lp:mysql-replication-listener

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

2 recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

73. By Mats Kindahl on 2013-04-20

Releasing version 0.1.0

72. By Mats Kindahl on 2013-04-20

Fixing README file and CMake files.

Updates the README file with instructions for how to build using Hadoop version
1.1 and 1.0.

Update the packing information since it was not accurate.

Extend the CMakeFiles to contain proper version numbers.

71. By Mats Kindahl on 2013-04-20

Fixing copyrights in some files

70. By Shubhangi Garg <email address hidden> on 2013-04-15

BAPI 57: Bug: 'make test' reports test failure for Binlog Tests

Description:
'make test' fails since

1.BINLOG_TESTS are not correctly called in CMakeLists in tests directory.
CMakeLists.txt has the following code to add BINLOG_TESTS:

  set(MySQL_BINLOG_TESTS replaybinlog replay_sys_vars)
  add_test(SimpleTests ${MySQL_SIMPLE_TESTS})

replaybinlog, however takes MySQL uri as a command line argument.
Here 'replay_sys_vars' is passed as the uri to connect to, and hence
the test fails.

2. replaybinlog and replay_sys_vars need mtr executables to run correctly.

Fix:
Disable them for the first version and add a TODO.

69. By Shubhangi Garg <email address hidden> on 2013-04-15

BAPI 56:Bug: Row_event_iterator runs into infinite loop

Description:

For a row event, in case if the data type of a column is of types newly
implemented in 5.6; namely TIME2, DATETIME2 or TIMESTAMP2;
then Row_event_iterator() (which is responsible to iterate over one row),
is stuck in an infinite loop .

It happens so because it takes the size of the row equal to a variable length;
which is defined by each data type of the columns in a row.
In case if the data type is not defined, the default case sets
length = ~(uint_32)0 or (2^32 - 1); and hence the iteration never ends!

Fix:

Since we get the data type in the table map event itself,
as well as the metadata, the data_type for the column is correct.
We just require to get the length correctly.
Also,default case is changed to set length=UINT_MAX and an exception is
thrown in case an unrecognized field type is encountered.

68. By Shubhangi Garg <email address hidden> on 2013-04-07

    BAPI 57: Bug: Flag for row_events not printed correctly if more than one
                  flags are set

    Description:
    Since more than one flag can be set at a time, the previous implementation
    was not correct; which interpreted variable 'flag' as an integer and not a
    bitmask.
    Therefore, if more than one flags are set, the interpreted flag values were
    incorrect.

    Fix:
    Modified get_flag_string() to interpret flags by bit positions being set
    as 0 or 1.

67. By Neha Kumari <email address hidden> on 2013-03-29

post push fix

66. By Neha Kumari <email address hidden> on 2013-03-26

 Modified the method set_position and added one new method
 fetch_binlog_name_and_size().

65. By Shubhangi Garg <email address hidden> on 2013-03-25

    BAPI 53: binlog-browser wrong values for fake FDE

  binlog-browser gives negative value as start-position for fake Format Description
  Event, generated when slave disconnects and connects again.
  It can be reproduced by using --start-position with the browser,
  and the output will have the start position negative.

64. By Shubhangi Garg <email address hidden> on 2013-03-21

     BAPI 2:
     1. Disconnect Method to be made public
     2. Disconnec Method to be called from file_driver.cpp after finishing
        file read

The above functionality is required since with the current code in the API, any
user application can connect, and then keep waiting for the events to occur.
In order to exit, there was no other way except aborting the program.