Skip to content
Commits on Jan 13, 2012
  1. @josevalim

    Rack::BodyProxy should execute block even on failures.

    josevalim committed
    In general, Rack frameworks are moving logic to close hooks in order
    to support async behavior increasing the chance an exception will
    happen on close. Most servers will actually die if there is an exception
    on close, but such exceptions can also happen in the test environment.
    In such cases, we can accidentally leave a mutex locked, a database
    connection not collected and so forth, therefore, we need to ensure
    the block is called regardless closing the body failed.
Commits on Oct 1, 2011
  1. @raggi

    Simplify BodyProxy#close to avoid horrible loop conditions in streaming

    raggi committed
    scenarios, and save users from themselves
  2. @rkh
Commits on Jun 14, 2011
  1. @janx @rkh

    block should not be called if IOError raised

    janx committed with rkh
    Signed-off-by: Konstantin Haase <konstantin.mailinglists@googlemail.com>
  2. @rkh

    let Rack::BodyProxy raise an IOError (like IO and StringIO do) when c…

    rkh committed
    …alling #close twice. Related to #191.
  3. @janx @rkh

    fix BodyProxy#close

    janx committed with rkh
    Signed-off-by: Konstantin Haase <konstantin.mailinglists@googlemail.com>
Commits on Jun 9, 2011
  1. @rkh
Commits on May 31, 2011
  1. @rkh
Something went wrong with that request. Please try again.