Cannot run a specific test case when using table driven tests

[aaslamin]

Using the standard library, one is able to filter/run (a) specific (sub) test(s) by passing a pattern to the go test command - https://blog.golang.org/subtests

e.g. go test -run=myregexpattern

Adopting a table driven test style with Convey (example), when a test fails, one cannot run that specific test in isolation without commenting out test cases you are not interested in. This is not practical if your test table is huge. Ideally, I shouldn't need to touch the structure of my tests to run them in isolation.

I saw https://godoc.org/github.com/smartystreets/goconvey/convey#FocusConvey, but it's not useful for a table driven style.

Thoughts?

[mdwhatcott]

@aaslamin - The GoConvey test DSL and table-driven tests are quite different in their approach. I doubt I ever had hopes to combine or synthesize them. Could you post an example of what you are trying to accomplish?

[aaslamin]

Here is one @mdwhatcott

In the example below, I want to run the failing test case test case 3 in isolation so I can debug it. But, I cannot.

https://play.golang.org/p/K8X8xvkQdl2

package main

import (
    "fmt"
    "testing"

    . "github.com/smartystreets/goconvey/convey"
)

func SillyFunc(fail bool) error {
    if fail {
        return fmt.Errorf("error: i failed because i was told to do so")
    }

    return nil
}

func TestSillyFunc(t *testing.T) {

    testCases := map[string]struct {
        input       bool
        shouldError bool
    }{
        "test case 1": {
            input:       false,
            shouldError: false,
        },
        "test case 2": {
            input:       false,
            shouldError: false,
        },

        // notice how this test fails - I want to run this test case in isolation so I can debug it!

        "test case 3": {
            input:       true,
            shouldError: false,
        },
        "test case 4": {
            input:       false,
            shouldError: false,
        },
        "test case 5": {
            input:       false,
            shouldError: false,
        },
    }

    Convey("TestSilly", t, func() {
        for description, test := range testCases {
            Convey(description, func() {
                So(SillyFunc(test.input) != nil, ShouldEqual, test.shouldError)
            })
        }
    })
}

[aaslamin]

Output of go test -v

go test -v -count=1
=== RUN   TestSillyFunc

  TestSilly 
    test case 1 ✔
    test case 2 ✔
    test case 3 ✘
    test case 4 ✔
    test case 5 ✔

Desired goal

Be able to use the standard go CLI to run the failing test case (test case 3) in isolation via passing a pattern to the run flag as per documented in https://blog.golang.org/subtests

Other:

Also note that the test output without passing -v argument does not print any information regarding the test case that failed. In CI, you would have to pass the -v flag all the time which would cause a lot of noise as it will still print things even if a test suite passes.