Fix processing lsif dump with repeating lines with inVs

Rewrite processing of lines with "label":"contains" and "label":"item"
for appending for arrays instead override then when lines repeat few times
with equal outV.

changelogs/unreleased/305-lsif-arrays.yml

+---
+title: Fix processing lsif dump with repeating lines with inVs
+merge_request: 621
+author: Pavel Kuznetsov
+type: fixed

internal/lsif_transformer/parser/docs.go

@@ -138,7 +138,7 @@ func (d *Docs) addDocRanges(line []byte) error {
 		return err
 	}
 
-	d.DocRanges[docRange.OutV] = docRange.RangeIds
+	d.DocRanges[docRange.OutV] = append(d.DocRanges[docRange.OutV], docRange.RangeIds...)
 
 	return nil
 }

internal/lsif_transformer/parser/docs_test.go

@@ -35,11 +35,12 @@ func TestParseContainsLine(t *testing.T) {
 	require.NoError(t, err)
 	defer d.Close()
 
-	line := []byte(`{"id":"5","label":"contains","outV":"1", "inVs": ["2", "3"]}` + "\n")
+	data := []byte(`{"id":"5","label":"contains","outV":"1", "inVs": ["2", "3"]}` + "\n")
+	data = append(data, []byte(`{"id":"6","label":"contains","outV":"1", "inVs": [4]}`+"\n")...)
 
-	require.NoError(t, d.Parse(bytes.NewReader(line)))
+	require.NoError(t, d.Parse(bytes.NewReader(data)))
 
-	require.Equal(t, []Id{2, 3}, d.DocRanges[1])
+	require.Equal(t, []Id{2, 3, 4}, d.DocRanges[1])
 }
 
 func TestParsingVeryLongLine(t *testing.T) {

internal/lsif_transformer/parser/ranges_test.go

@@ -11,15 +11,20 @@ func TestRangesRead(t *testing.T) {
 	r, cleanup := setup(t)
 	defer cleanup()
 
-	firstRange := Range{Line: 1, Character: 2, RefId: 3}
+	firstRange := Range{Line: 1, Character: 2, RefId: 4}
 	rg, err := r.getRange(1)
 	require.NoError(t, err)
 	require.Equal(t, &firstRange, rg)
 
-	secondRange := Range{Line: 5, Character: 4, RefId: 3}
+	secondRange := Range{Line: 5, Character: 4, RefId: 4}
 	rg, err = r.getRange(2)
 	require.NoError(t, err)
 	require.Equal(t, &secondRange, rg)
+
+	thirdRange := Range{Line: 7, Character: 4, RefId: 4}
+	rg, err = r.getRange(3)
+	require.NoError(t, err)
+	require.Equal(t, &thirdRange, rg)
 }
 
 func TestSerialize(t *testing.T) {
@@ -30,7 +35,7 @@ func TestSerialize(t *testing.T) {
 
 	var buf bytes.Buffer
 	err := r.Serialize(&buf, []Id{1}, docs)
-	want := `[{"start_line":1,"start_char":2,"definition_path":"def-path#L2","hover":null,"references":[{"path":"ref-path#L6"}]}` + "\n]"
+	want := `[{"start_line":1,"start_char":2,"definition_path":"def-path#L2","hover":null,"references":[{"path":"ref-path#L6"},{"path":"ref-path#L8"}]}` + "\n]"
 
 	require.NoError(t, err)
 	require.Equal(t, want, buf.String())
@@ -42,9 +47,11 @@ func setup(t *testing.T) (*Ranges, func()) {
 
 	require.NoError(t, r.Read("range", []byte(`{"id":1,"label":"range","start":{"line":1,"character":2}}`)))
 	require.NoError(t, r.Read("range", []byte(`{"id":"2","label":"range","start":{"line":5,"character":4}}`)))
+	require.NoError(t, r.Read("range", []byte(`{"id":"3","label":"range","start":{"line":7,"character":4}}`)))
 
-	require.NoError(t, r.Read("item", []byte(`{"id":4,"label":"item","property":"definitions","outV":"3","inVs":[1],"document":"6"}`)))
-	require.NoError(t, r.Read("item", []byte(`{"id":"5","label":"item","property":"references","outV":3,"inVs":["2"],"document":"7"}`)))
+	require.NoError(t, r.Read("item", []byte(`{"id":5,"label":"item","property":"definitions","outV":"4","inVs":[1],"document":"6"}`)))
+	require.NoError(t, r.Read("item", []byte(`{"id":"6","label":"item","property":"references","outV":4,"inVs":["2"],"document":"7"}`)))
+	require.NoError(t, r.Read("item", []byte(`{"id":"7","label":"item","property":"references","outV":4,"inVs":["3"],"document":"7"}`)))
 
 	cleanup := func() {
 		require.NoError(t, r.Close())

internal/lsif_transformer/parser/references.go

@@ -53,11 +53,13 @@ func (r *References) Store(refId Id, references []Item) error {
 		return nil
 	}
 
-	err := r.Items.SetEntry(r.CurrentOffsetId, references)
+	items := append(r.getItems(refId), references...)
+	err := r.Items.SetEntry(r.CurrentOffsetId, items)
 	if err != nil {
 		return err
 	}
 
+	size = len(items)
 	r.Offsets.SetEntry(refId, ReferencesOffset{Id: r.CurrentOffsetId, Len: int32(size)})
 	r.CurrentOffsetId += Id(size)