[cache] Add cache validation by program version, based on yt-dlp

test/test_cache.py

@@ -3,17 +3,18 @@
 
 from __future__ import unicode_literals
 
-import shutil
-
 # Allow direct execution
 import os
 import sys
 import unittest
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
+import shutil
 
 from test.helper import FakeYDL
 from youtube_dl.cache import Cache
+from youtube_dl.utils import version_tuple
+from youtube_dl.version import __version__
 
 
 def _is_empty(d):
@@ -54,6 +55,17 @@ class TestCache(unittest.TestCase):
         self.assertFalse(os.path.exists(self.test_dir))
         self.assertEqual(c.load('test_cache', 'k.'), None)
 
+    def test_cache_validation(self):
+        ydl = FakeYDL({
+            'cachedir': self.test_dir,
+        })
+        c = Cache(ydl)
+        obj = {'x': 1, 'y': ['ä', '\\a', True]}
+        c.store('test_cache', 'k.', obj)
+        self.assertEqual(c.load('test_cache', 'k.', min_ver='1970.01.01'), obj)
+        new_version = '.'.join(('%d' % ((v + 1) if i == 0 else v, )) for i, v in enumerate(version_tuple(__version__)))
+        self.assertIs(c.load('test_cache', 'k.', min_ver=new_version), None)
+
 
 if __name__ == '__main__':
     unittest.main()

youtube_dl/cache.py

@@ -10,12 +10,21 @@ import traceback
 
 from .compat import compat_getenv
 from .utils import (
+    error_to_compat_str,
     expand_path,
+    is_outdated_version,
+    try_get,
     write_json_file,
 )
+from .version import __version__
 
 
 class Cache(object):
+
+    _YTDL_DIR = 'youtube-dl'
+    _VERSION_KEY = _YTDL_DIR + '_version'
+    _DEFAULT_VERSION = '2021.12.17'
+
     def __init__(self, ydl):
         self._ydl = ydl
 
@@ -23,7 +32,7 @@ class Cache(object):
         res = self._ydl.params.get('cachedir')
         if res is None:
             cache_root = compat_getenv('XDG_CACHE_HOME', '~/.cache')
-            res = os.path.join(cache_root, 'youtube-dl')
+            res = os.path.join(cache_root, self._YTDL_DIR)
         return expand_path(res)
 
     def _get_cache_fn(self, section, key, dtype):
@@ -50,13 +59,22 @@ class Cache(object):
             except OSError as ose:
                 if ose.errno != errno.EEXIST:
                     raise
-            write_json_file(data, fn)
+            write_json_file({self._VERSION_KEY: __version__, 'data': data}, fn)
         except Exception:
             tb = traceback.format_exc()
             self._ydl.report_warning(
                 'Writing cache to %r failed: %s' % (fn, tb))
 
-    def load(self, section, key, dtype='json', default=None):
+    def _validate(self, data, min_ver):
+        version = try_get(data, lambda x: x[self._VERSION_KEY])
+        if not version:  # Backward compatibility
+            data, version = {'data': data}, self._DEFAULT_VERSION
+        if not is_outdated_version(version, min_ver or '0', assume_new=False):
+            return data['data']
+        self._ydl.to_screen(
+            'Discarding old cache from version {version} (needs {min_ver})'.format(**locals()))
+
+    def load(self, section, key, dtype='json', default=None, min_ver=None):
         assert dtype in ('json',)
 
         if not self.enabled:
@@ -66,12 +84,12 @@ class Cache(object):
         try:
             try:
                 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
-                    return json.load(cachef)
+                    return self._validate(json.load(cachef), min_ver)
             except ValueError:
                 try:
                     file_size = os.path.getsize(cache_fn)
                 except (OSError, IOError) as oe:
-                    file_size = str(oe)
+                    file_size = error_to_compat_str(oe)
                 self._ydl.report_warning(
                     'Cache retrieval from %s failed (%s)' % (cache_fn, file_size))
         except IOError: