v1.2.6

What's Changed

• AddingPowerSupport_CI/Testing by @santosh653 in #37
• Change LICENSE back to the original MIT License by @cfiderer in #47
• Update README.md by @lukasmalkmus in #48
• RFC: Consume z.current prepared by Reader.Read in Reader.WriteTo by @mtrmac in #50

New Contributors

• @santosh653 made their first contribution in #37
• @cfiderer made their first contribution in #47
• @lukasmalkmus made their first contribution in #48
• @mtrmac made their first contribution in #50

Full Changelog: v1.2.5...v1.2.6

v1.2.5

reader: Fix missing buffers - #35

v1.2.4

Never return negative values from `Write`, as they can panic in `bufi…

…o` (#33)

* Never return negative values from `Write`, as they can panic in `bufio`

v1.2.3

pgzip: Fix an occasional segfault/panic with pgzip (#30)

This change fixes a occasional (5-10 out of 1000 test runs) segfault/panic on
sync.Pool struct access. Basically what happens there is that the original code
assigns a zero-value sync.Pool struct to the existing z.dstPool variable.
In general this kind of assignment is safe only if object is not used anymore /
in parallel by anything else. While is kind of true in a sense that none of the
goroutines are using it, in case of sync.Pool the garbage collector does
special management of the data with weak references and in right weather
conditions corrupts internal state (more precisely the sync.Pool.local might
become nil before reading it but after reading a nonzero sync.Pool.localSize;
leading to invalid memory reference in sync.Pool.Get() ).

v1.2.2

godoc: added new Default values for SetConcurrency (#29)

it seems that the default was changed and the function documentation was wrong

v1.2.1

v1.2.1

v1.1

Merge pull request #14 from klauspost/fix-truncated-gzip-input

Fix truncated gzip input

v1.0.1: Merge pull request #12 from klauspost/rewrite-error-handling

Rewrite error handling

v1.0: Update .travis.yml

Add OSX to test