Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Commits on Mar 5, 2015
  1. Matt Domsch

    Merge pull request #488 from glogiotatidis/master

    mdomsch authored
    If response['headers'] is empty, then don't update dict.
  2. Giorgos Logiotatidis

    If response['headers'] is empty, then don't update dict.

    glogiotatidis authored
    Ceph sometime returns
    
    {'status': 200, 'headers': {}, 'reason': '', 'data': ''}
    
    which chokes s3cmd because headers is an empty dictionary.
Commits on Mar 2, 2015
  1. Matt Domsch

    Sign x-emc-* headers (#486)

    mdomsch authored
    This allows s3cmd to work with EMC ECS S3-liek storage.
Commits on Feb 18, 2015
  1. Matt Domsch

    Merge pull request #468 from ppe/support-display_name-in-acl-revoke

    mdomsch authored
    Add support for display names in --acl-revoke (#223)
  2. Matt Domsch

    Merge pull request #480 from fviard/improve-errors-501-504

    mdomsch authored
    Improved handling for errors 501 and 504:
  3. Matt Domsch

    s3cmd help: --limit-rate works for uploads too

    mdomsch authored
    Not just for downloads (as help text indicated).  It works for uploads
    too.
  4. Matt Domsch

    Merge pull request #459 from thesues/throttle

    mdomsch authored
    add new option --limit-rate to throttle download/upload speed
  5. Matt Domsch

    Merge pull request #478 from petrus-jvrensburg/patch-1

    mdomsch authored
    Update README.md
  6. Matt Domsch

    Merge pull request #483 from mdomsch/bug/broken-pipe

    mdomsch authored
    Bug/broken pipe
Commits on Feb 14, 2015
  1. Matt Domsch

    accept 'us-east-1' as valid Config.bucket_location in [mb] command

    mdomsch authored
    The [mb] (create bucket) command would fail with an S3 error
    
    <Error>
      <Code>InvalidLocationConstraint</Code>
      <Message>The specified location-constraint is not valid</Message>
    
    when  Config.bucket_location = 'us-east-1', though that's accepted for
    all other commands.
    
    This patch accepts us-east-1 as valid for [mb] command as well as 'US'.
  2. Florent Viard

    Improved handling for errors 501 and 504:

    fviard authored
    Error 501: It means that a feature is not implemented in the server, so no need for useless retries.
    Error 504: Gateway timeout situation, exit code should indicate a server error.
Commits on Feb 11, 2015
  1. Matt Domsch
  2. Matt Domsch
  3. Matt Domsch

    Revert "signurl should honor Config.use_https"

    mdomsch authored
    This reverts commit 8788fe5.
    
    The above patch overloads the use of the use_https meaning, which
    isn't desirable.  S3's wildcard certs don't work well for https.  So
    if a caller cares that it's https and not http, they're welcome to
    make that change themselves.
  4. Matt Domsch
Commits on Feb 10, 2015
  1. Petrus Janse van Rensburg

    Update README.md

    petrus-jvrensburg authored
    Mention `sync` command.
Commits on Feb 9, 2015
  1. Matt Domsch

    fix signurl test

    mdomsch authored
    Now that it accepts https and http both, we need the test suite to
    recognize either is valid.
  2. Matt Domsch
  3. Matt Domsch

    signurl should honor Config.use_https

    mdomsch authored
    With Config.use_https = True, the resulting URL from 'signurl' should
    be https, not http.  With = False, it should be 'http'.
    
    SourceForge bug #163. https://sourceforge.net/p/s3tools/bugs/163/
Commits on Feb 8, 2015
  1. Matt Domsch

    Update version to 1.5.2

    mdomsch authored
  2. Matt Domsch

    remove ez_setup

    mdomsch authored
    ez_setup is annoying.  pip install s3cmd always complains that
    setuptools isn't new enough, and you've got to update that.  In
    Fedora and all sane build systems, the RPM package build can't use
    ez_setup as it downloads bits, which isn't allowed, so I had removed
    ez_setup frm there already.
    
    We aren't using any new features from setuptools that the base repo
    packages don't already provide.  pip already requires setuptools, so
    installing via pip we don't need to add anything.  We added the RPM
    BuildRequires: python-setuptools a few commits ago, so building there
    is fine.
Commits on Feb 7, 2015
  1. Matt Domsch

    Handle unvalidated SSL certificate

    mdomsch authored
    SSLSocket.getpeercert() can return {} meaning the library did not
    validate the SSL certificate.  We can't do the match_hostname() tests
    against this, so don't bother trying.
    
    This addresses #472 which was failing on Ubuntu 14.04 because it does
    not have a new-enough SSL library to validate the certificate.
Commits on Feb 6, 2015
  1. Matt Domsch

    make bundled s3cmd.spec.in closer to Fedora spec

    mdomsch authored
    Fedora/EPEL still drops ez_setup stuff as it isn't needed (we have a good
    enough python-setuptools in Fedora/EPEL already).  But I want to keep
    ez_setup in place for the pip installs.
  2. Matt Domsch

    setup.py: add install_requires python-magic

    mdomsch authored
    After this, pip installs will pick up python-magic automatically.
    This eliminates the regular warning, and adds functionality.
Commits on Feb 4, 2015
  1. Matt Domsch

    Update to version 1.5.1.2

    mdomsch authored
  2. Matt Domsch
  3. Matt Domsch

    Update version to 1.5.1.1

    mdomsch authored
  4. Matt Domsch

    include ez_setup.py in MANIFEST.in

    mdomsch authored
    fixes pip install s3cmd from PyPi
  5. Matt Domsch
  6. Matt Domsch
  7. Matt Domsch

    update NEWS for 1.5.1

    mdomsch authored
  8. Matt Domsch

    remove useless debug() and print() message

    mdomsch authored
    This may have been helpful during development, but it isn't any longer.
  9. Matt Domsch

    Fix indentation causing failures

    mdomsch authored
    Commit 41787ac accidentally over-indented its change.  This was
    leading to test suite failures, valid failures.  Specifically, the
    test suite would fail when removing buckets that have objects in them
    (the first test of the suite).
    
    Reading that patch, it's obvious that the "if error_node is not None"
    test got indented one level too deep from where it was supposed to
    be.  Fixing that resolves the test suite failure again.
  10. Matt Domsch
  11. Matt Domsch
Something went wrong with that request. Please try again.