GitHub user chiastic-security opened a pull request:
https://github.com/apache/phoenix/pull/118
Correct two bugs in PMetaDataImpl
More entries than necessary could end up being evicted.
This PR also adds a new test to catch one of the bugs.
Previously, one of the bugs (`>` should have been `>=`) was being picked up
on Java 8, but not Java 7, because the traversal ordering of one of the JDK
collections happened to have changed. The bug was still present in Java 7, but
the collection ordering meant that the test didn't pick it up.
The second bug (`if` should have been `while`) is picked up by the new test
on both Java 7 and Java 8. The essence of the bug is that if several things are
queued for eviction, and something large and old gets added to the eviction
queue, the implementation was only considering removing one element from the
eviction queue, whereas it should have removed as many as possible from the
eviction queue while still leaving enough to clear space for the new element
being added to the table.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/chiastic-security/phoenix-for-cloudera master
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/phoenix/pull/118.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #118
----
commit def72e50806cd2c4ea50807a25667714099a4a5d
Author: James Heather <[email protected]>
Date: 2015-10-02T14:14:24Z
Correct two bugs in PMetaDataImpl, whereby more entries than necessary
could end up being evicted; also add a new test to catch one of the bugs.
Previously, one of the bugs (">" should have been ">=") was only being
picked up on Java 8, because the traversal ordering of one of the JDK
collections happened to have changed. The bug was still present in Java 7, but
the collection ordering meant that the test didn't pick it up.
The second bug ("if" should have been "while") is picked up by the new test
on both Java 7 and Java 8. The essence of the bug is that if several things are
queued for eviction, and something large and old gets added to the eviction
queue, the implementation was only considering removing one element from the
eviction queue, whereas it should have removed as many as possible from the
eviction queue while still leaving enough to clear space for the new element
being added to the table.
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
