assert.ErrorAs: confusing error message #1343
Description
Typically, the target passed to errors.As has a zero value. But if the errors.As call is wrapped by assert.ErrorAs, this can lead to a confusing error message if the assertion fails:
--- FAIL: TestCustomErr (0.00s)
main_test.go:32:
Error Trace: /.../main_test.go:32
Error: Should be in error chain:
expected: %!q(PANIC=Error method: runtime error: invalid memory address or nil pointer dereference)
in chain: "file does not exist"
Test: TestCustomErr
Messages: index 1
FAIL
For expected:, it's printing the target. But in my case the target can't be printed until it's assigned a value, and I think printing the target is not useful anyway - I think what it should be doing instead is printing the type of the target:
--- FAIL: TestCustomErr (0.00s)
main_test.go:32:
Error Trace: /.../main_test.go:32
Error: Should be in error chain:
expected: "*main.customErr"
in chain: "file does not exist"
Test: TestCustomErr
Messages: index 1
FAIL
Here's the code I used:
package main
import (
"errors"
"fmt"
"io/fs"
"testing"
"github.com/stretchr/testify/assert"
)
type customErr struct {
err error
}
func (ce customErr) Error() string {
return ce.err.Error()
}
func getErrs() []error {
return []error{
customErr{
err: errors.New(`something`),
},
fs.ErrNotExist,
}
}
func TestCustomErr(t *testing.T) {
for i, err := range getErrs() {
var aCustomErr customErr
assert.ErrorAs(t, err, &aCustomErr, fmt.Sprintf(`index %v`, i))
}
}I am happy to submit a PR to address this.