Daniel Fischer

lp:~dannythefool/+junk/trunk-outdated

Created by Mats Kindahl on 2011-10-10 and last modified on 2012-10-16

This is an outdated branch that's no longer being updated.

Get this branch:: bzr branch lp:~dannythefool/+junk/trunk-outdated

Only Daniel Fischer can upload to this branch. If you are Daniel Fischer please log in for upload directions.

Related bugs

Bug #1111: doesn't extract last track	Medium	Fix Released
Bug #3782: Conflict when converting my baz branches to bzr	Medium	Invalid

Link a bug report

Related blueprints

Branch information

Owner:: Daniel Fischer

Status:: Abandoned

Recent revisions

3782. By Gleb Shchepa <email address hidden> on 2012-05-03

WL#5855: Print EXPLAIN in JSON format

Minor update to make old Sun C++ compiler happy.

3781. By Tor Didriksen <email address hidden> on 2012-05-03

BUG#13983918 - FACTOR OUT MY_GET_ERR_MSG() FROM MY_ERROR() FOR GENERAL USE

Post-push fix for unit test,
adjust to different perror() messages on different platforms.

3780. By Frazer Clement <email address hidden> on 2012-05-03

Second part of not removing MySQL Cluster binary man pages

3779. By Alexander Barkov <email address hidden> on 2012-05-03

Bug#11765016 / 57926: ILLEGAL MIX OF COLLATIONS FOR OPERATION 'UNION' .. USING CONCAT/FUNCTION/
Not a bug. Adding coverage tests.

3778. By Frazer Clement <email address hidden> on 2012-05-03

Don't remove man pages for ndb* binaries in MySQL Cluster

3777. By Hemant Kumar on 2012-05-03

Added "--big-test" option corresponding to "embedded" run in default.weekly.

3776. By Frazer Clement <email address hidden> on 2012-05-03

WL#5917 Appending data to Rows_log_event

Overview
--------
This feature implements new 'version 2' Binlog row events for the existing
WRITE_ROW, UPDATE_ROW, DELETE_ROW events. Version 2 Binlog row events
become the default.

These new 'version 2' Binlog row events can transport extra information
*with* each row event.

A Master can optionally generate 'version 1' Binlog row events by setting the
--log-bin-use-v1-row-events=1 MySQLD option/variable. This is necessary
during an upgrade where the Master is upgraded before the Slave.

Motivation
----------
This feature is primarily motivated by WL5353 Reflecting GCI, which uses
the extra row event data to transport per-row event transaction identities.

Details
-------
- 3 new event types added
- Existing row events renamed with _v1 suffix
- log-bin-use-v1-row-events option added
- Existing row event classes reused
- Extra row info is one use case of new 'v2' extensible row
event format. Should be no need for 'v3'.
- Binlog injector Api enhanced to allow info to be passed in
- THD Api enhanced to expose info to storage engine running as
Slave SQL thread
- mysqlbinlog tool enhanced to show extra info in --verbose mode.
- New testcases : rpl_extra_row_data, log_bin_use_v1_row_events_basic
- Enhanced testcase : binlog_old_versions
- Various corrected / adjusted testcases

3775. By Thayumanavar <email address hidden> on 2012-05-03

Bug#12951609: 62282: LOCK_THREAD_COUNT IS USED FOR RANDOM FUNCTIONS
PROBLEM: LOCK_thread_count was used to protect sql_rand.LOCK_thread_count
is used for a different purpose and is held across at various places
which causes hot contention during connect times.LOCK_thread_count is not
meant to protect sql_rand.
FIX: Added a new global LOCK_sql_rand rather than use LOCK_THREAD_COUNT.