Home

Real-life setup and configuration

Paths and checkers

Internally, the library works with canonical type descriptions. A canonical type description is just a map with paths as keys and vectors of functions (checkers) as values.

Each path is used to descend into a whole value to discover a set of leaf values that are checked by the checkers.

• Path elements (what can be in the left-hand side)
• Checkers (what can be in the right-hand side)

Descending into structures

The happy case is when the path matches the whole value's structure. What happens, though, when the next path element is :a and the next bit of the structure isn't associative? Or if the next bit is a map but it doesn't have an :a key?

• Misshapen structures
• Truncated structures
• The problem of nil

Shorthand for condensed type descriptions

Although you can, and often do, use canonical type descriptions, you'll often instead compose them out of a sequence of condensed type descriptions, such as this:

(type! :X record? {:a even?} (requires :a :b))

• Rejecting nil and missing values
• Paths from maps where checkers belong
• Explicitly branching paths
• Describing the whole value

Some esoteric topics

• Rejecting extra keys
• Helpful error messages from anonymous checkers
• Behind the scenes: How checking and type definition works
• [[The implies predicate]]

Gotchas

• Two-dimensional arrays
• Redefining types used in definitions of other types

Glossary