Diffs are slightly off for assert.Equal on maps

[abourget]

It seems the map is rendered with no trailing , for the last element, so the diff catches that as a modified line.. but really the line isn't modified..

Diff:
            --- Expected
            +++ Actual
            @@ -1,5 +1,7 @@
            -(map[string]interface {}) (len=3) {
            +(map[string]interface {}) (len=5) {
              (string) (len=15) "installed.first": (string) (len=20) "2006-01-02T15:04:05Z",
              (string) (len=14) "installed.last": (string) (len=20) "2006-01-02T15:04:05Z",
            - (string) (len=23) "installed.windows.first": (string) (len=20) "2006-01-02T15:04:05Z"
            + (string) (len=23) "installed.windows.first": (string) (len=20) "2006-01-02T15:04:05Z",
            + (string) (len=22) "installed.windows.last": (string) (len=20) "2006-01-02T15:04:05Z",
            + (string) (len=30) "installed.windows.last_version": (string) (len=3) "8.1"
             }

See how installed.windows.first is different only in the trailing , !

[ernesto-jimenez]

thanks @abourget

I'll have to check there's an option in spew to print the trailing comma.

Would you be interested in doing a pull request with the enhancement?

[abourget]

yes :) I'll take a look into it...

[atombender]

Opened a PR for this in go-spew.

[georgelesica-wf]

I pinged the maintainer about the go-spew PR. If he agrees to merge it then we can close this, otherwise I'll probably go ahead and close it anyway unless someone has a clever idea that will resolve it without changing go-spew.

[atombender]

@georgelesica-wf Have you considered using Litter instead of Spew? We wrote Litter to replace Spew in our tests, as we use a lot of snapshot tests where we store results in files and compare them. Its output is more consistent and more readable, since it's basically just Go.

[ghostsquad]

seems that when comparing maps, order matters? is that the issue here?