frosty

Serialize native Nim types to strings, streams, or sockets.

Usage

There are two operations: freeze and thaw.

freeze[T](input: T; output: Socket or Stream or var string)

import frosty

var
  data = someArbitraryDataFactory()
  handle = openFileStream("somefile", fmWrite)
# write serialized data into the file handle
freeze(data, handle)
close handle

thaw[T](input: Socket or Stream or string; output: var T)

import frosty

var
  data: SomeArbitraryType
  handle = openFileStream("somefile", fmRead)
# read deserialized data from the file handle
thaw(handle, data)
close handle

freeze[T](input: T): string

thaw[T](input: Socket or Stream or string): T

import frosty

# adhoc serialization and deserialization
var brrr = freeze("my data")
assert thaw[string](brrr) == "my data"

Performance

Frosty can handle cyclic data structures, but not memory graphs of extreme size -- you can exhaust the stack because our traversal is implemented via recursion. This will be solved soon.

Benchmark vs. Flatty

The source to the following benchmark is found in the tests directory.

The benchmark compares frosty to https://github.com/treeform/flatty for a few static datatypes -- flatty does not work with cast objects yet.

Installation

$ nimph clone disruptek/frosty

or if you're still using Nimble like it's 2012,

$ nimble install https://github.com/disruptek/frosty

Options

• --define:frostySorted=on to use Miran's sorta B-Trees.
• --define:frostyNet=off to omit Socket support for platform reasons.
• --define:frostyDebug=on to emit some debugging output.

Documentation

The documentation employs Nim's runnableExamples feature to ensure that usage examples are guaranteed to be accurate. The documentation is rebuilt during the CI process and hosted on GitHub.

License

MIT