fix: use common lineHash to share indice between text1 and text2

Use common cache of line contents between two texts in `DiffLinesToChars` to get line diffs correctly.
In some cases, line diffs cannot be retrieved correctly in the standard way (https://github.com/google/diff-match-patch/wiki/Line-or-Word-Diffs#line-mode).
In the below case, we failed to get line diffs correctly before this fix.

```go:main.go
package main

import (
	"fmt"

	"github.com/sergi/go-diff/diffmatchpatch"
)

const (
	text1 = `hoge:
  step11:
  - arrayitem1
  - arrayitem2
  step12:
    step21: hoge
    step22: -93
fuga: flatitem
`
	text2 = `hoge:
  step11:
  - arrayitem4
  - arrayitem2
  - arrayitem3
  step12:
    step21: hoge
    step22: -92
fuga: flatitem
`
)

func main() {
	dmp := diffmatchpatch.New()
	a, b, c := dmp.DiffLinesToChars(text1, text2)
	diffs := dmp.DiffMain(a, b, false)
	diffs = dmp.DiffCharsToLines(diffs, c)
	// diffs = dmp.DiffCleanupSemantic(diffs)
	fmt.Println(diffs)
}
```

```text:output
[{Insert hoge:
  step11:
hoge:
} {Equal hoge:
} {Insert hoge:
} {Equal   step11:
} {Insert hoge:
} {Equal   - arrayitem1
} {Insert hoge:
} {Equal   - arrayitem2
} {Insert hoge:
} {Equal   step12:
} {Insert hoge:
} {Equal     step21: hoge
} {Insert hoge:
} {Equal     step22: -93
} {Delete fuga: flatitem
}]
```

Note: This fix corresponds to a javascript implementation.
(ref: https://github.com/google/diff-match-patch/blob/62f2e689f498f9c92dbc588c58750addec9b1654/javascript/diff_match_patch_uncompressed.js#L466)

Author: nrnrk

diffmatchpatch/diff.go

@@ -1313,17 +1313,17 @@ func (dmp *DiffMatchPatch) diffLinesToStrings(text1, text2 string) (string, stri
 	// '\x00' is a valid character, but various debuggers don't like it. So we'll insert a junk entry to avoid generating a null character.
 	lineArray := []string{""} // e.g. lineArray[4] == 'Hello\n'
 
+	lineHash := make(map[string]int)
 	//Each string has the index of lineArray which it points to
-	strIndexArray1 := dmp.diffLinesToStringsMunge(text1, &lineArray)
-	strIndexArray2 := dmp.diffLinesToStringsMunge(text2, &lineArray)
+	strIndexArray1 := dmp.diffLinesToStringsMunge(text1, &lineArray, lineHash)
+	strIndexArray2 := dmp.diffLinesToStringsMunge(text2, &lineArray, lineHash)
 
 	return intArrayToString(strIndexArray1), intArrayToString(strIndexArray2), lineArray
 }
 
 // diffLinesToStringsMunge splits a text into an array of strings, and reduces the texts to a []string.
-func (dmp *DiffMatchPatch) diffLinesToStringsMunge(text string, lineArray *[]string) []uint32 {
+func (dmp *DiffMatchPatch) diffLinesToStringsMunge(text string, lineArray *[]string, lineHash map[string]int) []uint32 {
 	// Walk the text, pulling out a substring for each line. text.split('\n') would would temporarily double our memory footprint. Modifying text would create many large strings to garbage collect.
-	lineHash := map[string]int{} // e.g. lineHash['Hello\n'] == 4
 	lineStart := 0
 	lineEnd := -1
 	strs := []uint32{}

diffmatchpatch/diff_test.go

@@ -318,6 +318,8 @@ func TestDiffLinesToChars(t *testing.T) {
 		{"a", "b", "1", "2", []string{"", "a", "b"}},
 		// Omit final newline.
 		{"alpha\nbeta\nalpha", "", "1,2,3", "", []string{"", "alpha\n", "beta\n", "alpha"}},
+		// Same lines in Text1 and Text2
+		{"abc\ndefg\n12345\n", "abc\ndef\n12345\n678", "1,2,3", "1,4,3,5", []string{"", "abc\n", "defg\n", "12345\n", "def\n", "678"}},
 	} {
 		actualChars1, actualChars2, actualLines := dmp.DiffLinesToChars(tc.Text1, tc.Text2)
 		assert.Equal(t, tc.ExpectedChars1, actualChars1, fmt.Sprintf("Test case #%d, %#v", i, tc))