sanitize names when using them as tag value

related: graphite-project/graphite-web#2458

Author: replay

lib/carbon/tests/test_util.py

@@ -4,6 +4,7 @@
 
 from carbon.util import parseDestinations
 from carbon.util import enableTcpKeepAlive
+from carbon.util import TaggedSeries
 
 
 class UtilTest(TestCase):
@@ -21,6 +22,33 @@ def setTcpKeepAlive(self, value):
         enableTcpKeepAlive(_Transport(), True, None)
         self.assertEquals(s.getsockopt(socket.SOL_TCP, socket.SO_KEEPALIVE), 1)
 
+    def test_sanitizing_name_as_tag_value(self):
+        test_cases = [
+            {
+                'original': "my~.test.abc",
+                'expected': "my.test.abc",
+            }, {
+                'original': "a.b.c",
+                'expected': "a.b.c",
+            }, {
+                'original': "~~a~~.~~~b~~~.~~~c~~~",
+                'expected': "a.b.c",
+            }, {
+                'original': "a.b.c~",
+                'expected': "a.b.c",
+            }, {
+                'original': "~a.b.c",
+                'expected': "a.b.c",
+            }, {
+                'original': "~a~",
+                'expected': "a",
+            },
+        ]
+
+        for test_case in test_cases:
+            result = TaggedSeries.sanitize_name_as_tag_value(test_case['original'])
+            self.assertEquals(result, test_case['expected'])
+
 
 # Destinations have the form:
 # <host> ::= <string without colons> | "[" <string> "]"

lib/carbon/util.py

@@ -365,7 +365,7 @@ def parse_openmetrics(cls, path):
       tags[m.group(1)] = m.group(2).replace(r'\"', '"').replace(r'\\', '\\')
       rawtags = rawtags[len(m.group(0)):]
 
-    tags['name'] = metric
+    tags['name'] = cls.sanitize_name_as_tag_value(metric)
     return cls(metric, tags)
 
   @classmethod
@@ -386,9 +386,14 @@ def parse_carbon(cls, path):
 
       tags[tag[0]] = tag[1]
 
-    tags['name'] = metric
+    tags['name'] = cls.sanitize_name_as_tag_value(metric)
     return cls(metric, tags)
 
+  @staticmethod
+  def sanitize_name_as_tag_value(name):
+    """take a metric name and sanitize it so it is guaranteed to be a valid tag value"""
+    return name.replace('~', '')
+
   @staticmethod
   def format(tags):
     return tags.get('name', '') + ''.join(sorted([