From 725874d1630c62afe225838b8cbbf6552225ca9d Mon Sep 17 00:00:00 2001
From: Jan Philipp Timme <jan.philipp@timme.it>
Date: Sun, 2 Jun 2019 19:45:28 +0200
Subject: [PATCH] Rewrite how registry objects get parsed for better
 compatibility

---
 generate.py | 51 +++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 37 insertions(+), 14 deletions(-)

diff --git a/generate.py b/generate.py
index 8b4900e..20eac54 100755
--- a/generate.py
+++ b/generate.py
@@ -52,27 +52,42 @@ def parse_route_objects(folder_path):
     for root, dirs, files in os.walk(folder_path):
         for filename in files:
             current_file = str(root) + str(filename)
-            route_object = read_route_object_file(current_file)
+            route_object = read_schema_object_file(current_file)
             route_objects.append(route_object)
     return route_objects
 
-def read_route_object_file(file_path):
+def read_schema_object_file(file_path):
     with open(file_path) as f:
         data = f.read()
-    route_object = {}
+    schema_object = {}
+    current_key = None
+    current_value = None
     for line in data.split('\n'):
-        if line.strip() == '': continue
-        items = line.split(':', 1)
-        key = items[0].strip()
-        value = items[1].strip()
-        if key in route_object.keys():
-            if isinstance(route_object[key], list):
-                route_object[key].append(value)
-            else:
-                route_object[key] = [route_object[key], value]
+        # Ignore completely empty lines
+        if line == '': continue
+        # If a new key/value pair comes up (meaning there's a : within the first 20 chars)
+        if ':' in line[0:20]:
+            # Write last processed key/value-pair into schema_object if needed
+            if current_key != None:
+                if current_key in schema_object.keys():
+                    if isinstance(schema_object[current_key], list):
+                        schema_object[current_key].append(current_value)
+                    else:
+                        schema_object[current_key] = [schema_object[current_key], current_value]
+                else:
+                    schema_object[current_key] = current_value
+            # Then begin parsing the next one
+            items = line.split(':', 1)
+            current_key = items[0].strip()
+            current_value = items[1].strip()
         else:
-            route_object[key] = value
-    return route_object
+            # This denotes a linebreak in a multiline value
+            if line[0] == '+':
+                current_value += '\n'
+            elif line.startswith(' ' * 20):
+                # This is an additional line in a multiline value
+                current_value += line[20:] + '\n'
+    return schema_object
 
 def create_roa_entries(route_objects, filter_rules, mode, f):
     for route_object in route_objects:
@@ -101,6 +116,9 @@ def create_roa_entries(route_objects, filter_rules, mode, f):
             f.write('# route object not permitted: ' + str(route_network) + "\n")
             f.write('# route object was denied by filter rule: ' + str(matching_filter_rule) + "\n")
             continue
+        # Check if min-length fits filter requirements
+        
+
         # Figure out max-length (filter rule wins over route object)
         if 'max-length' in route_object.keys():
             allowed_max_len = min(int(matching_filter_rule['max_length']), int(route_object['max-length']))
@@ -119,6 +137,11 @@ def create_roa_entries(route_objects, filter_rules, mode, f):
 if __name__ == '__main__':
     # Get registry path over commandline argument
     registry_path = sys.argv[1]
+
+    # Uncomment this block to test the parser against that evil file in the registry.
+    #print(read_schema_object_file(registry_path + '/data/schema/SCHEMA-SCHEMA'))
+    #exit
+
     # Read filter rules
     ipv4_filter_rules = parse_ipv4_filter_rules(registry_path)
     ipv6_filter_rules = parse_ipv6_filter_rules(registry_path)