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

annotate mercurial/hgweb/request.py @ 36884:ccb70a77f746

hgweb: refactor 304 handling code We had generic code in wsgirequest for handling HTTP 304 responses. We also had a special case for it in the catch all exception handler in the WSGI application. We only ever raise 304 in one place. So, we don't need to treat it specially in the catch all exception handler. But it is useful to validate behavior of 304 responses. We port the code that sends a 304 to use the new response API. We then move the code for screening 304 sanity into the new response API. As part of doing so, we discovered that we would send Content-Length: 0. This is not allowed. So, we fix our response code to not emit that header for empty response bodies. Differential Revision: https://phab.mercurial-scm.org/D2794

author	Gregory Szorc <gregory.szorc@gmail.com>
date	Sat, 10 Mar 2018 18:42:00 -0800
parents	97f44b0720e2
children	b2a3308d6a21

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 statusmessage,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	19 )
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	20
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	21 from ..thirdparty import (
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	22 attr,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	23 )
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	24 from .. import (
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	25 error,
34514 528b21b853aa request: coerce content-type to native str Augie Fackler <augie@google.com> parents: 34513 diff changeset	26 pycompat,
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	27 util,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	28 )
138 c77a679e9cfa Revamped templated hgweb mpm@selenic.com parents: 137 diff changeset	29
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	30 class multidict(object):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	31 """A dict like object that can store multiple values for a key.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	32
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	33 Used to store parsed request parameters.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	34
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	35 This is inspired by WebOb's class of the same name.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	36 """
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	37 def __init__(self):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	38 # Stores (key, value) 2-tuples. This isn't the most efficient. But we
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	39 # don't rely on parameters that much, so it shouldn't be a perf issue.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	40 # we can always add dict for fast lookups.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	41 self._items = []
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	42
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	43 def __getitem__(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	44 """Returns the last set value for a key."""
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	45 for k, v in reversed(self._items):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	46 if k == key:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	47 return v
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	48
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	49 raise KeyError(key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	50
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	51 def __setitem__(self, key, value):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	52 """Replace a values for a key with a new value."""
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	53 try:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	54 del self[key]
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	55 except KeyError:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	56 pass
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	57
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	58 self._items.append((key, value))
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	59
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	60 def __delitem__(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	61 """Delete all values for a key."""
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	62 oldlen = len(self._items)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	63
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	64 self._items[:] = [(k, v) for k, v in self._items if k != key]
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	65
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	66 if oldlen == len(self._items):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	67 raise KeyError(key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	68
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	69 def __contains__(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	70 return any(k == key for k, v in self._items)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	71
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	72 def __len__(self):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	73 return len(self._items)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	74
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	75 def get(self, key, default=None):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	76 try:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	77 return self.__getitem__(key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	78 except KeyError:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	79 return default
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	80
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	81 def add(self, key, value):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	82 """Add a new value for a key. Does not replace existing values."""
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	83 self._items.append((key, value))
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	84
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	85 def getall(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	86 """Obtains all values for a key."""
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	87 return [v for k, v in self._items if k == key]
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	88
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	89 def getone(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	90 """Obtain a single value for a key.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	91
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	92 Raises KeyError if key not defined or it has multiple values set.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	93 """
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	94 vals = self.getall(key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	95
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	96 if not vals:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	97 raise KeyError(key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	98
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	99 if len(vals) > 1:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	100 raise KeyError('multiple values for %r' % key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	101
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	102 return vals[0]
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	103
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	104 def asdictoflists(self):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	105 d = {}
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	106 for k, v in self._items:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	107 if k in d:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	108 d[k].append(v)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	109 else:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	110 d[k] = [v]
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	111
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	112 return d
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	113
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	114 @attr.s(frozen=True)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	115 class parsedrequest(object):
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	116 """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	117
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	118 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	119 """
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	120
36854 16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	121 # Request method.
16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	122 method = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	123 # Full URL for this request.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	124 url = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	125 # 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	126 baseurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	127 # 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	128 # 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	129 advertisedurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	130 advertisedbaseurl = attr.ib()
36873 a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	131 # URL scheme (part before ``://``). e.g. ``http`` or ``https``.
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	132 urlscheme = attr.ib()
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	133 # Value of REMOTE_USER, if set, or None.
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	134 remoteuser = attr.ib()
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	135 # Value of REMOTE_HOST, if set, or None.
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	136 remotehost = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	137 # WSGI application path.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	138 apppath = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	139 # 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	140 dispatchparts = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	141 # 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	142 dispatchpath = attr.ib()
36819 cfb9ef24968c hgweb: use parsed request to construct query parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36817 diff changeset	143 # 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	144 # 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	145 havepathinfo = attr.ib()
36874 8ddb5c354906 hgweb: expose repo name on parsedrequest Gregory Szorc <gregory.szorc@gmail.com> parents: 36873 diff changeset	146 # The name of the repository being accessed.
8ddb5c354906 hgweb: expose repo name on parsedrequest Gregory Szorc <gregory.szorc@gmail.com> parents: 36873 diff changeset	147 reponame = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	148 # 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	149 querystring = attr.ib()
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	150 # multidict of query string parameters.
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	151 qsparams = attr.ib()
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	152 # 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	153 # insensitive keys.
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	154 headers = attr.ib()
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	155 # Request body input stream.
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	156 bodyfh = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	157
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	158 def parserequestfromenv(env, bodyfh):
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	159 """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	160
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	161 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	162 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	163 """
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	164 # 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	165
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	166 # 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	167 # 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	168 # 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	169 #wsgiref.validate.check_environ(env)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	170
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	171 # 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	172 # (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	173 # 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	174 # 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	175 if pycompat.ispy3:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	176 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	177 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	178 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	179
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	180 # 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	181 # the environment variables.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	182 # 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	183 # how URLs are reconstructed.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	184 fullurl = env['wsgi.url_scheme'] + '://'
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	185 advertisedfullurl = fullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	186
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	187 def addport(s):
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	188 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	189 if env['SERVER_PORT'] != '443':
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	190 s += ':' + env['SERVER_PORT']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	191 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	192 if env['SERVER_PORT'] != '80':
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	193 s += ':' + env['SERVER_PORT']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	194
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	195 return s
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	196
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	197 if env.get('HTTP_HOST'):
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	198 fullurl += env['HTTP_HOST']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	199 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	200 fullurl += env['SERVER_NAME']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	201 fullurl = addport(fullurl)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	202
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	203 advertisedfullurl += env['SERVER_NAME']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	204 advertisedfullurl = addport(advertisedfullurl)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	205
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	206 baseurl = fullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	207 advertisedbaseurl = advertisedfullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	208
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	209 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	210 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	211 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	212 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	213
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	214 if env.get('QUERY_STRING'):
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	215 fullurl += '?' + env['QUERY_STRING']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	216 advertisedfullurl += '?' + env['QUERY_STRING']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	217
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	218 # 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	219 # 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	220 # 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	221 # 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	222 # 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	223 # the dispatch path.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	224
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	225 apppath = env['SCRIPT_NAME']
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	226
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	227 if env.get('REPO_NAME'):
36816 0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	228 if not apppath.endswith('/'):
0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	229 apppath += '/'
0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	230
0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	231 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	232
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	233 if 'PATH_INFO' in env:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	234 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	235
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	236 # Strip out repo parts.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	237 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	238 if dispatchparts[:len(repoparts)] == repoparts:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	239 dispatchparts = dispatchparts[len(repoparts):]
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	240 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	241 dispatchparts = []
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	242
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	243 dispatchpath = '/'.join(dispatchparts)
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	244
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	245 querystring = env.get('QUERY_STRING', '')
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	246
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	247 # 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	248 # a dict to facilitate fast lookup.
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	249 qsparams = multidict()
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	250 for k, v in util.urlreq.parseqsl(querystring, keep_blank_values=True):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	251 qsparams.add(k, v)
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	252
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	253 # 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	254 # 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	255 # 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	256 headers = []
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	257 for k, v in env.iteritems():
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	258 if k.startswith('HTTP_'):
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	259 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	260
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	261 headers = wsgiheaders.Headers(headers)
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	262
36853 ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	263 # 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	264 # 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	265 # 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	266 # bytes are available to read.
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	267 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	268 headers['Content-Length'] = env['CONTENT_LENGTH']
ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	269
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	270 # 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	271 # 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	272 #bodyfh = env['wsgi.input']
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	273 #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	274 # 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	275
36854 16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	276 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	277 url=fullurl, baseurl=baseurl,
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	278 advertisedurl=advertisedfullurl,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	279 advertisedbaseurl=advertisedbaseurl,
36873 a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	280 urlscheme=env['wsgi.url_scheme'],
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	281 remoteuser=env.get('REMOTE_USER'),
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	282 remotehost=env.get('REMOTE_HOST'),
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	283 apppath=apppath,
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	284 dispatchparts=dispatchparts, dispatchpath=dispatchpath,
36819 cfb9ef24968c hgweb: use parsed request to construct query parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36817 diff changeset	285 havepathinfo='PATH_INFO' in env,
36874 8ddb5c354906 hgweb: expose repo name on parsedrequest Gregory Szorc <gregory.szorc@gmail.com> parents: 36873 diff changeset	286 reponame=env.get('REPO_NAME'),
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	287 querystring=querystring,
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	288 qsparams=qsparams,
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	289 headers=headers,
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	290 bodyfh=bodyfh)
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	291
36881 16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	292 class offsettrackingwriter(object):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	293 """A file object like object that is append only and tracks write count.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	294
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	295 Instances are bound to a callable. This callable is called with data
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	296 whenever a ``write()`` is attempted.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	297
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	298 Instances track the amount of written data so they can answer ``tell()``
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	299 requests.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	300
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	301 The intent of this class is to wrap the ``write()`` function returned by
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	302 a WSGI ``start_response()`` function. Since ``write()`` is a callable and
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	303 not a file object, it doesn't implement other file object methods.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	304 """
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	305 def __init__(self, writefn):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	306 self._write = writefn
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	307 self._offset = 0
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	308
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	309 def write(self, s):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	310 res = self._write(s)
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	311 # Some Python objects don't report the number of bytes written.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	312 if res is None:
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	313 self._offset += len(s)
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	314 else:
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	315 self._offset += res
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	316
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	317 def flush(self):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	318 pass
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	319
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	320 def tell(self):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	321 return self._offset
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	322
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	323 class wsgiresponse(object):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	324 """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	325
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	326 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	327
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	328 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	329 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	330 issued.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	331
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	332 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	333 ``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	334 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	335 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	336 and returns the value from ``sendresponse()``.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	337 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	338
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	339 def __init__(self, req, startresponse):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	340 """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	341
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	342 ``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	343 ``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	344 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	345 self._req = req
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	346 self._startresponse = startresponse
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	347
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	348 self.status = None
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	349 self.headers = wsgiheaders.Headers([])
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	350
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	351 self._bodybytes = None
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	352 self._bodygen = None
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	353 self._bodywillwrite = False
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	354 self._started = False
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	355 self._bodywritefn = None
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	356
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	357 def _verifybody(self):
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	358 if (self._bodybytes is not None or self._bodygen is not None
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	359 or self._bodywillwrite):
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	360 raise error.ProgrammingError('cannot define body multiple times')
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	361
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	362 def setbodybytes(self, b):
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	363 """Define the response body as static bytes.
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	364
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	365 The empty string signals that there is no response body.
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	366 """
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	367 self._verifybody()
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	368 self._bodybytes = b
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	369 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	370
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	371 def setbodygen(self, gen):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	372 """Define the response body as a generator of bytes."""
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	373 self._verifybody()
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	374 self._bodygen = gen
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	375
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	376 def setbodywillwrite(self):
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	377 """Signal an intent to use write() to emit the response body.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	378
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	379 This is the least preferred way to send a body.
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	380
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	381 It is preferred for WSGI applications to emit a generator of chunks
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	382 constituting the response body. However, some consumers can't emit
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	383 data this way. So, WSGI provides a way to obtain a ``write(data)``
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	384 function that can be used to synchronously perform an unbuffered
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	385 write.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	386
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	387 Calling this function signals an intent to produce the body in this
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	388 manner.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	389 """
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	390 self._verifybody()
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	391 self._bodywillwrite = True
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	392
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	393 def sendresponse(self):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	394 """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	395
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	396 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	397 ``setbodybytes()`` or ``setbodygen()`` must be called.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	398
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	399 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	400 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	401 if self._started:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	402 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	403
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	404 self._started = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	405
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	406 if not self.status:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	407 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	408
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	409 if (self._bodybytes is None and self._bodygen is None
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	410 and not self._bodywillwrite):
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	411 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	412
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	413 # RFC 7232 Section 4.1 states that a 304 MUST generate one of
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	414 # {Cache-Control, Content-Location, Date, ETag, Expires, Vary}
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	415 # and SHOULD NOT generate other headers unless they could be used
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	416 # to guide cache updates. Furthermore, RFC 7230 Section 3.3.2
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	417 # states that no response body can be issued. Content-Length can
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	418 # be sent. But if it is present, it should be the size of the response
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	419 # that wasn't transferred.
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	420 if self.status.startswith('304 '):
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	421 # setbodybytes('') will set C-L to 0. This doesn't conform with the
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	422 # spec. So remove it.
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	423 if self.headers.get('Content-Length') == '0':
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	424 del self.headers['Content-Length']
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	425
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	426 # Strictly speaking, this is too strict. But until it causes
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	427 # problems, let's be strict.
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	428 badheaders = {k for k in self.headers.keys()
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	429 if k.lower() not in ('date', 'etag', 'expires',
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	430 'cache-control',
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	431 'content-location',
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	432 'vary')}
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	433 if badheaders:
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	434 raise error.ProgrammingError(
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	435 'illegal header on 304 response: %s' %
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	436 ', '.join(sorted(badheaders)))
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	437
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	438 if self._bodygen is not None or self._bodywillwrite:
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	439 raise error.ProgrammingError("must use setbodybytes('') with "
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	440 "304 responses")
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	441
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	442 # 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	443 # 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	444 # 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	445 # 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	446 #
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	447 # 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	448 # sending any response in some conditions.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	449 drain = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	450 close = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	451
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	452 # 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	453 # 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	454 # (httplib doesn't do this.)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	455 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	456 pass
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	457 # 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	458 # 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	459 # WSGI applications.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	460 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	461 pass
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	462 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	463 # 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	464 # 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	465 # 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	466 # 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	467 # 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	468 # state.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	469 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	470 close = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	471 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	472 # 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	473 # 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	474 # 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	475 # 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	476 # 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	477 drain = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	478
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	479 if close:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	480 self.headers['Connection'] = 'Close'
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	481
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	482 if drain:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	483 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	484 while True:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	485 chunk = self._req.bodyfh.read(32768)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	486 if not chunk:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	487 break
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	488
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	489 write = self._startresponse(pycompat.sysstr(self.status),
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	490 self.headers.items())
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	491
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	492 if self._bodybytes:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	493 yield self._bodybytes
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	494 elif self._bodygen:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	495 for chunk in self._bodygen:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	496 yield chunk
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	497 elif self._bodywillwrite:
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	498 self._bodywritefn = write
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	499 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	500 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	501
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	502 def getbodyfile(self):
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	503 """Obtain a file object like object representing the response body.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	504
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	505 For this to work, you must call ``setbodywillwrite()`` and then
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	506 ``sendresponse()`` first. ``sendresponse()`` is a generator and the
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	507 function won't run to completion unless the generator is advanced. The
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	508 generator yields not items. The easiest way to consume it is with
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	509 ``list(res.sendresponse())``, which should resolve to an empty list -
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	510 ``[]``.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	511 """
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	512 if not self._bodywillwrite:
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	513 raise error.ProgrammingError('must call setbodywillwrite() first')
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	514
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	515 if not self._started:
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	516 raise error.ProgrammingError('must call sendresponse() first; did '
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	517 'you remember to consume it since it '
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	518 'is a generator?')
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	519
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	520 assert self._bodywritefn
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	521 return offsettrackingwriter(self._bodywritefn)
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	522
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	523 class wsgirequest(object):
26132 9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	524 """Higher-level API for a WSGI request.
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	525
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	526 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	527 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	528 for obtaining request parameters, writing HTTP output, etc.
9df8c729e2e7 hgweb: add some documentation Gregory Szorc <gregory.szorc@gmail.com> parents: 25660 diff changeset	529 """
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	530 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	531 version = wsgienv[r'wsgi.version']
3673 eb0b4a2d70a9 white space and line break cleanups Thomas Arendsen Hein <thomas@intevation.de> parents: 2859 diff changeset	532 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	533 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	534 % version)
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	535
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	536 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	537
290fc4c3d1e0 hgweb: use a capped reader for WSGI input stream Gregory Szorc <gregory.szorc@gmail.com> parents: 36858 diff changeset	538 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	539 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	540 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	541 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	542
34512 482d6f6dba91 hgweb: when constructing or adding to a wsgi environ dict, use native strs Augie Fackler <augie@google.com> parents: 27046 diff changeset	543 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	544 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	545 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	546 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	547 self.env = wsgienv
36864 01f6bba64424 hgweb: remove support for POST form data (BC) Gregory Szorc <gregory.szorc@gmail.com> parents: 36863 diff changeset	548 self.req = parserequestfromenv(wsgienv, inp)
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	549 self.res = wsgiresponse(self.req, start_response)
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	550 self._start_response = start_response
5993 948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	551 self.server_write = None
2506 d0db3462d568 This patch make several WSGI related alterations. Eric Hopper <hopper@omnifarious.org> parents: 2466 diff changeset	552 self.headers = []
d0db3462d568 This patch make several WSGI related alterations. Eric Hopper <hopper@omnifarious.org> parents: 2466 diff changeset	553
18352 e33b9b92a200 hgweb: pass the actual response body to request.response, not just the length Mads Kiilerich <mads@kiilerich.com> parents: 18351 diff changeset	554 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	555 if not isinstance(type, str):
528b21b853aa request: coerce content-type to native str Augie Fackler <augie@google.com> parents: 34513 diff changeset	556 type = pycompat.sysstr(type)
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	557 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	558 self.headers.append((r'Content-Type', type))
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	559 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	560 filename = (filename.rpartition('/')[-1]
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	561 .replace('\\', '\\\\').replace('"', '\\"'))
764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	562 self.headers.append(('Content-Disposition',
764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	563 '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	564 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	565 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	566
5926 15ef6b9c1f2f hgweb: be sure to send a valid content-type for raw files Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5922 diff changeset	567 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	568 if not isinstance(v, str):
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	569 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	570
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	571 if isinstance(status, ErrorResponse):
18348 764a758780b6 hgweb: simplify wsgirequest header handling Mads Kiilerich <mads@kiilerich.com> parents: 18347 diff changeset	572 self.headers.extend(status.headers)
36288 a0a004b29a51 hgweb: correctly bytes-ify status, not string-ify Augie Fackler <augie@google.com> parents: 34722 diff changeset	573 status = statusmessage(status.code, pycompat.bytestr(status))
5993 948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	574 elif status == 200:
948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	575 status = '200 Script output follows'
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	576 elif isinstance(status, int):
956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	577 status = statusmessage(status)
956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	578
36861 2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	579 # 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	580 # 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	581 # (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	582 # 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	583 #
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	584 # 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	585 # sending any response in some conditions.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	586 drain = False
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	587 close = False
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	588
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	589 # 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	590 # 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	591 # the request. (httplib doesn't do this.)
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	592 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	593 pass
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	594 # 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	595 # 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	596 # WSGI applications.
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	597 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	598 pass
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	599 else:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	600 # 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	601 # 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	602 # 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	603 # 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	604 # 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	605 # state.
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	606 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	607 close = True
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	608 else:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	609 # 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	610 # 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	611 # 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	612 # 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	613 # 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	614 drain = True
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	615
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	616 if close:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	617 self.headers.append((r'Connection', r'Close'))
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	618
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	619 if drain:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	620 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	621 while True:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	622 chunk = self.req.bodyfh.read(32768)
36861 2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	623 if not chunk:
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	624 break
2cdf47e14c30 hgweb: refactor the request draining code Gregory Szorc <gregory.szorc@gmail.com> parents: 36860 diff changeset	625
36291 af0a19d8812b py3: get bytes-repr of network errors portably Augie Fackler <augie@google.com> parents: 36288 diff changeset	626 self.server_write = self._start_response(
af0a19d8812b py3: get bytes-repr of network errors portably Augie Fackler <augie@google.com> parents: 36288 diff changeset	627 pycompat.sysstr(status), self.headers)
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	628 self._start_response = None
956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	629 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	630 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	631 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	632 self.server_write = None
5888 956afc025c0f hgweb: separate out start_response() calling Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5887 diff changeset	633
5993 948a41e77902 hgweb: explicit response status Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5930 diff changeset	634 def write(self, thing):
18351 3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	635 if thing:
3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	636 try:
3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	637 self.server_write(thing)
25660 328739ea70c3 global: mass rewrite to use modern exception syntax Gregory Szorc <gregory.szorc@gmail.com> parents: 18352 diff changeset	638 except socket.error as inst:
18351 3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	639 if inst[0] != errno.ECONNRESET:
3fbdbeab38cc hgweb: don't pass empty response chunks on Mads Kiilerich <mads@kiilerich.com> parents: 18350 diff changeset	640 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	641
4246 cc81c512a531 avoid _wsgioutputfile <-> _wsgirequest circular reference Alexis S. L. Carvalho <alexis@cecm.usp.br> parents: 3673 diff changeset	642 def flush(self):
cc81c512a531 avoid _wsgioutputfile <-> _wsgirequest circular reference Alexis S. L. Carvalho <alexis@cecm.usp.br> parents: 3673 diff changeset	643 return None
cc81c512a531 avoid _wsgioutputfile <-> _wsgirequest circular reference Alexis S. L. Carvalho <alexis@cecm.usp.br> parents: 3673 diff changeset	644
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	645 def wsgiapplication(app_maker):
5887 41a3fce17625 hgweb: return iterable, add deprecation note Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5886 diff changeset	646 '''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	647 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	648 application = app_maker()
0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	649 def run_wsgi(env, respond):
5887 41a3fce17625 hgweb: return iterable, add deprecation note Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5886 diff changeset	650 return application(env, respond)
5760 0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	651 return run_wsgi

Mercurial > public > mercurial-scm > hg-stable

annotate mercurial/hgweb/request.py @ 36884:ccb70a77f746