encode filenames as bytes to handle all locale setups

This was failing on environments that did not have any LANG or LC_* locale
variables set.  This is a valid setup, and is common in headless setups, so
it needs to be handled.

This also adds a new pass of the test suite without the locale env vars set
so that this situation is also tests on gitlab-ci, not only gpjenkins.

The error this caused was:
UnicodeEncodeError: 'ascii' codec can't encode characters in position 6-18: ordinal not in range(128)

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 0d1713c0dd1b78f275c79345c7478314c101aefc..cbda46b7dca4978034a0ff10f40d4720b4c87c9f 100644 (file)
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -1,4 +1,4 @@
-image: fdroid/ci:server-20161223
+image: registry.gitlab.com/fdroid/ci-images:server-latest
 
 test:
   script:

diff --git a/fdroidserver/common.py b/fdroidserver/common.py
index 42918ed7f3a00d21ee4c40a724e7d44c6a2f4a2d..0c12c3f9b9875d9228cc54919964dcabdcb1da45 100644 (file)
--- a/fdroidserver/common.py
+++ b/fdroidserver/common.py
@@ -1700,7 +1700,8 @@ class KnownApks:
 
 def get_file_extension(filename):
     """get the normalized file extension, can be blank string but never None"""
-
+    if isinstance(filename, bytes):
+        filename = filename.decode('utf-8')
     return os.path.splitext(filename)[1].lower()[1:]
 
 
@@ -2333,15 +2334,17 @@ def get_per_app_repos():
 
 def is_repo_file(filename):
     '''Whether the file in a repo is a build product to be delivered to users'''
+    if isinstance(filename, str):
+        filename = filename.encode('utf-8', errors="surrogateescape")
     return os.path.isfile(filename) \
-        and not filename.endswith('.asc') \
-        and not filename.endswith('.sig') \
+        and not filename.endswith(b'.asc') \
+        and not filename.endswith(b'.sig') \
         and os.path.basename(filename) not in [
-            'index.jar',
-            'index_unsigned.jar',
-            'index.xml',
-            'index.html',
-            'index-v1.jar',
-            'index-v1.json',
-            'categories.txt',
+            b'index.jar',
+            b'index_unsigned.jar',
+            b'index.xml',
+            b'index.html',
+            b'index-v1.jar',
+            b'index-v1.json',
+            b'categories.txt',
         ]

diff --git a/fdroidserver/index.py b/fdroidserver/index.py
index c284ce50674824f6c02795fb95e727813544e0f3..b42c63d1e37513af8b4955c9f3391c9acf950472 100644 (file)
--- a/fdroidserver/index.py
+++ b/fdroidserver/index.py
@@ -436,14 +436,14 @@ def make_v0(apps, apks, repodir, repodict, requestsdict):
                 and common.config['make_current_version_link'] \
                 and repodir == 'repo':  # only create these
             namefield = common.config['current_version_name_source']
-            sanitized_name = re.sub('''[ '"&%?+=/]''', '', app.get(namefield))
-            apklinkname = sanitized_name + '.apk'
-            current_version_path = os.path.join(repodir, current_version_file)
+            sanitized_name = re.sub(b'''[ '"&%?+=/]''', b'', app.get(namefield).encode('utf-8'))
+            apklinkname = sanitized_name + b'.apk'
+            current_version_path = os.path.join(repodir, current_version_file).encode('utf-8', 'surrogateescape')
             if os.path.islink(apklinkname):
                 os.remove(apklinkname)
             os.symlink(current_version_path, apklinkname)
             # also symlink gpg signature, if it exists
-            for extension in ('.asc', '.sig'):
+            for extension in (b'.asc', b'.sig'):
                 sigfile_path = current_version_path + extension
                 if os.path.exists(sigfile_path):
                     siglinkname = apklinkname + extension

diff --git a/fdroidserver/update.py b/fdroidserver/update.py
index 6412029b249ba21abc409339ca5e76ed7fcb6425..0e42d68450fd4b6d056515bfdfa7967f53ac8f25 100644 (file)
--- a/fdroidserver/update.py
+++ b/fdroidserver/update.py
@@ -656,13 +656,15 @@ def scan_repo_files(apkcache, repodir, knownapks, use_date_from_file=False):
 
     cachechanged = False
     repo_files = []
+    repodir = repodir.encode('utf-8')
     for name in os.listdir(repodir):
         file_extension = common.get_file_extension(name)
         if file_extension == 'apk' or file_extension == 'obb':
             continue
         filename = os.path.join(repodir, name)
-        if filename.endswith('_src.tar.gz'):
-            logging.debug('skipping source tarball: ' + filename)
+        name_utf8 = name.decode('utf-8')
+        if filename.endswith(b'_src.tar.gz'):
+            logging.debug('skipping source tarball: ' + filename.decode('utf-8'))
             continue
         if not common.is_repo_file(filename):
             continue
@@ -683,34 +685,34 @@ def scan_repo_files(apkcache, repodir, knownapks, use_date_from_file=False):
                 else:
                     repo_file['added'] = datetime(*a[:6])
             if repo_file.get('hash') == shasum:
-                logging.debug("Reading " + name + " from cache")
+                logging.debug("Reading " + name_utf8 + " from cache")
                 usecache = True
             else:
                 logging.debug("Ignoring stale cache data for " + name)
 
         if not usecache:
-            logging.debug("Processing " + name)
+            logging.debug("Processing " + name_utf8)
             repo_file = collections.OrderedDict()
             # TODO rename apkname globally to something more generic
-            repo_file['name'] = name
-            repo_file['apkName'] = name
+            repo_file['name'] = name_utf8
+            repo_file['apkName'] = name_utf8
             repo_file['hash'] = shasum
             repo_file['hashType'] = 'sha256'
             repo_file['versionCode'] = 0
             repo_file['versionName'] = shasum
             # the static ID is the SHA256 unless it is set in the metadata
             repo_file['packageName'] = shasum
-            n = name.split('_')
+            n = name_utf8.split('_')
             if len(n) == 2:
                 packageName = n[0]
                 versionCode = n[1].split('.')[0]
-                if re.match(r'^-?[0-9]+$', versionCode) \
-                   and common.is_valid_package_name(name.split('_')[0]):
+                if re.match('^-?[0-9]+$', versionCode) \
+                   and common.is_valid_package_name(name_utf8.split('_')[0]):
                     repo_file['packageName'] = packageName
                     repo_file['versionCode'] = int(versionCode)
-            srcfilename = name + "_src.tar.gz"
+            srcfilename = name + b'_src.tar.gz'
             if os.path.exists(os.path.join(repodir, srcfilename)):
-                repo_file['srcname'] = srcfilename
+                repo_file['srcname'] = srcfilename.decode('utf-8')
             repo_file['size'] = stat.st_size
 
             apkcache[name] = repo_file