[tracker/poppler-glib: 15/16] tracker-extract,pdf: Improve logging



commit f5f1a2d0fb6d5d7e04d076b9f4efa4ba2910a295
Author: Aleksander Morgado <aleksander lanedo com>
Date:   Tue Feb 22 15:08:33 2011 +0100

    tracker-extract,pdf: Improve logging

 src/tracker-extract/tracker-extract-pdf.c |   15 ++++++++++++---
 1 files changed, 12 insertions(+), 3 deletions(-)
---
diff --git a/src/tracker-extract/tracker-extract-pdf.c b/src/tracker-extract/tracker-extract-pdf.c
index 9de3b3c..673b51f 100644
--- a/src/tracker-extract/tracker-extract-pdf.c
+++ b/src/tracker-extract/tracker-extract-pdf.c
@@ -189,13 +189,14 @@ extract_content (PopplerDocument *document,
 	gint n_pages, i = 0;
 	GString *string;
 	GTimer *timer;
+	gsize remaining_bytes = n_bytes;
 
 	n_pages = poppler_document_get_n_pages (document);
 	string = g_string_new ("");
 	timer = g_timer_new ();
 
 	while (i < n_pages &&
-	       n_bytes > 0 &&
+	       remaining_bytes > 0 &&
 	       g_timer_elapsed (timer, NULL) < 5) {
 		PopplerPage *page;
 		gsize written_bytes;
@@ -212,18 +213,26 @@ extract_content (PopplerDocument *document,
 		}
 
 		if (tracker_text_validate_utf8 (text,
-		                                MIN (strlen (text), n_bytes),
+		                                MIN (strlen (text), remaining_bytes),
 		                                &string,
 		                                &written_bytes)) {
 			g_string_append_c (string, ' ');
 		}
 
-		n_bytes -= written_bytes;
+		remaining_bytes -= written_bytes;
+
+		g_debug ("Extracted %" G_GSIZE_FORMAT " bytes from page %d, "
+		         "%" G_GSIZE_FORMAT " bytes remaining",
+		         written_bytes, i, remaining_bytes);
 
 		g_free (text);
 		g_object_unref (page);
 	}
 
+	g_debug ("Content extraction finished: %d/%d pages indexed in %lf seconds, "
+	         "%" G_GSIZE_FORMAT " bytes extracted",
+	         i, n_pages, g_timer_elapsed (timer, NULL), (n_bytes - remaining_bytes));
+
 	g_timer_destroy (timer);
 
 	return g_string_free (string, FALSE);



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]