Dynamic type change, Faster Decoding, io.Reader ingestion

• (Breaking) Dynamic type moved map[string]interface{} to []byte type allowing the user to choose the decode method
• Moved from generic JSON decoding to a more allocation efficient method with a fork of stdlib/encoding/json
• Most tests made to be parallel
• ingest.FromReader() added to allow writing from anything implementing io.Reader for blobstore ingestion to Kusto
• ingest client now caches mapping references for 5 minutes before needing another check on if the mapping reference exists
• Fixed bug that was preventing ingestion on Windows paths
• Added more file format support for ingestion

v0.1.3

Removing the kusto/ preface

Performance enhancements, end to end tests, remove logging

This release makes a variety of performance enhancements, speeding decodes 2x. It now decodes faster than the server sends (verified against other language clients).

Merged our internal end to end test.

Removed logging lines except for when using Unsafe, which can now be suppressed.

Provide detection of non-transient errors

This release adds errors.Retry(), which will take an error and determine if can be retried or not.

In addition, this does some cleanup to errors that are being returned, doc fixes and other minor misc. improvements.

Fix for v1 framing DataType translation bug

See issue: #11

Initial Release

Summary

This is the initial 0.1.0 Release, developed within Microsoft Azure DevOps and migrated to Github for public release.

This includes support for:

• Kusto Queries
• Kusto Management
• Kusto Ingestion

Authors

• John Doak
• Daniel Dubovski

Additional Thanks

Kusto Dev:

• Vladik Branevich for letting Daniel and I get this work completed.
• Ziv Caspi - For help in the early days with some of my streaming questions and pointing me in the right direction for query parameters.

Cloud Advocacy:

• Aaron Wislang - Aaron helped review all the documentation, tested the code and compiled the doc examples that weren't stored as compilable examples. His feedback was invaluable.

Container Compute Team:

• David Justice - David would let me bounce ideas and questions off of him any time of the day (or night). He would point me to people who could help me and was both helpful and insightful.

Process Engineering:

• Less Lincoln/Bink Naughton - They kept pressure off me to allow this to happen. This is a little outside my day job and when I needed to really crank down on this for a couple weeks to ship, they didn't hesitate to let me. They bought into helping the entire Azure community and helping our colleagues.

If I missed anyone, please let me apologize, it was not intentional.