From 4766c457c07ba5d275cd4ede18f693fe79483d83 Mon Sep 17 00:00:00 2001
From: Jakub Onderka <jakub.onderka@gmail.com>
Date: Fri, 22 May 2020 19:23:08 +0200
Subject: [PATCH] Validate values in CI

---
 tools/validate_values.py | 78 ++++++++++++++++++++++++++++++++++++++++
 validate_all.sh          |  4 +++
 2 files changed, 82 insertions(+)
 create mode 100644 tools/validate_values.py

diff --git a/tools/validate_values.py b/tools/validate_values.py
new file mode 100644
index 0000000..b7cce97
--- /dev/null
+++ b/tools/validate_values.py
@@ -0,0 +1,78 @@
+#!/usr/bin/env python3
+import re
+import json
+import sys
+from ipaddress import ip_network
+from pathlib import Path
+from typing import List, Iterator, Optional, Tuple
+
+HOSTNAME_RE = re.compile("(?!-)[A-Z\d-]{1,63}(?<!-)$", re.IGNORECASE)
+
+
+class InvalidListValue:
+    def __init__(self, file_path: Path, value: str, error: Optional[str] = None):
+        self.file_path = file_path
+        self.value = value
+        self.error = error
+
+    def __str__(self):
+        if self.error:
+            return "{}: {}".format(self.file_path, self.error)
+        else:
+            return "{}: Invalid value '{}'".format(self.file_path, self.value)
+
+
+def is_valid_hostname(hostname: str):
+    if len(hostname) > 255:
+        raise ValueError("Hostname {0!r} is too long (maximum is 255 chars)".format(hostname))
+
+    if hostname[-1] == ".":
+        hostname = hostname[:-1]  # strip exactly one dot from the right, if present
+
+    if hostname[0] == ".":
+        hostname = hostname[1:]  # strip exactly one dot from the left, if present
+
+    if not all(HOSTNAME_RE.match(x) for x in hostname.split(".")):
+        raise ValueError("Hostname {0!r} contains invalid characters".format(hostname))
+
+
+def is_valid_regexp(regexp: str):
+    try:
+        re.compile(regexp)
+    except re.error:
+        raise ValueError("Regexp {0!r} is not valid".format(regexp))
+
+
+def validate(values: List[str], func) -> Iterator[Tuple[str, ValueError]]:
+    for value in values:
+        try:
+            func(value)
+        except ValueError as e:
+            yield value, e
+
+
+def validate_file(p: Path) -> Iterator[InvalidListValue]:
+    invalid_values = []
+    with p.open() as f:
+        warninglist = json.load(f, encoding="utf-8")
+        if warninglist["type"] == "cidr":
+            invalid_values = validate(warninglist["list"], lambda value: ip_network(value, strict=True))
+        elif warninglist["type"] == "regexp":
+            invalid_values = validate(warninglist["list"], lambda value: is_valid_regexp(value))
+    # Disabled, because current lists contains invalid domains
+    #    elif warninglist["type"] == "hostname":
+    #        invalid_values = validate_hostnames(warninglist["list"])
+
+    for (value, exception) in invalid_values:
+        yield InvalidListValue(p, value, str(exception))
+
+
+if __name__ == "__main__":
+    invalid_values = []
+    for p in Path('../lists/').glob('*/*.json'):
+        invalid_values.extend(validate_file(p))
+
+    if invalid_values:
+        for invalid_value in invalid_values:
+            print(invalid_value, file=sys.stderr)
+        sys.exit(1)
diff --git a/validate_all.sh b/validate_all.sh
index a43b849..a21bcc4 100755
--- a/validate_all.sh
+++ b/validate_all.sh
@@ -42,3 +42,7 @@ do
   echo ''
 done
 
+pushd tools
+python3 validate_values.py
+popd
+