add support for Struct in Map

[svyotov]

This fixes #103 also fixes #90 however I am not sure it is enough as I see other issues with map.

[vdemeester]

@svyotov thanks for the PR. It would be better to add a test that validates that behaviour.
Also, I checked-out that PR and ran the tests, and I think it's breaking current behaviour

--- FAIL: TestMapsWithOverwrite (0.00s)
    mergo_test.go:372: Test failed:
        got  :
        map[string]mergo.simpleTest{"d":mergo.simpleTest{Value:61}, "a":mergo.simpleTest{Value:0}, "b":mergo.simpleTest
{Value:42}, "c":mergo.simpleTest{Value:13}}

        want :
        map[string]mergo.simpleTest{"a":mergo.simpleTest{Value:16}, "b":mergo.simpleTest{Value:0}, "c":mergo.simpleTest
{Value:12}, "d":mergo.simpleTest{Value:61}, "e":mergo.simpleTest{Value:14}}

=== RUN   TestMaps
--- FAIL: TestMaps (0.00s)
    mergo_test.go:402: Test failed:
        got  :
        map[string]mergo.simpleTest{"a":mergo.simpleTest{Value:0}, "b":mergo.simpleTest{Value:42}, "c":mergo.simpleTest
{Value:13}, "d":mergo.simpleTest{Value:61}}

        want :
        map[string]mergo.simpleTest{"b":mergo.simpleTest{Value:42}, "c":mergo.simpleTest{Value:13}, "d":mergo.simpleTes
t{Value:61}, "e":mergo.simpleTest{Value:14}, "a":mergo.simpleTest{Value:0}}

[svyotov]

@vdemeester I have added a unit test which hopefully is correct and better shows the issue, however I have not managed to make the code work yet.

[svyotov]

@vdemeester go test ./... is passing for me, however I have just started understanding reflect in Golang so I appreciate there might be a much better way of doing this. I changed some of the unit tests as isNil returns true for int when equal to 0 and some of the changes should be done.
If you wish to not use this PR and close it that is fine by me, though IMO at the very least issue90_test.go should be reused as the tests there is failing with the code currently on master.

If you have pointers regarding how to fix this/ improve this I would love to hear from you and see if I can make it into a new PR.

[svyotov]

Any thoughts on this PR @vdemeester, @imdario?

[darccio]

I'll check it this evening.

[vdemeester]

• unsafe usage worries me a bit, I wonder if it impacts performances. We should create some benchmark tests to make sure we don't regress too much.
• It adds some complexity to the code (that seems to be required… but stiff… 😅)

The test is actually green. I'll fix the CI in a separate PR so that we can trust the GIthub checks 👼

[boosh]

Hopefully this'll also fix #14

[darccio]

Thanks @svyotov. After a long wait, I've decided to go ahead.