second try - handle CS_PREPARE_TO_COMMIT explicitly


storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp:
  handle CS_PREPARE_TO_COMMIT explictly

  add assertion


storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  Add assertion that we dont try to read key from lqh on scan (as it does not have key...)

  also close scan which are in "delivered" state, as it's impossible to release locks afterwards


storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp:
  also close scan which are in "delivered" state, as it's impossible to release locks afterwards

  store new op when converting WRITE to UPDATE so that ACCKEYCONF gets correct


storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  store new op when converting WRITE to UPDATE so that ACCKEYCONF gets correct

into  perch.ndb.mysql.com:/home/jonas/src/51-jonas


storage/ndb/test/ndbapi/testNodeRestart.cpp:
  Auto merged

into  perch.ndb.mysql.com:/home/jonas/src/51-work


storage/ndb/src/kernel/blocks/ERROR_codes.txt:
  Auto merged
storage/ndb/src/kernel/blocks/dbdih/DbdihMain.cpp:
  Auto merged
storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp:
  Auto merged
storage/ndb/test/ndbapi/testNodeRestart.cpp:
  Auto merged
storage/ndb/test/run-test/daily-basic-tests.txt:
  merge

  Dont be too aggressive in Dbtc::nodeFailCheckTransaction
  let it timeout by 1, so that it does not assert that it has waited too long
  old impl. set timeotu value to 0, making timeout = (ctcTimer - 0) which could be quite big.


ndb/src/kernel/blocks/ERROR_codes.txt:
  error codes
ndb/src/kernel/blocks/dbdih/DbdihMain.cpp:
  2 new error inserts
  7030 - delay in GCP_PREPARE until checkLocalNodefailComplete is true
  7031 - delay in GCP_PREPARE and die
ndb/src/kernel/blocks/dbtc/DbtcMain.cpp:
  Dont set api con timer to 0, as this might trigger asserion in timeoutfound lab if state == PREPARE_TO_COMMIT
ndb/test/ndbapi/testNodeRestart.cpp:
  testcase
ndb/test/run-test/daily-basic-tests.txt:
  autotest

  print some in case of error


storage/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp:
  print scan state in case of error

  fix return value of "testBasic -n InsertError2 T1"


storage/ndb/test/ndbapi/testBasic.cpp:
  fix return value

  fix testNodeRestart -n Bug18612 when running with only 1 node group


storage/ndb/test/ndbapi/testNodeRestart.cpp:
  This test needs 2 node groups.

  remove too strong assertion


storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  remove too strong assertion

  fix so that getElement read localkey from lockowner instead of from page
  plus some cleanups


storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  1) Use OP_INITIAL instead of ~0
  2) Use JBB instead of JBA (once that I temporary changed...)
  3) Add more validation to validate_lock_queue (insert/delete)
  4) make getElement read localkey from lockowner instead of from page

  increase max time of MassiveRollback


storage/ndb/test/run-test/daily-basic-tests.txt:
  increase max time

  fix scan bugs introduced by acc modifications
  add more error testcases


storage/ndb/src/kernel/blocks/dbacc/Dbacc.hpp:
  remove unused state
storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  1) remove unused state
  2) Fix abort of running lock owner
  3) Fix abort of running op in parallell queue (especially scans)
storage/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp:
  remove some printouts
  add some jams
  fix so that close tupscan, can not acciently start acc scan in queue
    (NOTE limits #tupscans to 12 which is not really necessary...but the fix was easy)
storage/ndb/src/kernel/blocks/dbtup/DbtupScan.cpp:
  Use abort of locks when closing/blocked
    as Dbacc gets annoyed when committing an op with state running
storage/ndb/src/kernel/blocks/dbtux/DbtuxScan.cpp:
  Use abort of locks when closing/blocked
    as Dbacc gets annoyed when committing an op with state running
storage/ndb/test/include/HugoOperations.hpp:
  new method
storage/ndb/test/ndbapi/testBasic.cpp:
  add more test cases
storage/ndb/test/ndbapi/testScan.cpp:
  add more testcases
storage/ndb/test/run-test/daily-basic-tests.txt:
  add more testcases
storage/ndb/test/src/HugoOperations.cpp:
  add more testcases

  bug#19928 and bug#19929
  fix to critical bugs in tup scan
  that affected lcp,backup and opt. nr


storage/ndb/src/kernel/blocks/dbtup/DbtupScan.cpp:
  1) dont let dirty read scan find uncommitted inserts
  2) force opt. nr scan to wait for locked rows
  3) when finding LCP keep record, use accOpPtr -1, so that it will not be committed towards ACC

  introduce acc per row logical mutex to fix difficult error handling cases
  


storage/ndb/src/kernel/blocks/dbacc/Dbacc.hpp:
  1) Fix per row mutex so that only 1 op at a time is running on a row
  2) Change TUP_ALLOC/DEALLOC semantic, so that a new record will be allocated if LQ = { T1(DEL) - T2(INS) }
  3) Rewrite lock queus to be O(1) in all cases but a few abort cases where we scan parallell queue
  4) Impl. a validate_lock_queue/dump_lock_queue test framework
storage/ndb/src/kernel/blocks/dbacc/DbaccInit.cpp:
  1) Fix per row mutex so that only 1 op at a time is running on a row
  2) Change TUP_ALLOC/DEALLOC semantic, so that a new record will be allocated if LQ = { T1(DEL) - T2(INS) }
  3) Rewrite lock queus to be O(1) in all cases but a few abort cases where we scan parallell queue
  4) Impl. a validate_lock_queue/dump_lock_queue test framework
storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  1) Fix per row mutex so that only 1 op at a time is running on a row
  2) Change TUP_ALLOC/DEALLOC semantic, so that a new record will be allocated if LQ = { T1(DEL) - T2(INS) }
  3) Rewrite lock queus to be O(1) in all cases but a few abort cases where we scan parallell queue
  4) Impl. a validate_lock_queue/dump_lock_queue test framework
storage/ndb/src/kernel/blocks/dblqh/Dblqh.hpp:
  1) impl. a new read key from operation record needed by acc
  2) expand TRACE_OP toolkit
  3) impl. ACCKEY_ORD as needed by ACC changes
storage/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp:
  1) impl. a new read key from operation record needed by acc
  2) expand TRACE_OP toolkit
  3) impl. ACCKEY_ORD as needed by ACC changes
storage/ndb/src/kernel/blocks/dbtup/Dbtup.hpp:
  remove unused states/methods
storage/ndb/src/kernel/blocks/dbtup/DbtupExecQuery.cpp:
  remove extremly tricky code that handles disk_insert_but_no_mem_insert
    that is no long needed with current acc changes
storage/ndb/src/kernel/blocks/dbtup/DbtupGen.cpp:
  remove unused states/methods
storage/ndb/test/ndbapi/testOperations.cpp:
  renable last 3 lock upgrade testcases since they now pass

  fix crashing dump numbers


storage/ndb/include/kernel/signaldata/DumpStateOrd.hpp:
  fix crashing dump numbers

  Add mutex surronding sessions, as ndb_mgmd now actively tries to go and "purge stale sessions"


storage/ndb/include/util/SocketServer.hpp:
  Add mutex surronding sessions, as ndb_mgmd now actively tries to go and  "purge stale sessions"
storage/ndb/src/common/util/SocketServer.cpp:
  Add mutex surronding sessions, as ndb_mgmd now actively tries to go and  "purge stale sessions"
storage/ndb/src/mgmsrv/Services.cpp:
  Add mutex surronding sessions, as ndb_mgmd now actively tries to go and  "purge stale sessions"

  Fix recursive mutex lock in drop index (ndbapi programs only)


storage/ndb/src/ndbapi/NdbDictionaryImpl.cpp:
  Fix recursive mutex lock in drop index (ndbapi programs only)
storage/ndb/src/ndbapi/NdbDictionaryImpl.hpp:
  Fix recursive mutex lock in drop index (ndbapi programs only)

into  perch.ndb.mysql.com:/home/jonas/src/51-jonas

into  eel.(none):/home/jonas/src/51-jonas

into  perch.ndb.mysql.com:/home/jonas/src/51-jonas


storage/ndb/src/kernel/blocks/dbtc/Dbtc.hpp:
  Auto merged
storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp:
  Auto merged

storage/ndb/src/kernel/blocks/dbacc/Dbacc.hpp:
  more unused variables
storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  more unused variables

storage/ndb/src/kernel/blocks/dbacc/Dbacc.hpp:
  remove unused variables
storage/ndb/src/kernel/blocks/dbacc/DbaccMain.cpp:
  remove unused variables

into  perch.ndb.mysql.com:/home/jonas/src/mysql-5.1-new


storage/ndb/src/kernel/blocks/dbtc/Dbtc.hpp:
  Auto merged
storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp:
  Auto merged

into  mysql.com:/home/mydev/mysql-5.1-bug10405


storage/myisam/mi_check.c:
  Auto merged

into  mysql.com:/home/my/mysql-5.1


mysql-test/mysql-test-run.pl:
  Auto merged
sql/ha_ndbcluster.cc:
  Auto merged

(Fixes core dump in rpl_failed_optimize.test)
Ensure we end active transcations if we do an admin command (like optimize, repair etc)


mysql-test/extra/rpl_tests/rpl_failed_optimize.test:
  Added extra test + drop of table at end of test
mysql-test/lib/mtr_report.pl:
  Fail if mysqld asserts or prints stack
mysql-test/mysql-test-run.sh:
  Fail if mysqld asserts or prints stack
mysql-test/r/exampledb.result:
  Cleanup of events_tests (as this caused a lot of problems if it didn't work)
mysql-test/r/innodb.result:
  Extra test to see that we can do an optimize table on an active transaction
mysql-test/r/rpl_failed_optimize.result:
  Added extra test + drop of table at end of test
mysql-test/t/exampledb.test:
  Cleanup of events_tests (as this caused a lot of problems if it didn't work)
mysql-test/t/innodb.test:
  Extra test to see that we can do an optimize table on an active transaction
sql/handler.cc:
  Don't use row level logging on optimize or repair table.
sql/log.cc:
  Simplify code (no logic changes)
sql/mysql_priv.h:
  Added prototype
sql/sql_base.cc:
  Better name for define
sql/sql_class.cc:
  Indentation fix
sql/sql_parse.cc:
  Make end_active_trans() global
sql/sql_table.cc:
  Ensure we end active transcations if we do an admin command (like optimize, repair etc)

into  mysql.com:/home/ndbdev/jmiller/clones/mysql-5.1-new

    fixed memleak in ndbcluster_end


mysql-test/mysql-test-run.pl:
  do not install ndb slave if no slaves
sql/ha_ndbcluster.cc:
  fixed memleak in ndbcluster_end

into  mysql.com:/home/my/mysql-5.1


sql/ha_ndbcluster.cc:
  Auto merged
mysql-test/t/disabled.def:
  Manual merge

into  mysql.com:/home/mydev/mysql-5.1-bug10405

into  poseidon.ndb.mysql.com:/home/tomas/mysql-5.1-new-ndb