• BREAKING CHANGE: Removed deprecated FromBytes. Use OpenBytes instead.
• Fixed verifier metadata error message to require a non-empty map for the
  database description. GitHub #187.
• Introduces the v2 API with Reader.Lookup(ip).Decode(...), netip.Addr
  support, custom decoder interfaces, and richer error reporting.
• See MIGRATION.md for guidance on upgrading projects from v1 to v2.

• Replaced runtime.SetFinalizer with runtime.AddCleanup for resource
  cleanup in Go 1.24+. This provides more reliable finalization behavior and
  better garbage collection performance.

• SECURITY: Fixed integer overflow vulnerability in search tree size
  calculation that could potentially allow malformed databases to trigger
  security issues.
• SECURITY: Enhanced bounds checking in tree traversal functions to return
  proper errors instead of silent failures when encountering malformed
  databases.
• Added validation for invalid prefixes in NetworksWithin to prevent
  unexpected behavior with malformed input.
• Added SkipEmptyValues() option for Networks and NetworksWithin to skip
  networks whose data is an empty map or empty array. This is useful for
  databases that store empty maps or arrays for records without meaningful
  data. GitHub #172.
• Optimized custom unmarshaler type assertion to use Go 1.25's
  reflect.TypeAssert when available, reducing allocations in reflection code
  paths.
• Improved memory mapping implementation by using SyscallConn() instead of
  Fd() to avoid side effects and prepare for Go 1.25+ Windows I/O
  enhancements. Pull request by database64128. GitHub #179.
• Added OpenBytes function for better API discoverability and consistency
  with Open(). FromBytes is now deprecated and will be removed in a future
  version.

2.0.0-beta.8

• Fixed "no next offset available" error that occurred when using custom
  unmarshalers that decode container types (maps, slices) in struct fields.
  The reflection decoder now correctly calculates field positions when
  advancing to the next field after custom unmarshaling.

2.0.0-beta.7

• Update capitalization of "uint" in ReadUInt* to match KindUint* as well
  as the Go standard library.

2.0.0-beta.6

• Update capitalization of "uint" in ReadUInt* to match KindUint* as well
  as the Go standard library.

2.0.0-beta.5

• Added Offset() method to Decoder to get the current database offset. This
  enables custom unmarshalers to implement caching for improved performance when
  loading databases with duplicate data structures.
• Fixed infinite recursion in pointer-to-pointer data structures, which are
  invalid per the MaxMind DB specification.

• BREAKING CHANGE: Removed experimental deserializer interface and
  supporting code. Applications using this interface should migrate to the
  Unmarshaler interface by implementing UnmarshalMaxMindDB(d *Decoder) error
  instead.
• Open and FromBytes now accept options.
• BREAKING CHANGE: IncludeNetworksWithoutData and IncludeAliasedNetworks
  now return a NetworksOption rather than being one themselves. These must now
  be called as functions: Networks(IncludeAliasedNetworks()) instead of
  Networks(IncludeAliasedNetworks). This was done to improve the documentation
  organization.
• Added Unmarshaler interface to allow custom decoding implementations for
  performance-critical applications. Types implementing
  UnmarshalMaxMindDB(d *Decoder) error will automatically use custom decoding
  logic instead of reflection, following the same pattern as
  json.Unmarshaler.
• Added public Decoder type and Kind constants in mmdbdata package for
  manual decoding. Decoder provides methods like ReadMap(), ReadSlice(),
  ReadString(), ReadUInt32(), PeekKind(), etc. Kind type includes
  helper methods String(), IsContainer(), and IsScalar() for type
  introspection. The main maxminddb package re-exports these types for
  backward compatibility. NewDecoder() supports an options pattern for
  future extensibility.
• Enhanced UnmarshalMaxMindDB to work with nested struct fields, slice
  elements, and map values. The custom unmarshaler is now called recursively
  for any type that implements the Unmarshaler interface, similar to
  encoding/json.
• Improved error messages to include byte offset information and, for the
  reflection-based API, path information for nested structures using JSON
  Pointer format. For example, errors may now show "at offset 1234, path
  /city/names/en" or "at offset 1234, path /list/0/name" instead of just the
  underlying error message.
• PERFORMANCE: Added string interning optimization that reduces allocations
  while maintaining thread safety. Reduces allocation count from 33 to 10 per
  operation in downstream libraries. Uses a fixed 512-entry cache with per-entry
  mutexes for bounded memory usage (~8KB) while minimizing lock contention.

• Open will now fall back to loading the database in memory if the file-system does not support mmap. Pull request by database64128. GitHub #163.
• Made significant improvements to the Windows memory-map handling. . GitHub #162.
• Fix an integer overflow on large databases when using a 32-bit architecture. See ipinfo/mmdbctl#33.

• Allow negative indexes for arrays when using DecodePath. #152
• Add IncludeNetworksWithoutData option for Networks and NetworksWithin. #155 and #156