rescribe: Add context cancelling to extractPdfImgs(), so it's no longer possible to get the gui into a bad state by cancelling before startProcess began (hopefully)

2 files changed, 21 insertions, 8 deletions

diff --git a/cmd/rescribe/gui.go b/cmd/rescribe/gui.go
index 92d8f5b..f4ac9bd 100644
--- a/cmd/rescribe/gui.go
+++ b/cmd/rescribe/gui.go
@@ -380,11 +380,13 @@ func startGui(log log.Logger, cmd string, training string, tessdir string) error
 			progressBar.SetValue(0.1)
 
 			if strings.HasSuffix(dir.Text, ".pdf") && !f.IsDir() {
-				bookdir, err = extractPdfImgs(bookdir)
+				bookdir, err = extractPdfImgs(ctx, bookdir)
 				if err != nil {
-					msg := fmt.Sprintf("Error opening PDF: %v\n", bookdir, err)
-					dialog.ShowError(errors.New(msg), myWindow)
-					fmt.Fprintf(os.Stderr, msg)
+					if !strings.HasSuffix(err.Error(), "context canceled") {
+						msg := fmt.Sprintf("Error opening PDF %s: %v\n", bookdir, err)
+						dialog.ShowError(errors.New(msg), myWindow)
+						fmt.Fprintf(os.Stderr, msg)
+					}
 
 					progressBar.SetValue(0.0)
 					gobtn.SetText("Process OCR")
diff --git a/cmd/rescribe/main.go b/cmd/rescribe/main.go
index d4acfc1..41dc83c 100644
--- a/cmd/rescribe/main.go
+++ b/cmd/rescribe/main.go
@@ -264,13 +264,15 @@ These training files are included in rescribe, and are always available:
 		log.Fatalln("Error opening book file/dir:", err)
 	}
 
+	var ctx context.Context
+
 	// try opening as a PDF, and extracting
 	if !fi.IsDir() {
 		if flag.NArg() < 2 {
 			savedir = strings.TrimSuffix(bookdir, ".pdf")
 		}
 
-		bookdir, err = extractPdfImgs(bookdir)
+		bookdir, err = extractPdfImgs(ctx, bookdir)
 		if err != nil {
 			log.Fatalln("Error opening file as PDF:", err)
 		}
@@ -285,8 +287,6 @@ These training files are included in rescribe, and are always available:
 		ispdf = true
 	}
 
-	var ctx context.Context
-
 	err = startProcess(ctx, *verboselog, tessCommand, bookdir, bookname, trainingName, savedir, tessdir)
 	if err != nil {
 		log.Fatalln(err)
@@ -306,7 +306,7 @@ These training files are included in rescribe, and are always available:
 
 // extractPdfImgs extracts all images embedded in a PDF to a
 // temporary directory, which is returned on success.
-func extractPdfImgs(path string) (string, error) {
+func extractPdfImgs(ctx context.Context, path string) (string, error) {
 	defer func() {
 		// unfortunately the pdf library will panic if it sees an encoding
 		// it can't decode, so recover from that and give a warning
@@ -334,6 +334,11 @@ func extractPdfImgs(path string) (string, error) {
 	}
 
 	for pgnum := 1; pgnum <= p.NumPage(); pgnum++ {
+		select {
+		case <-ctx.Done():
+			return "", ctx.Err()
+		default:
+		}
 		if p.Page(pgnum).V.IsNull() {
 			continue
 		}
@@ -376,6 +381,12 @@ func extractPdfImgs(path string) (string, error) {
 	}
 	// TODO: check for places where there are multiple images per page, and only keep largest ones where that's the case
 
+	select {
+	case <-ctx.Done():
+		return "", ctx.Err()
+	default:
+	}
+
 	return tempdir, nil
 }