Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
branch: master
Commits on Nov 19, 2013
  1. Merge pull request #58 from vills/fix_quoting

    authored
    fix requests without object_name
  2. fix requests without object_name

    Виль Суркин authored
  3. Merge pull request #57 from vills/fix_quoting

    authored
    prevent double-quoting object name
Commits on Nov 15, 2013
  1. preven double-quoting object name

    Vil Surkin authored
Commits on Sep 13, 2013
  1. Merge pull request #56 from bloodeagle40234/fix-signature

    authored
    fix signature generation
Commits on Sep 11, 2013
  1. fix signature generation

    Kota Tsuyuzaki authored
    In canonical_string, query string must
    be lexicographically sorted by query name.
    This patch enables to work it correctly,
    and adds a unit test for checking it.
Commits on Jul 30, 2013
  1. Merge pull request #55 from bloodeagle40234/add_entity_too_large_error

    authored
    Add a error of EntityTooLarge
Commits on Jul 29, 2013
  1. Add a error of EntityTooLarge

    Kota Tsuyuzaki authored
    This fix makes to return 'EntityTooLarge' response,
    when the object which user try to upload exceeds the maximum
    allowed object size of swift.
    
    Signed-off-by: Kota Tsuyuzaki <tsuyuzaki.kota@lab.ntt.co.jp>
Commits on Jun 28, 2013
  1. Merge pull request #54 from zaitcev/master

    authored
    Undo part of patch by Willys
Commits on Jun 25, 2013
  1. @zaitcev

    Undo part of patch by Willys

    zaitcev authored
    A regression in tests was introduced by commit
    1216309, intended to
    fix issue #49.
    
    Willys' code actually seems to make sense, but I chose
    to undo it because
     - It appears to be extraneous to the issue #49
     - It does not fix the tests or remove the now-unused constant
    
    Someone should re-fix it properly. But for now I prefer tests
    continue to pass.
Commits on Jun 19, 2013
  1. Merge pull request #53 from vitoordaz/master

    authored
    canonical_string refactoring + fix typing in middleware comment
Commits on Jun 18, 2013
  1. @vitoordaz
Commits on Jun 3, 2013
  1. @wyllys66

    Issue #49 - fixed Date and ExpireDate checking logic

    wyllys66 authored
    Cleaned up PEP8 and Flake8 formatting errors.
Commits on May 1, 2013
  1. @wyllys66
Commits on Apr 16, 2013
  1. @morucci

    Take care of 403 Forbidden HTTP error

    morucci authored
    Return AccessDenied error to client
    instead of default InvalidURI when subsequents
    request in pipeline return 403 error.
Commits on Apr 6, 2013
  1. @zaitcev

    Content-Range uses space after bytes

    zaitcev authored
    Our test expected "Content-Range: bytes=0-N", but this is incorrect.
    The RFC-2616 14.16 clearly says:
      bytes-unit SP byte-range-resp-spec "/" ( instance-length | "*" )
    
    This is different from Ranges where bytes=0-N is specified.
    The wonders of Web standards!
    
    If a version of swob sent the equal sign previously, it was wrong,
    but now corrected.
Commits on Mar 15, 2013
  1. Merge pull request #37 from bloodeagle40234/raw-path-signature

    authored
    fix signature bug to use RAW_PATH_INFO
Commits on Feb 27, 2013
  1. fix signature bug to use RAW_PATH_INFO

    Kota Tsuyuzaki authored
    This fixes signature creation to use RAW_PATH_INFO.
    
    Swift3 could not create correct signature in case of
    using escaped character(e.g. %2F, %2D) in PATH_INFO,
    because env['PATH_INFO'] was decoded(unescaped) by
    eventlet.wsgi before a request arrived at swift3.
    It caused signature mismatch and authentication failure.
    
    This change enables swift3 to create signature from
    RAW_PATH_INFO and fixes that bug.
    
    Note: This patch works well only in later version than
          eventlet 0.9.17, because older version does not
          have RAW_PATH_INFO variable.
          When using older version, swift3 works in the same
          way as ever(use req.path of swob).
    
    Signed-off-by: Kota Tsuyuzaki <tsuyuzaki.kota@lab.ntt.co.jp>
Commits on Feb 7, 2013
  1. Update README about keystone usage

    authored
    Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Commits on Feb 1, 2013
  1. fix a bug for multi object deletion

    Kota Tsuyuzaki authored
    This fix a bug for deleting multiple objects.
    This change enables swift3 to receive request body
    correctly in order to prevent that swift3 does not
    return response for multi object deletion request.
    
    Signed-off-by: Kota Tsuyuzaki <tsuyuzaki.kota@lab.ntt.co.jp>
Commits on Dec 21, 2012
  1. @kazum

    Add basic support for multi object deletion

    kazum authored
    This adds basic support for deleting multiple objects from a bucket
    using a single POST request.
    
    TODOs:
     - support for deleting the specific version of the object.
     - return NoSuchBucked error when the specified bucket does not exist.
     - return MissingRequestBodyError error when the request body is
       empty.
     - return MalformedXML error when the provided XML is not well-formed.
    
    Signed-off-by: MORITA Kazutaka <morita.kazutaka@lab.ntt.co.jp>
Commits on Dec 10, 2012
  1. @creiht

    Updated HEAD and ACL calls to objects to work correctly

    creiht authored committed
Commits on Dec 9, 2012
  1. Swift3 1.7

    authored
    Added a versin tag for distribution packages.
    
    Likely Swift3 1.7 works with Swift 1.7.0 and the previous versions
    however not with the following verions (1.7.1 and so on).
    
    Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Commits on Oct 29, 2012
  1. @vitoordaz
  2. @vitoordaz

    bucket location stub

    vitoordaz authored
  3. @vitoordaz
Commits on Oct 28, 2012
  1. @vitoordaz
  2. @vitoordaz

    fix bucket creation. REQUEST_METHOD PUT not translated to POST if req…

    vitoordaz authored committed
    …uest has X_AMZ_ACL header
Commits on Oct 22, 2012
  1. @kvdveer

    Stub out ?versioning command.

    kvdveer authored
    This is necessary for CyberDuck, which issues the versioning command and errors
    out if that doesn't work. This implementation simply returns that versioning
    hasn't been configured.
Commits on Oct 17, 2012
  1. @vitoordaz

    webob => swob

    vitoordaz authored
Commits on Oct 15, 2012
  1. @fifieldt

    Fixes PEP8 violations

    fifieldt authored committed
    This commit changes whitespace, comma placements and line lengths to
    fix PEP8 violations in middleware.py and test_swift3.py
  2. @fifieldt

    Fixes object listing with special key names

    fifieldt authored committed
    swift (which prevents storing these keys on its own API) will store
    special characters in percent-encoded form, meaning that with the S3
    API, when keys were listed, they appeared as percent-encoded, rather
    than the original name.
    
    This change uses unquote to resolve the percent-encodings to the
    original name when listing keys in the bucket.
    
    These test cases now pass:
    tests.functional.test_s3.test_bucket_create_special_key_names
  3. @fifieldt

    Adds basic set XML ACL support

    fifieldt authored committed
    Extends the swift_acl_translate method for basic XML setting
    
    Test case now passes:
    functional.test_s3.test_bucket_acl_xml_fullcontrol
    
    Adds URIs to get_acl method, so test case now passes:
    functional.test_s3.test_bucket_acl_canned
  4. @fifieldt

    Add a method to validate bucket names

    fifieldt authored committed
    This change adds a method validate_bucket_name which can be used
    to determine if the bucket names meet S3 standards or not.
    
    It is currently unused, but in a future commit it will be called
    from BucketController's PUT method.
  5. @fifieldt

    WIP - acls and test cases

    fifieldt authored committed
    This patch is a functional work in progress. It
    * introduces get_acls based on real acl values in swift
    * implements a 'private' set acl
    * adds err_responses for MissingContentLength, BadDigest
    * adds a POST handler for BucketController (that only returns 501, for now)
    
    the get_acl function in particular isn't pretty, so this is a work in
    progress. Committing as it is passing 13 more testcases than previous
    attempts.
Something went wrong with that request. Please try again.