X-Git-Url: https://code.grnet.gr/git/pithos/blobdiff_plain/5635f9ef95acb792845c34d73d1b871feac72173..297513ba94f542505f7e58529cd4d2d47cd6f531:/pithos/api/util.py diff --git a/pithos/api/util.py b/pithos/api/util.py index 0c60b3a..296be84 100644 --- a/pithos/api/util.py +++ b/pithos/api/util.py @@ -35,76 +35,132 @@ from functools import wraps from time import time from traceback import format_exc from wsgiref.handlers import format_date_time +from binascii import hexlify, unhexlify +from datetime import datetime, tzinfo, timedelta from django.conf import settings from django.http import HttpResponse +from django.utils import simplejson as json from django.utils.http import http_date, parse_etags +from django.utils.encoding import smart_str -from pithos.api.compat import parse_http_date_safe -from pithos.api.faults import (Fault, NotModified, BadRequest, ItemNotFound, LengthRequired, - PreconditionFailed, ServiceUnavailable) -from pithos.backends import backend +from pithos.api.compat import parse_http_date_safe, parse_http_date +from pithos.api.faults import (Fault, NotModified, BadRequest, Unauthorized, Forbidden, ItemNotFound, + Conflict, LengthRequired, PreconditionFailed, RequestEntityTooLarge, + RangeNotSatisfiable, ServiceUnavailable) +from pithos.backends import connect_backend +from pithos.backends.base import NotAllowedError, QuotaError -import datetime import logging import re - +import hashlib +import uuid +import decimal logger = logging.getLogger(__name__) -def printable_meta_dict(d): +class UTC(tzinfo): + def utcoffset(self, dt): + return timedelta(0) + + def tzname(self, dt): + return 'UTC' + + def dst(self, dt): + return timedelta(0) + +def json_encode_decimal(obj): + if isinstance(obj, decimal.Decimal): + return str(obj) + raise TypeError(repr(obj) + " is not JSON serializable") + +def isoformat(d): + """Return an ISO8601 date string that includes a timezone.""" + + return d.replace(tzinfo=UTC()).isoformat() + +def rename_meta_key(d, old, new): + if old not in d: + return + d[new] = d[old] + del(d[old]) + +def printable_header_dict(d): """Format a meta dictionary for printing out json/xml. - Convert all keys to lower case and replace dashes to underscores. - Change 'modified' key from backend to 'last_modified' and format date. + Convert all keys to lower case and replace dashes with underscores. + Format 'last_modified' timestamp. """ - if 'modified' in d: - d['last_modified'] = datetime.datetime.fromtimestamp(int(d['modified'])).isoformat() - del(d['modified']) + + d['last_modified'] = isoformat(datetime.fromtimestamp(d['last_modified'])) return dict([(k.lower().replace('-', '_'), v) for k, v in d.iteritems()]) -def format_meta_key(k): - """Convert underscores to dashes and capitalize intra-dash strings""" +def format_header_key(k): + """Convert underscores to dashes and capitalize intra-dash strings.""" return '-'.join([x.capitalize() for x in k.replace('_', '-').split('-')]) -def get_meta_prefix(request, prefix): - """Get all prefix-* request headers in a dict. Reformat keys with format_meta_key()""" +def get_header_prefix(request, prefix): + """Get all prefix-* request headers in a dict. Reformat keys with format_header_key().""" + prefix = 'HTTP_' + prefix.upper().replace('-', '_') - return dict([(format_meta_key(k[5:]), v) for k, v in request.META.iteritems() if k.startswith(prefix)]) - -def get_account_meta(request): - """Get metadata from an account request""" - meta = get_meta_prefix(request, 'X-Account-Meta-') - return meta - -def put_account_meta(response, meta): - """Put metadata in an account response""" - response['X-Account-Container-Count'] = meta['count'] - response['X-Account-Bytes-Used'] = meta['bytes'] - if 'modified' in meta: - response['Last-Modified'] = http_date(int(meta['modified'])) + # TODO: Document or remove '~' replacing. + return dict([(format_header_key(k[5:]), v.replace('~', '')) for k, v in request.META.iteritems() if k.startswith(prefix) and len(k) > len(prefix)]) + +def get_account_headers(request): + meta = get_header_prefix(request, 'X-Account-Meta-') + groups = {} + for k, v in get_header_prefix(request, 'X-Account-Group-').iteritems(): + n = k[16:].lower() + if '-' in n or '_' in n: + raise BadRequest('Bad characters in group name') + groups[n] = v.replace(' ', '').split(',') + while '' in groups[n]: + groups[n].remove('') + return meta, groups + +def put_account_headers(response, meta, groups, policy): + if 'count' in meta: + response['X-Account-Container-Count'] = meta['count'] + if 'bytes' in meta: + response['X-Account-Bytes-Used'] = meta['bytes'] + response['Last-Modified'] = http_date(int(meta['modified'])) for k in [x for x in meta.keys() if x.startswith('X-Account-Meta-')]: - response[k.encode('utf-8')] = meta[k].encode('utf-8') - -def get_container_meta(request): - """Get metadata from a container request""" - meta = get_meta_prefix(request, 'X-Container-Meta-') - return meta - -def put_container_meta(response, meta): - """Put metadata in a container response""" - response['X-Container-Object-Count'] = meta['count'] - response['X-Container-Bytes-Used'] = meta['bytes'] - if 'modified' in meta: - response['Last-Modified'] = http_date(int(meta['modified'])) + response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) + if 'until_timestamp' in meta: + response['X-Account-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) + for k, v in groups.iteritems(): + k = smart_str(k, strings_only=True) + k = format_header_key('X-Account-Group-' + k) + v = smart_str(','.join(v), strings_only=True) + response[k] = v + for k, v in policy.iteritems(): + response[smart_str(format_header_key('X-Account-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) + +def get_container_headers(request): + meta = get_header_prefix(request, 'X-Container-Meta-') + policy = dict([(k[19:].lower(), v.replace(' ', '')) for k, v in get_header_prefix(request, 'X-Container-Policy-').iteritems()]) + return meta, policy + +def put_container_headers(request, response, meta, policy): + if 'count' in meta: + response['X-Container-Object-Count'] = meta['count'] + if 'bytes' in meta: + response['X-Container-Bytes-Used'] = meta['bytes'] + response['Last-Modified'] = http_date(int(meta['modified'])) for k in [x for x in meta.keys() if x.startswith('X-Container-Meta-')]: - response[k.encode('utf-8')] = meta[k].encode('utf-8') - response['X-Container-Object-Meta'] = [x[14:] for x in meta['object_meta'] if x.startswith('X-Object-Meta-')] - -def get_object_meta(request): - """Get metadata from an object request""" - meta = get_meta_prefix(request, 'X-Object-Meta-') + response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) + l = [smart_str(x, strings_only=True) for x in meta['object_meta'] if x.startswith('X-Object-Meta-')] + response['X-Container-Object-Meta'] = ','.join([x[14:] for x in l]) + response['X-Container-Block-Size'] = request.backend.block_size + response['X-Container-Block-Hash'] = request.backend.hash_algorithm + if 'until_timestamp' in meta: + response['X-Container-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) + for k, v in policy.iteritems(): + response[smart_str(format_header_key('X-Container-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) + +def get_object_headers(request): + meta = get_header_prefix(request, 'X-Object-Meta-') if request.META.get('CONTENT_TYPE'): meta['Content-Type'] = request.META['CONTENT_TYPE'] if request.META.get('HTTP_CONTENT_ENCODING'): @@ -113,22 +169,79 @@ def get_object_meta(request): meta['Content-Disposition'] = request.META['HTTP_CONTENT_DISPOSITION'] if request.META.get('HTTP_X_OBJECT_MANIFEST'): meta['X-Object-Manifest'] = request.META['HTTP_X_OBJECT_MANIFEST'] - return meta + return meta, get_sharing(request), get_public(request) -def put_object_meta(response, meta): - """Put metadata in an object response""" +def put_object_headers(response, meta, restricted=False): response['ETag'] = meta['hash'] response['Content-Length'] = meta['bytes'] response['Content-Type'] = meta.get('Content-Type', 'application/octet-stream') response['Last-Modified'] = http_date(int(meta['modified'])) - for k in [x for x in meta.keys() if x.startswith('X-Object-Meta-')]: - response[k.encode('utf-8')] = meta[k].encode('utf-8') - for k in ('Content-Encoding', 'Content-Disposition', 'X-Object-Manifest'): - if k in meta: - response[k] = meta[k] + if not restricted: + response['X-Object-Modified-By'] = smart_str(meta['modified_by'], strings_only=True) + response['X-Object-Version'] = meta['version'] + response['X-Object-Version-Timestamp'] = http_date(int(meta['version_timestamp'])) + for k in [x for x in meta.keys() if x.startswith('X-Object-Meta-')]: + response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) + for k in ('Content-Encoding', 'Content-Disposition', 'X-Object-Manifest', + 'X-Object-Sharing', 'X-Object-Shared-By', 'X-Object-Allowed-To', + 'X-Object-Public'): + if k in meta: + response[k] = smart_str(meta[k], strings_only=True) + else: + for k in ('Content-Encoding', 'Content-Disposition'): + if k in meta: + response[k] = meta[k] + +def update_manifest_meta(request, v_account, meta): + """Update metadata if the object has an X-Object-Manifest.""" + + if 'X-Object-Manifest' in meta: + hash = '' + bytes = 0 + try: + src_container, src_name = split_container_object_string('/' + meta['X-Object-Manifest']) + objects = request.backend.list_objects(request.user, v_account, + src_container, prefix=src_name, virtual=False) + for x in objects: + src_meta = request.backend.get_object_meta(request.user, + v_account, src_container, x[0], x[1]) + hash += src_meta['hash'] + bytes += src_meta['bytes'] + except: + # Ignore errors. + return + meta['bytes'] = bytes + md5 = hashlib.md5() + md5.update(hash) + meta['hash'] = md5.hexdigest().lower() + +def update_sharing_meta(request, permissions, v_account, v_container, v_object, meta): + if permissions is None: + return + allowed, perm_path, perms = permissions + if len(perms) == 0: + return + ret = [] + r = ','.join(perms.get('read', [])) + if r: + ret.append('read=' + r) + w = ','.join(perms.get('write', [])) + if w: + ret.append('write=' + w) + meta['X-Object-Sharing'] = '; '.join(ret) + if '/'.join((v_account, v_container, v_object)) != perm_path: + meta['X-Object-Shared-By'] = perm_path + if request.user != v_account: + meta['X-Object-Allowed-To'] = allowed + +def update_public_meta(public, meta): + if not public: + return + meta['X-Object-Public'] = public def validate_modification_preconditions(request, meta): - """Check that the modified timestamp conforms with the preconditions set""" + """Check that the modified timestamp conforms with the preconditions set.""" + if 'modified' not in meta: return # TODO: Always return? @@ -136,86 +249,101 @@ def validate_modification_preconditions(request, meta): if if_modified_since is not None: if_modified_since = parse_http_date_safe(if_modified_since) if if_modified_since is not None and int(meta['modified']) <= if_modified_since: - raise NotModified('Object has not been modified') + raise NotModified('Resource has not been modified') if_unmodified_since = request.META.get('HTTP_IF_UNMODIFIED_SINCE') if if_unmodified_since is not None: if_unmodified_since = parse_http_date_safe(if_unmodified_since) if if_unmodified_since is not None and int(meta['modified']) > if_unmodified_since: - raise PreconditionFailed('Object has been modified') + raise PreconditionFailed('Resource has been modified') def validate_matching_preconditions(request, meta): - """Check that the ETag conforms with the preconditions set""" - if 'hash' not in meta: - return # TODO: Always return? + """Check that the ETag conforms with the preconditions set.""" + + hash = meta.get('hash', None) if_match = request.META.get('HTTP_IF_MATCH') - if if_match is not None and if_match != '*': - if meta['hash'] not in [x.lower() for x in parse_etags(if_match)]: - raise PreconditionFailed('Object Etag does not match') + if if_match is not None: + if hash is None: + raise PreconditionFailed('Resource does not exist') + if if_match != '*' and hash not in [x.lower() for x in parse_etags(if_match)]: + raise PreconditionFailed('Resource ETag does not match') if_none_match = request.META.get('HTTP_IF_NONE_MATCH') if if_none_match is not None: - if if_none_match == '*' or meta['hash'] in [x.lower() for x in parse_etags(if_none_match)]: - raise NotModified('Object Etag matches') - -def copy_or_move_object(request, src_path, dest_path, move=False): - """Copy or move an object""" - if type(src_path) == str: - parts = src_path.split('/') - if len(parts) < 3 or parts[0] != '': - raise BadRequest('Invalid X-Copy-From or X-Move-From header') - src_container = parts[1] - src_name = '/'.join(parts[2:]) - elif type(src_path) == tuple and len(src_path) == 2: - src_container, src_name = src_path - if type(dest_path) == str: - parts = dest_path.split('/') - if len(parts) < 3 or parts[0] != '': - raise BadRequest('Invalid Destination header') - dest_container = parts[1] - dest_name = '/'.join(parts[2:]) - elif type(dest_path) == tuple and len(dest_path) == 2: - dest_container, dest_name = dest_path - - meta = get_object_meta(request) - # Keep previous values of 'Content-Type' (if a new one is absent) and 'hash'. - try: - src_meta = backend.get_object_meta(request.user, src_container, src_name) - except NameError: - raise ItemNotFound('Container or object does not exist') - if 'Content-Type' in meta and 'Content-Type' in src_meta: - del(src_meta['Content-Type']) - for k in ('Content-Type', 'hash'): - if k in src_meta: - meta[k] = src_meta[k] + # TODO: If this passes, must ignore If-Modified-Since header. + if hash is not None: + if if_none_match == '*' or hash in [x.lower() for x in parse_etags(if_none_match)]: + # TODO: Continue if an If-Modified-Since header is present. + if request.method in ('HEAD', 'GET'): + raise NotModified('Resource ETag matches') + raise PreconditionFailed('Resource exists or ETag matches') + +def split_container_object_string(s): + if not len(s) > 0 or s[0] != '/': + raise ValueError + s = s[1:] + pos = s.find('/') + if pos == -1: + raise ValueError + return s[:pos], s[(pos + 1):] + +def copy_or_move_object(request, src_account, src_container, src_name, dest_account, dest_container, dest_name, move=False): + """Copy or move an object.""" + meta, permissions, public = get_object_headers(request) + src_version = request.META.get('HTTP_X_SOURCE_VERSION') try: if move: - backend.move_object(request.user, src_container, src_name, dest_container, dest_name, meta, replace_meta=True) + version_id = request.backend.move_object(request.user, src_account, src_container, src_name, + dest_account, dest_container, dest_name, + meta, False, permissions) else: - backend.copy_object(request.user, src_container, src_name, dest_container, dest_name, meta, replace_meta=True) - except NameError: + version_id = request.backend.copy_object(request.user, src_account, src_container, src_name, + dest_account, dest_container, dest_name, + meta, False, permissions, src_version) + except NotAllowedError: + raise Forbidden('Not allowed') + except (NameError, IndexError): raise ItemNotFound('Container or object does not exist') + except ValueError: + raise BadRequest('Invalid sharing header') + except AttributeError, e: + raise Conflict('\n'.join(e.data) + '\n') + except QuotaError: + raise RequestEntityTooLarge('Quota exceeded') + if public is not None: + try: + request.backend.update_object_public(request.user, dest_account, dest_container, dest_name, public) + except NotAllowedError: + raise Forbidden('Not allowed') + except NameError: + raise ItemNotFound('Object does not exist') + return version_id + +def get_int_parameter(p): + if p is not None: + try: + p = int(p) + except ValueError: + return None + if p < 0: + return None + return p def get_content_length(request): - content_length = request.META.get('CONTENT_LENGTH') - if not content_length: - raise LengthRequired('Missing Content-Length header') - try: - content_length = int(content_length) - if content_length < 0: - raise ValueError - except ValueError: - raise BadRequest('Invalid Content-Length header') + content_length = get_int_parameter(request.META.get('CONTENT_LENGTH')) + if content_length is None: + raise LengthRequired('Missing or invalid Content-Length header') return content_length def get_range(request, size): - """Parse a Range header from the request + """Parse a Range header from the request. Either returns None, when the header is not existent or should be ignored, or a list of (offset, length) tuples - should be further checked. """ + ranges = request.META.get('HTTP_RANGE', '').replace(' ', '') if not ranges.startswith('bytes='): return None @@ -247,7 +375,7 @@ def get_range(request, size): return ret def get_content_range(request): - """Parse a Content-Range header from the request + """Parse a Content-Range header from the request. Either returns None, when the header is not existent or should be ignored, or an (offset, length, total) tuple - check as length, total may be None. @@ -275,39 +403,112 @@ def get_content_range(request): total = int(total) else: total = None - if (upto and offset > upto) or \ - (total and offset >= total) or \ - (total and upto and upto >= total): + if (upto is not None and offset > upto) or \ + (total is not None and offset >= total) or \ + (total is not None and upto is not None and upto >= total): return None - if not upto: + if upto is None: length = None else: length = upto - offset + 1 return (offset, length, total) +def get_sharing(request): + """Parse an X-Object-Sharing header from the request. + + Raises BadRequest on error. + """ + + permissions = request.META.get('HTTP_X_OBJECT_SHARING') + if permissions is None: + return None + + # TODO: Document or remove '~' replacing. + permissions = permissions.replace('~', '') + + ret = {} + permissions = permissions.replace(' ', '') + if permissions == '': + return ret + for perm in (x for x in permissions.split(';')): + if perm.startswith('read='): + ret['read'] = list(set([v.replace(' ','').lower() for v in perm[5:].split(',')])) + if '' in ret['read']: + ret['read'].remove('') + if '*' in ret['read']: + ret['read'] = ['*'] + if len(ret['read']) == 0: + raise BadRequest('Bad X-Object-Sharing header value') + elif perm.startswith('write='): + ret['write'] = list(set([v.replace(' ','').lower() for v in perm[6:].split(',')])) + if '' in ret['write']: + ret['write'].remove('') + if '*' in ret['write']: + ret['write'] = ['*'] + if len(ret['write']) == 0: + raise BadRequest('Bad X-Object-Sharing header value') + else: + raise BadRequest('Bad X-Object-Sharing header value') + + # Keep duplicates only in write list. + dups = [x for x in ret.get('read', []) if x in ret.get('write', []) and x != '*'] + if dups: + for x in dups: + ret['read'].remove(x) + if len(ret['read']) == 0: + del(ret['read']) + + return ret + +def get_public(request): + """Parse an X-Object-Public header from the request. + + Raises BadRequest on error. + """ + + public = request.META.get('HTTP_X_OBJECT_PUBLIC') + if public is None: + return None + + public = public.replace(' ', '').lower() + if public == 'true': + return True + elif public == 'false' or public == '': + return False + raise BadRequest('Bad X-Object-Public header value') + def raw_input_socket(request): - """Return the socket for reading the rest of the request""" + """Return the socket for reading the rest of the request.""" + server_software = request.META.get('SERVER_SOFTWARE') - if not server_software: - if 'wsgi.input' in request.environ: - return request.environ['wsgi.input'] - raise ServiceUnavailable('Unknown server software') - if server_software.startswith('WSGIServer'): - return request.environ['wsgi.input'] - elif server_software.startswith('mod_python'): + if server_software and server_software.startswith('mod_python'): return request._req + if 'wsgi.input' in request.environ: + return request.environ['wsgi.input'] raise ServiceUnavailable('Unknown server software') -MAX_UPLOAD_SIZE = 10 * (1024 * 1024) # 10MB +MAX_UPLOAD_SIZE = 5 * (1024 * 1024 * 1024) # 5GB -def socket_read_iterator(sock, length=0, blocksize=4096): - """Return a maximum of blocksize data read from the socket in each iteration +def socket_read_iterator(request, length=0, blocksize=4096): + """Return a maximum of blocksize data read from the socket in each iteration. Read up to 'length'. If 'length' is negative, will attempt a chunked read. The maximum ammount of data read is controlled by MAX_UPLOAD_SIZE. """ + + sock = raw_input_socket(request) if length < 0: # Chunked transfers + # Small version (server does the dechunking). + if request.environ.get('mod_wsgi.input_chunked', None) or request.META['SERVER_SOFTWARE'].startswith('gunicorn'): + while length < MAX_UPLOAD_SIZE: + data = sock.read(blocksize) + if data == '': + return + yield data + raise BadRequest('Maximum size is reached') + + # Long version (do the dechunking). data = '' while length < MAX_UPLOAD_SIZE: # Get chunk size. @@ -334,39 +535,42 @@ def socket_read_iterator(sock, length=0, blocksize=4096): while chunk_length > 0: chunk = sock.read(min(chunk_length, blocksize)) chunk_length -= len(chunk) - length += len(chunk) + if length > 0: + length += len(chunk) data += chunk if len(data) >= blocksize: ret = data[:blocksize] data = data[blocksize:] yield ret sock.read(2) # CRLF - # TODO: Raise something to note that maximum size is reached. + raise BadRequest('Maximum size is reached') else: if length > MAX_UPLOAD_SIZE: - # TODO: Raise something to note that maximum size is reached. - pass + raise BadRequest('Maximum size is reached') while length > 0: data = sock.read(min(length, blocksize)) + if not data: + raise BadRequest() length -= len(data) yield data class ObjectWrapper(object): - """Return the object's data block-per-block in each iteration + """Return the object's data block-per-block in each iteration. Read from the object using the offset and length provided in each entry of the range list. """ - def __init__(self, v_account, v_container, v_object, ranges, size, hashmap, boundary): - self.v_account = v_account - self.v_container = v_container - self.v_object = v_object + def __init__(self, backend, ranges, sizes, hashmaps, boundary): + self.backend = backend self.ranges = ranges - self.size = size - self.hashmap = hashmap + self.sizes = sizes + self.hashmaps = hashmaps self.boundary = boundary + self.size = sum(self.sizes) - self.block_index = -1 + self.file_index = 0 + self.block_index = 0 + self.block_hash = -1 self.block = '' self.range_index = -1 @@ -377,17 +581,25 @@ class ObjectWrapper(object): def part_iterator(self): if self.length > 0: - # Get the block for the current offset. - bi = int(self.offset / backend.block_size) - if self.block_index != bi: + # Get the file for the current offset. + file_size = self.sizes[self.file_index] + while self.offset >= file_size: + self.offset -= file_size + self.file_index += 1 + file_size = self.sizes[self.file_index] + + # Get the block for the current position. + self.block_index = int(self.offset / self.backend.block_size) + if self.block_hash != self.hashmaps[self.file_index][self.block_index]: + self.block_hash = self.hashmaps[self.file_index][self.block_index] try: - self.block = backend.get_block(self.hashmap[bi]) + self.block = self.backend.get_block(self.block_hash) except NameError: raise ItemNotFound('Block does not exist') - self.block_index = bi + # Get the data from the block. - bo = self.offset % backend.block_size - bl = min(self.length, backend.block_size - bo) + bo = self.offset % self.backend.block_size + bl = min(self.length, len(self.block) - bo) data = self.block[bo:bo + bl] self.offset += bl self.length -= bl @@ -410,6 +622,7 @@ class ObjectWrapper(object): if self.range_index < len(self.ranges): # Part header. self.offset, self.length = self.ranges[self.range_index] + self.file_index = 0 if self.range_index > 0: out.append('') out.append('--' + self.boundary) @@ -425,6 +638,87 @@ class ObjectWrapper(object): out.append('') return '\r\n'.join(out) +def object_data_response(request, sizes, hashmaps, meta, public=False): + """Get the HttpResponse object for replying with the object's data.""" + + # Range handling. + size = sum(sizes) + ranges = get_range(request, size) + if ranges is None: + ranges = [(0, size)] + ret = 200 + else: + check = [True for offset, length in ranges if + length <= 0 or length > size or + offset < 0 or offset >= size or + offset + length > size] + if len(check) > 0: + raise RangeNotSatisfiable('Requested range exceeds object limits') + ret = 206 + if_range = request.META.get('HTTP_IF_RANGE') + if if_range: + try: + # Modification time has passed instead. + last_modified = parse_http_date(if_range) + if last_modified != meta['modified']: + ranges = [(0, size)] + ret = 200 + except ValueError: + if if_range != meta['hash']: + ranges = [(0, size)] + ret = 200 + + if ret == 206 and len(ranges) > 1: + boundary = uuid.uuid4().hex + else: + boundary = '' + wrapper = ObjectWrapper(request.backend, ranges, sizes, hashmaps, boundary) + response = HttpResponse(wrapper, status=ret) + put_object_headers(response, meta, public) + if ret == 206: + if len(ranges) == 1: + offset, length = ranges[0] + response['Content-Length'] = length # Update with the correct length. + response['Content-Range'] = 'bytes %d-%d/%d' % (offset, offset + length - 1, size) + else: + del(response['Content-Length']) + response['Content-Type'] = 'multipart/byteranges; boundary=%s' % (boundary,) + return response + +def put_object_block(request, hashmap, data, offset): + """Put one block of data at the given offset.""" + + bi = int(offset / request.backend.block_size) + bo = offset % request.backend.block_size + bl = min(len(data), request.backend.block_size - bo) + if bi < len(hashmap): + hashmap[bi] = request.backend.update_block(hashmap[bi], data[:bl], bo) + else: + hashmap.append(request.backend.put_block(('\x00' * bo) + data[:bl])) + return bl # Return ammount of data written. + +def hashmap_hash(request, hashmap): + """Produce the root hash, treating the hashmap as a Merkle-like tree.""" + + def subhash(d): + h = hashlib.new(request.backend.hash_algorithm) + h.update(d) + return h.digest() + + if len(hashmap) == 0: + return hexlify(subhash('')) + if len(hashmap) == 1: + return hashmap[0] + + s = 2 + while s < len(hashmap): + s = s * 2 + h = [unhexlify(x) for x in hashmap] + h += [('\x00' * len(h[0]))] * (s - len(hashmap)) + while len(h) > 1: + h = [subhash(h[x] + h[x + 1]) for x in range(0, len(h), 2)] + return hexlify(h[0]) + def update_response_headers(request, response): if request.serialization == 'xml': response['Content-Type'] = 'application/xml; charset=UTF-8' @@ -432,14 +726,17 @@ def update_response_headers(request, response): response['Content-Type'] = 'application/json; charset=UTF-8' elif not response['Content-Type']: response['Content-Type'] = 'text/plain; charset=UTF-8' - + + if not response.has_header('Content-Length') and not (response.has_header('Content-Type') and response['Content-Type'].startswith('multipart/byteranges')): + response['Content-Length'] = len(response.content) + if settings.TEST: response['Date'] = format_date_time(time()) def render_fault(request, fault): if settings.DEBUG or settings.TEST: fault.details = format_exc(fault) - + request.serialization = 'text' data = '\n'.join((fault.message, fault.details)) + '\n' response = HttpResponse(data, status=fault.code) @@ -447,10 +744,11 @@ def render_fault(request, fault): return response def request_serialization(request, format_allowed=False): - """Return the serialization format requested + """Return the serialization format requested. Valid formats are 'text' and 'json', 'xml' if 'format_allowed' is True. """ + if not format_allowed: return 'text' @@ -469,15 +767,18 @@ def request_serialization(request, format_allowed=False): return 'text' -def api_method(http_method=None, format_allowed=False): - """Decorator function for views that implement an API method""" +def api_method(http_method=None, format_allowed=False, user_required=True): + """Decorator function for views that implement an API method.""" + def decorator(func): @wraps(func) def wrapper(request, *args, **kwargs): try: if http_method and request.method != http_method: raise BadRequest('Method not allowed.') - + if user_required and getattr(request, 'user', None) is None: + raise Unauthorized('Access denied') + # The args variable may contain up to (account, container, object). if len(args) > 1 and len(args[1]) > 256: raise BadRequest('Container name too large.') @@ -486,9 +787,8 @@ def api_method(http_method=None, format_allowed=False): # Fill in custom request variables. request.serialization = request_serialization(request, format_allowed) - # TODO: Authenticate. - request.user = "test" - + request.backend = connect_backend() + response = func(request, *args, **kwargs) update_response_headers(request, response) return response @@ -498,5 +798,8 @@ def api_method(http_method=None, format_allowed=False): logger.exception('Unexpected error: %s' % e) fault = ServiceUnavailable('Unexpected error') return render_fault(request, fault) + finally: + if getattr(request, 'backend', None) is not None: + request.backend.wrapper.conn.close() return wrapper return decorator