Merge pull request 'fix(docs): pair comment ids with the document order (#92)' (#93) from fix/docs-comment-id#92 into main

docx_comment.go

@@ -65,6 +65,7 @@ func InjectDocxComments(src []byte, author string, comments []DocxComment) ([]by
 	}
 
 	used := make([]bool, len(comments))
+	applied := make([]DocxComment, 0, len(comments))
 	id := 0
 	newDoc := docxParaRe.ReplaceAllStringFunc(string(doc), func(p string) string {
 		txt := docxPlain(p)
@@ -73,6 +74,9 @@ func InjectDocxComments(src []byte, author string, comments []DocxComment) ([]by
 				continue
 			}
 			used[i] = true
+			// Record in document order so comments.xml id N matches the marker
+			// with id N (the spec order may differ from the document order).
+			applied = append(applied, c)
 			p = docxAddMarkers(p, id)
 			id++
 			break
@@ -80,12 +84,6 @@ func InjectDocxComments(src []byte, author string, comments []DocxComment) ([]by
 		return p
 	})
 
-	applied := make([]DocxComment, 0, len(comments))
-	for i, c := range comments {
-		if used[i] {
-			applied = append(applied, c)
-		}
-	}
 	parts["word/document.xml"] = []byte(newDoc)
 	parts["word/comments.xml"] = docxBuildComments(author, applied)
 	parts["[Content_Types].xml"] = docxAddContentType(parts["[Content_Types].xml"])

docx_comment_test.go

@@ -4,6 +4,7 @@ import (
 	"archive/zip"
 	"bytes"
 	"io"
+	"regexp"
 	"strings"
 	"testing"
 )
@@ -124,6 +125,48 @@ func TestInjectDocxCommentsNoMatch(t *testing.T) {
 	}
 }
 
+// TestInjectDocxCommentsOrder guards the marker-id ↔ comments.xml pairing when
+// the spec order differs from the document order.
+func TestInjectDocxCommentsOrder(t *testing.T) {
+	src := buildTestDocx(t)
+	spec := []DocxComment{
+		{Anchor: "Nothing relevant", Text: "comment B"},
+		{Anchor: "The term is two years", Text: "comment A"},
+	}
+	out, applied, err := InjectDocxComments(src, "A", spec)
+	if err != nil || applied != 2 {
+		t.Fatalf("inject: applied=%d err=%v", applied, err)
+	}
+	doc := unzipPart(t, out, "word/document.xml")
+	cmts := unzipPart(t, out, "word/comments.xml")
+
+	// id -> text from comments.xml
+	byID := map[string]string{}
+	for _, m := range regexp.MustCompile(`(?s)<w:comment w:id="(\d+)"[^>]*>(.*?)</w:comment>`).FindAllStringSubmatch(cmts, -1) {
+		txt := strings.TrimSpace(stripTags(m[2]))
+		byID[m[1]] = txt
+	}
+	// paragraph that starts with the term text must point at "comment A"
+	for _, p := range regexp.MustCompile(`(?s)<w:p\b[^>]*>.*?</w:p>`).FindAllString(doc, -1) {
+		if !strings.Contains(stripTags(p), "The term is two years") {
+			continue
+		}
+		m := regexp.MustCompile(`<w:commentRangeStart w:id="(\d+)"/>`).FindStringSubmatch(p)
+		if m == nil {
+			t.Fatal("no commentRangeStart in the term paragraph")
+		}
+		if got := byID[m[1]]; got != "comment A" {
+			t.Fatalf("term paragraph points at id %s = %q, want \"comment A\"", m[1], got)
+		}
+		return
+	}
+	t.Fatal("term paragraph not found")
+}
+
+func stripTags(s string) string {
+	return regexp.MustCompile(`<[^>]*>`).ReplaceAllString(s, "")
+}
+
 func TestParseDocxComments(t *testing.T) {
 	got, err := ParseDocxComments([]byte(`[{"anchor":"a","text":"b"}]`))
 	if err != nil {