Merge pull request #248 from harshavardhana/pr_out_add_simpler_code_f…

…or_is_valid_bucket_name_is_valid_url_use_self_endpoint_url_not_self_url

Add simpler code for is_valid_bucket_name, is_valid_url, use self._endpoint_url not self._url

minio/_compat.py → minio/compat.py

@@ -16,26 +16,19 @@
 import sys
 
 try:
-    from urllib.parse import urlparse as compat_urllib_parse
+    from urllib.parse import urlsplit
 except ImportError:  # python 2
-    from urlparse import urlparse as compat_urllib_parse
+    from urlparse import urlsplit
 
 try:
-    from urllib.request import pathname2url as compat_pathname2url
+    from urllib.request import pathname2url as urlencode
 except ImportError:  # python 2
-    from urllib import pathname2url as compat_pathname2url
+    from urllib import pathname2url as urlencode
 
 try:
-    from urllib.request import url2pathname as compat_url2pathname
+    from urllib.request import url2pathname as urldecode
 except ImportError:  # python 2
-    from urllib import url2pathname as compat_url2pathname
-
-
-def urlencode(text):
-    if sys.version_info < (3, 0):
-        return compat_url2pathname(text.encode('utf-8'))
-    else:
-        return compat_url2pathname(text)
+    from urllib import url2pathname as urldecode
 
 
 strtype = None

minio/error.py

@@ -13,11 +13,33 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+class InvalidURLError(Exception):
+    def __init__(self, message, **kwargs):
+        self.message = message
+        super(InvalidURLError, self).__init__(**kwargs)
+
+    def __str__(self):
+        string_format = 'InvalidURLError: message: {0}'
+        return string_format.format(self.message)
+
 class InvalidBucketError(Exception):
-    pass
+    def __init__(self, message, **kwargs):
+        self.message = message
+        super(InvalidBucketError, self).__init__(**kwargs)
+
+    def __str__(self):
+        string_format = 'InvalidBucketError: message: {0}'
+        return string_format.format(self.message)
 
 class InvalidArgumentError(Exception):
-    pass
+    def __init__(self, message, **kwargs):
+        self.message = message
+        super(InvalidArgumentError, self).__init__(**kwargs)
+
+    def __str__(self):
+        string_format = 'InvalidArgumentError: message: {0}'
+        return string_format.format(self.message)
+
 
 class ResponseError(Exception):
     def __init__(self, code, message, request_id, host_id, resource, xml=None,

minio/generators.py

@@ -22,7 +22,7 @@ class ListObjectsIterator(object):
     def __init__(self, client, url, bucket, prefix,
                  recursive, access_key, secret_key):
         self._http = client
-        self._url = url
+        self._endpoint_url = url
         self._bucket = bucket
         self._prefix = prefix
         self._recursive = recursive
@@ -67,7 +67,7 @@ def _fetch(self):
         if self._marker is not None:
             query['marker'] = self._marker
 
-        url = get_target_url(self._url, bucket=self._bucket, query=query)
+        url = get_target_url(self._endpoint_url, bucket=self._bucket, query=query)
 
         method = 'GET'
         headers = {}
@@ -89,7 +89,7 @@ def __init__(self, client, url, bucket, key=None,
                  access_key=None, secret_key=None):
         # from user
         self._http = client
-        self._url = url
+        self._endpoint_url = url
         self._bucket = bucket
         self._key = key
         self._access_key = access_key
@@ -145,7 +145,7 @@ def _fetch(self):
         if self._upload_id_marker is not None:
             query['upload-id-marker'] = self._upload_id_marker
 
-        url = get_target_url(self._url, bucket=self._bucket, query=query)
+        url = get_target_url(self._endpoint_url, bucket=self._bucket, query=query)
 
         method = 'GET'
         headers = {}
@@ -167,7 +167,7 @@ def __init__(self, client, url, bucket, key, upload_id,
                  access_key=None, secret_key=None):
         # from user
         self._http = client
-        self._url = url
+        self._endpoint_url = url
         self._bucket = bucket
         self._key = key
         self._upload_id = upload_id
@@ -218,8 +218,8 @@ def _fetch(self):
         if self._part_marker is not None:
             query['part-number-marker'] = self._part_marker
 
-        url = get_target_url(self._url, bucket=self._bucket, key=self._key,
-                             query=query)
+        url = get_target_url(self._endpoint_url, bucket=self._bucket,
+                             key=self._key, query=query)
 
         method = 'GET'
         headers = {}

minio/helpers.py

@@ -17,13 +17,13 @@
 Helper functions
 """
 
-import cgi
 import collections
 import binascii
 import hashlib
 import re
 
-from ._compat import strtype, compat_pathname2url
+from .compat import urlsplit, strtype, urlencode
+from .error import InvalidBucketError, InvalidURLError
 
 def get_region(hostname):
     """
@@ -70,7 +70,7 @@ def get_target_url(url, bucket=None, key=None, query=None):
             if ordered_query[component_key] is not None:
                 single_component.append('=')
                 single_component.append(
-                    compat_pathname2url(cgi.escape(str(ordered_query[component_key]))).replace('/', '%2F'))
+                    urlencode(str(ordered_query[component_key])).replace('/', '%2F'))
             query_components.append(''.join(single_component))
 
         query_string = '&'.join(query_components)
@@ -80,38 +80,55 @@ def get_target_url(url, bucket=None, key=None, query=None):
 
     return ''.join(url_components)
 
-def is_valid_url(url):
+def is_valid_url(endpoint_url):
     """
-    validate a given url
+    Verify the endpoint_url is valid.
+    :type endpoint_url: string
+    :param endpoint_url: An endpoint_url.  Must have at least a scheme
+        and a hostname.
+    :return: True if the endpoint url is valid. False otherwise.
     """
-    if not isinstance(url, strtype):
+    if not isinstance(endpoint_url, strtype):
         raise TypeError('url')
 
-    regex = re.compile(
-        r'^(?:http)s?://' # http:// or https://
-        r'(?:(?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+(?:[A-Z]{2,6}\.?|[A-Z0-9-]{2,}\.?)|' # domain...
-        r'localhost|' # localhost...
-        r'\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}|' # ...or ipv4
-        r'\[?[A-F0-9]*:[A-F0-9:]+\]?)' # ...or ipv6
-        r'(?::\d+)?' # optional port
-        r'(?:/?|[/?]\S+)$', re.IGNORECASE)
-
-    if not regex.match(url):
-        raise ValueError('url')
-
-def is_valid_bucket_name(bucketname):
-    """
-    validate a given bucketname
-    """
-    is_non_empty_string(bucketname)
-    if len(bucketname) < 3 or len(bucketname) > 63:
-        raise ValueError('bucket')
-    if '/' in bucketname:
-        raise ValueError('bucket')
-    if not re.match("^[a-z0-9]+[a-z0-9\\-]*[a-z0-9]+$", bucketname):
-        raise ValueError('bucket')
-    if re.match("/[0-9]+\\.[0-9]+\\.[0-9]+\\.[0-9]+/", bucketname):
-        raise ValueError('bucket')
+    parts = urlsplit(endpoint_url)
+    hostname = parts.hostname
+    if hostname is None:
+        raise InvalidURLError('url')
+    if len(hostname) > 255:
+        raise InvalidURLError('url')
+    if hostname[-1] == ".":
+        hostname = hostname[:-1]
+    allowed = re.compile(
+        "^((?!-)[A-Z\d-]{1,63}(?<!-)\.)*((?!-)[A-Z\d-]{1,63}(?<!-))$",
+        re.IGNORECASE)
+    if not allowed.match(hostname):
+        raise InvalidURLError('url')
+
+def is_valid_bucket_name(bucket_name):
+    """
+    Check to see if the ``bucket_name`` complies with the
+    restricted DNS naming conventions necessary to allow
+    access via virtual-hosting style.
+
+    Even though "." characters are perfectly valid in this DNS
+    naming scheme, we are going to punt on any name containing a
+    "." character because these will cause SSL cert validation
+    problems if we try to use virtual-hosting style addressing.
+    """
+    validbucket = re.compile('[a-z0-9][a-z0-9\-]*[a-z0-9]')
+    if '.' in bucket_name:
+        raise InvalidBucketError('bucket')
+    n = len(bucket_name)
+    if n < 3 or n > 63:
+        # Wrong length
+        raise InvalidBucketError('bucket')
+    if n == 1:
+        if not bucket_name.isalnum():
+            raise InvalidBucketError('bucket')
+    match = validbucket.match(bucket_name)
+    if match is None or match.end() != len(bucket_name):
+        raise InvalidBucketError('bucket')
 
 def is_non_empty_string(input_string):
     """
@@ -127,7 +144,7 @@ def encode_object_key(key):
     url encode object key
     """
     is_non_empty_string(key)
-    return compat_pathname2url(key)
+    return urlencode(key)
 
 def get_sha256(content):
     """
@@ -139,7 +156,7 @@ def get_sha256(content):
     hasher.update(content)
     return hasher.digest()
 
-def get_md5(content):
+def get_md5(cntoent):
     """
     calculate md5 for given content
     """