Apache Release 1.2.0-incubating

Removes a check that required the runtime JDK to be JDK-8. This does not change the requirement that JDK-8 is required for full runtime operation. However in environments that do not use the advanced off-heap (direct) capabilities of Memory this module may still work in JDK versions later than JDK-8. However, there is no guarantee of this as this code has not been fully refactored and tested to operate in these later versions. Caveat Emptor.

This release was created for the convenience of some major users that need to do testing in versions later than JDK-8, and do not use the advanced features of Memory.

There are no other significant changes in this version compared to the previous 1.1.0-incubating release.

1.2.0-incubating-RC1

Release Candidate

Apache Release 1.1.0-incubating

This release came from RC2, which fixed some problems in the structure of the POM file of
RC1. There are no other changes.

1.1.0-incubating-RC1

This candidate was not released. There were some structural problems in the POM file.

Apache Release 1.0.0-incubating

This is the first release as Apache DataSketches-memory (incubating). This is a work-in-progress as it doesn't have all the licensing details quite right. But the code is good and passes all the tests. There are no significant code changes from the previous release. There will be another release coming with some speed improvements. So we suggest that you wait for the next release.

memory-0.12.2

Added new functionality.

XxHash is now not only available from Memory, it also is directly available for hashing primitive arrays. See class XxHash64.

Added xxHash, WritableMemory.allocate(capacity, ByteOrder)

• Added xxHash(...) that is accessible via Memory or Buffer and also replaces the slow hashCode().

• Added WritableMemory.allocate(capacity, ByteOrder)

• Added LGTM integration and badges.

• Javadoc improvements & code cleanup.

Improved performance, expanded API

memory-0.12.0

[maven-release-plugin] copy for tag memory-0.12.0

Added Non-Native Endianness

• Byte arrays can be wrapped with optional ByteOrder specified with non-native endianness. Any putting or getting of multibyte primitives or primitive arrays into this byte array will be automatically byte-swapped.

• The endianness of ByteBuffers will be detected when wrapping the ByteBuffer. Any putting or getting of primitives will be automatically byte-swapped.

• This release includes some minor API changes to make them more understandable.

• This release includes some bug fixes.

Added PutCharsToUtf8() and GetCharsFromUtf8()

• Added PutCharsToUtf8() and GetCharsFromUtf8()

These methods have been specifically designed to avoid garbage generation during the conversions. The speeds are comparable with the Java built-in methods.

• Made the bounds checking more conservative when putting and getting arrays, where the cost of the bounds check is negligible when compared to the cost of internal array copies.

• Cleaned up some of the object construction cost of the ResourceState class.