags='DOTALL';
+SELECT @@default_regex_flags;
+SET @@default_regex_flags='DOTALL,DUPNAMES,EXTENDED,EXTRA,MULTILINE,UNGREEDY';
+SELECT @@default_regex_flags;
+SET @@default_regex_flags=DEFAULT;
+
+SET @@global.default_regex_flags='MULTILINE';
+SELECT @@session.default_regex_flags;
+connect (con1,localhost,root,,);
+--echo # connection con1
+connection con1;
+SELECT @@session.default_regex_flags;
+connection default;
+--echo # connection default
+disconnect con1;
+SELECT @@session.default_regex_flags;
+SET @@global.default_regex_flags=DEFAULT;
+
+SET default_regex_flags=DEFAULT;
diff --git a/mysql-test/t/func_regexp_pcre.test b/mysql-test/t/func_regexp_pcre.test
index 4892d9931da..7e8df840239 100644
--- a/mysql-test/t/func_regexp_pcre.test
+++ b/mysql-test/t/func_regexp_pcre.test
@@ -349,3 +349,51 @@ DROP TABLE t1;
 
 
 SELECT REGEXP_SUBSTR('See https://mariadb.org/en/foundation/ for details', 'https?://[^/]*');
+
+
+--echo #
+--echo # MDEV-6027 RLIKE: "." no longer matching new line
+--echo #
+SELECT  'cat and\ndog'  RLIKE 'cat.*dog';
+SELECT  'cat and\r\ndog'  RLIKE 'cat.*dog';
+SELECT 'a\nb' RLIKE 'a.b';
+SELECT 'a\nb' RLIKE '(?-s)a.b';
+SET default_regex_flags='DOTALL';
+SELECT @@default_regex_flags;
+SELECT  'cat and\ndog'  RLIKE 'cat.*dog';
+SELECT  'cat and\r\ndog'  RLIKE 'cat.*dog';
+SELECT 'a\nb' RLIKE 'a.b';
+SELECT 'a\nb' RLIKE '(?-s)a.b';
+SET default_regex_flags=DEFAULT;
+
+--error ER_REGEXP_ERROR
+SELECT REGEXP_SUBSTR('Monday Mon','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$');
+SET default_regex_flags='DUPNAMES';
+SELECT REGEXP_SUBSTR('Monday Mon','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$');
+SELECT REGEXP_SUBSTR('Tuesday Tue','^((?<DN>Mon|Fri|Sun)day|(?<DN>Tue)sday).*(?P=DN)$');
+SET default_regex_flags=DEFAULT;
+
+SELECT 'AB' RLIKE 'A B';
+SELECT 'AB' RLIKE 'A# this is a comment\nB';
+SET default_regex_flags='EXTENDED';
+SELECT 'AB' RLIKE 'A B';
+SELECT 'AB' RLIKE 'A# this is a comment\nB';
+SET default_regex_flags=DEFAULT;
+
+SELECT 'Aq' RLIKE 'A\\q';
+SET default_regex_flags='EXTRA';
+--error ER_REGEXP_ERROR
+SELECT 'Aq' RLIKE 'A\\q';
+SET default_regex_flags=DEFAULT;
+
+SELECT 'a\nb\nc' RLIKE '^b$';
+SET default_regex_flags='MULTILINE';
+SELECT 'a\nb\nc' RLIKE '^b$';
+SET default_regex_flags=DEFAULT;
+
+SELECT REGEXP_SUBSTR('abc','.+');
+SELECT REGEXP_REPLACE('abc','^(.*)(.*)$','\\1/\\2');
+SET default_regex_flags='UNGREEDY';
+SELECT REGEXP_SUBSTR('abc','.+');
+SELECT REGEXP_REPLACE('abc','^(.*)(.*)$','\\1/\\2');
+SET default_regex_flags=DEFAULT;
diff --git a/sql/item_cmpfunc.cc b/sql/item_cmpfunc.cc
index 4b9eb37488e..5028e18a6d7 100644
--- a/sql/item_cmpfunc.cc
+++ b/sql/item_cmpfunc.cc
@@ -5055,6 +5055,11 @@ bool Item_func_like::find_selective_predicates_list_processor(uchar *arg)
 }
 
 
+int Regexp_processor_pcre::default_regex_flags()
+{
+  return default_regex_flags_pcre(current_thd);
+}
+
 
 /**
   Convert string to lib_charset, if needed.
diff --git a/sql/item_cmpfunc.h b/sql/item_cmpfunc.h
index 4b3acf83f85..bf28b00c908 100644
--- a/sql/item_cmpfunc.h
+++ b/sql/item_cmpfunc.h
@@ -1513,9 +1513,10 @@ public:
     m_library_charset(&my_charset_utf8_general_ci),
     m_subpatterns_needed(0)
   {}
+  int default_regex_flags();
   void init(CHARSET_INFO *data_charset, int extra_flags, uint nsubpatterns)
   {
-    m_library_flags= extra_flags |
+    m_library_flags= default_regex_flags() | extra_flags |
                     (data_charset != &my_charset_bin ?
                      (PCRE_UTF8 | PCRE_UCP) : 0) |
                     ((data_charset->state &
diff --git a/sql/set_var.h b/sql/set_var.h
index 28562b72dac..564dac587d9 100644
--- a/sql/set_var.h
+++ b/sql/set_var.h
@@ -337,6 +337,7 @@ bool fix_delay_key_write(sys_var *self, THD *thd, enum_var_type type);
 
 ulonglong expand_sql_mode(ulonglong sql_mode);
 bool sql_mode_string_representation(THD *thd, ulonglong sql_mode, LEX_STRING *ls);
+int default_regex_flags_pcre(const THD *thd);
 
 extern sys_var *Sys_autocommit_ptr;
 
diff --git a/sql/sql_class.h b/sql/sql_class.h
index 801fe2d0206..d050a7b3704 100644
--- a/sql/sql_class.h
+++ b/sql/sql_class.h
@@ -516,6 +516,7 @@ typedef struct system_variables
   ulonglong join_buff_size;
   ulonglong sortbuff_size;
   ulonglong group_concat_max_len;
+  ulonglong default_regex_flags;
   ha_rows select_limit;
   ha_rows max_join_size;
   ha_rows expensive_subquery_limit;
diff --git a/sql/sys_vars.cc b/sql/sys_vars.cc
index d22e8498642..44f47646e38 100644
--- a/sql/sys_vars.cc
+++ b/sql/sys_vars.cc
@@ -4568,6 +4568,46 @@ static Sys_var_set Sys_log_slow_filter(
        log_slow_filter_names,
        DEFAULT(MAX_SET(array_elements(log_slow_filter_names)-1)));
 
+static const char *default_regex_flags_names[]= 
+{
+  "DOTALL",    // (?s)  . matches anything including NL
+  "DUPNAMES",  // (?J)  Allow duplicate names for subpatterns
+  "EXTENDED",  // (?x)  Ignore white space and # comments
+  "EXTRA",     // (?X)  extra features (e.g. error on unknown escape character)
+  "MULTILINE", // (?m)  ^ and $ match newlines within data
+  "UNGREEDY",  // (?U)  Invert greediness of quantifiers
+  0
+};
+static const int default_regex_flags_to_pcre[]=
+{
+  PCRE_DOTALL,
+  PCRE_DUPNAMES,
+  PCRE_EXTENDED,
+  PCRE_EXTRA,
+  PCRE_MULTILINE,
+  PCRE_UNGREEDY,
+  0
+};
+int default_regex_flags_pcre(const THD *thd)
+{
+  ulonglong src= thd->variables.default_regex_flags;
+  int i, res;
+  for (i= res= 0; default_regex_flags_to_pcre[i]; i++)
+  {
+    if (src & (1 << i))
+      res|= default_regex_flags_to_pcre[i];
+  }
+  return res;
+}
+static Sys_var_set Sys_default_regex_flags(
+       "default_regex_flags",
+       "Default flags for the regex library. "
+       "Syntax: default-regex-flags='[flag[,flag[,flag...]]]'. "
+       "See the manual for the complete list of valid flags",
+       SESSION_VAR(default_regex_flags), CMD_LINE(REQUIRED_ARG),
+       default_regex_flags_names,
+       DEFAULT(0));
+
 static Sys_var_ulong Sys_log_slow_rate_limit(
        "log_slow_rate_limit",
        "Write to slow log every #th slow query. Set to 1 to log everything. "