Commits on Feb 8, 2013
  1. @raggi

    Bump version number

    raggi committed Feb 7, 2013
  2. @raggi
  3. @raggi

    Prevent symlink path traversals

     * Closes CVE-2013-0262
    raggi committed Feb 6, 2013
  4. @TimMoore @raggi

    Stop overwriting existing pidfiles.

    A race condition can arise when two servers are started simultaneously. Both
    instances may complete the check for an existing pidfile before either one
    writes it.
    
    Now the pidfile is opened with ::File::EXCL, which raises an error if the file
    already exists. This error is handled by retrying the check and the write.
    TimMoore committed with raggi Jan 30, 2013
  5. @raggi

    Use secure_compare for hmac comparison

     * Closes CVE-2013-0263
    raggi committed Feb 7, 2013
  6. @raggi

    Add secure_compare to Rack::Utils

    Conflicts:
    	test/spec_utils.rb
    raggi committed Feb 6, 2013
  7. @raggi
  8. @raggi

    Fix a bug where host matching occurs out of order

     * Closes #504
    raggi committed Feb 6, 2013
  9. @raggi
  10. @jordimassaguerpla @raggi

    add license information to gemspec

    this way it can be used with rubygems.org API
    jordimassaguerpla committed with raggi Feb 6, 2013
  11. @oscardelben @raggi

    Fix reference link

    oscardelben committed with raggi Feb 2, 2013
  12. @oscardelben @raggi

    Document Rack::Deflater

    oscardelben committed with raggi Feb 2, 2013
  13. @raggi

    Request#trusted_proxy? no longer accepts lines

     * Closes #508
     * Adds some limited coverage. More issues highlighted - incomplete local ips.
    raggi committed Feb 2, 2013
Commits on Feb 7, 2013
  1. @spastorino @raggi
Commits on Jan 28, 2013
  1. @raggi
  2. @raggi
  3. @raggi

    Merge pull request #500 from aocole/patch-1

    Changing incorrect documentation
    raggi committed Jan 28, 2013
  4. @raggi

    Remove specific version code from Lint

     * Too easy to miss during updates
     * Required format unchanged
     * Closes #501
    raggi committed Jan 28, 2013
  5. @raggi

    Reimplement keys and values on SessionHash

     * Basic additional APIs to simplify requirements for Rails and Devise
    raggi committed Jan 28, 2013
Commits on Jan 25, 2013
  1. @aocole

    Changing incorrect documentation

    The original comment on set_session said to return true or false
    depending on whether the session was saved or not. In reality, this
    method MUST return the session id in order for #commit_session to set
    the cookie data properly.
    aocole committed Jan 24, 2013
Commits on Jan 22, 2013
  1. @raggi

    Merge branch 'lint-headerhash' of git://bogomips.org/rack

    * 'lint-headerhash' of git://bogomips.org/rack:
      lint: avoid TypeError on non-Hash-like response headers
    raggi committed Jan 22, 2013
  2. @rkh

    Merge pull request #499 from barttenbrinke/master

    Added specific test when X-Forwarded-For is 'unknown'
    rkh committed Jan 22, 2013
  3. @barttenbrinke
  4. lint: avoid TypeError on non-Hash-like response headers

    According to SPEC (and check_headers), Response headers need only
    respond to #each.  Thus, check_hijack_response should rely on
    Rack::Utils::HeaderHash if it wishes to access the headers in a
    hash-like fashion.
    Eric Wong committed Jan 22, 2013
  5. @raggi
  6. @raggi

    Update README for 1.5.0 release

    raggi committed Jan 21, 2013
  7. @raggi

    Switch to RFC 2822 expires

    raggi committed Jan 21, 2013
Commits on Jan 21, 2013
  1. @raggi

    Fix a long standing misnomer for date formats

    References #414
    raggi committed Jan 21, 2013
  2. @raggi

    Use Dir.tmpdir instead of hardcoded /tmp

    Closes #492
    raggi committed Jan 21, 2013
  3. @raggi

    Don't modify the middleware hash in Rack::Server

     * Closes #498
    raggi committed Jan 21, 2013
  4. @raggi

    Merge pull request #496 from homakov/patch-3

    Remove never called string
    raggi committed Jan 21, 2013
Commits on Jan 14, 2013
  1. @homakov

    Remove never called string

    Previous check `p.empty?` makes sure that p contains at least 1 symbol.
    After `.split('=', 2)` k or v or both will turn into some string which means `k || v` will always return true and `next` will never be called.
    homakov committed Jan 14, 2013
Commits on Jan 13, 2013
  1. @raggi

    Update to 1.5.0.beta.2

    raggi committed Jan 13, 2013
  2. @raggi

    Update README security notes

    raggi committed Jan 13, 2013
  3. @raggi