Revision history for Data-Consumer
0.17 Wed Nov 1 2017
Fix a bug with ignored in MySQL2, thanks to Maksym Davydov
0.16 Fri Jun 24 2016
The deprecated warning behavior of Data::Consumer::MySQL was
a major blocker to upgrading Data::Consumer, so it has been
changed to be opt in. See the docs for Data::Consumer::MySQL.
0.15 Fri May 8 2015
Fork Data::Consumer::MySQL into Data::Consumer::MySQL2. There
were and are subtle issues with the implementation of DC::MySQL.
As we have a lot of code using it the easiest way is to fork
the module and fix it in the new version. *Most* old code should
be unaffected. However if you are processing very quickly, with
many workers, on a newer MySQL especially, then it is possible
that two workers might process the same record twice due to a subtle
race condition in MySQL query plan evaluation. The new version in
MySQL2 should be robust to this problem.
0.14 Sun Oct 12 2014
Reset the 'fail' flag before processing any new item
0.13 Mon Jul 18 2011
Fixed bug in leave() implementation: The status for 'unprocessed'
could previously not be '0'.
0.12 Fri Mar 27 2011
Note in documentation that an explicit lock_prefix should almost
always be specified.
0.11 Fri Mar 27 2011
Fix SELECT GET_LOCK order-of-evaluation issue.
This is an important bug fix if you're running Data::Consumer in
a pipeline-fashion with multiple consumers working on items in
various stages. If in doubt, upgrade.
0.10
Lots of changes. Fixed an important bug and disabled the broken sweep
functionality for now. Also silence some warnings on older perls.
0.09
Added support for 'leave', 'ignore', 'fail' and 'halt'. They probably
could be better documented and tested.
0.08
Pretty close to ready for release. Cleaned up pod and perltidied
the code (my current parameters of choice are in comments in .pm
files). Spent some time on the make process to have configurable
mysql arguments, which default to off for automated processed.
0.01 2008-02-10
Initial creation. Support Mysql and File based data resources.