This module provides Template Haskell functionality to generate data types with automatic Pack/Unpack instances. It uses a simple grammar to define data types and their field mappings.

Usage

Define data types using the pkmacro quasi-quoter. The syntax is indentation-insensitive:

-- First, set up default instances
setdefaultderives  -- Sets up Generic, Pack, and Unpack derives

-- Define a newtype wrapper with Pack/Unpack instances
newtype AAA = AAA Int32
  deriving stock (Generic)
  deriving newtype (Pack, Unpack)

[pkmacro|
-- Regular data type with two fields
data A {
  f1 :: Int32,                -- Regular field
  f2 :: Int32 via AAA,       -- Field with custom serialization
}

-- Data type with one field and explicit deriving
data B {
  f3 :: Int32,
  deriving (Generic, Show)
}

-- Empty data type (creates constructor with no fields)
data C {}
|]

The grammar supports:

• Empty data types (no fields)
• Custom serialization via via clause
• Multiple data types in one block
• Indentation-insensitive syntax
• Comments (both -- and {- -} style)

Syntax

The full syntax for data type definitions is:

data TypeName {
  field1 :: Type1 [via Type2],     -- Field with optional via clause
  field2 :: Type3,                 -- Regular field
  [deriving (Class1, Class2)]      -- Optional proper deriving clause
  [and shadow deriving (Pack, Unpack) with (Class3, Class4)] -- Optional shadow deriving
}

Elements in square brackets are optional.

• The via clause specifies a different type to use for serialization
• deriving adds instances to the main data type
• shadow deriving adds instances to the generated shadow type used for serialization
• Multiple data types can be defined in a single quasi-quoter block

Type Syntax

Types can include:

• Simple types: Int32, Text, etc.
• Parameterized types: Maybe a, [Int]
• Type applications: a @k
• Promoted types: 'True, 'Just
• Type literals: "hello", 123
• Parenthesized types: (a, b), (Either a b)

See: Pack, Unpack, setdefaultderives, addproperderives, and addshadowderives.