diff --git a/tests/add_test.py b/tests/add_test.py
index f6e3e68..a7492d6 100755
--- a/tests/add_test.py
+++ b/tests/add_test.py
@@ -4,57 +4,56 @@ import os, sys, re
 
 args = sys.argv
 
-if not len(args) in (3, 4):
-	print "Usage: add_test.py <rawdata file> <report file> [tests stash file]"
-	print "If the last is not specified, `tests_stash.txt' is assumed"
+if len(args) != 3:
+	print "Usage: create_test.py <rawdata file> <report file>"
+	print "The .test file is written to stdout."
 	sys.exit(1)
 
-rawdata = []
-for line in open(args[1], "rt").readlines():
-	lookfor = ["basic_cpuid", "ext_cpuid", "intel_fn4", "intel_fn11"]
-	good = False
-	for match in lookfor:
-		if line.find(match) != -1:
-			good = True
-			break
-	if good:
-		rawdata.append(line.strip())
+def readRawFile():
+	rawdata = []
+	for line in open(args[1], "rt").readlines():
+		lookfor = ["basic_cpuid", "ext_cpuid", "intel_fn4", "intel_fn11"]
+		good = False
+		for match in lookfor:
+			if line.find(match) != -1:
+				good = True
+				break
+		if good:
+			rawdata.append(line.strip())
+	return rawdata
 
-repdata = []
-rexp = re.compile('(-?[0-9]+).*')
-for line in open(args[2], "rt").readlines():
-	s = line.strip()
-	if s.find(":") == -1:
-		continue
-	numeric = ["family", "model", "stepping", "ext_family", "ext_model",
-	           "num_cores", "num_logical", "L1 D cache", "L1 I cache",
-	           "L2 cache", "L3 cache", "L1D assoc.", "L2 assoc.",
-	           "L3 assoc.", "L1D line sz", "L2 line sz", "L3 line sz"]
-	field = s[:s.find(":")].strip()
-	if field in numeric:
-		value = s[s.find(":")+1:].strip()
-		if not rexp.match(value):
-			raise "Bad format of value: [%s]" % s
-		repdata.append(rexp.findall(value)[0])
-	if field == "code name":
-		value = s[s.find("`") + 1: s.find("'")]
-		repdata.append(value)
-	if field == "features":
-		value = s[s.find(":") + 2:]
-		repdata.append(value)
-	if field == "SSE units":
-		value = s[s.find(":") + 2:]
-		# the value here is something like "XX bits (authoritative)". We remove the "bits" part:
-		i = value.find("bits")
-		if i != -1:
-			value = value[:i] + value[i + 5:]
-		repdata.append(value)
+def readResultFile():
+	repdata = []
+	rexp = re.compile('(-?[0-9]+).*')
+	for line in open(args[2], "rt").readlines():
+		s = line.strip()
+		if s.find(":") == -1:
+			continue
+		numeric = ["family", "model", "stepping", "ext_family", "ext_model",
+				"num_cores", "num_logical", "L1 D cache", "L1 I cache",
+				"L2 cache", "L3 cache", "L1D assoc.", "L2 assoc.",
+				"L3 assoc.", "L1D line sz", "L2 line sz", "L3 line sz"]
+		field = s[:s.find(":")].strip()
+		if field in numeric:
+			value = s[s.find(":")+1:].strip()
+			if not rexp.match(value):
+				raise "Bad format of value: [%s]" % s
+			repdata.append(rexp.findall(value)[0])
+		if field == "code name":
+			value = s[s.find("`") + 1: s.find("'")]
+			repdata.append(value)
+		if field == "features":
+			value = s[s.find(":") + 2:]
+			repdata.append(value)
+		if field == "SSE units":
+			value = s[s.find(":") + 2:]
+			# the value here is something like "XX bits (authoritative)". We remove the "bits" part:
+			i = value.find("bits")
+			if i != -1:
+				value = value[:i] + value[i + 5:]
+			repdata.append(value)
+	return repdata
 
-stash = "tests_stash.txt"
-if len(args) == 4:
-	stash = args[3]
-fout = open(stash, "at")
 delimiter = "-" * 80
-lines = rawdata + [delimiter] + repdata + [delimiter]
-fout.writelines(map(lambda s: s + "\n", lines))
-fout.close()
+lines = readRawFile() + [delimiter] + readResultFile()
+sys.stdout.writelines(map(lambda s: s + "\n", lines))