test(chromium): cover header and footer image deduplication

test/integration/scenario/images.go

@@ -0,0 +1,72 @@
+package scenario
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+)
+
+func (s *scenario) thePdfShouldReuseImageXObjects(ctx context.Context, name string, images, pages int) error {
+	path := filepath.Join(s.workdir, s.resp.Header().Get("Gotenberg-Trace"), name)
+	if _, err := os.Stat(path); err != nil {
+		return fmt.Errorf("stat PDF %q: %w", path, err)
+	}
+
+	cmd := []string{"pdfimages", "-list", filepath.Base(path)}
+	output, err := execCommandInIntegrationToolsContainer(ctx, cmd, path)
+	if err != nil {
+		return fmt.Errorf("exec %q: %w", cmd, err)
+	}
+
+	return checkImageXObjectReuse(output, images, pages)
+}
+
+func checkImageXObjectReuse(output string, images, pages int) error {
+	lines := strings.Split(strings.TrimSpace(output), "\n")
+	if len(lines) < 2 || !strings.HasPrefix(strings.TrimSpace(lines[0]), "page") {
+		return fmt.Errorf("invalid pdfimages -list output: %q", output)
+	}
+
+	objects := make(map[[2]int]struct{})
+	imagesPerPage := make(map[int]int)
+	for _, line := range lines[2:] {
+		fields := strings.Fields(line)
+		if len(fields) < 12 {
+			return fmt.Errorf("invalid pdfimages -list row: %q", line)
+		}
+
+		page, err := strconv.Atoi(fields[0])
+		if err != nil {
+			return fmt.Errorf("parse image page %q: %w", fields[0], err)
+		}
+		if page < 1 || page > pages {
+			return fmt.Errorf("image on page %d outside expected range 1-%d", page, pages)
+		}
+
+		// pdfimages lists appearances, including repeated references to the
+		// same image. Columns 11 and 12 identify the underlying PDF object.
+		var ref [2]int
+		for i := range ref {
+			ref[i], err = strconv.Atoi(fields[10+i])
+			if err != nil {
+				return fmt.Errorf("parse image object ID in %q: %w", line, err)
+			}
+		}
+		objects[ref] = struct{}{}
+		imagesPerPage[page]++
+	}
+
+	for page := 1; page <= pages; page++ {
+		if imagesPerPage[page] != images {
+			return fmt.Errorf("expected %d image appearance(s) on page %d, but actual is %d", images, page, imagesPerPage[page])
+		}
+	}
+	if len(objects) != images {
+		return fmt.Errorf("expected %d unique image XObject(s) shared across %d page(s), but actual is %d", images, pages, len(objects))
+	}
+
+	return nil
+}

test/integration/scenario/images_test.go

@@ -0,0 +1,106 @@
+package scenario
+
+import (
+	"fmt"
+	"strings"
+	"testing"
+)
+
+func TestCheckImageXObjectReuse(t *testing.T) {
+	t.Parallel()
+
+	const header = `page   num  type   width height color comp bpc  enc interp  object ID x-ppi y-ppi size ratio
+--------------------------------------------------------------------------------------------
+`
+	imageRow := func(page, object, generation int) string {
+		return fmt.Sprintf("%d %d image 32 32 rgb 3 8 image no %d %d 96 96 3104B 101%%\n", page, page-1, object, generation)
+	}
+
+	tests := []struct {
+		name    string
+		output  string
+		pages   int
+		wantErr string
+	}{
+		{
+			name:   "one page baseline",
+			output: header + imageRow(1, 7, 0),
+			pages:  1,
+		},
+		{
+			name:   "three pages share one image object",
+			output: header + imageRow(1, 7, 0) + imageRow(2, 7, 0) + imageRow(3, 7, 0),
+			pages:  3,
+		},
+		{
+			name:    "regression embeds a different object per page",
+			output:  header + imageRow(1, 7, 0) + imageRow(2, 8, 0) + imageRow(3, 9, 0),
+			pages:   3,
+			wantErr: "expected 1 unique image XObject(s) shared across 3 page(s), but actual is 3",
+		},
+		{
+			name:    "object generation is part of its identity",
+			output:  header + imageRow(1, 7, 0) + imageRow(2, 7, 1) + imageRow(3, 7, 0),
+			pages:   3,
+			wantErr: "expected 1 unique image XObject(s) shared across 3 page(s), but actual is 2",
+		},
+		{
+			name:    "image missing from a page",
+			output:  header + imageRow(1, 7, 0) + imageRow(3, 7, 0),
+			pages:   3,
+			wantErr: "expected 1 image appearance(s) on page 2, but actual is 0",
+		},
+		{
+			name:    "no image cannot pass",
+			output:  header,
+			pages:   1,
+			wantErr: "expected 1 image appearance(s) on page 1, but actual is 0",
+		},
+		{
+			name:    "repeated image only on first page cannot pass",
+			output:  header + imageRow(1, 7, 0) + imageRow(1, 7, 0) + imageRow(1, 7, 0),
+			pages:   3,
+			wantErr: "expected 1 image appearance(s) on page 1, but actual is 3",
+		},
+		{
+			name:    "image on unexpected page",
+			output:  header + imageRow(1, 7, 0) + imageRow(2, 7, 0),
+			pages:   1,
+			wantErr: "image on page 2 outside expected range 1-1",
+		},
+		{
+			name:    "tool error cannot pass",
+			output:  "Syntax Error: Couldn't find trailer dictionary",
+			pages:   1,
+			wantErr: "invalid pdfimages -list output:",
+		},
+		{
+			name:    "malformed image row",
+			output:  header + "1 0 image\n",
+			pages:   1,
+			wantErr: "invalid pdfimages -list row:",
+		},
+		{
+			name:    "inline image has no reusable object ID",
+			output:  header + "1 0 image 32 32 rgb 3 8 image no [inline] - 96 96 3104B 101%\n",
+			pages:   1,
+			wantErr: "parse image object ID",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+			err := checkImageXObjectReuse(tt.output, 1, tt.pages)
+			if tt.wantErr == "" {
+				if err != nil {
+					t.Fatalf("unexpected error: %v", err)
+				}
+				return
+			}
+			if err == nil || !strings.Contains(err.Error(), tt.wantErr) {
+				t.Fatalf("expected error containing %q, got %v", tt.wantErr, err)
+			}
+		})
+	}
+}

test/integration/scenario/scenario.go

@@ -1799,6 +1799,7 @@ func InitializeScenario(ctx *godog.ScenarioContext) {
 	ctx.Then(`^the "([^"]*)" PDF (should|should NOT) have the following content at page (\d+):$`, s.thePdfShouldHaveTheFollowingContentAtPage)
 	ctx.Then(`^the "([^"]*)" PDF (should|should NOT) have content matching "([^"]*)" at page (\d+)$`, s.thePdfShouldHaveContentMatchingAtPage)
 	ctx.Then(`^the "([^"]*)" PDF should have (\d+) image\(s\)$`, s.thePdfShouldHaveImages)
+	ctx.Then(`^the "([^"]*)" PDF should reuse (\d+) image XObject\(s\) across (\d+) page\(s\)$`, s.thePdfShouldReuseImageXObjects)
 	ctx.Then(`^the "([^"]*)" image should be (\d+)x(\d+) pixels$`, s.theImageShouldBePixels)
 	ctx.Then(`^the "([^"]*)" image pixel at (\d+),(\d+) should be "([^"]*)"$`, s.theImagePixelShouldBe)
 	ctx.Then(`^the "([^"]*)" file size should be (less|greater) than (\d+) KB$`, s.theFileSizeShouldBe)