mercurial-scm/hg-stable: mercurial/hgweb/request.py annotate

annotate mercurial/hgweb/request.py @ 36867:a88d68dc3ee8

hgweb: create dedicated type for WSGI responses We have refactored the request side of WSGI processing into a dedicated type. Now let's do the same thing for the response side. We invent a ``wsgiresponse`` type. It takes an instance of a request (for consulation) and the WSGI application's "start_response" handler. The type basically allows setting the HTTP status line, response headers, and the response body. The WSGI application calls sendresponse() to start sending output. Output is emitted as a generator to be fed through the WSGI application. According to PEP 3333, this is the preferred way for output to be transmitted. (Our legacy ``wsgirequest`` exposed a write() to send data. We do not wish to support this API because it isn't recommended by PEP 3333.) The wire protocol code has been ported to use the new API. Differential Revision: https://phab.mercurial-scm.org/D2775

author	Gregory Szorc <gregory.szorc@gmail.com>
date	Sat, 10 Mar 2018 11:23:05 -0800
parents	422be99519e5
children	ec0af9c59270

rev	line source
2391 d351a3be3371 Fixing up comment headers for split up code. Eric Hopper <hopper@omnifarious.org> parents: 2355 diff changeset	1 # hgweb/request.py - An http request from either CGI or the standalone server.
131 c9d51742471c moving hgweb to mercurial subdir jake@edge2.net parents: diff changeset	2 #
238 3b92f8fe47ae hgweb.py: kill #! line, clean up copyright notice mpm@selenic.com parents: 222 diff changeset	3 # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net>
2859 345bac2bc4ec update copyrights. Vadim Gelfer <vadim.gelfer@gmail.com> parents: 2535 diff changeset	4 # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com>
131 c9d51742471c moving hgweb to mercurial subdir jake@edge2.net parents: diff changeset	5 #
8225 46293a0c7e9f updated license to be explicit about GPL version 2 Martin Geisler <mg@lazybytes.net> parents: 7742 diff changeset	6 # This software may be used and distributed according to the terms of the
10263 25e572394f5c Update license to GPLv2+ Matt Mackall <mpm@selenic.com> parents: 10261 diff changeset	7 # GNU General Public License version 2 or any later version.
131 c9d51742471c moving hgweb to mercurial subdir jake@edge2.net parents: diff changeset	8
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	9 from __future__ import absolute_import
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	10
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	11 import errno
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	12 import socket
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	13 import wsgiref.headers as wsgiheaders
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	14 #import wsgiref.validate
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	15
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	16 from .common import (
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	17 ErrorResponse,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	18 HTTP_NOT_MODIFIED,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	19 statusmessage,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	20 )
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	21
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	22 from ..thirdparty import (
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	23 attr,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	24 )
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	25 from .. import (
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	26 error,
34514 528b21b853aa request: coerce content-type to native str Augie Fackler <augie@google.com> parents: 34513 diff changeset	27 pycompat,
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	28 util,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	29 )
138 c77a679e9cfa Revamped templated hgweb mpm@selenic.com parents: 137 diff changeset	30
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	31 @attr.s(frozen=True)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	32 class parsedrequest(object):
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	33 """Represents a parsed WSGI request.
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	34
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	35 Contains both parsed parameters as well as a handle on the input stream.
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	36 """
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	37
36854 16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	38 # Request method.
16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	39 method = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	40 # Full URL for this request.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	41 url = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	42 # URL without any path components. Just <proto>://<host><port>.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	43 baseurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	44 # Advertised URL. Like ``url`` and ``baseurl`` but uses SERVER_NAME instead
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	45 # of HTTP: Host header for hostname. This is likely what clients used.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	46 advertisedurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	47 advertisedbaseurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	48 # WSGI application path.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	49 apppath = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	50 # List of path parts to be used for dispatch.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	51 dispatchparts = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	52 # URL path component (no query string) used for dispatch.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	53 dispatchpath = attr.ib()
36819 cfb9ef24968c hgweb: use parsed request to construct query parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36817 diff changeset	54 # Whether there is a path component to this request. This can be true
cfb9ef24968c hgweb: use parsed request to construct query parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36817 diff changeset	55 # when ``dispatchpath`` is empty due to REPO_NAME muckery.
cfb9ef24968c hgweb: use parsed request to construct query parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36817 diff changeset	56 havepathinfo = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	57 # Raw query string (part after "?" in URL).
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	58 querystring = attr.ib()
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	59 # List of 2-tuples of query string arguments.
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	60 querystringlist = attr.ib()
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	61 # Dict of query string arguments. Values are lists with at least 1 item.
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	62 querystringdict = attr.ib()
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	63 # wsgiref.headers.Headers instance. Operates like a dict with case
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	64 # insensitive keys.
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	65 headers = attr.ib()
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	66 # Request body input stream.
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	67 bodyfh = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	68
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	69 def parserequestfromenv(env, bodyfh):
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	70 """Parse URL components from environment variables.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	71
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	72 WSGI defines request attributes via environment variables. This function
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	73 parses the environment variables into a data structure.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	74 """
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	75 # PEP-0333 defines the WSGI spec and is a useful reference for this code.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	76
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	77 # We first validate that the incoming object conforms with the WSGI spec.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	78 # We only want to be dealing with spec-conforming WSGI implementations.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	79 # TODO enable this once we fix internal violations.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	80 #wsgiref.validate.check_environ(env)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	81
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	82 # PEP-0333 states that environment keys and values are native strings
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	83 # (bytes on Python 2 and str on Python 3). The code points for the Unicode
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	84 # strings on Python 3 must be between \00000-\000FF. We deal with bytes
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	85 # in Mercurial, so mass convert string keys and values to bytes.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	86 if pycompat.ispy3:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	87 env = {k.encode('latin-1'): v for k, v in env.iteritems()}
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	88 env = {k: v.encode('latin-1') if isinstance(v, str) else v
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	89 for k, v in env.iteritems()}
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	90
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	91 # https://www.python.org/dev/peps/pep-0333/#environ-variables defines
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	92 # the environment variables.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	93 # https://www.python.org/dev/peps/pep-0333/#url-reconstruction defines
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	94 # how URLs are reconstructed.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	95 fullurl = env['wsgi.url_scheme'] + '://'
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	96 advertisedfullurl = fullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	97
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	98 def addport(s):
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	99 if env['wsgi.url_scheme'] == 'https':
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	100 if env['SERVER_PORT'] != '443':
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	101 s += ':' + env['SERVER_PORT']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	102 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	103 if env['SERVER_PORT'] != '80':
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	104 s += ':' + env['SERVER_PORT']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	105
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	106 return s
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	107
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	108 if env.get('HTTP_HOST'):
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	109 fullurl += env['HTTP_HOST']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	110 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	111 fullurl += env['SERVER_NAME']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	112 fullurl = addport(fullurl)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	113
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	114 advertisedfullurl += env['SERVER_NAME']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	115 advertisedfullurl = addport(advertisedfullurl)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	116
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	117 baseurl = fullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	118 advertisedbaseurl = advertisedfullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	119
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	120 fullurl += util.urlreq.quote(env.get('SCRIPT_NAME', ''))
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	121 advertisedfullurl += util.urlreq.quote(env.get('SCRIPT_NAME', ''))
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	122 fullurl += util.urlreq.quote(env.get('PATH_INFO', ''))
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	123 advertisedfullurl += util.urlreq.quote(env.get('PATH_INFO', ''))
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	124
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	125 if env.get('QUERY_STRING'):
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	126 fullurl += '?' + env['QUERY_STRING']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	127 advertisedfullurl += '?' + env['QUERY_STRING']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	128
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	129 # When dispatching requests, we look at the URL components (PATH_INFO
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	130 # and QUERY_STRING) after the application root (SCRIPT_NAME). But hgwebdir
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	131 # has the concept of "virtual" repositories. This is defined via REPO_NAME.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	132 # If REPO_NAME is defined, we append it to SCRIPT_NAME to form a new app
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	133 # root. We also exclude its path components from PATH_INFO when resolving
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	134 # the dispatch path.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	135
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	136 apppath = env['SCRIPT_NAME']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	137
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	138 if env.get('REPO_NAME'):
36816 0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	139 if not apppath.endswith('/'):
0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	140 apppath += '/'
0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	141
0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	142 apppath += env.get('REPO_NAME')
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	143
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	144 if 'PATH_INFO' in env:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	145 dispatchparts = env['PATH_INFO'].strip('/').split('/')
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	146
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	147 # Strip out repo parts.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	148 repoparts = env.get('REPO_NAME', '').split('/')
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	149 if dispatchparts[:len(repoparts)] == repoparts:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	150 dispatchparts = dispatchparts[len(repoparts):]
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	151 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	152 dispatchparts = []
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	153
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	154 dispatchpath = '/'.join(dispatchparts)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	155
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	156 querystring = env.get('QUERY_STRING', '')
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	157
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	158 # We store as a list so we have ordering information. We also store as
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	159 # a dict to facilitate fast lookup.
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	160 querystringlist = util.urlreq.parseqsl(querystring, keep_blank_values=True)
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	161
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	162 querystringdict = {}
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	163 for k, v in querystringlist:
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	164 if k in querystringdict:
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	165 querystringdict[k].append(v)
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	166 else:
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	167 querystringdict[k] = [v]
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	168
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	169 # HTTP_* keys contain HTTP request headers. The Headers structure should
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	170 # perform case normalization for us. We just rewrite underscore to dash
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	171 # so keys match what likely went over the wire.
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	172 headers = []
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	173 for k, v in env.iteritems():
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	174 if k.startswith('HTTP_'):
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	175 headers.append((k[len('HTTP_'):].replace('_', '-'), v))
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	176
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	177 headers = wsgiheaders.Headers(headers)
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	178
36853 ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	179 # This is kind of a lie because the HTTP header wasn't explicitly
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	180 # sent. But for all intents and purposes it should be OK to lie about
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	181 # this, since a consumer will either either value to determine how many
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	182 # bytes are available to read.
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	183 if 'CONTENT_LENGTH' in env and 'HTTP_CONTENT_LENGTH' not in env:
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	184 headers['Content-Length'] = env['CONTENT_LENGTH']
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	185
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	186 # TODO do this once we remove wsgirequest.inp, otherwise we could have
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	187 # multiple readers from the underlying input stream.
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	188 #bodyfh = env['wsgi.input']
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	189 #if 'Content-Length' in headers:
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	190 # bodyfh = util.cappedreader(bodyfh, int(headers['Content-Length']))
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	191
36854 16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	192 return parsedrequest(method=env['REQUEST_METHOD'],
16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	193 url=fullurl, baseurl=baseurl,
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	194 advertisedurl=advertisedfullurl,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	195 advertisedbaseurl=advertisedbaseurl,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	196 apppath=apppath,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	197 dispatchparts=dispatchparts, dispatchpath=dispatchpath,
36819 cfb9ef24968c hgweb: use parsed request to construct query parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36817 diff changeset	198 havepathinfo='PATH_INFO' in env,
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	199 querystring=querystring,
3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	200 querystringlist=querystringlist,
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	201 querystringdict=querystringdict,
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	202 headers=headers,
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	203 bodyfh=bodyfh)
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	204
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	205 class wsgiresponse(object):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	206 """Represents a response to a WSGI request.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	207
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	208 A response consists of a status line, headers, and a body.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	209
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	210 Consumers must populate the ``status`` and ``headers`` fields and
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	211 make a call to a ``setbody*()`` method before the response can be
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	212 issued.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	213
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	214 When it is time to start sending the response over the wire,
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	215 ``sendresponse()`` is called. It handles emitting the header portion
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	216 of the response message. It then yields chunks of body data to be
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	217 written to the peer. Typically, the WSGI application itself calls
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	218 and returns the value from ``sendresponse()``.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	219 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	220
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	221 def __init__(self, req, startresponse):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	222 """Create an empty response tied to a specific request.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	223
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	224 ``req`` is a ``parsedrequest``. ``startresponse`` is the
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	225 ``start_response`` function passed to the WSGI application.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	226 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	227 self._req = req
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	228 self._startresponse = startresponse
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	229
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	230 self.status = None
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	231 self.headers = wsgiheaders.Headers([])
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	232
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	233 self._bodybytes = None
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	234 self._bodygen = None
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	235 self._started = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	236
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	237 def setbodybytes(self, b):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	238 """Define the response body as static bytes."""
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	239 if self._bodybytes is not None or self._bodygen is not None:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	240 raise error.ProgrammingError('cannot define body multiple times')
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	241
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	242 self._bodybytes = b
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	243 self.headers['Content-Length'] = '%d' % len(b)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	244
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	245 def setbodygen(self, gen):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	246 """Define the response body as a generator of bytes."""
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	247 if self._bodybytes is not None or self._bodygen is not None:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	248 raise error.ProgrammingError('cannot define body multiple times')
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	249
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	250 self._bodygen = gen
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	251
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	252 def sendresponse(self):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	253 """Send the generated response to the client.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	254
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	255 Before this is called, ``status`` must be set and one of
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	256 ``setbodybytes()`` or ``setbodygen()`` must be called.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	257
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	258 Calling this method multiple times is not allowed.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	259 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	260 if self._started:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	261 raise error.ProgrammingError('sendresponse() called multiple times')
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	262
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	263 self._started = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	264
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	265 if not self.status:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	266 raise error.ProgrammingError('status line not defined')
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	267
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	268 if self._bodybytes is None and self._bodygen is None:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	269 raise error.ProgrammingError('response body not defined')
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	270
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	271 # Various HTTP clients (notably httplib) won't read the HTTP response
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	272 # until the HTTP request has been sent in full. If servers (us) send a
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	273 # response before the HTTP request has been fully sent, the connection
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	274 # may deadlock because neither end is reading.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	275 #
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	276 # We work around this by "draining" the request data before
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	277 # sending any response in some conditions.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	278 drain = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	279 close = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	280
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	281 # If the client sent Expect: 100-continue, we assume it is smart enough
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	282 # to deal with the server sending a response before reading the request.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	283 # (httplib doesn't do this.)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	284 if self._req.headers.get('Expect', '').lower() == '100-continue':
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	285 pass
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	286 # Only tend to request methods that have bodies. Strictly speaking,
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	287 # we should sniff for a body. But this is fine for our existing
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	288 # WSGI applications.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	289 elif self._req.method not in ('POST', 'PUT'):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	290 pass
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	291 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	292 # If we don't know how much data to read, there's no guarantee
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	293 # that we can drain the request responsibly. The WSGI
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	294 # specification only says that servers should ensure the
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	295 # input stream doesn't overrun the actual request. So there's
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	296 # no guarantee that reading until EOF won't corrupt the stream
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	297 # state.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	298 if not isinstance(self._req.bodyfh, util.cappedreader):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	299 close = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	300 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	301 # We /could/ only drain certain HTTP response codes. But 200 and
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	302 # non-200 wire protocol responses both require draining. Since
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	303 # we have a capped reader in place for all situations where we
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	304 # drain, it is safe to read from that stream. We'll either do
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	305 # a drain or no-op if we're already at EOF.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	306 drain = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	307
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	308 if close:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	309 self.headers['Connection'] = 'Close'
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	310
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	311 if drain:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	312 assert isinstance(self._req.bodyfh, util.cappedreader)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	313 while True:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	314 chunk = self._req.bodyfh.read(32768)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	315 if not chunk:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	316 break
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	317
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	318 self._startresponse(pycompat.sysstr(self.status), self.headers.items())
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	319 if self._bodybytes:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	320 yield self._bodybytes
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	321 elif self._bodygen:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	322 for chunk in self._bodygen:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	323 yield chunk
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	324 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	325 error.ProgrammingError('do not know how to send body')
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	326
5566 d74fc8dec2b4 Less indirection in the WSGI web interface. This simplifies some code, and makes it more compliant with WSGI. Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5563 diff changeset	327 class wsgirequest(object):
26132 9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	328 """Higher-level API for a WSGI request.
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	329
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	330 WSGI applications are invoked with 2 arguments. They are used to
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	331 instantiate instances of this class, which provides higher-level APIs
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	332 for obtaining request parameters, writing HTTP output, etc.
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	333 """
5566 d74fc8dec2b4 Less indirection in the WSGI web interface. This simplifies some code, and makes it more compliant with WSGI. Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5563 diff changeset	334 def __init__(self, wsgienv, start_response):
34512 482d6f6dba91 hgweb: when constructing or adding to a wsgi environ dict, use native strs Augie Fackler <augie@google.com> parents: 27046 diff changeset	335 version = wsgienv[r'wsgi.version']
3673 eb0b4a2d70a9 white space and line break cleanups Thomas Arendsen Hein <thomas@intevation.de> parents: 2859 diff changeset	336 if (version < (1, 0)) or (version >= (2, 0)):
4633 ff7253a0d1da Cleanup of whitespace, indentation and line continuation. Thomas Arendsen Hein <thomas@intevation.de> parents: 4250 diff changeset	337 raise RuntimeError("Unknown and unsupported WSGI version %d.%d"
2506 d0db3462d568 This patch make several WSGI related alterations. Eric Hopper <hopper@omnifarious.org> parents: 2466 diff changeset	338 % version)
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	339
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	340 inp = wsgienv[r'wsgi.input']
36860 290fc4c3d1e0 hgweb: use a capped reader for WSGI input stream Gregory Szorc <gregory.szorc@gmail.com> parents: 36858 diff changeset	341
290fc4c3d1e0 hgweb: use a capped reader for WSGI input stream Gregory Szorc <gregory.szorc@gmail.com> parents: 36858 diff changeset	342 if r'HTTP_CONTENT_LENGTH' in wsgienv:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	343 inp = util.cappedreader(inp, int(wsgienv[r'HTTP_CONTENT_LENGTH']))
36860 290fc4c3d1e0 hgweb: use a capped reader for WSGI input stream Gregory Szorc <gregory.szorc@gmail.com> parents: 36858 diff changeset	344 elif r'CONTENT_LENGTH' in wsgienv:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	345 inp = util.cappedreader(inp, int(wsgienv[r'CONTENT_LENGTH']))
36860 290fc4c3d1e0 hgweb: use a capped reader for WSGI input stream Gregory Szorc <gregory.szorc@gmail.com> parents: 36858 diff changeset	346
34512 482d6f6dba91 hgweb: when constructing or adding to a wsgi environ dict, use native strs Augie Fackler <augie@google.com> parents: 27046 diff changeset	347 self.err = wsgienv[r'wsgi.errors']
482d6f6dba91 hgweb: when constructing or adding to a wsgi environ dict, use native strs Augie Fackler <augie@google.com> parents: 27046 diff changeset	348 self.threaded = wsgienv[r'wsgi.multithread']
482d6f6dba91 hgweb: when constructing or adding to a wsgi environ dict, use native strs Augie Fackler <augie@google.com> parents: 27046 diff changeset	349 self.multiprocess = wsgienv[r'wsgi.multiprocess']
482d6f6dba91 hgweb: when constructing or adding to a wsgi environ dict, use native strs Augie Fackler <augie@google.com> parents: 27046 diff changeset	350 self.run_once = wsgienv[r'wsgi.run_once']
2506 d0db3462d568 This patch make several WSGI related alterations. Eric Hopper <hopper@omnifarious.org> parents: 2466 diff changeset	351 self.env = wsgienv
36864 01f6bba64424 hgweb: remove support for POST form data (BC) Gregory Szorc <gregory.szorc@gmail.com> parents: 36863 diff changeset	352 self.req = parserequestfromenv(wsgienv, inp)
36865 422be99519e5 hgweb: remove support for short query string based aliases (BC) Gregory Szorc <gregory.szorc@gmail.com> parents: 36864 diff changeset	353 self.form = self.req.querystringdict
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	354 self.res = wsgiresponse(self.req, start_response)
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	355 self._start_response = start_response
5993 948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	356 self.server_write = None
2506 d0db3462d568 This patch make several WSGI related alterations. Eric Hopper <hopper@omnifarious.org> parents: 2466 diff changeset	357 self.headers = []
d0db3462d568 This patch make several WSGI related alterations. Eric Hopper <hopper@omnifarious.org> parents: 2466 diff changeset	358
18352 e33b9b92a200 hgweb: pass the actual response body to request.response, not just the length Mads Kiilerich <mads@kiilerich.com> parents: 18351 diff changeset	359 def respond(self, status, type, filename=None, body=None):
34514 528b21b853aa request: coerce content-type to native str Augie Fackler <augie@google.com> parents: 34513 diff changeset	360 if not isinstance(type, str):
528b21b853aa request: coerce content-type to native str Augie Fackler <augie@google.com> parents: 34513 diff changeset	361 type = pycompat.sysstr(type)
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	362 if self._start_response is not None:
34722 95be8928d6b2 hgweb: fill in content-type and content-length as native strings Augie Fackler <augie@google.com> parents: 34514 diff changeset	363 self.headers.append((r'Content-Type', type))
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	364 if filename:
26846 7c1b4840c2cd hgweb: replace some str.split() calls by str.partition() or str.rpartition() Anton Shestakov <av6@dwimlabs.net> parents: 26200 diff changeset	365 filename = (filename.rpartition('/')[-1]
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	366 .replace('\\', '\\\\').replace('"', '\\"'))
764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	367 self.headers.append(('Content-Disposition',
764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	368 'inline; filename="%s"' % filename))
18352 e33b9b92a200 hgweb: pass the actual response body to request.response, not just the length Mads Kiilerich <mads@kiilerich.com> parents: 18351 diff changeset	369 if body is not None:
34722 95be8928d6b2 hgweb: fill in content-type and content-length as native strings Augie Fackler <augie@google.com> parents: 34514 diff changeset	370 self.headers.append((r'Content-Length', str(len(body))))
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	371
5926 15ef6b9c1f2f hgweb: be sure to send a valid content-type for raw files Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5922 diff changeset	372 for k, v in self.headers:
15ef6b9c1f2f hgweb: be sure to send a valid content-type for raw files Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5922 diff changeset	373 if not isinstance(v, str):
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	374 raise TypeError('header value must be string: %r' % (v,))
5926 15ef6b9c1f2f hgweb: be sure to send a valid content-type for raw files Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5922 diff changeset	375
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	376 if isinstance(status, ErrorResponse):
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	377 self.headers.extend(status.headers)
12739 8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	378 if status.code == HTTP_NOT_MODIFIED:
8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	379 # RFC 2616 Section 10.3.5: 304 Not Modified has cases where
8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	380 # it MUST NOT include any headers other than these and no
8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	381 # body
8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	382 self.headers = [(k, v) for (k, v) in self.headers if
8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	383 k in ('Date', 'ETag', 'Expires',
8dcd3203a261 hgweb: don't send a body or illegal headers during 304 response Augie Fackler <durin42@gmail.com> parents: 10263 diff changeset	384 'Cache-Control', 'Vary')]
36288 a0a004b29a51 hgweb: correctly bytes-ify status, not string-ify Augie Fackler <augie@google.com> parents: 34722 diff changeset	385 status = statusmessage(status.code, pycompat.bytestr(status))
5993 948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	386 elif status == 200:
948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	387 status = '200 Script output follows'
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	388 elif isinstance(status, int):
956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	389 status = statusmessage(status)
956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	390
36861 2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	391 # Various HTTP clients (notably httplib) won't read the HTTP
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	392 # response until the HTTP request has been sent in full. If servers
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	393 # (us) send a response before the HTTP request has been fully sent,
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	394 # the connection may deadlock because neither end is reading.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	395 #
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	396 # We work around this by "draining" the request data before
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	397 # sending any response in some conditions.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	398 drain = False
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	399 close = False
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	400
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	401 # If the client sent Expect: 100-continue, we assume it is smart
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	402 # enough to deal with the server sending a response before reading
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	403 # the request. (httplib doesn't do this.)
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	404 if self.env.get(r'HTTP_EXPECT', r'').lower() == r'100-continue':
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	405 pass
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	406 # Only tend to request methods that have bodies. Strictly speaking,
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	407 # we should sniff for a body. But this is fine for our existing
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	408 # WSGI applications.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	409 elif self.env[r'REQUEST_METHOD'] not in (r'POST', r'PUT'):
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	410 pass
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	411 else:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	412 # If we don't know how much data to read, there's no guarantee
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	413 # that we can drain the request responsibly. The WSGI
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	414 # specification only says that servers should ensure the
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	415 # input stream doesn't overrun the actual request. So there's
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	416 # no guarantee that reading until EOF won't corrupt the stream
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	417 # state.
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	418 if not isinstance(self.req.bodyfh, util.cappedreader):
36861 2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	419 close = True
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	420 else:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	421 # We /could/ only drain certain HTTP response codes. But 200
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	422 # and non-200 wire protocol responses both require draining.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	423 # Since we have a capped reader in place for all situations
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	424 # where we drain, it is safe to read from that stream. We'll
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	425 # either do a drain or no-op if we're already at EOF.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	426 drain = True
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	427
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	428 if close:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	429 self.headers.append((r'Connection', r'Close'))
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	430
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	431 if drain:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	432 assert isinstance(self.req.bodyfh, util.cappedreader)
36861 2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	433 while True:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	434 chunk = self.req.bodyfh.read(32768)
36861 2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	435 if not chunk:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	436 break
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	437
36291 af0a19d8812b py3: get bytes-repr of network errors portably Augie Fackler <augie@google.com> parents: 36288 diff changeset	438 self.server_write = self._start_response(
af0a19d8812b py3: get bytes-repr of network errors portably Augie Fackler <augie@google.com> parents: 36288 diff changeset	439 pycompat.sysstr(status), self.headers)
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	440 self._start_response = None
956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	441 self.headers = []
18352 e33b9b92a200 hgweb: pass the actual response body to request.response, not just the length Mads Kiilerich <mads@kiilerich.com> parents: 18351 diff changeset	442 if body is not None:
e33b9b92a200 hgweb: pass the actual response body to request.response, not just the length Mads Kiilerich <mads@kiilerich.com> parents: 18351 diff changeset	443 self.write(body)
e33b9b92a200 hgweb: pass the actual response body to request.response, not just the length Mads Kiilerich <mads@kiilerich.com> parents: 18351 diff changeset	444 self.server_write = None
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	445
5993 948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	446 def write(self, thing):
18351 3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	447 if thing:
3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	448 try:
3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	449 self.server_write(thing)
25660 328739ea70c3 global: mass rewrite to use modern exception syntax Gregory Szorc <gregory.szorc@gmail.com> parents: 18352 diff changeset	450 except socket.error as inst:
18351 3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	451 if inst[0] != errno.ECONNRESET:
3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	452 raise
1159 b6f5a947e62e Change use of global sys.stdout, sys.stdin os.environ to a hgrequest object. Vincent Wagelaar <vincent@ricardis.tudelft.nl> parents: 1143 diff changeset	453
4246 cc81c512a531 avoid _wsgioutputfile <-> _wsgirequest circular reference Alexis S. L. Carvalho <alexis@cecm.usp.br> parents: 3673 diff changeset	454 def flush(self):
cc81c512a531 avoid _wsgioutputfile <-> _wsgirequest circular reference Alexis S. L. Carvalho <alexis@cecm.usp.br> parents: 3673 diff changeset	455 return None
cc81c512a531 avoid _wsgioutputfile <-> _wsgirequest circular reference Alexis S. L. Carvalho <alexis@cecm.usp.br> parents: 3673 diff changeset	456
5566 d74fc8dec2b4 Less indirection in the WSGI web interface. This simplifies some code, and makes it more compliant with WSGI. Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5563 diff changeset	457 def wsgiapplication(app_maker):
5887 41a3fce17625 hgweb: return iterable, add deprecation note Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5886 diff changeset	458 '''For compatibility with old CGI scripts. A plain hgweb() or hgwebdir()
41a3fce17625 hgweb: return iterable, add deprecation note Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5886 diff changeset	459 can and should now be used as a WSGI application.'''
5760 0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	460 application = app_maker()
0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	461 def run_wsgi(env, respond):
5887 41a3fce17625 hgweb: return iterable, add deprecation note Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5886 diff changeset	462 return application(env, respond)
5760 0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	463 return run_wsgi

Mercurial > public > mercurial-scm > hg-stable

annotate mercurial/hgweb/request.py @ 36867:a88d68dc3ee8