scrubcsv 1.0.0

Added

• Official 1.0.0 release, because this program has been used in production for
  years.
• --clean-column-names=stable provides a new column-name-cleaning algorithm.
  This converts ASCII characters to lowercase, converts spaces to _, and
  verifies that the resulting column name is a unique, valid C identifier. The
  goal of this new column name cleaner is to make our output column names
  easily predictable.
• --reserve-column-names=REGEX will cause scrubcsv to fail if it would
  generate output columns matching REGEX.
• Binary builds for more platforms, including M1.

hashcsv 1.0.1

Added

• Build new binaries, including Mac M1.

geochunk 1.0.0

Added

• Official 1.0.0 release, because this program has been used in production for
  years.
• Binary builds for more platforms, including M1.

fixed2csv 1.0.0

Added

• Official 1.0.0 release, because this program has been used in production for
  years.
• Binary builds for more platforms, including M1.

catcsv 1.0.0

Added

• Official 1.0.0 release, because this program has been used in production for
  years.
• Binary builds for more platforms, including M1.

hashcsv 1.0.0

Changed

• Stabilized CLI interface.
• Set up automatic binary releases using GitHub Actions.

hashcsv v0.1.1 (no binary release)

Changed

• Inner loop optimized, raising throughput from ~256 KiB/s to 65 MiB/s. I had suspected that two memory allocations per CSV row were slowing us down, but I am surprised that eliminating them improved performance by a factor of 250×. As always, malloc is highly expensive in hot loops.