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

annotate mercurial/hgweb/request.py @ 50766:0caa653856a9 draft stable

hgweb: encode WSGI environment using the ISO-8859-1 codec The WSGI specification (PEP 3333) specifies that on Python 3 all strings passed by the server must be of type str with code points encodable using the ISO 8859-1 codec. For some reason, I introduced a bug in 2632c1ed8f34 by applying the reverse change. Maybe I got confused because PEP 3333 says that arbitrary operating system environment variables may be contained in the WSGI environment and therefore we need to handle the WSGI environment variables like we would handle operating system environment variables. The bug mentioned in the previous paragraph and fixed by this changeset manifested e.g. in the path of the URL being encoded in the wrong way. Browsers encode non-ASCII bytes with the percent-encoding. WSGI servers will decode the percent-encoded bytes and pass them to the application as strings where each byte is mapped to the corresponding code point with the same ordinal (i.e. it is decoded using the ISO-8859-1 codec). Mercurial uses the bytes type for these strings (which makes much more sense), so we need to encode it again using the ISO-8859-1 codec. If we use another codec, it can result in nonsense.

author	Manuel Jacob <me@manueljacob.de>
date	Mon, 07 Aug 2023 23:12:02 +0200
parents	fda5a4b853ab
children

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>
46819 d4ba4d51f85f contributor: change mentions of mpm to olivia Rapha?l Gom?s <rgomes@octobus.net> parents: 45957 diff changeset	4 # Copyright 2005, 2006 Olivia Mackall <olivia@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
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	10 # import wsgiref.validate
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	11
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	12 from ..thirdparty import attr
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	13 from .. import (
44825 2632c1ed8f34 hgweb: encode WSGI environment like OS environment Manuel Jacob <me@manueljacob.de> parents: 44824 diff changeset	14 encoding,
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	15 error,
34514 528b21b853aa request: coerce content-type to native str Augie Fackler <augie@google.com> parents: 34513 diff changeset	16 pycompat,
27046 37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	17 util,
37fcfe52c68c hgweb: use absolute_import Yuya Nishihara <yuya@tcha.org> parents: 26846 diff changeset	18 )
46907 ffd3e823a7e5 urlutil: extract `url` related code from `util` into the new module Pierre-Yves David <pierre-yves.david@octobus.net> parents: 46819 diff changeset	19 from ..utils import (
ffd3e823a7e5 urlutil: extract `url` related code from `util` into the new module Pierre-Yves David <pierre-yves.david@octobus.net> parents: 46819 diff changeset	20 urlutil,
ffd3e823a7e5 urlutil: extract `url` related code from `util` into the new module Pierre-Yves David <pierre-yves.david@octobus.net> parents: 46819 diff changeset	21 )
138 c77a679e9cfa Revamped templated hgweb mpm@selenic.com parents: 137 diff changeset	22
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	23
49037 642e31cb55f0 py3: use class X: instead of class X(object): Gregory Szorc <gregory.szorc@gmail.com> parents: 49004 diff changeset	24 class multidict:
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	25 """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	26
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	27 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	28
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	29 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	30 """
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	31
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	32 def __init__(self):
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	33 self._items = {}
36868 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 def __getitem__(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	36 """Returns the last set value for a key."""
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	37 return self._items[key][-1]
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	38
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	39 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	40 """Replace a values for a key with a new value."""
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	41 self._items[key] = [value]
36868 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 __delitem__(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	44 """Delete all values for a key."""
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	45 del self._items[key]
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	46
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	47 def __contains__(self, key):
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	48 return key in self._items
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	49
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	50 def __len__(self):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	51 return len(self._items)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	52
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	53 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	54 try:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	55 return self.__getitem__(key)
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	56 except KeyError:
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	57 return default
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	58
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	59 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	60 """Add a new value for a key. Does not replace existing values."""
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	61 self._items.setdefault(key, []).append(value)
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	62
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	63 def getall(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	64 """Obtains all values for a key."""
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	65 return self._items.get(key, [])
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	66
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	67 def getone(self, key):
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	68 """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	69
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	70 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	71 """
37000 44467a4d472f hgweb: refactor multirequest to be a dict of lists Gregory Szorc <gregory.szorc@gmail.com> parents: 36917 diff changeset	72 vals = self._items[key]
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	73
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	74 if len(vals) > 1:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	75 raise KeyError(b'multiple values for %r' % key)
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	76
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	77 return vals[0]
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	78
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	79 def asdictoflists(self):
49004 f254fc73d956 global: bulk replace simple pycompat.iteritems(x) with x.items() Gregory Szorc <gregory.szorc@gmail.com> parents: 48996 diff changeset	80 return {k: list(v) for k, v in self._items.items()}
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	81
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	82
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	83 @attr.s(frozen=True)
49037 642e31cb55f0 py3: use class X: instead of class X(object): Gregory Szorc <gregory.szorc@gmail.com> parents: 49004 diff changeset	84 class parsedrequest:
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	85 """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	86
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	87 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	88 """
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	89
36854 16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	90 # Request method.
16292bbda39c hgweb: store and use request method on parsed request Gregory Szorc <gregory.szorc@gmail.com> parents: 36853 diff changeset	91 method = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	92 # Full URL for this request.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	93 url = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	94 # 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	95 baseurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	96 # 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	97 # 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	98 advertisedurl = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	99 advertisedbaseurl = attr.ib()
36873 a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	100 # 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	101 urlscheme = attr.ib()
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	102 # 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	103 remoteuser = attr.ib()
a755fd3b7146 hgweb: expose URL scheme and REMOTE_* attributes Gregory Szorc <gregory.szorc@gmail.com> parents: 36872 diff changeset	104 # 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	105 remotehost = attr.ib()
36905 e67a2e05fa8a hgweb: clarify that apppath begins with a forward slash Gregory Szorc <gregory.szorc@gmail.com> parents: 36904 diff changeset	106 # Relative WSGI application path. If defined, will begin with a
e67a2e05fa8a hgweb: clarify that apppath begins with a forward slash Gregory Szorc <gregory.szorc@gmail.com> parents: 36904 diff changeset	107 # ``/``.
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	108 apppath = attr.ib()
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	109 # 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	110 dispatchparts = attr.ib()
36904 d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	111 # URL path component (no query string) used for dispatch. Can be
d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	112 # ``None`` to signal no path component given to the request, an
d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	113 # empty string to signal a request to the application's root URL,
d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	114 # or a string not beginning with ``/`` containing the requested
d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	115 # path under the application.
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	116 dispatchpath = attr.ib()
36874 8ddb5c354906 hgweb: expose repo name on parsedrequest Gregory Szorc <gregory.szorc@gmail.com> parents: 36873 diff changeset	117 # The name of the repository being accessed.
8ddb5c354906 hgweb: expose repo name on parsedrequest Gregory Szorc <gregory.szorc@gmail.com> parents: 36873 diff changeset	118 reponame = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	119 # 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	120 querystring = attr.ib()
36868 ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	121 # 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	122 qsparams = attr.ib()
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	123 # 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	124 # insensitive keys.
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	125 headers = attr.ib()
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	126 # Request body input stream.
da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	127 bodyfh = attr.ib()
36915 84110a1d0f7d hgweb: store the raw WSGI environment dict Gregory Szorc <gregory.szorc@gmail.com> parents: 36914 diff changeset	128 # WSGI environment dict, unmodified.
84110a1d0f7d hgweb: store the raw WSGI environment dict Gregory Szorc <gregory.szorc@gmail.com> parents: 36914 diff changeset	129 rawenv = attr.ib()
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	130
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	131
37818 877185de62cf hgweb: reuse body file object when hgwebdir calls hgweb (issue5851) Gregory Szorc <gregory.szorc@gmail.com> parents: 37616 diff changeset	132 def parserequestfromenv(env, reponame=None, altbaseurl=None, bodyfh=None):
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	133 """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	134
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	135 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	136 parses the environment variables into a data structure.
36903 d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	137
d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	138 If ``reponame`` is defined, the leading path components matching that
d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	139 string are effectively shifted from ``PATH_INFO`` to ``SCRIPT_NAME``.
d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	140 This simulates the world view of a WSGI application that processes
d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	141 requests from the base URL of a repo.
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	142
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	143 If ``altbaseurl`` (typically comes from ``web.baseurl`` config option)
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	144 is defined, it is used - instead of the WSGI environment variables - for
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	145 constructing URL components up to and including the WSGI application path.
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	146 For example, if the current WSGI application is at ``/repo`` and a request
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	147 is made to ``/rev/@`` with this argument set to
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	148 ``http://myserver:9000/prefix``, the URL and path components will resolve as
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	149 if the request were to ``http://myserver:9000/prefix/rev/@``. In other
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	150 words, ``wsgi.url_scheme``, ``SERVER_NAME``, ``SERVER_PORT``, and
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	151 ``SCRIPT_NAME`` are all effectively replaced by components from this URL.
37818 877185de62cf hgweb: reuse body file object when hgwebdir calls hgweb (issue5851) Gregory Szorc <gregory.szorc@gmail.com> parents: 37616 diff changeset	152
877185de62cf hgweb: reuse body file object when hgwebdir calls hgweb (issue5851) Gregory Szorc <gregory.szorc@gmail.com> parents: 37616 diff changeset	153 ``bodyfh`` can be used to specify a file object to read the request body
877185de62cf hgweb: reuse body file object when hgwebdir calls hgweb (issue5851) Gregory Szorc <gregory.szorc@gmail.com> parents: 37616 diff changeset	154 from. If not defined, ``wsgi.input`` from the environment dict is used.
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	155 """
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	156 # PEP 3333 defines the WSGI spec and is a useful reference for this code.
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	157
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	158 # 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	159 # 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	160 # TODO enable this once we fix internal violations.
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	161 # wsgiref.validate.check_environ(env)
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	162
48996 7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	163 # PEP-0333 states that environment keys and values are native strings.
7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	164 # The code points for the Unicode strings on Python 3 must be between
7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	165 # \00000-\000FF. We deal with bytes in Mercurial, so mass convert string
7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	166 # keys and values to bytes.
7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	167 def tobytes(s):
7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	168 if not isinstance(s, str):
7eebe5630bcc hgweb: remove Python 3 conditional Gregory Szorc <gregory.szorc@gmail.com> parents: 48966 diff changeset	169 return s
50766 0caa653856a9 hgweb: encode WSGI environment using the ISO-8859-1 codec Manuel Jacob <me@manueljacob.de> parents: 49928 diff changeset	170 return s.encode('iso8859-1')
44825 2632c1ed8f34 hgweb: encode WSGI environment like OS environment Manuel Jacob <me@manueljacob.de> parents: 44824 diff changeset	171
49004 f254fc73d956 global: bulk replace simple pycompat.iteritems(x) with x.items() Gregory Szorc <gregory.szorc@gmail.com> parents: 48996 diff changeset	172 env = {tobytes(k): tobytes(v) for k, v in env.items()}
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	173
37616 5e81cf9651c1 hgweb: fallback to checking wsgireq.env for REPO_NAME for 3rd party hosting Matt Harbison <matt_harbison@yahoo.com> parents: 37606 diff changeset	174 # Some hosting solutions are emulating hgwebdir, and dispatching directly
5e81cf9651c1 hgweb: fallback to checking wsgireq.env for REPO_NAME for 3rd party hosting Matt Harbison <matt_harbison@yahoo.com> parents: 37606 diff changeset	175 # to an hgweb instance using this environment variable. This was always
5e81cf9651c1 hgweb: fallback to checking wsgireq.env for REPO_NAME for 3rd party hosting Matt Harbison <matt_harbison@yahoo.com> parents: 37606 diff changeset	176 # checked prior to d7fd203e36cc; keep doing so to avoid breaking them.
5e81cf9651c1 hgweb: fallback to checking wsgireq.env for REPO_NAME for 3rd party hosting Matt Harbison <matt_harbison@yahoo.com> parents: 37606 diff changeset	177 if not reponame:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	178 reponame = env.get(b'REPO_NAME')
37616 5e81cf9651c1 hgweb: fallback to checking wsgireq.env for REPO_NAME for 3rd party hosting Matt Harbison <matt_harbison@yahoo.com> parents: 37606 diff changeset	179
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	180 if altbaseurl:
46907 ffd3e823a7e5 urlutil: extract `url` related code from `util` into the new module Pierre-Yves David <pierre-yves.david@octobus.net> parents: 46819 diff changeset	181 altbaseurl = urlutil.url(altbaseurl)
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	182
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	183 # 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	184 # the environment variables.
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	185 # 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	186 # how URLs are reconstructed.
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	187 fullurl = env[b'wsgi.url_scheme'] + b'://'
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	188
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	189 if altbaseurl and altbaseurl.scheme:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	190 advertisedfullurl = altbaseurl.scheme + b'://'
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	191 else:
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	192 advertisedfullurl = fullurl
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	193
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	194 def addport(s, port):
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	195 if s.startswith(b'https://'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	196 if port != b'443':
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	197 s += b':' + port
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	198 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	199 if port != b'80':
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	200 s += b':' + port
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	201
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	202 return s
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	203
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	204 if env.get(b'HTTP_HOST'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	205 fullurl += env[b'HTTP_HOST']
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	206 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	207 fullurl += env[b'SERVER_NAME']
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	208 fullurl = addport(fullurl, env[b'SERVER_PORT'])
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	209
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	210 if altbaseurl and altbaseurl.host:
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	211 advertisedfullurl += altbaseurl.host
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	212
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	213 if altbaseurl.port:
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	214 port = altbaseurl.port
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	215 elif altbaseurl.scheme == b'http' and not altbaseurl.port:
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	216 port = b'80'
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	217 elif altbaseurl.scheme == b'https' and not altbaseurl.port:
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	218 port = b'443'
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	219 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	220 port = env[b'SERVER_PORT']
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	221
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	222 advertisedfullurl = addport(advertisedfullurl, port)
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	223 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	224 advertisedfullurl += env[b'SERVER_NAME']
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	225 advertisedfullurl = addport(advertisedfullurl, env[b'SERVER_PORT'])
36814 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 baseurl = fullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	228 advertisedbaseurl = advertisedfullurl
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	229
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	230 fullurl += util.urlreq.quote(env.get(b'SCRIPT_NAME', b''))
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	231 fullurl += util.urlreq.quote(env.get(b'PATH_INFO', b''))
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	232
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	233 if altbaseurl:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	234 path = altbaseurl.path or b''
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	235 if path and not path.startswith(b'/'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	236 path = b'/' + path
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	237 advertisedfullurl += util.urlreq.quote(path)
219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	238 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	239 advertisedfullurl += util.urlreq.quote(env.get(b'SCRIPT_NAME', b''))
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	240
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	241 advertisedfullurl += util.urlreq.quote(env.get(b'PATH_INFO', b''))
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	242
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	243 if env.get(b'QUERY_STRING'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	244 fullurl += b'?' + env[b'QUERY_STRING']
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	245 advertisedfullurl += b'?' + env[b'QUERY_STRING']
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	246
36903 d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	247 # If ``reponame`` is defined, that must be a prefix on PATH_INFO
d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	248 # that represents the repository being dispatched to. When computing
d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	249 # the dispatch info, we ignore these leading path components.
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	250
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	251 if altbaseurl:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	252 apppath = altbaseurl.path or b''
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	253 if apppath and not apppath.startswith(b'/'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	254 apppath = b'/' + apppath
36906 219b23359f4c hgweb: support constructing URLs from an alternate base URL Gregory Szorc <gregory.szorc@gmail.com> parents: 36905 diff changeset	255 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	256 apppath = env.get(b'SCRIPT_NAME', b'')
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	257
36903 d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	258 if reponame:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	259 repoprefix = b'/' + reponame.strip(b'/')
36816 0031e972ded2 hgweb: use the parsed application path directly Gregory Szorc <gregory.szorc@gmail.com> parents: 36814 diff changeset	260
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	261 if not env.get(b'PATH_INFO'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	262 raise error.ProgrammingError(b'reponame requires PATH_INFO')
36903 d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	263
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	264 if not env[b'PATH_INFO'].startswith(repoprefix):
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	265 raise error.ProgrammingError(
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	266 b'PATH_INFO does not begin with repo '
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	267 b'name: %s (%s)' % (env[b'PATH_INFO'], reponame)
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	268 )
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	269
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	270 dispatchpath = env[b'PATH_INFO'][len(repoprefix) :]
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	271
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	272 if dispatchpath and not dispatchpath.startswith(b'/'):
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	273 raise error.ProgrammingError(
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	274 b'reponame prefix of PATH_INFO does '
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	275 b'not end at path delimiter: %s (%s)'
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	276 % (env[b'PATH_INFO'], reponame)
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	277 )
36903 d7fd203e36cc hgweb: refactor repository name URL parsing Gregory Szorc <gregory.szorc@gmail.com> parents: 36902 diff changeset	278
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	279 apppath = apppath.rstrip(b'/') + repoprefix
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	280 dispatchparts = dispatchpath.strip(b'/').split(b'/')
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	281 dispatchpath = b'/'.join(dispatchparts)
36904 d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	282
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	283 elif b'PATH_INFO' in env:
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	284 if env[b'PATH_INFO'].strip(b'/'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	285 dispatchparts = env[b'PATH_INFO'].strip(b'/').split(b'/')
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	286 dispatchpath = b'/'.join(dispatchparts)
36904 d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	287 else:
d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	288 dispatchparts = []
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	289 dispatchpath = b''
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	290 else:
69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	291 dispatchparts = []
36904 d0b0fedbfb53 hgweb: change how dispatch path is reported Gregory Szorc <gregory.szorc@gmail.com> parents: 36903 diff changeset	292 dispatchpath = None
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	293
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	294 querystring = env.get(b'QUERY_STRING', b'')
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	295
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	296 # 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	297 # 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	298 qsparams = multidict()
ec0af9c59270 hgweb: use a multidict for holding query string parameters Gregory Szorc <gregory.szorc@gmail.com> parents: 36867 diff changeset	299 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	300 qsparams.add(k, v)
36817 3c15b84ab66c hgweb: teach WSGI parser about query strings Gregory Szorc <gregory.szorc@gmail.com> parents: 36816 diff changeset	301
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	302 # 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	303 # 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	304 # 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	305 headers = []
49004 f254fc73d956 global: bulk replace simple pycompat.iteritems(x) with x.items() Gregory Szorc <gregory.szorc@gmail.com> parents: 48996 diff changeset	306 for k, v in env.items():
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	307 if k.startswith(b'HTTP_'):
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	308 headers.append((k[len(b'HTTP_') :].replace(b'_', b'-'), v))
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	309
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	310 from . import wsgiheaders # avoid cycle
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	311
36822 f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	312 headers = wsgiheaders.Headers(headers)
f9078c6caeb6 hgweb: parse and store HTTP request headers Gregory Szorc <gregory.szorc@gmail.com> parents: 36819 diff changeset	313
36853 ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	314 # 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	315 # 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	316 # 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	317 # bytes are available to read.
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	318 if b'CONTENT_LENGTH' in env and b'HTTP_CONTENT_LENGTH' not in env:
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	319 headers[b'Content-Length'] = env[b'CONTENT_LENGTH']
36853 ed0456fde625 hgweb: handle CONTENT_LENGTH Gregory Szorc <gregory.szorc@gmail.com> parents: 36822 diff changeset	320
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	321 if b'CONTENT_TYPE' in env and b'HTTP_CONTENT_TYPE' not in env:
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	322 headers[b'Content-Type'] = env[b'CONTENT_TYPE']
37052 55e901396005 hgweb: also set Content-Type header Gregory Szorc <gregory.szorc@gmail.com> parents: 37000 diff changeset	323
37818 877185de62cf hgweb: reuse body file object when hgwebdir calls hgweb (issue5851) Gregory Szorc <gregory.szorc@gmail.com> parents: 37616 diff changeset	324 if bodyfh is None:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	325 bodyfh = env[b'wsgi.input']
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	326 if b'Content-Length' in headers:
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	327 bodyfh = util.cappedreader(
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	328 bodyfh, int(headers[b'Content-Length'] or b'0')
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	329 )
36863 da4e2f87167d hgweb: expose input stream on parsed WSGI request object Gregory Szorc <gregory.szorc@gmail.com> parents: 36862 diff changeset	330
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	331 return parsedrequest(
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	332 method=env[b'REQUEST_METHOD'],
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	333 url=fullurl,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	334 baseurl=baseurl,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	335 advertisedurl=advertisedfullurl,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	336 advertisedbaseurl=advertisedbaseurl,
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	337 urlscheme=env[b'wsgi.url_scheme'],
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	338 remoteuser=env.get(b'REMOTE_USER'),
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	339 remotehost=env.get(b'REMOTE_HOST'),
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	340 apppath=apppath,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	341 dispatchparts=dispatchparts,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	342 dispatchpath=dispatchpath,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	343 reponame=reponame,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	344 querystring=querystring,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	345 qsparams=qsparams,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	346 headers=headers,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	347 bodyfh=bodyfh,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	348 rawenv=env,
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	349 )
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	350
36814 69b2d0900cd7 hgweb: parse WSGI request into a data structure Gregory Szorc <gregory.szorc@gmail.com> parents: 36736 diff changeset	351
49037 642e31cb55f0 py3: use class X: instead of class X(object): Gregory Szorc <gregory.szorc@gmail.com> parents: 49004 diff changeset	352 class offsettrackingwriter:
36881 16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	353 """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	354
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	355 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	356 whenever a ``write()`` is attempted.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	357
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	358 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	359 requests.
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	360
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	361 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	362 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	363 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	364 """
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	365
36881 16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	366 def __init__(self, writefn):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	367 self._write = writefn
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	368 self._offset = 0
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	369
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	370 def write(self, s):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	371 res = self._write(s)
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	372 # 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	373 if res is None:
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	374 self._offset += len(s)
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	375 else:
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	376 self._offset += res
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	377
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	378 def flush(self):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	379 pass
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	380
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	381 def tell(self):
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	382 return self._offset
16499427f6de hgweb: refactor fake file object proxy for archiving Gregory Szorc <gregory.szorc@gmail.com> parents: 36874 diff changeset	383
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	384
49037 642e31cb55f0 py3: use class X: instead of class X(object): Gregory Szorc <gregory.szorc@gmail.com> parents: 49004 diff changeset	385 class wsgiresponse:
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	386 """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	387
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	388 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	389
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	390 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	391 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	392 issued.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	393
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	394 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	395 ``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	396 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	397 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	398 and returns the value from ``sendresponse()``.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	399 """
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 def __init__(self, req, startresponse):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	402 """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	403
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	404 ``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	405 ``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	406 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	407 self._req = req
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	408 self._startresponse = startresponse
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	409
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	410 self.status = None
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	411 from . import wsgiheaders # avoid cycle
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	412
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	413 self.headers = wsgiheaders.Headers([])
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	414
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	415 self._bodybytes = None
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	416 self._bodygen = None
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	417 self._bodywillwrite = False
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	418 self._started = False
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	419 self._bodywritefn = None
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	420
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	421 def _verifybody(self):
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	422 if (
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	423 self._bodybytes is not None
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	424 or self._bodygen is not None
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	425 or self._bodywillwrite
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	426 ):
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	427 raise error.ProgrammingError(b'cannot define body multiple times')
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	428
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	429 def setbodybytes(self, b):
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	430 """Define the response body as static bytes.
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	431
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	432 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	433 """
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	434 self._verifybody()
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	435 self._bodybytes = b
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	436 self.headers[b'Content-Length'] = b'%d' % len(b)
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	437
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	438 def setbodygen(self, gen):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	439 """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	440 self._verifybody()
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	441 self._bodygen = gen
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	442
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	443 def setbodywillwrite(self):
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	444 """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	445
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	446 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	447
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	448 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	449 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	450 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	451 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	452 write.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	453
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	454 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	455 manner.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	456 """
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	457 self._verifybody()
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	458 self._bodywillwrite = True
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	459
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	460 def sendresponse(self):
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	461 """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	462
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	463 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	464 ``setbodybytes()`` or ``setbodygen()`` must be called.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	465
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	466 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	467 """
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	468 if self._started:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	469 raise error.ProgrammingError(
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	470 b'sendresponse() called multiple times'
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	471 )
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	472
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	473 self._started = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	474
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	475 if not self.status:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	476 raise error.ProgrammingError(b'status line not defined')
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	477
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	478 if (
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	479 self._bodybytes is None
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	480 and self._bodygen is None
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	481 and not self._bodywillwrite
49928 fda5a4b853ab hgweb: skip body creation of HEAD for most requests Joerg Sonnenberger <joerg@bec.de> parents: 49037 diff changeset	482 and self._req.method != b'HEAD'
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	483 ):
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	484 raise error.ProgrammingError(b'response body not defined')
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	485
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	486 # 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	487 # {Cache-Control, Content-Location, Date, ETag, Expires, Vary}
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	488 # 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	489 # 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	490 # 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	491 # 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	492 # that wasn't transferred.
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	493 if self.status.startswith(b'304 '):
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	494 # 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	495 # spec. So remove it.
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	496 if self.headers.get(b'Content-Length') == b'0':
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	497 del self.headers[b'Content-Length']
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	498
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	499 # 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	500 # problems, let's be strict.
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	501 badheaders = {
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	502 k
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	503 for k in self.headers.keys()
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	504 if k.lower()
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	505 not in (
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	506 b'date',
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	507 b'etag',
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	508 b'expires',
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	509 b'cache-control',
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	510 b'content-location',
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	511 b'content-security-policy',
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	512 b'vary',
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	513 )
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	514 }
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	515 if badheaders:
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	516 raise error.ProgrammingError(
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	517 b'illegal header on 304 response: %s'
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	518 % b', '.join(sorted(badheaders))
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	519 )
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	520
ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	521 if self._bodygen is not None or self._bodywillwrite:
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	522 raise error.ProgrammingError(
43117 8ff1ecfadcd1 cleanup: join string literals that are already on one line Martin von Zweigbergk <martinvonz@google.com> parents: 43106 diff changeset	523 b"must use setbodybytes('') with 304 responses"
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	524 )
36884 ccb70a77f746 hgweb: refactor 304 handling code Gregory Szorc <gregory.szorc@gmail.com> parents: 36882 diff changeset	525
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	526 # 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	527 # 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	528 # 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	529 # 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	530 #
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	531 # 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	532 # sending any response in some conditions.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	533 drain = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	534 close = False
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	535
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	536 # 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	537 # 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	538 # (httplib doesn't do this.)
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	539 if self._req.headers.get(b'Expect', b'').lower() == b'100-continue':
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	540 pass
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	541 # 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	542 # 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	543 # WSGI applications.
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	544 elif self._req.method not in (b'POST', b'PUT'):
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	545 pass
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	546 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	547 # 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	548 # 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	549 # 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	550 # 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	551 # 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	552 # state.
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	553 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	554 close = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	555 else:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	556 # 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	557 # 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	558 # 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	559 # 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	560 # 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	561 drain = True
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	562
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	563 if close:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	564 self.headers[b'Connection'] = b'Close'
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	565
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	566 if drain:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	567 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	568 while True:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	569 chunk = self._req.bodyfh.read(32768)
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	570 if not chunk:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	571 break
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	572
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	573 strheaders = [
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	574 (pycompat.strurl(k), pycompat.strurl(v))
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	575 for k, v in self.headers.items()
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	576 ]
2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	577 write = self._startresponse(pycompat.sysstr(self.status), strheaders)
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	578
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	579 if self._bodybytes:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	580 yield self._bodybytes
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	581 elif self._bodygen:
a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	582 for chunk in self._bodygen:
40434 6107d4549fcc hgweb: cast bytearray to bytes Gregory Szorc <gregory.szorc@gmail.com> parents: 37828 diff changeset	583 # PEP-3333 says that output must be bytes. And some WSGI
6107d4549fcc hgweb: cast bytearray to bytes Gregory Szorc <gregory.szorc@gmail.com> parents: 37828 diff changeset	584 # implementations enforce this. We cast bytes-like types here
6107d4549fcc hgweb: cast bytearray to bytes Gregory Szorc <gregory.szorc@gmail.com> parents: 37828 diff changeset	585 # for convenience.
6107d4549fcc hgweb: cast bytearray to bytes Gregory Szorc <gregory.szorc@gmail.com> parents: 37828 diff changeset	586 if isinstance(chunk, bytearray):
6107d4549fcc hgweb: cast bytearray to bytes Gregory Szorc <gregory.szorc@gmail.com> parents: 37828 diff changeset	587 chunk = bytes(chunk)
6107d4549fcc hgweb: cast bytearray to bytes Gregory Szorc <gregory.szorc@gmail.com> parents: 37828 diff changeset	588
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	589 yield chunk
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	590 elif self._bodywillwrite:
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	591 self._bodywritefn = write
49928 fda5a4b853ab hgweb: skip body creation of HEAD for most requests Joerg Sonnenberger <joerg@bec.de> parents: 49037 diff changeset	592 elif self._req.method == b'HEAD':
fda5a4b853ab hgweb: skip body creation of HEAD for most requests Joerg Sonnenberger <joerg@bec.de> parents: 49037 diff changeset	593 pass
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	594 else:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	595 error.ProgrammingError(b'do not know how to send body')
36867 a88d68dc3ee8 hgweb: create dedicated type for WSGI responses Gregory Szorc <gregory.szorc@gmail.com> parents: 36865 diff changeset	596
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	597 def getbodyfile(self):
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	598 """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	599
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	600 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	601 ``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	602 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	603 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	604 ``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	605 ``[]``.
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	606 """
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	607 if not self._bodywillwrite:
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	608 raise error.ProgrammingError(b'must call setbodywillwrite() first')
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	609
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	610 if not self._started:
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	611 raise error.ProgrammingError(
43077 687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	612 b'must call sendresponse() first; did '
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	613 b'you remember to consume it since it '
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals Augie Fackler <augie@google.com> parents: 43076 diff changeset	614 b'is a generator?'
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	615 )
36882 97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	616
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	617 assert self._bodywritefn
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	618 return offsettrackingwriter(self._bodywritefn)
97f44b0720e2 hgweb: port archive command to modern response API Gregory Szorc <gregory.szorc@gmail.com> parents: 36881 diff changeset	619
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	620
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	621 def wsgiapplication(app_maker):
45957 89a2afe31e82 formating: upgrade to black 20.8b1 Augie Fackler <raf@durin42.com> parents: 44825 diff changeset	622 """For compatibility with old CGI scripts. A plain hgweb() or hgwebdir()
89a2afe31e82 formating: upgrade to black 20.8b1 Augie Fackler <raf@durin42.com> parents: 44825 diff changeset	623 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	624 application = app_maker()
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	625
5760 0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	626 def run_wsgi(env, respond):
5887 41a3fce17625 hgweb: return iterable, add deprecation note Dirkjan Ochtman <dirkjan@ochtman.nl> parents: 5886 diff changeset	627 return application(env, respond)
43076 2372284d9457 formatting: blacken the codebase Augie Fackler <augie@google.com> parents: 40434 diff changeset	628
5760 0145f9afb0e7 Removed tabs and trailing whitespace in python files Thomas Arendsen Hein <thomas@intevation.de> parents: 5566 diff changeset	629 return run_wsgi

Mercurial > public > mercurial-scm > hg-stable

annotate mercurial/hgweb/request.py @ 50766:0caa653856a9 draft stable