regexs for getting packageName and versionCode from filenames

This is useful for parsing APK files, which can include packageName,
versionCode, and optionally 7 char signing key ID (i.e. <sig>).
This also can set the packageName and versionCoe for non APK files, so
that it is easy to assign them to metadata files, and to allow for
upgrades by setting the versionCode in the filename.

fdroidserver/common.py

@@ -55,6 +55,8 @@ from .asynchronousfilereader import AsynchronousFileReader
 
 # A signature block file with a .DSA, .RSA, or .EC extension
 CERT_PATH_REGEX = re.compile(r'^META-INF/.*\.(DSA|EC|RSA)$')
+APK_NAME_REGEX = re.compile(r'^([a-zA-Z][\w.]*)_(-?[0-9]+)_?([0-9a-f]{7})?\.apk')
+STANDARD_FILE_NAME_REGEX = re.compile(r'^(\w[\w.]*)_(-?[0-9]+)\.\w+')
 
 XMLElementTree.register_namespace('android', 'http://schemas.android.com/apk/res/android')
 

fdroidserver/update.py

@@ -853,14 +853,10 @@ def scan_repo_files(apkcache, repodir, knownapks, use_date_from_file=False):
             # the static ID is the SHA256 unless it is set in the metadata
             repo_file['packageName'] = shasum
 
-            n = name_utf8.rsplit('_', maxsplit=1)
-            if len(n) == 2:
-                packageName = n[0]
-                versionCode = n[1].split('.')[0]
-                if re.match('^-?[0-9]+$', versionCode) \
-                   and common.is_valid_package_name(n[0]):
-                    repo_file['packageName'] = packageName
-                    repo_file['versionCode'] = int(versionCode)
+            m = common.STANDARD_FILE_NAME_REGEX.match(name_utf8)
+            if m:
+                repo_file['packageName'] = m.group(1)
+                repo_file['versionCode'] = int(m.group(2))
             srcfilename = name + b'_src.tar.gz'
             if os.path.exists(os.path.join(repodir, srcfilename)):
                 repo_file['srcname'] = srcfilename.decode('utf-8')

tests/common.TestCase

@@ -281,6 +281,70 @@ class CommonTest(unittest.TestCase):
                 key = "val"
                 """))
 
+    def test_apk_name_regex(self):
+        good = [
+            'urzipπÇÇπÇÇ现代汉语通用字българскиعربي1234ö_-123456.apk',
+            'urzipπÇÇπÇÇ现代汉语通用字българскиعربي1234ö_123456_abcdef0.apk',
+            'urzip_-123456.apk',
+            'a0_0.apk',
+            'Z0_0.apk',
+            'a0_0_abcdef0.apk',
+            'a_a_a_a_0_abcdef0.apk',
+            'a_____0.apk',
+            'a_____123456_abcdef0.apk',
+            'org.fdroid.fdroid_123456.apk',
+            # valid, but "_99999" is part of packageName rather than versionCode
+            'org.fdroid.fdroid_99999_123456.apk',
+            # should be valid, but I can't figure out the regex since \w includes digits
+            # 'πÇÇπÇÇ现代汉语通用字българскиعربي1234ö_0_123bafd.apk',
+        ]
+        for name in good:
+            m = fdroidserver.common.APK_NAME_REGEX.match(name)
+            self.assertIsNotNone(m)
+            self.assertIn(m.group(2), ('-123456', '0', '123456'))
+            self.assertIn(m.group(3), ('abcdef0', None))
+
+        bad = [
+            'urzipπÇÇπÇÇ现代汉语通用字българскиعربي1234ö_123456_abcdefg.apk',
+            'urzip-_-198274.apk',
+            'urzip-_0_123bafd.apk',
+            'no spaces allowed_123.apk',
+            '0_0.apk',
+            '0_0_abcdef0.apk',
+        ]
+        for name in bad:
+            self.assertIsNone(fdroidserver.common.APK_NAME_REGEX.match(name))
+
+    def test_standard_file_name_regex(self):
+        good = [
+            'urzipπÇÇπÇÇ现代汉语通用字българскиعربي1234ö_-123456.mp3',
+            'urzipπÇÇπÇÇ现代汉语通用字българскиعربي1234ö_123456.mov',
+            'Document_-123456.pdf',
+            'WTF_0.MOV',
+            'Z0_0.ebk',
+            'a_a_a_a_0.txt',
+            'org.fdroid.fdroid.privileged.ota_123456.zip',
+            'πÇÇπÇÇ现代汉语通用字българскиعربي1234ö_0.jpeg',
+            'a_____0.PNG',
+            # valid, but "_99999" is part of packageName rather than versionCode
+            'a_____99999_123456.zip',
+            'org.fdroid.fdroid_99999_123456.zip',
+        ]
+        for name in good:
+            m = fdroidserver.common.STANDARD_FILE_NAME_REGEX.match(name)
+            self.assertIsNotNone(m)
+            self.assertIn(m.group(2), ('-123456', '0', '123456'))
+
+        bad = [
+            'urzipπÇÇπÇÇ现代汉语通用字българскиعربي1234ö_abcdefg.JPEG',
+            'urzip-_-198274.zip',
+            'urzip-_123bafd.pdf',
+            'no spaces allowed_123.foobar',
+            'a_____0.',
+        ]
+        for name in bad:
+            self.assertIsNone(fdroidserver.common.STANDARD_FILE_NAME_REGEX.match(name))
+
 
 if __name__ == "__main__":
     parser = optparse.OptionParser()

tests/stats/known_apks.txt

@@ -1,4 +1,4 @@
-fake.ota.update_1234.zip 897a92a4ccff4f415f6ba275b2af16d4ecaee60a983b215bddcb9f8964e7a24c 2016-03-10
+fake.ota.update_1234.zip fake.ota.update 2016-03-10
 obb.main.oldversion_1444412523.apk obb.main.oldversion 2013-12-31
 obb.main.twoversions_1101613.apk obb.main.twoversions 2015-10-12
 obb.main.twoversions_1101615.apk obb.main.twoversions 2016-01-01