Bring coverage up to 100% line coverage

[zimmski]

There are some lines that still do not have test cases https://coveralls.io/github/sergi/go-diff It would be nice to bring the coverage up to a real 100% and then stay there by making decreasing coverage a PR failure.

[maksimov]

I'm learning Go and I thought the best way to do it would be to improve some tests. I brought coverage up a notch from 99.3% to 99.4%, but then I started experimenting with a test for patch.go:546 and I've added this test-case (never mind the expected value, I was going to correct it):
{"@@ -0,0 +0,0 @@\n+abc\n", "Bad patch mode"}

Now my test fails like this:

panic: runtime error: invalid memory address or nil pointer dereference [recovered]
	panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x20 pc=0x1398840]

goroutine 51 [running]:
testing.tRunner.func1(0xc4201c40e0)
	/usr/local/go/src/testing/testing.go:644 +0x55d
panic(0x13f3de0, 0x1620fb0)
	/usr/local/go/src/runtime/panic.go:489 +0x2ee
github.com/maksimov/go-diff/diffmatchpatch.TestPatchFromText(0xc4201c40e0)
	/Users/stas/Fun/TryGo/src/github.com/maksimov/go-diff/diffmatchpatch/patch_test.go:81 +0xb90
testing.tRunner(0xc4201c40e0, 0x145ace8)
	/usr/local/go/src/testing/testing.go:679 +0x229
created by testing.(*T).Run
	/usr/local/go/src/testing/testing.go:716 +0x53d

Do you think it's a genuine candidate for code improvement or am I being silly with my test case?
Thanks!

[maksimov]

Yes, I guess it's my test case. The panic is in the test code.

[zimmski]

I did not try it but when I guess that the test panics because "err" is nil so the "Error" call fails because of that. I think it is NEVER silly to improve a project ;-) To be honest, if you are just learning Go it would be more fun to write new code than adding test cases. Finding test cases for this project is especially hard, there aren't many instances left and the easy ones are already gone.

[zimmski]

Maybe #54 is more fun since you would write new code, must test it, and then have to benchmark your solution. The marker is here

go-diff/diffmatchpatch/diff.go

Line 459 in 431343a

// TODO research and benchmark this, why is it not activated? https://github.com/sergi/go-diff/issues/54

But beware that if the solution has bad benchmark results, we cannot merge it.

[zimmski]

I looked through the open issues and I think #49 could be also mildly interesting since the whole project is missing example for every exported API endpoint. #57 and #58 could be also interesting, at least there is new code involved. However, if you are just learning Go I am a little lost on what might be the most interesting and easiest issue for this project.

[maksimov]

Thanks very much for your help, @zimmski! I'll try a few things and see how I get on.

[zimmski]

@maksimov I think with your changes the bar has been raised high enough to close this issue