rapid

package module
v1.1.0 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Aug 8, 2023 License: MPL-2.0 Imports: 29 Imported by: 42

README

rapid PkgGoDev CI

Rapid is a Go library for property-based testing.

Rapid checks that properties you define hold for a large number of automatically generated test cases. If a failure is found, rapid automatically minimizes the failing test case before presenting it.

Features

  • Imperative Go API with type-safe data generation using generics
  • Data generation biased to explore "small" values and edge cases more thoroughly
  • Fully automatic minimization of failing test cases
  • Persistence and automatic re-running of minimized failing test cases
  • Support for state machine ("stateful" or "model-based") testing
  • No dependencies outside the Go standard library

Examples

Here is what a trivial test using rapid looks like (playground):

package rapid_test

import (
	"sort"
	"testing"

	"pgregory.net/rapid"
)

func TestSortStrings(t *testing.T) {
	rapid.Check(t, func(t *rapid.T) {
		s := rapid.SliceOf(rapid.String()).Draw(t, "s")
		sort.Strings(s)
		if !sort.StringsAreSorted(s) {
			t.Fatalf("unsorted after sort: %v", s)
		}
	})
}

More complete examples:

Comparison

Rapid aims to bring to Go the power and convenience Hypothesis brings to Python.

Compared to testing.F.Fuzz, rapid shines in generating complex structured data, including state machine tests, but lacks coverage-guided feedback and mutations. Note that with MakeFuzz, any rapid test can be used as a fuzz target for the standard fuzzer.

Compared to gopter, rapid provides a much simpler API (queue test in rapid vs gopter), is much smarter about data generation and is able to minimize failing test cases fully automatically, without any user code.

As for testing/quick, it lacks both convenient data generation facilities and any form of test case minimization, which are two main things to look for in a property-based testing library.

FAQ

What is property-based testing?

Suppose we've written arithmetic functions add, subtract and multiply and want to test them. Traditional testing approach is example-based — we come up with example inputs and outputs, and verify that the system behavior matches the examples:

func TestArithmetic_Example(t *testing.T) {
	t.Run("add", func(t *testing.T) {
		examples := [][3]int{
			{0, 0, 0},
			{0, 1, 1},
			{2, 2, 4},
			// ...
		}
		for _, e := range examples {
			if add(e[0], e[1]) != e[2] {
				t.Fatalf("add(%v, %v) != %v", e[0], e[1], e[2])
			}
		}
	})
	t.Run("subtract", func(t *testing.T) { /* ... */ })
	t.Run("multiply", func(t *testing.T) { /* ... */ })
}

In comparison, with property-based testing we define higher-level properties that should hold for arbitrary input. Each time we run a property-based test, properties are checked on a new set of pseudo-random data:

func TestArithmetic_Property(t *testing.T) {
	rapid.Check(t, func(t *rapid.T) {
		var (
			a = rapid.Int().Draw(t, "a")
			b = rapid.Int().Draw(t, "b")
			c = rapid.Int().Draw(t, "c")
		)
		if add(a, 0) != a {
			t.Fatalf("add() does not have 0 as identity")
		}
		if add(a, b) != add(b, a) {
			t.Fatalf("add() is not commutative")
		}
		if add(a, add(b, c)) != add(add(a, b), c) {
			t.Fatalf("add() is not associative")
		}
		if multiply(a, add(b, c)) != add(multiply(a, b), multiply(a, c)) {
			t.Fatalf("multiply() is not distributive over add()")
		}
		// ...
	})
}

Property-based tests are more powerful and concise than example-based ones — and are also much more fun to write. As an additional benefit, coming up with general properties of the system often improves the design of the system itself.

What properties should I test?

As you've seen from the examples above, it depends on the system you are testing. Usually a good place to start is to put yourself in the shoes of your user and ask what are the properties the user will rely on (often unknowingly or implicitly) when building on top of your system. That said, here are some broadly applicable and often encountered properties to keep in mind:

  • function does not panic on valid input data
  • behavior of two algorithms or data structures is identical
  • all variants of the decode(encode(x)) == x roundtrip
How does rapid work?

At its core, rapid does a fairly simple thing: generates pseudo-random data based on the specification you provide, and check properties that you define on the generated data.

Checking is easy: you simply write if statements and call something like t.Fatalf when things look wrong.

Generating is a bit more involved. When you construct a Generator, nothing happens: Generator is just a specification of how to Draw the data you want. When you call Draw, rapid will take some bytes from its internal random bitstream, use them to construct the value based on the Generator specification, and track how the random bytes used correspond to the value (and its subparts). This knowledge about the structure of the values being generated, as well as their relationship with the parts of the bitstream allows rapid to intelligently and automatically minify any failure found.

What about fuzzing?

Property-based testing focuses on quick feedback loop: checking the properties on a small but diverse set of pseudo-random inputs in a fractions of a second.

In comparison, fuzzing focuses on slow, often multi-day, brute force input generation that maximizes the coverage.

Both approaches are useful. Property-based tests are used alongside regular example-based tests during development, and fuzzing is used to search for edge cases and security vulnerabilities. With MakeFuzz, any rapid test can be used as a fuzz target.

Usage

Just run go test as usual, it will pick up also all rapid tests.

There are a number of optional flags to influence rapid behavior, run go test -args -h and look at the flags with the -rapid. prefix. You can then pass such flags as usual. For example:

go test -rapid.checks=10_000

Status

Rapid is stable: tests using rapid should continue to work with all future rapid releases with the same major version. Possible exceptions to this rule are API changes that replace the concrete type of parameter with an interface type, or other similar mostly non-breaking changes.

License

Rapid is licensed under the Mozilla Public License Version 2.0.

Documentation

Overview

Package rapid implements utilities for property-based testing.

Check verifies that properties you define hold for a large number of automatically generated test cases. If a failure is found, rapid fails the current test and presents an automatically minimized version of the failing test case.

T.Repeat is used to construct state machine (sometimes called "stateful" or "model-based") tests.

Generators

Primitives:

Collections:

User-defined types:

Other:

Index

Examples

Constants

This section is empty.

Variables

This section is empty.

Functions

func Check 

func Check(t TB, prop func(*T))

Check fails the current test if rapid can find a test case which falsifies prop.

Property is falsified in case of a panic or a call to *T.Fatalf, *T.Fatal, *T.Errorf, *T.Error, *T.FailNow or *T.Fail.

Example (ParseDate)

Rename to TestParseDate(t *testing.T) to make an actual (failing) test. 

package main

import (
	"fmt"
	"strconv"
	"testing"

	"pgregory.net/rapid"
)

// ParseDate parses dates in the YYYY-MM-DD format.
func ParseDate(s string) (int, int, int, error) {
	if len(s) != 10 {
		return 0, 0, 0, fmt.Errorf("%q has wrong length: %v instead of 10", s, len(s))
	}

	if s[4] != '-' || s[7] != '-' {
		return 0, 0, 0, fmt.Errorf("'-' separators expected in %q", s)
	}

	y, err := strconv.Atoi(s[0:4])
	if err != nil {
		return 0, 0, 0, fmt.Errorf("failed to parse year: %v", err)
	}

	m, err := strconv.Atoi(s[6:7])
	if err != nil {
		return 0, 0, 0, fmt.Errorf("failed to parse month: %v", err)
	}

	d, err := strconv.Atoi(s[8:10])
	if err != nil {
		return 0, 0, 0, fmt.Errorf("failed to parse day: %v", err)
	}

	return y, m, d, nil
}

func testParseDate(t *rapid.T) {
	y := rapid.IntRange(0, 9999).Draw(t, "y")
	m := rapid.IntRange(1, 12).Draw(t, "m")
	d := rapid.IntRange(1, 31).Draw(t, "d")

	s := fmt.Sprintf("%04d-%02d-%02d", y, m, d)

	y_, m_, d_, err := ParseDate(s)
	if err != nil {
		t.Fatalf("failed to parse date %q: %v", s, err)
	}

	if y_ != y || m_ != m || d_ != d {
		t.Fatalf("got back wrong date: (%d, %d, %d)", y_, m_, d_)
	}
}

// Rename to TestParseDate(t *testing.T) to make an actual (failing) test.
func main() {
	var t *testing.T
	rapid.Check(t, testParseDate)
}

Output:

func ID added in v0.5.0 

func ID[V any](v V) V

ID returns its argument as is. ID is a helper for use with SliceOfDistinct and similar functions.

func MakeCheck 

func MakeCheck(prop func(*T)) func(*testing.T)

MakeCheck is a convenience function for defining subtests suitable for *testing.T.Run. It allows you to write this: 

t.Run("subtest name", rapid.MakeCheck(func(t *rapid.T) {
    // test code
}))

instead of this: 

t.Run("subtest name", func(t *testing.T) {
    rapid.Check(t, func(t *rapid.T) {
        // test code
    })
})

func MakeFuzz added in v0.5.2 

func MakeFuzz(prop func(*T)) func(*testing.T, []byte)

MakeFuzz creates a fuzz target for *testing.F.Fuzz: 

func FuzzFoo(f *testing.F) {
    f.Fuzz(rapid.MakeFuzz(func(t *rapid.T) {
        // test code
    }))
}

func StateMachineActions added in v0.6.0 

func StateMachineActions(sm StateMachine) map[string]func(*T)

StateMachineActions creates an actions map for *T.Repeat from methods of a StateMachine type instance using reflection.

Types

type Generator 

type Generator[V any] struct {
	// contains filtered or unexported fields
}

Generator describes a generator of values of type V.

func Bool added in v0.3.6 

func Bool() *Generator[bool]

func Byte 

func Byte() *Generator[byte]

func ByteMax 

func ByteMax(max byte) *Generator[byte]

func ByteMin 

func ByteMin(min byte) *Generator[byte]

func ByteRange 

func ByteRange(min byte, max byte) *Generator[byte]

func Custom 

func Custom[V any](fn func(*T) V) *Generator[V]

Custom creates a generator which produces results of calling fn. In fn, values should be generated by calling other generators; it is invalid to return a value from fn without using any other generator. Custom is a primary way of creating user-defined generators.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	type point struct {
		x int
		y int
	}

	gen := rapid.Custom(func(t *rapid.T) point {
		return point{
			x: rapid.IntRange(-100, 100).Draw(t, "x"),
			y: rapid.IntRange(-100, 100).Draw(t, "y"),
		}
	})

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

{-1 23}
{-3 -50}
{0 94}
{-2 -50}
{11 -57}

func Deferred added in v0.5.0 

func Deferred[V any](fn func() *Generator[V]) *Generator[V]

Deferred creates a generator which defers calling fn until attempting to produce a value. This allows to define recursive generators.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func recursive() *rapid.Generator[any] {
	return rapid.OneOf(
		rapid.Bool().AsAny(),
		rapid.SliceOfN(rapid.Deferred(recursive), 1, 2).AsAny(),
	)
}

func main() {
	gen := recursive()
	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

[[[[false] false]]]
false
[[true [[[true]]]]]
true
true

func Float32 

func Float32() *Generator[float32]

Float32 is a shorthand for Float32Range(-math.MaxFloat32, math.MaxFloat32).

func Float32Max 

func Float32Max(max float32) *Generator[float32]

Float32Max is a shorthand for Float32Range(-math.MaxFloat32, max).

func Float32Min 

func Float32Min(min float32) *Generator[float32]

Float32Min is a shorthand for Float32Range(min, math.MaxFloat32).

func Float32Range 

func Float32Range(min float32, max float32) *Generator[float32]

Float32Range creates a generator of 32-bit floating-point numbers in range [min, max]. Both min and max can be infinite.

func Float64 

func Float64() *Generator[float64]

Float64 is a shorthand for Float64Range(-math.MaxFloat64, math.MaxFloat64).

func Float64Max 

func Float64Max(max float64) *Generator[float64]

Float64Max is a shorthand for Float64Range(-math.MaxFloat64, max).

func Float64Min 

func Float64Min(min float64) *Generator[float64]

Float64Min is a shorthand for Float64Range(min, math.MaxFloat64).

func Float64Range 

func Float64Range(min float64, max float64) *Generator[float64]

Float64Range creates a generator of 64-bit floating-point numbers in range [min, max]. Both min and max can be infinite.

func Int 

func Int() *Generator[int]

func Int16 

func Int16() *Generator[int16]

func Int16Max 

func Int16Max(max int16) *Generator[int16]

func Int16Min 

func Int16Min(min int16) *Generator[int16]

func Int16Range 

func Int16Range(min int16, max int16) *Generator[int16]

func Int32 

func Int32() *Generator[int32]

func Int32Max 

func Int32Max(max int32) *Generator[int32]

func Int32Min 

func Int32Min(min int32) *Generator[int32]

func Int32Range 

func Int32Range(min int32, max int32) *Generator[int32]

func Int64 

func Int64() *Generator[int64]

func Int64Max 

func Int64Max(max int64) *Generator[int64]

func Int64Min 

func Int64Min(min int64) *Generator[int64]

func Int64Range 

func Int64Range(min int64, max int64) *Generator[int64]

func Int8 

func Int8() *Generator[int8]

func Int8Max 

func Int8Max(max int8) *Generator[int8]

func Int8Min 

func Int8Min(min int8) *Generator[int8]

func Int8Range 

func Int8Range(min int8, max int8) *Generator[int8]

func IntMax 

func IntMax(max int) *Generator[int]

func IntMin 

func IntMin(min int) *Generator[int]

func IntRange 

func IntRange(min int, max int) *Generator[int]

func Just 

func Just[V any](val V) *Generator[V]

Just creates a generator which always produces the given value. Just(val) is a shorthand for SampledFrom([]V{val}).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.Just(42)

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

42
42
42
42
42

func Make added in v0.5.0 

func Make[V any]() *Generator[V]

Make creates a generator of values of type V, using reflection to infer the required structure.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.Make[map[int]bool]()

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

map[-433:true -261:false -53:false -23:false 1:true 184:false]
map[-3:true 0:true]
map[4:true]
map[-359:true -154:true -71:true -17:false -1:false 590:false 22973756520:true]
map[]
Example (Tree) 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

type nodeValue int

type tree struct {
	Value       nodeValue
	Left, Right *tree
}

func (t *tree) String() string {
	if t == nil {
		return "nil"
	}
	return fmt.Sprintf("(%s %v %s)", t.Left.String(), t.Value, t.Right.String())
}

func main() {
	gen := rapid.Make[*tree]()

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

(nil 1 (nil 184 nil))
(((nil -1 (((((nil -485 ((nil -2 ((((nil -5 nil) -9898554875447 nil) -34709387 ((nil 50440 nil) 113 (((((nil -442 nil) -66090341586 nil) 179745 nil) 494 (((nil -2 nil) 543360606020 nil) 15261837 nil)) -1778 nil))) -21034573818 nil)) -5 nil)) 15606609 nil) 882666 (nil 3 nil)) -12 (nil -2 ((nil 1 nil) -2 (((nil 11 nil) -187307 ((nil -198 (nil -6895 nil)) 12027 (nil -539313 nil))) 1532 (nil 6 nil))))) 1745354 nil)) -2 nil) -3 nil)
nil
(((nil -15 (nil 6598 nil)) -131 (nil 317121006373596 ((nil 14 ((nil -9223372036854775808 nil) 1 nil)) 14668 nil))) 590 nil)
nil

func Map added in v0.5.4 

func Map[U any, V any](g *Generator[U], fn func(U) V) *Generator[V]

Map creates a generator producing fn(u) for each u produced by g.

Example 
package main

import (
	"fmt"
	"strconv"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.Map(rapid.Int(), strconv.Itoa)
	for i := 0; i < 5; i++ {
		fmt.Printf("%#v\n", gen.Example(i))
	}
}

Output:

"-3"
"-186981"
"4"
"-2"
"43"

func MapOf 

func MapOf[K comparable, V any](key *Generator[K], val *Generator[V]) *Generator[map[K]V]

MapOf is a shorthand for MapOfN(key, val, -1, -1).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.MapOf(rapid.Int(), rapid.StringMatching(`[a-z]+`))

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

map[1:nhlgqwasbggbaociac 561860:r]
map[-3752:pizpv -3:bacuabp 0:bi]
map[-33086515648293:gewf -264276:b -1313:a -258:v -4:b -2:fdhbzcz 4:ubfsdbowrja 1775:tcozav 8334:lvcprss 376914:braigey]
map[-350:h 590:coaaamcasnapgaad]
map[]

func MapOfN 

func MapOfN[K comparable, V any](key *Generator[K], val *Generator[V], minLen int, maxLen int) *Generator[map[K]V]

MapOfN creates a map[K]V generator. If minLen >= 0, generated maps have minimum length of minLen. If maxLen >= 0, generated maps have maximum length of maxLen. MapOfN panics if maxLen >= 0 and minLen > maxLen.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.MapOfN(rapid.Int(), rapid.StringMatching(`[a-z]+`), 5, 5)

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

map[-130450326583:bd -2983:bbdbcs 1:nhlgqwasbggbaociac 31:kmdnpmcbuagzr 561860:r]
map[-82024404:d -3752:pizpv -3:bacuabp 0:bi 179745:rzkneb]
map[-33086515648293:gewf -258:v 4:ubfsdbowrja 1775:tcozav 8334:lvcprss]
map[-4280678227:j -25651:aafmd -3308:o -350:h 590:coaaamcasnapgaad]
map[-9614404661322:gsb -378:y 2:paai 4629136912:otg 1476419818092:qign]

func MapOfNValues 

func MapOfNValues[K comparable, V any](val *Generator[V], minLen int, maxLen int, keyFn func(V) K) *Generator[map[K]V]

MapOfNValues creates a map[K]V generator, where keys are generated by applying keyFn to values. If minLen >= 0, generated maps have minimum length of minLen. If maxLen >= 0, generated maps have maximum length of maxLen. MapOfNValues panics if maxLen >= 0 and minLen > maxLen.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.MapOfNValues(rapid.StringMatching(`[a-z]+`), 5, 5, func(s string) int { return len(s) })

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

map[1:s 2:dr 3:anc 7:xguehfc 11:sbggbaociac]
map[1:b 2:bp 4:ydag 5:jarxz 6:ebzkwa]
map[1:j 3:gjl 5:eeeqa 7:stcozav 9:fxmcadagf]
map[2:ub 8:waraafmd 10:bfiqcaxazu 16:rjgqimcasnapgaad 17:gckfbljafcedhcvfc]
map[1:k 2:ay 3:wzb 4:dign 7:faabhcb]

func MapOfValues 

func MapOfValues[K comparable, V any](val *Generator[V], keyFn func(V) K) *Generator[map[K]V]

MapOfValues is a shorthand for MapOfNValues(val, -1, -1, keyFn).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.MapOfValues(rapid.StringMatching(`[a-z]+`), func(s string) int { return len(s) })

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

map[2:dr 7:xguehfc 11:sbggbaociac]
map[2:bp 5:jarxz 6:ebzkwa]
map[1:j 2:aj 3:gjl 4:vayt 5:eeeqa 6:riacaa 7:stcozav 8:mfdhbzcz 9:fxmcadagf 10:bgsbraigey 15:gxongygnxqlovib]
map[2:ub 8:waraafmd 10:bfiqcaxazu 16:rjgqimcasnapgaad 17:gckfbljafcedhcvfc]
map[]

func OneOf 

func OneOf[V any](gens ...*Generator[V]) *Generator[V]

OneOf creates a generator which produces each value by selecting one of gens and producing a value from it. OneOf panics if gens is empty.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.OneOf(rapid.Int32Range(1, 10).AsAny(), rapid.Float32Range(100, 1000).AsAny())

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

997.0737
10
475.3125
2
9

func Permutation added in v0.5.3 

func Permutation[S ~[]E, E any](slice S) *Generator[S]

Permutation creates a generator which produces permutations of the given slice.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.Permutation([]int{1, 2, 3})

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

[2 3 1]
[3 2 1]
[2 1 3]
[3 2 1]
[1 2 3]

func Ptr 

func Ptr[E any](elem *Generator[E], allowNil bool) *Generator[*E]

Ptr creates a *E generator. If allowNil is true, Ptr can return nil pointers.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.Ptr(rapid.Int(), true)

	for i := 0; i < 5; i++ {
		v := gen.Example(i)
		if v == nil {
			fmt.Println("<nil>")
		} else {
			fmt.Println("(*int)", *v)
		}
	}
}

Output:

(*int) 1
(*int) -3
<nil>
(*int) 590
<nil>

func Rune 

func Rune() *Generator[rune]

Rune creates a rune generator. Rune is equivalent to RuneFrom with default set of runes and tables.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.Rune()

	for i := 0; i < 25; i++ {
		if i%5 == 0 {
			fmt.Println()
		} else {
			fmt.Print(" ")
		}
		fmt.Printf("%q", gen.Example(i))
	}
}

Output:

'\n' '\x1b' 'A' 'a' '*'
'0' '@' '?' '\'' '\ue05d'
'<' '%' '!' '\u0604' 'A'
'%' '╷' '~' '!' '/'
'\u00ad' '𝪪' '@' '҈' ' '

func RuneFrom 

func RuneFrom(runes []rune, tables ...*unicode.RangeTable) *Generator[rune]

RuneFrom creates a rune generator from provided runes and tables. RuneFrom panics if both runes and tables are empty. RuneFrom panics if tables contain an empty table.

Example 
package main

import (
	"fmt"
	"unicode"

	"pgregory.net/rapid"
)

func main() {
	gens := []*rapid.Generator[rune]{
		rapid.RuneFrom([]rune{'A', 'B', 'C'}),
		rapid.RuneFrom(nil, unicode.Cyrillic, unicode.Greek),
		rapid.RuneFrom([]rune{'⌘'}, &unicode.RangeTable{
			R32: []unicode.Range32{{0x1F600, 0x1F64F, 1}},
		}),
	}

	for _, gen := range gens {
		for i := 0; i < 5; i++ {
			if i > 0 {
				fmt.Print(" ")
			}
			fmt.Printf("%q", gen.Example(i))
		}
		fmt.Println()
	}
}

Output:

'A' 'A' 'A' 'B' 'A'
'Ͱ' 'Ѥ' 'Ͱ' 'ͱ' 'Ϳ'
'😀' '⌘' '😀' '😁' '😋'

func SampledFrom 

func SampledFrom[S ~[]E, E any](slice S) *Generator[E]

SampledFrom creates a generator which produces values from the given slice. SampledFrom panics if slice is empty.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.SampledFrom([]int{1, 2, 3})

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

2
3
2
3
1

func SliceOf 

func SliceOf[E any](elem *Generator[E]) *Generator[[]E]

SliceOf is a shorthand for SliceOfN(elem, -1, -1).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.SliceOf(rapid.Int())

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

[1 -1902 7 -236 14 -433 -1572631 -1 4219826 -50 1414 -3890044391133 -9223372036854775808 5755498240 -10 680558 10 -80458281 0 -27]
[-3 -2 -1 -3 -2172865589 -5 -2 -2503553836720]
[4 308 -2 21 -5843 3 1 78 6129321692 -59]
[590 -131 -15 -769 16 -1 14668 14 -1 -58784]
[]

func SliceOfBytesMatching 

func SliceOfBytesMatching(expr string) *Generator[[]byte]

SliceOfBytesMatching creates a UTF-8 byte slice generator matching the provided syntax.Perl regular expression.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.SliceOfBytesMatching(`[CAGT]+`)

	for i := 0; i < 5; i++ {
		fmt.Printf("%q\n", gen.Example(i))
	}
}

Output:

"CCTTGAGAGCGATACGGAAG"
"GCAGAACT"
"AACCGTCGAG"
"GGGAAAAGAT"
"AGTG"

func SliceOfDistinct 

func SliceOfDistinct[E any, K comparable](elem *Generator[E], keyFn func(E) K) *Generator[[]E]

SliceOfDistinct is a shorthand for SliceOfNDistinct(elem, -1, -1, keyFn).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.SliceOfDistinct(rapid.IntMin(0), func(i int) int { return i % 2 })

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

[1]
[2 1]
[4 1]
[590]
[]

func SliceOfN 

func SliceOfN[E any](elem *Generator[E], minLen int, maxLen int) *Generator[[]E]

SliceOfN creates a []E generator. If minLen >= 0, generated slices have minimum length of minLen. If maxLen >= 0, generated slices have maximum length of maxLen. SliceOfN panics if maxLen >= 0 and minLen > maxLen.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.SliceOfN(rapid.Int(), 5, 5)

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

[1 -1902 7 -236 14]
[-3 -2 -1 -3 -2172865589]
[4 308 -2 21 -5843]
[590 -131 -15 -769 16]
[4629136912 270 141395 -129322425838843911 -7]

func SliceOfNDistinct 

func SliceOfNDistinct[E any, K comparable](elem *Generator[E], minLen int, maxLen int, keyFn func(E) K) *Generator[[]E]

SliceOfNDistinct creates a []E generator. Elements of each generated slice are distinct according to keyFn. If minLen >= 0, generated slices have minimum length of minLen. If maxLen >= 0, generated slices have maximum length of maxLen. SliceOfNDistinct panics if maxLen >= 0 and minLen > maxLen. ID helper can be used as keyFn to generate slices of distinct comparable elements.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.SliceOfNDistinct(rapid.IntMin(0), 2, 2, func(i int) int { return i % 2 })

	for i := 0; i < 5; i++ {
		fmt.Println(gen.Example(i))
	}
}

Output:

[4219826 49]
[2 1]
[4 1]
[0 58783]
[4629136912 141395]

func String 

func String() *Generator[string]

String is a shorthand for StringOf(Rune()).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.String()

	for i := 0; i < 5; i++ {
		fmt.Printf("%q\n", gen.Example(i))
	}
}

Output:

"\n߾⃝?\rA�֍"
"\u2006𑰼"
"A￠\u0603ᾢ"
"+^#.[#৲"
""

func StringMatching 

func StringMatching(expr string) *Generator[string]

StringMatching creates a UTF-8 string generator matching the provided syntax.Perl regular expression.

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.StringMatching(`\(?([0-9]{3})\)?([ .-]?)([0-9]{3})([ .-]?)([0-9]{4})`)

	for i := 0; i < 5; i++ {
		fmt.Printf("%q\n", gen.Example(i))
	}
}

Output:

"(532) 649-9610"
"901)-5783983"
"914.444.1575"
"(316 696.3584"
"816)0861080"

func StringN 

func StringN(minRunes int, maxRunes int, maxLen int) *Generator[string]

StringN is a shorthand for StringOfN(Rune(), minRunes, maxRunes, maxLen).

Example 
package main

import (
	"fmt"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.StringN(5, 5, -1)

	for i := 0; i < 5; i++ {
		fmt.Printf("%q\n", gen.Example(i))
	}
}

Output:

"\n߾⃝?\r"
"\u2006𑰼#`\x1b"
"A￠\u0603ᾢÉ"
"+^#.["
".A<a¤"

func StringOf 

func StringOf(elem *Generator[rune]) *Generator[string]

StringOf is a shorthand for StringOfN(elem, -1, -1, -1).

Example 
package main

import (
	"fmt"
	"unicode"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.StringOf(rapid.RuneFrom(nil, unicode.Tibetan))

	for i := 0; i < 5; i++ {
		fmt.Printf("%q\n", gen.Example(i))
	}
}

Output:

"༁༭༇ཬ༆༐༖ༀྸ༁༆༎ༀ༁ཱི༂༨ༀ༂"
"༂༁ༀ༂༴ༀ༁ྵ"
"ༀ༴༁༅ན༃༁༎ྼ༄༽"
"༎༂༎ༀༀༀཌྷ༂ༀྥ"
""

func StringOfN 

func StringOfN(elem *Generator[rune], minRunes int, maxRunes int, maxLen int) *Generator[string]

StringOfN creates a UTF-8 string generator. If minRunes >= 0, generated strings have minimum minRunes runes. If maxRunes >= 0, generated strings have maximum maxRunes runes. If maxLen >= 0, generates strings have maximum length of maxLen. StringOfN panics if maxRunes >= 0 and minRunes > maxRunes. StringOfN panics if maxLen >= 0 and maxLen < maxRunes.

Example 
package main

import (
	"fmt"
	"unicode"

	"pgregory.net/rapid"
)

func main() {
	gen := rapid.StringOfN(rapid.RuneFrom(nil, unicode.ASCII_Hex_Digit), 6, 6, -1)

	for i := 0; i < 5; i++ {
		fmt.Printf("%q\n", gen.Example(i))
	}
}

Output:

"1D7B6a"
"2102e0"
"0e15c3"
"E2E000"
"aEd623"

func Uint 

func Uint() *Generator[uint]

func Uint16 

func Uint16() *Generator[uint16]

func Uint16Max 

func Uint16Max(max uint16) *Generator[uint16]

func Uint16Min 

func Uint16Min(min uint16) *Generator[uint16]

func Uint16Range 

func Uint16Range(min uint16, max uint16) *Generator[uint16]

func Uint32 

func Uint32() *Generator[uint32]

func Uint32Max 

func Uint32Max(max uint32) *Generator[uint32]

func Uint32Min 

func Uint32Min(min uint32) *Generator[uint32]

func Uint32Range 

func Uint32Range(min uint32, max uint32) *Generator[uint32]

func Uint64 

func Uint64() *Generator[uint64]

func Uint64Max 

func Uint64Max(max uint64) *Generator[uint64]

func Uint64Min 

func Uint64Min(min uint64) *Generator[uint64]

func Uint64Range 

func Uint64Range(min uint64, max uint64) *Generator[uint64]

func Uint8 

func Uint8() *Generator[uint8]

func Uint8Max 

func Uint8Max(max uint8) *Generator[uint8]

func Uint8Min 

func Uint8Min(min uint8) *Generator[uint8]

func Uint8Range 

func Uint8Range(min uint8, max uint8) *Generator[uint8]

func UintMax 

func UintMax(max uint) *Generator[uint]

func UintMin 

func UintMin(min uint) *Generator[uint]

func UintRange 

func UintRange(min uint, max uint) *Generator[uint]

func Uintptr 

func Uintptr() *Generator[uintptr]

func UintptrMax 

func UintptrMax(max uintptr) *Generator[uintptr]

func UintptrMin 

func UintptrMin(min uintptr) *Generator[uintptr]

func UintptrRange 

func UintptrRange(min uintptr, max uintptr) *Generator[uintptr]

func (*Generator[V]) AsAny added in v0.5.0 

func (g *Generator[V]) AsAny() *Generator[any]

AsAny creates a generator producing values from g converted to any.

func (*Generator[V]) Draw 

func (g *Generator[V]) Draw(t *T, label string) V

Draw produces a value from the generator.

func (*Generator[V]) Example 

func (g *Generator[V]) Example(seed ...int) V

Example produces an example value from the generator. If seed is provided, value is produced deterministically based on seed. Example should only be used for examples; always use *Generator.Draw in property-based tests.

func (*Generator[V]) Filter 

func (g *Generator[V]) Filter(fn func(V) bool) *Generator[V]

Filter creates a generator producing only values from g for which fn returns true.

func (*Generator[V]) String 

func (g *Generator[V]) String() string

type StateMachine 

type StateMachine interface {
	// Check is ran after every action and should contain invariant checks.
	//
	// All other public methods should have a form ActionName(t *rapid.T)
	// and are used as possible actions. At least one action has to be specified.
	Check(*T)
}

type T 

type T struct {
	// contains filtered or unexported fields
}

T is similar to testing.T, but with extra bookkeeping for property-based tests.

For tests to be reproducible, they should generally run in a single goroutine. If concurrency is unavoidable, methods on *T, such as *testing.T.Helper and *T.Errorf, are safe for concurrent calls, but *Generator.Draw from a given *T is not.

func (*T) Error 

func (t *T) Error(args ...any)

Error is equivalent to T.Log followed by T.Fail.

func (*T) Errorf 

func (t *T) Errorf(format string, args ...any)

Errorf is equivalent to T.Logf followed by T.Fail.

func (*T) Fail 

func (t *T) Fail()

func (*T) FailNow 

func (t *T) FailNow()

func (*T) Failed 

func (t *T) Failed() bool

func (*T) Fatal 

func (t *T) Fatal(args ...any)

Fatal is equivalent to T.Log followed by T.FailNow.

func (*T) Fatalf 

func (t *T) Fatalf(format string, args ...any)

Fatalf is equivalent to T.Logf followed by T.FailNow.

func (*T) Log 

func (t *T) Log(args ...any)

func (*T) Logf 

func (t *T) Logf(format string, args ...any)

func (*T) Repeat added in v0.7.0 

func (t *T) Repeat(actions map[string]func(*T))

Repeat executes a random sequence of actions (often called a "state machine" test). actions[""], if set, is executed before/after every other action invocation and should only contain invariant checking code.

For complex state machines, it can be more convenient to specify actions as methods of a special state machine type. In this case, StateMachineActions can be used to create an actions map from state machine methods using reflection.

Example (Queue)

Rename to TestQueue(t *testing.T) to make an actual (failing) test. 

package main

import (
	"testing"

	"pgregory.net/rapid"
)

// Queue implements integer queue with a fixed maximum size.
type Queue struct {
	buf []int
	in  int
	out int
}

func NewQueue(n int) *Queue {
	return &Queue{
		buf: make([]int, n+1),
	}
}

// Precondition: Size() > 0.
func (q *Queue) Get() int {
	i := q.buf[q.out]
	q.out = (q.out + 1) % len(q.buf)
	return i
}

// Precondition: Size() < n.
func (q *Queue) Put(i int) {
	q.buf[q.in] = i
	q.in = (q.in + 1) % len(q.buf)
}

func (q *Queue) Size() int {
	return (q.in - q.out) % len(q.buf)
}

func testQueue(t *rapid.T) {
	n := rapid.IntRange(1, 1000).Draw(t, "n") // maximum queue size
	q := NewQueue(n)                          // queue being tested
	var state []int                           // model of the queue

	t.Repeat(map[string]func(*rapid.T){
		"get": func(t *rapid.T) {
			if q.Size() == 0 {
				t.Skip("queue empty")
			}

			i := q.Get()
			if i != state[0] {
				t.Fatalf("got invalid value: %v vs expected %v", i, state[0])
			}
			state = state[1:]
		},
		"put": func(t *rapid.T) {
			if q.Size() == n {
				t.Skip("queue full")
			}

			i := rapid.Int().Draw(t, "i")
			q.Put(i)
			state = append(state, i)
		},
		"": func(t *rapid.T) {
			if q.Size() != len(state) {
				t.Fatalf("queue size mismatch: %v vs expected %v", q.Size(), len(state))
			}
		},
	})
}

// Rename to TestQueue(t *testing.T) to make an actual (failing) test.
func main() {
	var t *testing.T
	rapid.Check(t, testQueue)
}

Output:

func (*T) Skip 

func (t *T) Skip(args ...any)

Skip is equivalent to T.Log followed by T.SkipNow.

func (*T) SkipNow 

func (t *T) SkipNow()

SkipNow marks the current test case as invalid (except in T.Repeat actions, where it marks current action as non-applicable instead). If too many test cases are skipped, rapid will mark the test as failing due to inability to generate enough valid test cases.

Prefer *Generator.Filter to SkipNow, and prefer generators that always produce valid test cases to Filter.

func (*T) Skipf 

func (t *T) Skipf(format string, args ...any)

Skipf is equivalent to T.Logf followed by T.SkipNow.

type TB added in v0.4.8 

type TB interface {
	Helper()
	Name() string
	Logf(format string, args ...any)
	Log(args ...any)
	Skipf(format string, args ...any)
	Skip(args ...any)
	SkipNow()
	Errorf(format string, args ...any)
	Error(args ...any)
	Fatalf(format string, args ...any)
	Fatal(args ...any)
	FailNow()
	Fail()
	Failed() bool
}

TB is a common interface between *testing.T, *testing.B and *T.

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.