Hello Kudu Jenkins, Wenzhe Zhou,
I'd like you to reexamine a change. Please visit
http://gerrit.cloudera.org:8080/17376
to look at the new patch set (#4).
Change subject: [util] Add special handling for empty strings in FastHash
......................................................................
[util] Add special handling for empty strings in FastHash
Impala uses special handling for empty strings while hashing
https://github.com/apache/impala/blob/master/be/src/runtime/raw-value.inline.h#L352
This leads to discrepancy between hash values for empty
strings between Impala and Kudu.
This change basically matches the behavior with Impala.
Since FastHash is not used for any persistent data structures
so there are no backward compatibility issues that needs special
handling during upgrade. However if there is a mismatch
between Kudu client and Kudu server versions then
results are unexpected. This issue should be mentioned
in the Kudu release notes.
Discovered bugs in the Java implementation when a negative
byte is promoted to a long and left shifted. Discovery
was because of the magic value for empty strings/buffer.
Thankfully the Java implementation of the FastHash
is not used currently.
ALso fixed a bug in predicate test where flush would be missed
due to increase in number of values which affected TSAN.
Change-Id: Idb2d401ab692decd5bc03b48a96d710546c0039e
---
M java/kudu-client/src/main/java/org/apache/kudu/util/HashUtil.java
M java/kudu-client/src/test/java/org/apache/kudu/util/TestFashHash.java
M src/kudu/client/CMakeLists.txt
M src/kudu/client/predicate-test.cc
M src/kudu/util/hash_util-test.cc
M src/kudu/util/hash_util.h
6 files changed, 91 insertions(+), 28 deletions(-)
git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/76/17376/4
--
To view, visit http://gerrit.cloudera.org:8080/17376
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: Idb2d401ab692decd5bc03b48a96d710546c0039e
Gerrit-Change-Number: 17376
Gerrit-PatchSet: 4
Gerrit-Owner: Bankim Bhavsar <ban...@cloudera.com>
Gerrit-Reviewer: Bankim Bhavsar <ban...@cloudera.com>
Gerrit-Reviewer: Kudu Jenkins (120)
Gerrit-Reviewer: Wenzhe Zhou <wz...@cloudera.com>
