Release Name: 0.90.0

Notes:
Rainbows! is an HTTP server for sleepy Rack applications.  It is based on
Unicorn, but designed to handle applications that expect long
request/response times and/or slow clients.  For Rack applications not
heavily bound by slow external network dependencies, consider Unicorn
instead as it simpler and easier to debug.

* http://rainbows.rubyforge.org/
* rainbows-talk@rubyforge.org
* git://git.bogomips.org/rainbows.git

Changes:

This release should fix ThreadSpawn green thread blocking issues
under MRI 1.8.  Excessive socket closing is avoided when using
Thread* models with Sunshowers (or clients disconnecting
during uploads).

There is a new RevFiberSpawn concurrency model which combines
Rev with the traditional FiberSpawn model.



Changes:
 Documentation/comparison.haml       |   23 +++++-
 GIT-VERSION-GEN                     |    2 +-
 GNUmakefile                         |    4 +-
 README                              |    3 +-
 Rakefile                            |    4 +-
 TODO                                |    6 +-
 bin/rainbows                        |    2 +-
 lib/rainbows.rb                     |   49 ++++++++---
 lib/rainbows/app_pool.rb            |    2 +-
 lib/rainbows/base.rb                |   23 ++---
 lib/rainbows/const.rb               |   10 +-
 lib/rainbows/dev_fd_response.rb     |    2 +-
 lib/rainbows/error.rb               |   11 ++-
 lib/rainbows/ev_core.rb             |    4 -
 lib/rainbows/event_machine.rb       |    2 +-
 lib/rainbows/fiber/base.rb          |    5 +-
 lib/rainbows/fiber/io.rb            |   24 +++--
 lib/rainbows/fiber/rev.rb           |  160 +++++++++++++++++++++++++++++++++++
 lib/rainbows/rev/core.rb            |    2 +-
 lib/rainbows/rev_fiber_spawn.rb     |   29 ++++++
 lib/rainbows/revactor.rb            |   13 +---
 lib/rainbows/thread_pool.rb         |    8 ++
 t/GNUmakefile                       |    1 +
 t/simple-http_RevFiberSpawn.ru      |   10 ++
 t/simple-http_Revactor.ru           |    1 -
 t/simple-http_ThreadPool.ru         |    1 -
 t/simple-http_ThreadSpawn.ru        |    1 -
 t/sleep.ru                          |    9 +--
 t/t9000.ru                          |    9 +--
 t/worker-follows-master-to-death.ru |    7 +--
 30 files changed, 325 insertions(+), 102 deletions(-)

commit 0df219ef8f9a514553e13f2f98575896f6cc7c39
Author: Eric Wong 
Date:   Tue Dec 22 13:38:58 2009 -0800

    Rainbows! 0.90.0
    
    This release should fix ThreadSpawn green thread blocking issues
    under MRI 1.8.  Excessive socket closing is avoided when using
    Thread* models with Sunshowers (or clients disconnecting
    during uploads).
    
    There is a new RevFiberSpawn concurrency model which combines
    Rev with the traditional FiberSpawn model.

commit 8a15fc14067e85c419a31e6d70c2491e65d88e60
Author: Eric Wong 
Date:   Tue Dec 22 13:30:19 2009 -0800

    avoid setting "rainbows.autochunk" by default
    
    No point in becoming the straw that causes a rehash
    since hardly anybody uses it.

commit ca4c0b2938669f96eb0d71a0911275abb247e418
Author: Eric Wong 
Date:   Tue Dec 22 13:29:09 2009 -0800

    base: fix constant resolution under 1.8 for 1.8 bugfix

commit 505e0f7e54e6b0ea0823af92d339a4bffdb51dc1
Author: Eric Wong 
Date:   Tue Dec 22 13:11:32 2009 -0800

    update TODO

commit fa622de470d475f0afc94cb619cc69e7e127830c
Author: Eric Wong 
Date:   Tue Dec 22 01:22:32 2009 -0800

    common Rainbows.sleep(nr) method
    
    We'll export this across the board to all Rack applications
    to sleep with.  This provides the optimum method of sleeping
    regardless of the concurrency model you choose.  This method
    is still highly not recommended for pure event-driven models
    like Rev or EventMachine (but the threaded/fiber/actor-based
    variants are fine).

commit cdac4e6b8847754421c6f65baab2ac9a105d746a
Author: Eric Wong 
Date:   Tue Dec 22 12:45:03 2009 -0800

    cleanup in Rainbows.accept definition

commit dc5a3bc59b104086ae80116e597f51f99d8f1e24
Author: Eric Wong 
Date:   Tue Dec 22 12:37:31 2009 -0800

    fix README speling fail

commit 451ca6997b4f298b436605b7f0af75f369320425
Author: Eric Wong 
Date:   Tue Dec 22 12:46:26 2009 -0800

    fix Ruby 1.8 detection for (possible) green thread fix
    
    Thanks to Ben Sandofsky for the extra set of eyes

commit 537de8098e649406e7f60c89211a345ed852ab0b
Author: Eric Wong 
Date:   Tue Dec 22 00:54:19 2009 -0800

    bin/rainbows: avoid confusing people with shebang
    
    Or it'll confuse them more...  Really, it does not
    matter what the shebang points to as long as long as
    setup.rb can deal with it.

commit cc7beb5dafa71d8918f3129e7fb516f1e20388c0
Author: Eric Wong 
Date:   Tue Dec 22 00:49:37 2009 -0800

    README: add RevFiberSpawn

commit afb658c11af2120d239b852a24f945b1c23eb625
Author: Eric Wong 
Date:   Tue Dec 22 00:42:48 2009 -0800

    tests: enable RevFiberSpawn for 1.9

commit 4cb46cee74ccdf1ce2a9b8fa03459fb4879838f5
Author: Eric Wong 
Date:   Tue Dec 22 00:41:17 2009 -0800

    doc: update comparison for RevFiberSpawn

commit 19960488441651d689259071fa1be4f6957d681b
Author: Eric Wong 
Date:   Tue Dec 22 00:32:04 2009 -0800

    new RevFiberSpawn concurrency model
    
    This is like the traditional FiberSpawn, but more scalable (but
    not necessarily faster) as it can use epoll or kqueue.

commit ee7fe220ccbc991e1e7cbe982caf48e3303274c7
Author: Eric Wong 
Date:   Mon Dec 21 16:31:28 2009 -0800

    possible MRI 1.8 thread fix to avoid blocking accept()
    
    Under MRI 1.8, listen sockets do not appear to have the
    nonblocking I/O flag on by default, nor does it set the
    nonblocking I/O flag when calling #accept (but it does
    when using #accept_nonblock, of course).
    
    Normally this is not a problem even when using green threads
    since MRI will internally select(2) on the file descriptor
    before attempting a blocking (and immediately successful)
    accept(2).
    
    However, when sharing a listen descriptor across multiple
    processes, spurious wakeups are likely to occur, causing
    multiple processes may be woken up when a single client
    connects.
    
    This causes a problem because accept(2)-ing on multiple
    threads/processes for a single connection causes blocking accepts in
    multiple processes, leading to stalled green threads.
    
    This is not an issue under 1.9 where a blocking accept() call
    unlocks the GVL to let other threads run.

commit e962d665c1e27de8d930c3acc51502fa4d147532
Author: Eric Wong 
Date:   Mon Dec 21 11:53:42 2009 -0800

    GNUmakefile: push tgz to Rubyforge
    
    Oops.

commit 7cb8867021fa0d5a9be77f0bd24869fbf75356d0
Author: Eric Wong 
Date:   Sat Dec 19 23:38:09 2009 -0800

    Thread*: avoid double close of client socket
    
    TeeInput may explicitly close on client disconnects to
    avoid error messages being written to the socket, likewise
    with "hack.io" users.

commit 8abddc7b98f24ab603414b74cac9b1730dd277d7
Author: Eric Wong 
Date:   Sat Dec 19 22:55:37 2009 -0800

    error: (trivial) remove unnecessary whitespace

commit af0497398961b6556cae25c8ddd55327a4ad5104
Author: Eric Wong 
Date:   Wed Dec 16 23:59:29 2009 -0800

    join_threads method is exclusive to ThreadPool

commit 5dea67ecf3dd868e7b375312cdef2c4651b11437
Author: Eric Wong 
Date:   Wed Dec 16 23:57:08 2009 -0800

    cleanup: consolidate write_nonblock error handling

commit 52ad3f270e758e5bfdfe965cbecbd20d5048062f
Author: Eric Wong 
Date:   Wed Dec 16 02:16:23 2009 -0800

    fiber/io: split out wait_readable and wait_writable methods
    
    This makes them easier to override in subclasses.

commit 2ccce76d94046d017a9b02a0b033e3fff9483e98
Author: Eric Wong 
Date:   Wed Dec 16 00:50:23 2009 -0800

    get rid of the magic EvCore.setup
    
    It gets in the way of Rev/EM-based models that won't use EvCore.
    It doesn't actually do anything useful except making an extra
    layer of indirection to follow.

commit 51e656e3db593c662e7abdf969dd79570525aac9
Author: Eric Wong 
Date:   Sun Dec 13 23:14:47 2009 -0800

    share some constants with Unicorn
    
    Make RACK_DEFAULTS == Unicorn::HttpRequest::DEFAULTS
    and LOCALHOST == Unicorn::HttpRequest::LOCALHOST
    
    No point in having a duplicate objects, and it also makes it
    easier to share runtime constant modifications between servers.

commit f4798796921b20130baf57e5accb3be7efaef861
Author: Eric Wong 
Date:   Sun Dec 13 15:19:57 2009 -0800

    Rakefile: fix fm_update task