Cubdb Versions Save

Changes:

• [fix] When an update funcyion (such as get_and_update/3) does not change any entry, avoid any disk write, including the transaction header.

Changes:

• get_and_update/3 avoids unnecessary disk write when the value is unchanged

Changes:

• [fix] fix race condition during compaction

Changes:

• [breaking change] CubDB.put_new/3 returns {:error, :exists} instead of :exists when the key already exists.

Changes:

• CubDB.put_and_delete_multi/3 to atomically put and delete entries
• CubDB.put_new/3 to put an entry only if the key does not exist yet
• More efficient implementation of CubDB.put_multi/2 and CubDB.delete_multi/2
• CubDB.get_multi/2 does not block writers

First release candidate of CubDB 1.0.0 🚀

Changes:

• [breaking change] better defaults for CubDB:
  • auto_file_sync now defaults to true (slower but durable)
  • auto_compact now defaults to true
• [breaking change] select/2 now takes the timeout as an option instead of an additional argument
• [breaking change] get_and_update_multi/4 now takes the timeout as an option instead of an additional argument
• Better internal handling of timeouts that ensures cleanup of resources on the callee side

Changes:

• Better error handling when initializing CubDB
• Do not crash main CubDB process if compaction process crashes

Changes:

• Add CubDB.stop/3 function to stop the process
• Prevent opening multiple CubDB processes on the same data file
• Documentation improvements

Changes:

• Add How To section to documentation
• Documentation improvements

Changes:

• Documentation improvements