handle binary files, more progress line fixes

1 files changed, 15 insertions, 5 deletions

diff --git a/suspicious b/suspicious
index fe09204..dedf43f 100755
--- a/suspicious
+++ b/suspicious
@@ -302,17 +302,27 @@ for file in filelist:
 		continue
 	opened +=1
 	now = datetime.datetime.now()
-	estimate = (((now - start) / (opened + skipped)) * len(filelist)) 
 	if options.display_progress: 
-		if len(file)> 50:
+		frac_done = (opened + skipped)*1.0/len(filelist)
+		est = ((now - start) / (opened + skipped)) * len(filelist)
+		est_hr, est_rem = divmod(est.total_seconds(), 3600)
+		est_min, est_sec = divmod(est_rem, 60)
+		if len(file)> 52:
 			prog_file = file.split('/')[0] + "/.../" + file.split('/')[-1]
+			if len(prog_file) > 52:
+				prog_file = prog_file[0:52]
 		else:
 			prog_file = file
-		print('\r' + " " * len(progresstext) + '\r', end='', file=sys.stderr)
-		progresstext = str(((opened + skipped)*1.0/len(filelist))*100)[:5] + '% '+ " time left:" + str(estimate).split('.')[0] + ' ' + prog_file + '\r'
+		progresstext = \
+			"{:6.1%}  time left: {:02.0f}:{:02.0f}:{:02.0f} {:52s}\r" \
+			.format(frac_done, est_hr, est_min, est_sec, prog_file)
 		print(progresstext, end='', file=sys.stderr)
 	sys.stdout.flush()
-	filecontents = f.read()
+	try:
+		filecontents = f.read()
+	except UnicodeDecodeError:
+		print("possible binary: " + file)
+		continue
 	datasize += len(filecontents)		
 	filenamescore = scoretext(wordlist, file, options.maxwholewordlength)
 	filecontentsscore = scoretext(wordlist, filecontents, options.maxwholewordlength)