"Limits the number of worker processes to 1" hasn't been true since
06a4dc6 (fix one error in gunicorn_paster, global conf was ignored,
2010-06-22), although it was true when the line was added in 3c7d532
(Large refactor of the documentation and website, 2010-05-22).
"changes some error handling that's sent to clients" hasn't been true
since feb86d3 (don't display the traceback in the HTTP response,
2013-09-27).
The only remaining actions that --debug had were disabling --preload
and hiding debug-level config logging. The former seems useless (just
disable --preload directly) and the latter at doesn't seem useful
enough for a new setting (just turn down --log-level). With this
commit, --preload always works and you always get debug-level config
logging.
I left a stub Debug entry in gunicorn.config, which we can leave in
place while folks convert any gunicorn scripts and configurations to
drop --debug. When the time comes, we can just remove that entry. I
also the boolean-config tests to use --preload, since that will still
be around after we remove the dummy Debug entry.
Fixes#700.
Until now, only the documentation in source form was packaged.
This changeset improves the RPM generator so that the compiled HTML
version of the documentation is packaged. Also, other informative
documents (README, LICENSE, NOTICE, THANKS) and an example server
configuration file have been included.
Finally, the Sphinx documentation's conf.py has been improved so
that it is easier to compile the docs using the default theme.
Also simplifies the environment handling in the gevent_pywsgi
server so that it also has this key. An added side effect is
that the gunicorn FileWrapper gets set for the gevent_pywsgi
worker, too.
Fixes#486
By if the options ` --error-logfile` is set to '-' then wsgi.errors are
returned to sys.stderr, if any log file is passed either using the
log-config file or the option, Errors are written to the files.
By default no error is returned.
This changes allows a use to set the driver used to send the data over a
unix socket. 'dgram' for a dgram driver or 'stream' for a stream driver.
'stream' is the default. The change is documented in the settings
documentation.
fix#671
Gunicorn should generally only bother writing its own log and let
the application handle the way it want to log its own errors.
Now the log_file option will be overriden by the gunicorn options
`--error-logfile` and `--access-logfile` if they are given.
Gunicorn shouln't override the SERVER_* by the Host header. The client
will take care about it during the URL reconstruction if needed.
Since the spec don't support unix sockets, Gunicorn is using the HOST
heeader when available to create viable SERVER_* if possible so the
application and framworks will be happy. When the Host Header is not
available (SocketPath, '') is returned.
fix#628
``send_special_chunks`` sends the request data in two chunks, one having a length of 1 byte, which ensures there is no CRLF, and a second chunk containing the rest of the request data. Practically, this forces ``gunicorn.http.message.Request.read_line()`` to get the full request data with two buffer reads and reveal possible bugs related to the internal buffer mechanism. See #670 for more information.
The remote address should return the direct client addr not a forwarded
header.
This is a breaking change. The main problem with such changes is the way
the application or framework will handle the URL completion. Indeed most
of them are only expecting a TCP socket.
fix#633
