rainbows 0.94.0 → 0.95.0

data/ChangeLog

@@ -1,392 +1,560 @@
-ChangeLog from git://git.bogomips.org/rainbows.git (v0.92.0..v0.94.0)
+ChangeLog from git://git.bogomips.org/rainbows.git (v0.94.0..v0.95.0)
 
-    commit 6ada0eb9916bbcebbcf2af843c1f5fb11904caee
+    commit 5308c68e7e5b9d464f9f8c4252beff2d964bca1b
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 08:24:33 2010 +0000
+    Date:   Sat Jul 10 06:13:21 2010 +0000
     
-        Rainbows! 0.94.0 - one eight ate my homework!
+        Rainbows! 0.95.0 - sendfile() support!
         
-        This release fixes corrupted large response bodies for Ruby 1.8
-        users with the WriterThreadSpawn and WriterThreadPool models
-        introduced in 0.93.0.  This bug did not affect Ruby 1.9 users
-        nor the users of any older concurrency models.
+        In addition to the 1.9-only IO.copy_stream, the new sendfile
+        1.0.0 gem may optionally be used with most concurrency models
+        (even under 1.8).
         
-        There is also a strange new Rainbows::Sendfile middleware.  It
-        is used to negate the effect of Rack::Contrib::Sendfile, if that
-        makes sense.  See the RDoc or
-        http://rainbows.rubyforge.org/Rainbows/Sendfile.html for all the
-        gory details.
+        See http://rainbows.rubyforge.org/Static_Files.html for more info
         
-        Finally, the RDoc for our test suite is on the website:
+        Other changes:
         
-          http://rainbows.rubyforge.org/Test_Suite.html
+        * 1.9 encoding bugfix for (Rev)FiberSpawn and FiberPool
+        * fixed potential rack.input corruption with Revactor
+        * ThreadPool graceful shutdown no longer blocks until timeout
+        * optional ServerToken middleware for to display Server: header
+        * Dependencies bumped to Rack 1.1+ and Unicorn 1.1.0+
+        * numerous internal cleanups, small bugfixes and speedups
+        * more concise website oriented at users
+    
+    commit d3b97d1114e2d23a9231fe889cd5bf6927d451ba
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Sat Jul 10 08:04:48 2010 +0000
+    
+        doc: avoid documenting internals on RDoc website
         
-        I wrote this document back when the project started but
-        completely forgot to tell RDoc about it.  Personally, this
-        test suite is one of my favorite parts of the project.
+        Since we suck at building websites, we just rely on RDoc as a
+        website builder.  And since Rainbows! is an application server
+        (and not a programming library), our internal API should be of
+        little interest to end users.
+        
+        Anybody interested in Rainbows! (or any other project) internals
+        should be reading the source.
     
-    commit 761b4a346865267c3fef79a62b9bebcbf6277319
+    commit 9838b614621bbbff27a91166406d833be85adbbd
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 07:48:08 2010 +0000
+    Date:   Sat Jul 10 07:26:33 2010 +0000
     
-        sendfile: update RDoc for this middleware.
+        doc: Static_Files documentation
+        
+        Some folks may be interested in setting up Rainbows!
+        as a static file server.
     
-    commit 8ac80b06424350fd6088c1363661a12620040ae8
+    commit 59f10a6dcf895b728bbc34546aebdb8bf25a1b0b
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 07:11:47 2010 +0000
+    Date:   Fri Jul 9 08:45:03 2010 +0000
     
-        doc: update test suite README and link with RDoc
+        event_machine: rename response_write => em_write_response
         
-        Apparently this document was completely forgotten over the
-        course of development and never properly put on the website.
+        For consistency with rev_write_response (and the existing
+        "write_response").
+    
+    commit e1c10949271a9cd3cbad86d05ce0436f2e79bc65
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Fri Jul 9 08:30:15 2010 +0000
+    
+        http_response: allow Cramp 0.11 to send proper WS headers
         
-        The test suite is one of my favorite parts of Rainbows! and
-        allows us to test the server as real clients would hit it.
+        Cramp needs to override our normal header sending for
+        (old) WebSockets connections.
     
-    commit 9dd7e943707608d9c98240b33c859d3c09fac577
+    commit 9abb447a573751b87e26ce35f4f05d58290f4c41
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 06:54:52 2010 +0000
+    Date:   Thu Jul 8 23:21:18 2010 +0000
     
-        cleanup makefiles, move publish_doc into main
+        restore Rainbows::HttpResponse.write for Cramp
         
-        Since the docs for this project are hosted on Rubyforge.org
-        (currently Apache), it can't use the nginx "gzip_static on"
-        configuration I normally use on on *.bogomips.org.
+        Cramp monkey patches Rainbows internals for WebSockets
+        support and we forgot about it.  Add a new integration
+        test to ensure this continues to work in the future
+        (and force us to update the test for newer Cramp).
+    
+    commit 235985c6c4b37ed9a170c38052db3ef0772b1527
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Thu Jul 8 19:26:33 2010 +0000
+    
+        dev: bump isolate dependency to 2.1.0
+    
+    commit 90933787f1a3cf5b8e2d0497b56a9541e2e492ad
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Thu Jul 8 09:33:21 2010 +0000
+    
+        TODO: documentation needs work, so document it :P
+    
+    commit 772c7a486583630627f5de58aef82423487592e4
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Thu Jul 8 09:29:18 2010 +0000
+    
+        bump TCP_DEFER_ACCEPT default value
         
-        I never used the irb/sh wrappers in local.mk, either, and
-        the isolate bits have all been moved inside t/
+        Unicorn 1.1.0 lets us change this default, and we need it higher
+        to avoid wasting workers against stupidly (or maliciously) slow
+        clients.
+    
+    commit b07c3e00b390eebde32c762fd17ab387490c7552
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Thu Jul 8 09:25:18 2010 +0000
     
-    commit 93842bb5ef9e5da1d750e41614a99024dc7a0399
+        bump unicorn dependencies
+    
+    commit 13baf8bd44683a376b79c01c3abbaa71e5b86a71
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 06:36:28 2010 +0000
+    Date:   Wed Jul 7 08:30:22 2010 +0000
     
-        tests: make -C $MODEL.tNNNN-foo.sh work again out-of-the box
+        stream_file: split this out from rev/sendfile
         
-        It was missing dependencies on the first run
+        Other concurrency models will eventually be able to use it, too.
     
-    commit f66c59d7b259e5784c9da87df1a6f1dbca160577
+    commit 9ce364b079efdb3bcef504a84e2460a9b2b4b9ad
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 01:25:29 2010 +0000
+    Date:   Tue Jul 6 10:33:01 2010 +0000
     
-        tests: cleanup isolate usage
+        byte_slice: String#[range] is faster than String#slice!(range)
         
-        We don't need tmp/ in our top level directory
+        rb_str_slice_bang() allocates a new string internally and calls
+        rb_str_aref_m() AND rb_str_aset_m(), too.  String#[] just
+        calls rb_str_aref_m() directly, so it's a much quicker code
+        path.  Also, "[]" methods dispatch faster under 1.9, too.
     
-    commit e7690caedf80a1af2b6145d1b8f108ea42ee9a29
+    commit 9ab356d3969779d7f8f4be6579bfa7570c703f21
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 00:34:01 2010 +0000
+    Date:   Tue Jul 6 10:19:27 2010 +0000
+    
+        doc: update TODO
     
-        tests: disable async_sinatra test for Ruby 1.9.2
+    commit d2f240d7ea11738ba889668a669ffb7845d06e7b
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Tue Jul 6 09:58:40 2010 +0000
+    
+        fix string slicing under 1.9 after short writes
         
-        Sinatra 0.9.4 does not appear to be compatible with
-        Ruby 1.9.2...
+        Fortunately this only affects the hardly-used FiberSpawn and
+        FiberPool concurrency models, and also unreleased revisions of
+        Rev.  1.9 encoding is tricky to handle right when doing I/O in
+        Ruby...
     
-    commit 8e9144341c6169abd2fa6ec20343e3152534b38b
+    commit 0283928fca45a012b31662b30b49990dac2aae18
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 00:32:47 2010 +0000
+    Date:   Tue Jul 6 08:53:54 2010 +0000
     
-        local.mk.sample: add a hint for TMPDIR
+        rev/client: more aggressive non-blocking write
         
-        It makes tests run significantly faster on a beefy box.
+        non-blocking write() may cause kernel buffers to be allocated
+        behind the scenes, so retry the write() even if it's short
+        because it may succeed the next time around.
     
-    commit 7ca6682fac2a30a69834b46a678ec956db4a45e9
+    commit 8b65858a864aa0ecfa24ccf8f910c36af0ec1ad6
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 00:21:58 2010 +0000
+    Date:   Tue Jul 6 04:01:30 2010 +0000
     
-        local.mk.sample: update for running 1.9.2 (preview3) tests
+        cleanup error handling for aborted downloads
         
-        Now that 1.9.2 preview3 is available and usable, it'll
-        be added to the list of Rubies we run and officially
-        support.
+        We shouldn't ever spew errors to the stderr/logger
+        on client disconnects (ECONNRESET/EPIPE/etc...).
     
-    commit 50f764fcbc0ef84fd0589d16179fed637d817b19
+    commit d0a1fcaf25b10ff1d6894d50fa981f56169195f3
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 00:20:59 2010 +0000
+    Date:   Tue Jul 6 02:05:01 2010 +0000
     
-        Rakefile: explain ourselves when Isolating
+        rev: avoid unnecessary seeking when using sendfile
         
-        Yes there's a reason for everything we do :>
+        There's no need to ever change the underlying offset of a file
+        descriptor when using sendfile(), so don't.  This allows us to
+        avoid contention in the kernel/filesystem and eventually reuse
+        the same filesystem file descriptor for serving multiple
+        requests.
     
-    commit e93d93d7deb29b09dd68abfddd8a34cec8fed9b3
+    commit 70ad994ae8a854477e45c877cd0e9dda41389f86
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri Jun 4 00:16:51 2010 +0000
+    Date:   Mon Jul 5 08:07:03 2010 +0000
     
-        test-lib: set NO_PROXY to counter http_proxy
+        rev: properly enable post-fork sendfile() usage
         
-        Some testers (like myself) use http_proxy when isolating
-        gems to avoid wasting bandwidth, but we don't proxy requests
-        to localhost.
+        We need to load sendfile-using parts after the
+        "sendfile" library is loaded.
     
-    commit 6d62a7de86369e046e6f6b221153d867df11e74c
+    commit 39b178cdebe275cbc8ce19cf269bea7cd15ff4ca
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu Jun 3 17:15:18 2010 -0700
+    Date:   Sun Jul 4 22:16:52 2010 +0000
     
-        depend on Isolate ~> 2.0.2 for development
+        refactor response body handling for sendfile(2)
+        
+        This hopefully allows the "sendfile" gem to be required
+        anywhere in the Rainbows!/Unicorn config file, and not
+        have to be required via RUBYOPT or the '-r' command-line
+        switch.
         
-        It's useful given all the Gems we support but don't have hard
-        installation dependencies on.
+        We also modularize HttpResponse and avoids singleton methods
+        in the response path.  This (hopefully) makes it easier for
+        individual concurrency models to share code and override
+        individual methods.
     
-    commit cf122e941e10c2812b7ba5e75c053a28950ddcb6
+    commit 75f5aa9a0d6b37a94afbea3121fc2c16e70a2b1d
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu Jun 3 15:23:01 2010 -0700
+    Date:   Sun Jul 4 22:15:18 2010 +0000
     
-        add Rainbows::Sendfile middleware
+        rev/client: s/sendfile/rev_sendfile/
         
-        This lets most concurrency models understand and process
-        X-Sendfile efficiently with IO.copy_stream under Ruby 1.9.
-        EventMachine can take advantage of this middleware under
-        both Ruby 1.8 and Ruby 1.9.
+        Avoid confusing people with an overloaded method name
     
-    commit 5a0a190253434d3eef5332d979d56f3e16e876bd
+    commit a5dc800d2e4b3a86466a03f45f64b97d29e4b11e
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu Jun 3 14:55:37 2010 -0700
+    Date:   Sun Jul 4 08:13:24 2010 +0000
     
-        writer_thread_*: fix static file response under 1.8
+        revactor: properly zero buffer on zero reads
         
-        While these models are designed to work with IO.copy_stream
-        under Ruby 1.9, it should be possible to run them under Ruby
-        1.8 without returning corrupt responses.  The large file
-        response test is beefed up to compare SHA1 checksums of
-        the served file, not just sizes.
+        While we're at it, fix a comment, too.
     
-    commit b6e354116d4da5c7c6de59f87ab5815f21e202ed
+    commit e867a6d1a612baf7975b14f9bffa14ea8b05f35c
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu Jun 3 07:52:34 2010 +0000
+    Date:   Sun Jul 4 07:40:58 2010 +0000
     
-        test-lib: remove ancient wait_for_pid() function
+        thread_pool: force threads to wakeup on shutdown
         
-        It hasn't been used since the first month of development
-        when we started using FIFOs
+        We may use a blocking accept() loop if there is
+        only a single listener.  In that case threads may
+        not be able to exit if a SIGQUIT is received, so
+        force them to run when joining.
     
-    commit c3e22e4c2ad11785b5355a98355fd31d709976d5
+    commit 7e3de7a12629423e911dae8955e7cc2035d6c401
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 28 04:03:32 2010 +0000
+    Date:   Wed Jun 30 09:15:20 2010 +0000
     
-        Rainbows! 0.93.0 - MOAR!!!1
+        rev: add sendfile support
         
-        In our race to have more concurrency options than real sites
-        using this server, we've added two new and fully supported
-        concurrency models: WriterThreadSpawn and WriterThreadPool
+        This should improve performance for static file responses.
+    
+    commit 280c6816ac4ebbfae8298871e125376d30ef5679
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Wed Jun 30 08:13:07 2010 +0000
+    
+        dev_fd_response: avoid redeclaring Rainbows module
         
-        They're both designed to for serving large static files and work
-        best with IO.copy_stream (sendfile!) under Ruby 1.9.  They may
-        also be used to dynamically generate long running, streaming
-        responses after headers are sent (use "proxy_buffering off" with
-        nginx).
+        slowly cleaning up the generated RDoc
+    
+    commit 7495e0503f22f3cb89efcc3ae382f6544a6582e1
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Wed Jun 30 08:06:06 2010 +0000
+    
+        deferred_response: remove needless include
+    
+    commit c305d5efcd177d6516c7f0dd4f441b2605bdcdaa
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 28 09:02:34 2010 +0000
+    
+        TODO: extra item for byte-range responses
+    
+    commit 8175a52c67fb9dfc9c04a7b0597b680699f43deb
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 28 08:06:32 2010 +0000
+    
+        add preliminary sendfile (1.0.0) gem support
         
-        Unlike most concurrency options in Rainbows!, these are designed
-        to run behind nginx (or haproxy if you don't support POST/PUT
-        requests) and are vulnerable to slow client denial of service
-        attacks.
+        This still needs work and lots of cleanup, but the basics are
+        there.  The sendfile 1.0.0 RubyGem is now safe to use under MRI
+        1.8, and is superior to current (1.9.2-preview3) versions of
+        IO.copy_stream for static files in that it supports more
+        platforms and doesn't truncate large files on 32-bit platforms.
+    
+    commit 86e9c7013308d77def5fe41b52a35dea60c7361c
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 28 06:29:54 2010 +0000
+    
+        fiber/base: reuse process_client logic in base
         
-        I floated the idea of doing something along these lines back in
-        the early days of Unicorn, but deemed it too dangerous for some
-        applications.  But nothing is too dangerous for Rainbows!  So
-        here they are now for your experimentation.
+        This fleshes out Rainbows::Fiber::IO with a few
+        more methods for people using it.
+    
+    commit b75ad975856a681fe96651592b2bb24cc044fd50
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 28 05:20:17 2010 +0000
+    
+        (style) prefer "until" instead of "while !"
+    
+    commit eaf6cca3e1d38b7307c65c257b080041e65fd215
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 28 05:14:23 2010 +0000
+    
+        revactor: constant/namespace cleanups
     
-    commit f1cb1ce4e76ca13c9eb0bf8cfb4a51bd9f5a58c1
+    commit 5cde25e2fb6f7de9143a40da5bad4b91b582fcab
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Sat May 29 04:20:25 2010 +0000
+    Date:   Mon Jun 28 00:30:34 2010 +0000
     
-        WriterThreadSpawn: fix stupid local variable error
+        base: constant/namespace cleanup
     
-    commit 93fc16377eca1a1d5103fd91a576ff6c469f4ab1
+    commit 63d95ac64949a642596946b2b44b2d0bb7b9fefb
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 28 03:58:22 2010 +0000
+    Date:   Sun Jun 27 23:58:00 2010 +0000
     
-        doc: update comparison
+        http_response: make this a module, not a class
         
-        I still have a hard time keeping track of what's capable of
-        what.
+        No point in using a class here, there's no object
+    
+    commit 8756eac8c0e993caa6f704e27985405f6c9dfa7d
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Sun Jun 27 23:51:55 2010 +0000
+    
+        http_response: (nitpick) simplify conditional logic
     
-    commit 49c45c6ea09464862779b09ad7f22c30977a2060
+    commit 43745f26e201655f98351684e1d357cbd8c15b05
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 28 03:42:15 2010 +0000
+    Date:   Sun Jun 27 23:36:51 2010 +0000
     
-        tests: t0005: fudge the file malloc limit a bit
+        http_response: cleanup documentation
         
-        it's hard to make this test reliable, but try to
-        add a small fudge factor based on the MRI default
-        GC malloc limit.
+        No point in documenting our internals and overwhelming
+        users.
     
-    commit 86d5209d0ebdd597ffa19bff5f227ca10ba383c6
+    commit 4442a5f7d517645957d5a78911fec48010d3e6c7
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 28 03:07:37 2010 +0000
+    Date:   Thu Jun 24 06:48:54 2010 +0000
     
-        dev: isolate: bump cramp to 0.11
+        sendfile: update docs for Rack::Sendfile
+        
+        Rack::Contrib::Sendfile moved into Rack in December 2009.
     
-    commit 371f05992110616412c57a26fb6688dc2111c65d
+    commit 23551fb5241630e1c8d1ea3788defb4b5bcbe832
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 26 22:50:15 2010 +0000
+    Date:   Mon Jun 21 08:03:01 2010 +0000
     
-        writer_thread_spawn: worker_connections limits thread spawned
+        add Rainbows::Configurator module
         
-        This should be logical, since we keep the connection alive
-        when writing in our writer threads.
+        Make it easier to link to the Rainbows! configuration
+        documentation without anchors.  This also reduces the
+        amount of code we spew into Unicorn::Configurator.
     
-    commit 904526129a2377306a735c73b26699583b3b99c5
+    commit c9a573e669e62e824f8e3cf37cd94162cd7a4170
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 26 22:49:49 2010 +0000
+    Date:   Mon Jun 21 07:15:40 2010 +0000
     
-        thread_spawn: document why we sleep instead of Thread.pass
+        tests: enable ActorSpawn test for rbx
     
-    commit 4ccaf0a931a0731d1dca2710b86864075eb511a1
+    commit 4a132122eb4f931fa721a4ff5d2204486c0c3579
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 26 22:36:39 2010 +0000
+    Date:   Mon Jun 21 07:15:22 2010 +0000
     
-        writer_thread_pool: update documentation for new defaults
+        rev: documentation update for #write method
     
-    commit 58661617ab802010ecbc45ce3afbca1d63cb9189
+    commit f64a6781cbfbde40e4be6334520b1ff06a5e5626
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 26 22:20:57 2010 +0000
+    Date:   Mon Jun 21 05:54:06 2010 +0000
     
-        add WriterThreadSpawn concurrency model
+        dev_fd_response: disable under Rubinius for now
+        
+        We can't use it effectively in Rubinius yet, and it's
+        broken due to the issue described in:
+          http://github.com/evanphx/rubinius/issues/379
     
-    commit c6ecda097af9cc559b2d38b01ae23daf733b3786
+    commit a4a3fbb761e5ae4f36e9dd84ad933c6df63975ba
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 26 22:20:03 2010 +0000
+    Date:   Mon Jun 21 05:48:45 2010 +0000
     
-        gemspec: bump Unicorn dependency
+        dev_fd_response: cleanup and reorganization
         
-        no major internal changes until 2.0.0+
+        There's no need to #dup the middleware object, just use
+        a custom Rainbows::DevFdResponse::Body object.
+    
+    commit f2b75a59b54ec93e7fe0031d79127e2ed960725b
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 21 04:38:17 2010 +0000
+    
+        test_isolate: bump unicorn dependency
     
-    commit b37fa6ad928fb2b5dc55126c5ad4caa55ee3193b
+    commit 212f60d37fdfe684dbaed047391b4276a30d200b
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 26 21:32:13 2010 +0000
+    Date:   Mon Jun 21 03:24:06 2010 +0000
     
-        writer_thread_pool: remove single thread optimization
+        test: no need for IO#sync=true in async-response tests
         
-        It's not worth the trouble and testability since having
-        a single thread tends to bottleneck if there's a bad
-        client.
+        We only read from the IO pipe and never write to it
     
-    commit f741974675c066b245e781c0c248b1395188334b
+    commit 5470d0688ec610861241408790d22b583e211ac5
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Sat May 22 22:47:34 2010 +0000
+    Date:   Mon Jun 21 03:11:29 2010 +0000
     
-        tests: ignore trash/ directory entirely
+        tests: avoid embedded command-line switches in tests
         
-        This allows it to be a symlink to /dev/shm or similar
+        They're ugly and potentially non-portable to other servers.
+        They also make Unicorn + Rubinius unhappy, which makes us
+        unhappy as well.
     
-    commit e948380c2a28f12e7210130bc83fb0ae0e6fcac9
+    commit 51b2d65b081b0e802acdd25295e04889bbccaddc
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 14:22:08 2010 -0700
+    Date:   Mon Jun 21 03:00:30 2010 +0000
     
-        tests: chunked client_body_max_size fails with old curl
+        rev: restore Ruby 1.8.6 compatibility
+    
+    commit 82c93e28e743480d59c81406a3aaad3b6ba4269f
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Mon Jun 21 02:47:31 2010 +0000
+    
+        bump async_sinatra test dependencies
         
-        curl < 7.18.0 did not check for errors when doing chunked
-        uploads.  Unfortunately some distros are slow moving and
-        bundle ancient versions of curl.
+        Everything should be working under Ruby 1.9.2(-preview3) now.
     
-    commit bfe34128aed4ec696bb939c19e86a742297af830
+    commit 679a4b4673f66a299322facfe2acb8776bf61c00
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 11:48:28 2010 -0700
+    Date:   Fri Jun 18 09:44:57 2010 +0000
     
-        local.mk.sample: fix isolate 2.0.0 path :x
+        rev: cleanup write method
+        
+        I was originally experimenting with setsockopt to increase the
+        kernel buffer sizes in a loop, but the benefits were negligible
+        at best.
     
-    commit ac92728df5a673076e3fedd1436cef7c24e534b5
+    commit e5bd756d0a8ab73cab16facbd6f1b2275f6dfa21
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 03:40:49 2010 -0700
+    Date:   Fri Jun 18 08:59:01 2010 +0000
     
-        tests: t0004: beef up test for kills
+        rev: avoid extraneous data copy for small responses
         
-        since we don't set maximum time boundaries, just rely on
-        the logs to properly log the dead processes.
+        For small responses that can fit inside a kernel socket
+        buffer, copying that data into an IO::Buffer object is
+        a waste of precious memory bandwidth.
     
-    commit d1b2756e6a5b7b9e1c47861f230ec9effe6ca382
+    commit b290a65eb735aa89bb071e27f8421cc864c78e8c
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 03:27:06 2010 -0700
+    Date:   Fri Jun 18 08:10:55 2010 +0000
     
-        tests: t0004: disable upper limit check
+        fiber/{base,io}: use arrays instead of hashes for r/w accounting
         
-        since we've already waited for time to elapse, there's no
-        point in watching the upper limit here
+        This gives a tiny performance improvement to the FiberSpawn and
+        FiberPool concurrency models.
     
-    commit 94307f744608ad6fd4f8d0ab57d23cf405917b15
+    commit b5a24fdbf928b675e478987393212651ef5909d6
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 03:06:41 2010 -0700
+    Date:   Fri Jun 18 08:10:00 2010 +0000
     
-        tests: t0005: minimize writes
+        fiber/base: more accurate sleep timeout
         
-        This test can cause a lot of I/O, especially when
-        run in parallel.  Just rely on the fixed rsha1 code
-        to compute the SHA1 of the response.
+        Not that many people will actually call Rainbows.sleep
+        outside of tests...
     
-    commit abc207b2918606867094f2820bab58223e99aac4
+    commit fd1ec829c8c941143c8ae70b18e51838bc51da73
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 03:05:36 2010 -0700
+    Date:   Fri Jun 18 06:47:52 2010 +0000
     
-        tests: fix sha1sum usage
+        avoid needless HeaderHash#to_hash calls
         
-        non-random_blob arguments weren't being taken into account
-        correctly :x
+        HeaderHash objects can only be used as headers without
+        violating Rack::Lint in Rack 1.1.0 or later.
     
-    commit 9d17cc820044e797a12146730dc260e60302022a
+    commit ed9f3e90939cfa9efcc6e5e9382f1e1f40c49834
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 20 02:53:22 2010 -0700
+    Date:   Fri Jun 18 06:41:58 2010 +0000
     
-        tests: t0005 - disable concurrency for this test
+        prefer Array#[] lookup to Array#first/Array#last
         
-        slow test runners can buffer us and bloat memory usage
-        unpredictably when tests are run under load
+        Array#[] lookups are slightly faster under both rbx and 1.9,
+        and easier to read.
     
-    commit 6048147422f99f654b8a400575c58d39351c4699
+    commit 0e4014888f4f7f17d45c96677d672a538c6e6d94
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Wed May 19 20:15:15 2010 -0700
+    Date:   Wed Jun 16 09:53:32 2010 +0000
     
-        tests: t0007: robustness fix
+        test_isolate: fix ruby engine usage
         
-        On busy sytems, this timing-sensitive test is likely to fail,
-        so give it some extra slack
+        Oops, looks like 1.9.1 exports the RUBY_ENGINE constant.
     
-    commit d266f4da075b4b972ee97589ac58ce1d08022f84
+    commit 0a3e9cacd7ca2feb91bae23b0c678d5e986f866f
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Tue May 18 14:07:54 2010 -0700
+    Date:   Fri Jun 11 08:47:41 2010 +0000
     
-        doc: FAQ: clarify Rails support
+        update test infrastructure to support Rubinius
+        
+        Rubinius still has a few issues that prevent 100% support,
+        but it basically works if log rotation or USR2 upgrades aren't
+        required.  Tickets for all known issues for Rubinius have
+        been filed on the project's issue tracker.
+        
+        * rbx does not support -i/-p yet, so rely on MRI for that
+        * "io/nonblock" is missing
+        * avoiding any optional Gems for now (EM, Rev, etc..)
+    
+    commit 37687e027233072582c28bdd1530047a40ef7869
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Fri Jun 11 02:27:49 2010 +0000
+    
+        alt working_directory test from Unicorn
+    
+    commit aa402ca5c52ccf836d03e08a82c5a8392fca8f15
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Fri Jun 11 02:24:00 2010 +0000
+    
+        rely on Unicorn 0.991.0 for tests
+    
+    commit 0c08b142544e717ece5865869e14ead670f64eec
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Thu Jun 10 18:27:19 2010 -0700
+    
+        bin/rainbows: sync up with latest Unicorn
+        
+        This will allow us to be working_directory-aware as
+        far as config.ru goes.
     
-    commit 48313f442f53dfdb4ba24bbb12fc91f5c47a24ec
+    commit cf47d3e51d7b0aecdda8b0f27631161977900a47
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Mon May 17 11:36:56 2010 -0700
+    Date:   Thu Jun 10 18:22:25 2010 -0700
     
-        WriterThreadPool: default concurrency to 20 pools
+        add Rainbows::ServerToken middleware
         
-        Idle threads are cheap enough and having responses
-        queued up with a single slow client on a large response
-        is bad.
+        Some folks can now show off their Rainbows! installation
     
-    commit 2260d56329b036bcce7eb69831d3e480334d283c
+    commit b73218100032ead24360d33361fefa914074f30a
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 14 18:12:51 2010 -0700
+    Date:   Thu Jun 10 18:21:26 2010 -0700
     
-        add WriterThreadPool concurrency model
+        my-tap-lib: fix race condition in verbose mode
         
-        This is based on an idea I originally had for Unicorn but never
-        implemented in Unicorn since the benefits were unproven and the
-        risks were too high.
+        Don't try to redirect until we know our FIFO consumers are
+        ready for us.  This only seems to happen with bash and not
+        ksh...
+        
+        commit d0a0883eaaeec37800ca5cd07647b7b66a00c453 in Unicorn
     
-    commit 01ee5327c018f64fa4b6aa767c0328c56f699170
+    commit d65fb5896d4c9d019b31834c131b5fb0306ba10e
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 14 15:52:58 2010 -0700
+    Date:   Wed Jun 9 21:44:43 2010 +0000
     
-        add Rainbows::QueuePool helper
+        http_server: use super instead of alias_method
         
-        It'll be useful later on for a variety of things!
+        duh!
     
-    commit 5f854e876ba55180936ebfc82002df9bd882a4ae
+    commit 3b4fa90bee97d2346644db004c37db5cb29cd9cf
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 14 15:52:17 2010 -0700
+    Date:   Tue Jun 8 09:52:27 2010 +0000
     
-        test: do not let async-response run with Base
+        gemspec: bump unicorn dependency to 0.990.0
     
-    commit 67662fe0714cbcc89d732d4a04df975b0de2ca36
+    commit 6565060915266f053b5eb1c29a1cfbf825d430b0
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Fri May 14 14:18:36 2010 -0700
+    Date:   Mon Jun 7 11:30:36 2010 -0700
     
-        update isolate config for Isolate 2.0.0
+        README: update with Zbatery info
     
-    commit 65d5e99ee231624c66a26b70296a659d551c0266
+    commit e812eb864f88d88bb5e6bfdba617b03db2f07bfe
     Author: Eric Wong <normalperson@yhbt.net>
-    Date:   Thu May 6 22:22:56 2010 -0700
+    Date:   Sun Jun 6 02:33:22 2010 +0000
+    
+        tests: t9001: avoid needless filesystem activity
     
-        dev: isolate: bump unicorn dependency
+    commit b1315983056f91467b336aeb426759debc006a4e
+    Author: Eric Wong <normalperson@yhbt.net>
+    Date:   Sun Jun 6 02:14:58 2010 +0000
+    
+        centralize body => IO conversion logic
+        
+        Since EventMachine and Rev shared the same logic for optimizing
+        and avoiding extra file opens for IO/File-ish response bodies,
+        so centralize that.
+        
+        For Ruby 1.9 users, we've also enabled this logic so ThreadPool,
+        ThreadSpawn, WriterThreadPool, and WriterThreadSpawn can take
+        advantage of Rainbows::DevFdResponse-generated bodies while
+        proxying sockets.