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

  Make sure that SUB_GCP_COMPLETE_ACK is sent uncond, failure to do so can lead to problems...


storage/ndb/src/kernel/blocks/ndbfs/AsyncFile.cpp:
  retry on O_DIRECT
storage/ndb/src/ndbapi/TransporterFacade.cpp:
  Make sure that SUB_GCP_COMPLETE_ACK is sent uncond

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


storage/ndb/src/ndbapi/TransporterFacade.cpp:
  Auto merged
storage/ndb/src/ndbapi/TransporterFacade.hpp:
  Auto merged
storage/ndb/src/ndbapi/Ndbif.cpp:
  merge

  post merge


ndb/src/ndbapi/TransporterFacade.hpp:
  merge fix

into  perch.ndb.mysql.com:/home/jonas/src/50-work


ndb/src/ndbapi/Ndbif.cpp:
  Auto merged
ndb/src/ndbapi/TransporterFacade.cpp:
  Auto merged
ndb/src/ndbapi/TransporterFacade.hpp:
  Auto merged

  fix api_signal_log print of uncond signals


ndb/src/ndbapi/Ndbif.cpp:
  Send TC_COMMIT_ACK uncond
ndb/src/ndbapi/TransporterFacade.cpp:
  Fix printout of uncond signals
ndb/src/ndbapi/TransporterFacade.hpp:
  Fix printout of uncond signals

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

  Add abort in reception of invalid/inconsistent SUB_GCP_COMPLETE_REP
  as proceeding (in release wo/ asserts) make mysqld crash in mysterious ways


storage/ndb/src/ndbapi/NdbEventOperationImpl.cpp:
  Add abort in reception of invalid/inconsistent SUB_GCP_COMPLETE_REP
    as proceeding (in release wo/ asserts) make mysqld crash in mysterious ways

  handle open(O_DIRECT) -> EINVAL
  remove need for NDB_NO_ODIRECT


storage/ndb/src/kernel/blocks/ndbfs/AsyncFile.cpp:
  handle open(O_DIRECT) -> EINVAL

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


client/mysqltest.c:
  Auto merged
sql/ha_ndbcluster.cc:
  Auto merged
mysql-test/t/disabled.def:
  manual merge

  Fix 4 replica bug in packLqhKeyReq, problem introduced by opt. nr.


storage/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp:
  Fix packLqhKeyReqLab when having full AI in this and appending gci/rowid (used to casue 23+3 = 26 in signal length)
  Note: This is a very ineffectice patch, but only for 4 replicas
  This will be fixed when sending lqhkeyreq as long signal (or even just attrinfo as long)

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


server-tools/instance-manager/options.cc:
  Resolve conflict

into  mysql.com:/home/cps/mysql/devel/5.1-im-sort-options

server-tools/instance-manager/options.cc:
  sort options

into mysql.com:/Users/kent/mysql/bk/mysql-5.1-new


mysql-test/t/disabled.def:
  Auto merged

  Disabled 'rpl_deadlock_innodb', randomly failing


mysql-test/t/disabled.def:
  Disabled 'rpl_deadlock_innodb', randomly failing

mysql-test/mysql-test-run.pl:
  Support tracing the instance manager
server-tools/instance-manager/options.cc:
  Handle --debug/-#
server-tools/instance-manager/options.h:
  Handle --debug/-#

  To somplify merge later, make files more similar in format


mysql-test/t/disabled.def:
  To somplify merge later, make files more similar in format

freeing non freed objects on server shutdown
corrected timeout handling on schema events

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

mysql-test/include/master-slave-end.inc:
  New BitKeeper file ``mysql-test/include/master-slave-end.inc''
mysql-test/include/ndb_setup_slave.inc:
  New BitKeeper file ``mysql-test/include/ndb_setup_slave.inc''

    better replicaiton test cleanup to ensure better test predictability
+ some ndb test cleanup


client/mysqltest.c:
  increase save_master_pos timeout to 30 seconds for higher test predictability
mysql-test/extra/rpl_tests/rpl_ddl.test:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/extra/rpl_tests/rpl_row_UUID.test:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/r/rpl_ndb_dd_advance.result:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/r/rpl_ndb_sync.result:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/t/disabled.def:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/t/rpl_ndb_bank.test:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/t/rpl_ndb_basic.test:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/t/rpl_ndb_dd_advance.test:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/t/rpl_ndb_dd_basic.test:
  better replicaiton test cleanup to ensure better test predictability
mysql-test/t/rpl_ndb_sync.test:
  better replicaiton test cleanup to ensure better test predictability

into mysql.com:/home/hf/work/mysql-5.1.18115

client/mysql_upgrade bunary should be used instead


BitKeeper/deleted/.del-mysql_upgrade.sh~826da969ccf96ef:
  Delete: scripts/mysql_upgrade.sh
scripts/Makefile.am:
  mysql_upgrade.sh removed

into  mysql.com:/usr/home/bar/mysql-5.1-new

"upgrade" failed on windows, because of lower_case_table_name.
This fix changes letters to lower case, not to depend on OS.

Note, this change does not affect coverage of 
Bug N 17142 "Crash if create with encoded name",
which the test war originally written for.


mysql-test/r/upgrade.result:
  Changing "P" to "p" not to depend on lower_case_table_name
mysql-test/t/upgrade.test:
  Changing "P" to "p" not to depend on lower_case_table_name

into  mysql.com:/home/mysql_src/mysql-5.1-new

This is faster under gdb/valgrind (this should speed up work of people who have to run
a simple test 50 times to do debugging). Running without --big, which is the default, InnoDB tests
take at most 18MB so 128 was overkill. Running with --big, InnoDB tests take 450MB so 128
was not enough anyway.


mysql-test/mysql-test-run.pl:
  10M is faster to create than 128M (especially under gdb or valgrind) and is enough for most tests.

into  mysql.com:/usr/home/bar/mysql-5.1-new.b18285

into  neptunus.(none):/home/msvensson/mysql/mysql-5.1

into  neptunus.(none):/home/msvensson/mysql/mysql-5.1


mysql-test/lib/mtr_process.pl:
  Auto merged

 - Part 1, better handling of subprocess and stdout to be done


mysql-test/lib/mtr_process.pl:
  - Dont redirect stdout if ActiveState perl since fork creates a new thread and not another process.
  - Change waitpid to allow for negative pid as we get on Windows, strange but 
    that is how it is. It should never create a process with pid -1 though.
  - Remove an "if" for cygwin that does not do anything.
  - Add "print" in the potentially hanging while loop in "start_reap_all"
  - Dont use "getpgrp" on windows.
  - Remove "sleep 2" in mtr_exit
  - Remove the initial ping of mysqld's in 'mysql_admin_shutdown', instead just start one 
    mysqladmin for each expected mysqld server and wait for them to return.
    Then perform the ping of the mysqld server's. Those ping's normally return immediately.

BitKeeper/deleted/.del-mysql_test_run_new.dsp~9d8078f3f02fcc8e:
  Delete: VC++Files/mysql-test/mysql_test_run_new.dsp
BitKeeper/deleted/.del-mysql_test_run_new_ia64.dsp~e7ee71ec8d2995e3:
  Delete: VC++Files/mysql-test/mysql_test_run_new_ia64.dsp
BitKeeper/deleted/.del-mysql_test_run_new.vcproj~1ddaed30361efefe:
  Delete: VC++Files/mysql-test/mysql_test_run_new.vcproj
VC++Files/mysql.dsw:
  Remove reference to mysql_test_run_new
VC++Files/mysql.sln:
  Remove reference to mysql_test_run_new