- 11 Oct, 2014 16 commits
-
-
Sergei Golubchik authored
-
Sergei Golubchik authored
remove: * NDB from everywhere * IM from mtr-v1 * packaging/rpm-oel and packaging/rpm-uln * few unused spec files * plug.in file * .bzrignore
-
Sergei Golubchik authored
TOTAL_SSL_CONNECTIONS columns in CLIENT_STATISTICS and USER_STATISTICS
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
Extend existing plugins to support * SHOW QUERY_RESPONSE_TIME * FLUSH QUERY_RESPONSE_TIME * SHOW LOCALE move userstat tables to use the new API instead of hand-coded syntax
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
INFORMATION_SCHEMA.SYSTEM_VARIABLES
-
- 10 Oct, 2014 24 commits
-
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
1. @@boolean_var differs from SHOW VARIABLES 2. @@str_var ignored variable charset (which is wrong for path variables that use filesystem charset) 3. @@signed_int_var in the string context was printed as unsigned
-
Sergei Golubchik authored
-
Sergei Golubchik authored
it can never be set to NULL, so it should not use MY_I_S_MAYBE_NULL flag either
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
old fix only generated a warning for the *first* row in the output
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
replace flag SHOW_VALUE_IN_HELP (that's not really an independendent flag, as it should only be used for NO_CMD_LINE sysvars) with a special getopt.id type.
-
Sergei Golubchik authored
-
Sergei Golubchik authored
their only purpose was to distinguish between SHOW and SELECT and there are cleaner ways of doing it.
-
Sergei Golubchik authored
-
Sergei Golubchik authored
* ignore the OPTIMIZER_SWITCH_ENGINE_CONDITION_PUSHDOWN bit * issue a deprecation warning on 'engine_condition_pushdown=on' * remove unused remains of the old pre-5.5 engine_condition_pushdown variable
-
Sergei Golubchik authored
MDEV-6606 Server crashes in String::append on selecting sql_mode inside anonymous block
-
Sergei Golubchik authored
MDEV-6603 SBR failure upon executing a prepared statement with input placeholder under anonymous block Normally, Prepared_statement object rewrites the query on execution to replace ?-placeholders with values. The rewritten query may be written to logs (including binlog) or stored in the query cache. But for compound statements, the whole block is prepared and executed, while contained statements are logged individually. So it doesn't make sense to rewrite the original statement block. Instead, we need to rewrite every contained statement. SP is already doing it to replace SP variables with values. Let it rewrite PS parameters too in the same loop.
-
Sergei Golubchik authored
Fix query rewrites in PS code - it was memcpy-ing the same query tail many times. Instead use the same logic as in SP code, copy query pieces into the destination buffer. Extract this logic into a separate class Rewritable_query_parameter with Item_param and Item_splocal inheriting from it. Create a helper class Copy_query_with_rewrite that incapsulates the query rewriting logic, use it in SP and PS.
-
Sergei Golubchik authored
originally based on the patch by Antony T Curtis
-
Sergei Golubchik authored
originally based on the patch by Antony T Curtis
-
Sergei Golubchik authored
-
Sergei Golubchik authored
-