Add benchmark for ctags conversion (#679)

This change adds a benchmark for the conversion from ctags output to Zoekt
document data, plus a tiny optimization to presize the symbol slices.

build/ctags.go

@@ -112,8 +112,8 @@ func overlaps(symOffsets []zoekt.DocumentSection, start, end uint32) int {
 func tagsToSections(content []byte, tags []*ctags.Entry) ([]zoekt.DocumentSection, []*zoekt.Symbol, error) {
 	nls := newLinesIndices(content)
 	nls = append(nls, uint32(len(content)))
-	var symOffsets []zoekt.DocumentSection
-	var symMetaData []*zoekt.Symbol
+	symOffsets := make([]zoekt.DocumentSection, 0, len(tags))
+	symMetaData := make([]*zoekt.Symbol, 0, len(tags))
 
 	for _, t := range tags {
 		if t.Line <= 0 {

build/ctags_test.go

@@ -15,6 +15,7 @@
 package build
 
 import (
+	"os"
 	"reflect"
 	"testing"
 
@@ -229,3 +230,39 @@ func TestOverlaps(t *testing.T) {
 		})
 	}
 }
+
+func BenchmarkTagsToSections(b *testing.B) {
+	if checkCTags() == "" {
+		b.Skip("ctags not available")
+	}
+
+	file, err := os.ReadFile("./testdata/large_file.cc")
+	parser, err := ctags.NewParser(ctags.UniversalCTags, "universal-ctags")
+	if err != nil {
+		b.Fatal(err)
+	}
+
+	entries, err := parser.Parse("./testdata/large_file.cc", file)
+	if err != nil {
+		b.Fatal(err)
+	}
+
+	secs, _, err := tagsToSections(file, entries)
+	if err != nil {
+		b.Fatal(err)
+	}
+
+	if len(secs) != 439 {
+		b.Fatalf("got %d sections, want 439 sections", len(secs))
+	}
+
+	b.ResetTimer()
+	b.ReportAllocs()
+
+	for n := 0; n < b.N; n++ {
+		_, _, err := tagsToSections(file, entries)
+		if err != nil {
+			b.Fatal(err)
+		}
+	}
+}