1 files changed, 57 insertions, 0 deletions
diff --git a/internal/extract/real_test.go b/internal/extract/real_test.go
new file mode 100644
index 0000000..fd52f81
--- /dev/null
+++ b/internal/extract/real_test.go
@@ -0,0 +1,57 @@
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+package extract
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+// TestRealDocuments reads documents made by the programs people use -
+// LibreOffice 25.2 (docx, odt, doc, pdf, ods, xlsx, xls) and pandoc (pptx,
+// epub) - from invented text, not ones built by hand in a test (triage
+// 34s). The zip formats always run; a format that needs an external tool
+// runs when that tool is installed and is skipped otherwise.
+func TestRealDocuments(t *testing.T) {
+	e := newWithPath(os.Getenv("PATH"))
+	found := map[string]bool{}
+	for _, tool := range e.Tools() {
+		found[tool.Name] = tool.Path != ""
+	}
+	for _, c := range []struct {
+		file  string
+		tools []string // any one of them will do; none means Go reads it
+		want  string
+	}{
+		{"doc.docx", nil, "acme ltd"},
+		{"doc.odt", nil, "acme ltd"},
+		{"sheet.ods", nil, "acme ltd"},
+		{"sheet.xlsx", nil, "acme ltd"},
+		{"slides.pptx", nil, "acme ltd"},
+		{"book.epub", nil, "acme ltd"},
+		{"doc.pdf", []string{"pdftotext"}, "acme ltd"},
+		{"doc.doc", []string{"antiword", "catdoc"}, "acme ltd"},
+		{"sheet.xls", []string{"xls2csv"}, "acme ltd"},
+	} {
+		t.Run(c.file, func(t *testing.T) {
+			if len(c.tools) > 0 {
+				ok := false
+				for _, tool := range c.tools {
+					ok = ok || found[tool]
+				}
+				if !ok {
+					t.Skipf("none of %v installed", c.tools)
+				}
+			}
+			got, err := text(t, e, filepath.Join("testdata", "real", c.file), 0)
+			if err != nil {
+				t.Fatal(err)
+			}
+			if !strings.Contains(got, c.want) {
+				t.Errorf("text of %s lacks %q:\n%q", c.file, c.want, got)
+			}
+		})
+	}
+}