about summary refs log tree commit homepage
path: root/lib/rainbows/base.rb
DateCommit message (Expand)
2019-01-02quiet mismatched indentation warnings
2014-01-17unicorn 4.8 compatibility + test version bumps
2011-08-30avoid potential Logger deadlock in SIGQUIT and SIGUSR1
2011-06-27remove unused arg in Rainbows::Response.setup
2011-06-22Unicorn 4.x resync for ticker
2011-06-15remove IO_PURGATORY dependency
2011-06-09move reopen_worker_logs to base
2011-01-20merge rack_input into process_client
2011-01-06eliminate G constant and just use the Rainbows! module
2011-01-04globally refactor Range handling for responses
2010-12-08respect "rewindable_input false" in Unicorn config
2010-10-22code shuffling for kgio
2010-10-21unicorn 2.x updates + kgio
2010-09-28start using kgio library
2010-08-10doc: it's always "Rainbows!" with a bang(!)
2010-07-23rename parse_range => make_range!
2010-07-22enable Range: responses for static files for most models
2010-07-20rely on autoload for Base and TeeInput
2010-07-19refactor response handling for each concurrency model
2010-07-10doc: avoid documenting internals on RDoc website
2010-07-08restore Rainbows::HttpResponse.write for Cramp
2010-07-04refactor response body handling for sendfile(2)
2010-06-28add preliminary sendfile (1.0.0) gem support
2010-06-28fiber/base: reuse process_client logic in base
2010-06-28(style) prefer "until" instead of "while !"
2010-06-28base: constant/namespace cleanup
2010-06-11update test infrastructure to support Rubinius
2010-06-06centralize body => IO conversion logic
2010-05-26add WriterThreadSpawn concurrency model
2010-05-03cleanup request size limiting for TeeInput users
2010-05-03add client_max_body_size config directive
2010-04-27base: status == 100 check needs to_i conversion
2010-04-19use IO.copy_stream for Thread{Spawn,Pool} under 1.9
2010-03-28cleanup: avoid redundant REMOTE_ADDR logic
2009-12-22base: fix constant resolution under 1.8 for 1.8 bugfix
2009-12-22fix Ruby 1.8 detection for (possible) green thread fix
2009-12-21possible MRI 1.8 thread fix to avoid blocking accept()
2009-12-19Thread*: avoid double close of client socket
2009-12-16join_threads method is exclusive to ThreadPool
2009-12-16cleanup: consolidate write_nonblock error handling
2009-12-11env["hack.io"] for Fiber*, Revactor, Thread* models
2009-12-01more consistent code for worker timeout/exits
2009-11-30ThreadPool: no need to exit!
2009-11-29refactor threaded models to use blocking accept() if possible
2009-11-26cleanup and refactor error handling
2009-11-18make keepalive_timeout configurable
2009-11-18Thread*: start implementing keepalive timeout
2009-11-13gracefully exit workers if reopening logs fails
2009-11-11cleanup error handling pieces
2009-11-07remove unnecessary class variable