README
¶
layout: default title: testutil
testutil
Package testutil provides generic functions for testing in Go. Please consult the GoDoc page for specific details and visit ScientificGo to see more scientific packages for Go.
Installation
go get scientificgo.org/testutil
Usage
import "scientificgo.org/testutil"
Documentation
¶
Overview ¶
Package testutil provides generic functions for testing and benchmarking with go test, as well as type-agnostic implementations of Equal, Any and All.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func Test ¶
Test is a generic case-driven testing function that accepts a slice of cases, a numerical tolerance and either 1 or 2 functions to be tested. A sub-test is run for each case.
If 1 function is provided, then its output is tested against the outputs provided in each case.
If 2 functions are provided, then their respective outputs are compared, using the inputs provided in each case.
Types ¶
type Cases ¶
type Cases interface{}
Cases represents a generic data structure for table-driven testing. It is an alias for a slice of cases with the following structure:
[]struct {
Label string
In1 TIn1
⋮ ⋮
InN TInN
Out1 TOut1
⋮ ⋮
OutM TInM
}
where N and M are arbitrary.
For the common case of testing a numerical function of a single argument, this would take the simple form of:
[]struct {
Label string
In float64
Out float64
}{
{"SomeLabel", x, y},
{"AnotherLabel", x1, y1},
...
}
type EqualResult ¶ added in v0.1.7
type EqualResult struct {
// Ok is true if x equals y.
Ok bool
// Numerical is true if x and y are numerical.
Numerical bool
// RelativeError is the error in x relative to y if they are numerical.
// It is a complex number if x and y are complex numbers.
RelativeError reflect.Value
// AbsoluteError is the difference between x and y if they are numerical.
// It is a complex number if x and y are complex numbers.
AbsoluteError reflect.Value
// Position is the the first "location" that x does not equal y
// if x and y are structured data types.
//
// For slices and arrays it is the index of first element from x that does not equal y.
// For structs it is the index of the first field for which x does not equal y.
// For maps it is the index of the first key for which x does not equal y.
//
// If MissingValue is true, Position gives the index in y of the missing field or key.
Position int
// LengthMismatch is true if the number of elements, fields or keys in x
// differs from y for structured data types.
LengthMismatch bool
// MissingValue is true if x and y are maps or structs and x is missing one of the keys
// or fields in y.
MissingValue bool
}
EqualResult represents the result of an Equal comparison between arbitrary x and y.
func Equal ¶
func Equal(x, y, tolerance interface{}) EqualResult
Equal reports whether x (actual) is equal to y (expected).
For numerical types, x is equal to y if:
|x - y| < tolerance * |y|, for y ≠ 0 (relative error) |x| < tolerance, for y = 0 (absolute error)
For structured types (slice, array, struct, map), x equals y if every element/field/key of x equals that in y.
For func types, x equals y if x(args) equals y(args) for randomly generated args.
For other types x equals y if reflect.DeepEqual(x, y) is true.
Source Files