lp:~tsarev/percona-server/5.5_fix_bug_716210

Branch merges

Propose for merging

No branches dependent on this one.

Superseded for merging into lp:percona-server/5.5

Laurynas Biveinis (community): Needs Fixing on 2011-09-30

Alexey Kopytov: Pending requested 2011-09-29

Diff: 752 lines (+523/-92) (has conflicts)

1 file modified

patches/slow_extended.patch (+523/-92)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

152. By Oleg Tsarev on 2011-10-06

extend tests

151. By Oleg Tsarev on 2011-09-28

  * "all statistics variables used in slow_extended.patch and thd->send_row_count and thd->examined_row_count" == SESV (Slow Extended Statistic Variables)

  * Added to class THD (sql/sql_class.h) following methods:
         1) clear_slow_extended - clear SESV
         2) reset_sub_statement_state_slow_extended(Sub_statement_state *backup) - backup SESV to backup area
         3) restore_sub_statement_state_slow_extended(Sub_statement_state *backup) - restore SESV from backup area

  * Added to class Sub_statement_state (sql/sql_class.h) required for correct backup variables

  * Added to class THD (sql/sql_class.h) comments near the SESV with explanations of purpose of variables

  * Added call of clear_slow_extended to THD::init() (required on THD class initialisation and change_user)

  * Fix admin command processing in LOGGER::slow_log_print (sql/loc.cc). Before fix SESV don't clean correctly (because if(!query_length) never was true).
    Explanation:
      1) query and query_length according to comment in sql_class.h always correlated - if query == NULL => query_length == 0, and if query != NULL => query_length != 0
      2) After "if (!query) { ... }" query_lenght always not 0
      3) if (!query_length) (code before my fix) always trye

  * Fixed accoring to code policy code formating in "increment_thd_innodb_stats" (sql/sql_class.cc)

  * Fixed THD::reset_for_next_command (sql/sql_parse.cc) - replace variables cleaning by clear_slow_extended call

  * Fixed bug #716210: add thd->clear_slow_extended() to dispatch_command()
    Explanation: before dispatch_command mysql doesn't call "reset_for_next_command" and SESV inherited from previous query

150. By Oleg Tsarev on 2011-09-21

Fix bug #830199 - thd->last_errno was not updated on handle_condition (bug was introduced while porting slow_extended.patch 5.1 to 5.5)

149. By Oleg Tsarev on 2011-09-21

make regenerate

148. By Oleg Tsarev on 2011-09-21

Add test-case for bug #851011: SHOW SLAVE STATUS NOLOCK acquire lock on
  sql/slave.cc: 1650
    pthread_mutex_lock(&mi->run_lock);
Fix: wrap mutex lock/unlock by check to "UNLOCK" statement
Issue number: #12952.

147. By Alexey Kopytov on 2011-09-16

Automerge from the main branch.

146. By Alexey Kopytov on 2011-09-08

Bug #845019: Support for building without response time distribution is
broken

Do not call query_response_time_collect() if there is no support for
the response time distribution feature.

145. By Alexey Kopytov on 2011-09-07

Manual merge from trunk.

144. By Alexey Kopytov on 2011-09-06

Null merge from trunk.

143. By Alexey Kopytov on 2011-09-06

Manual merge from trunk.