ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.10 by root, Thu Jun 5 13:06:43 2008 UTC vs.
Revision 1.111 by root, Sun Apr 22 12:58:38 2012 UTC

…		…
3	AnyEvent::HTTP - simple but non-blocking HTTP/HTTPS client	3	AnyEvent::HTTP - simple but non-blocking HTTP/HTTPS client
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use AnyEvent::HTTP;	7	use AnyEvent::HTTP;
		8
		9	http_get "http://www.nethype.de/", sub { print $_[1] };
		10
		11	# ... do something else here
8		12
9	=head1 DESCRIPTION	13	=head1 DESCRIPTION
10		14
11	This module is an L<AnyEvent> user, you need to make sure that you use and	15	This module is an L<AnyEvent> user, you need to make sure that you use and
12	run a supported event loop.	16	run a supported event loop.
13		17
		18	This module implements a simple, stateless and non-blocking HTTP
		19	client. It supports GET, POST and other request methods, cookies and more,
		20	all on a very low level. It can follow redirects, supports proxies, and
		21	automatically limits the number of connections to the values specified in
		22	the RFC.
		23
		24	It should generally be a "good client" that is enough for most HTTP
		25	tasks. Simple tasks should be simple, but complex tasks should still be
		26	possible as the user retains control over request and response headers.
		27
		28	The caller is responsible for authentication management, cookies (if
		29	the simplistic implementation in this module doesn't suffice), referer
		30	and other high-level protocol details for which this module offers only
		31	limited support.
		32
14	=head2 METHODS	33	=head2 METHODS
15		34
16	=over 4	35	=over 4
17		36
18	=cut	37	=cut
19		38
20	package AnyEvent::HTTP;	39	package AnyEvent::HTTP;
21		40
22	use strict;	41	use common::sense;
23	no warnings;
24		42
25	use Carp;	43	use Errno ();
26		44
27	use AnyEvent ();	45	use AnyEvent 5.0 ();
28	use AnyEvent::Util ();	46	use AnyEvent::Util ();
29	use AnyEvent::Socket ();
30	use AnyEvent::Handle ();	47	use AnyEvent::Handle ();
31		48
32	use base Exporter::;	49	use base Exporter::;
33		50
34	our $VERSION = '1.0';	51	our $VERSION = '2.14';
35		52
36	our @EXPORT = qw(http_get http_request);	53	our @EXPORT = qw(http_get http_post http_head http_request);
37		54
38	our $USERAGENT = "Mozilla/5.0 (compatible; AnyEvent::HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)";	55	our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)";
39	our $MAX_RECURSE = 10;	56	our $MAX_RECURSE = 10;
40	our $MAX_PERSISTENT = 8;
41	our $PERSISTENT_TIMEOUT = 2;	57	our $PERSISTENT_TIMEOUT = 3;
42	our $TIMEOUT = 300;	58	our $TIMEOUT = 300;
43		59	our $MAX_PER_HOST = 4; # changing this is evil
44	# changing these is evil
45	our $MAX_PERSISTENT_PER_HOST = 2;
46	our $MAX_PER_HOST = 4; # not respected yet :(
47		60
48	our $PROXY;	61	our $PROXY;
		62	our $ACTIVE = 0;
49		63
50	my %KA_COUNT; # number of open keep-alive connections per host	64	my %KA_CACHE; # indexed by uhost currently, points to [$handle...] array
		65	my %CO_SLOT; # number of open connections, and wait queue, per host
51		66
52	=item http_get $url, key => value..., $cb->($data, $headers)	67	=item http_get $url, key => value..., $cb->($data, $headers)
53		68
54	Executes an HTTP-GET request. See the http_request function for details on	69	Executes an HTTP-GET request. See the http_request function for details on
55	additional parameters.	70	additional parameters and the return value.
56		71
57	=item http_head $url, key => value..., $cb->($data, $headers)	72	=item http_head $url, key => value..., $cb->($data, $headers)
58		73
59	Executes an HTTP-HEAD request. See the http_request function for details on	74	Executes an HTTP-HEAD request. See the http_request function for details
60	additional parameters.	75	on additional parameters and the return value.
61		76
62	=item http_post $url, $body, key => value..., $cb->($data, $headers)	77	=item http_post $url, $body, key => value..., $cb->($data, $headers)
63		78
64	Executes an HTTP-POST request with a request body of C<$bod>. See the	79	Executes an HTTP-POST request with a request body of C<$body>. See the
65	http_request function for details on additional parameters.	80	http_request function for details on additional parameters and the return
		81	value.
66		82
67	=item http_request $method => $url, key => value..., $cb->($data, $headers)	83	=item http_request $method => $url, key => value..., $cb->($data, $headers)
68		84
69	Executes a HTTP request of type C<$method> (e.g. C<GET>, C<POST>). The URL	85	Executes a HTTP request of type C<$method> (e.g. C<GET>, C<POST>). The URL
70	must be an absolute http or https URL.	86	must be an absolute http or https URL.
71		87
		88	When called in void context, nothing is returned. In other contexts,
		89	C<http_request> returns a "cancellation guard" - you have to keep the
		90	object at least alive until the callback get called. If the object gets
		91	destroyed before the callback is called, the request will be cancelled.
		92
72	The callback will be called with the response data as first argument	93	The callback will be called with the response body data as first argument
73	(or C<undef> if it wasn't available due to errors), and a hash-ref with	94	(or C<undef> if an error occured), and a hash-ref with response headers
74	response headers as second argument.	95	(and trailers) as second argument.
75		96
76	All the headers in that hash are lowercased. In addition to the response	97	All the headers in that hash are lowercased. In addition to the response
77	headers, the three "pseudo-headers" C<HTTPVersion>, C<Status> and	98	headers, the "pseudo-headers" (uppercase to avoid clashing with possible
78	C<Reason> contain the three parts of the HTTP Status-Line of the same	99	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the
		100	three parts of the HTTP Status-Line of the same name. If an error occurs
		101	during the body phase of a request, then the original C<Status> and
		102	C<Reason> values from the header are available as C<OrigStatus> and
		103	C<OrigReason>.
		104
		105	The pseudo-header C<URL> contains the actual URL (which can differ from
		106	the requested URL when following redirects - for example, you might get
		107	an error that your URL scheme is not supported even though your URL is a
		108	valid http URL because it redirected to an ftp URL, in which case you can
		109	look at the URL pseudo header).
		110
		111	The pseudo-header C<Redirect> only exists when the request was a result
		112	of an internal redirect. In that case it is an array reference with
		113	the C<($data, $headers)> from the redirect response. Note that this
		114	response could in turn be the result of a redirect itself, and C<<
		115	$headers->{Redirect}[1]{Redirect} >> will then contain the original
		116	response, and so on.
		117
79	name. If the server sends a header multiple lines, then their contents	118	If the server sends a header multiple times, then their contents will be
80	will be joined together with C<\x00>.	119	joined together with a comma (C<,>), as per the HTTP spec.
81		120
82	If an internal error occurs, such as not being able to resolve a hostname,	121	If an internal error occurs, such as not being able to resolve a hostname,
83	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<599>	122	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be
84	and the C<Reason> pseudo-header will contain an error message.	123	C<590>-C<599> and the C<Reason> pseudo-header will contain an error
		124	message. Currently the following status codes are used:
		125
		126	=over 4
		127
		128	=item 595 - errors during connection etsbalishment, proxy handshake.
		129
		130	=item 596 - errors during TLS negotiation, request sending and header processing.
		131
		132	=item 597 - errors during body receiving or processing.
		133
		134	=item 598 - user aborted request via C<on_header> or C<on_body>.
		135
		136	=item 599 - other, usually nonretryable, errors (garbled URL etc.).
		137
		138	=back
85		139
86	A typical callback might look like this:	140	A typical callback might look like this:
87		141
88	sub {	142	sub {
89	my ($body, $hdr) = @_;	143	my ($body, $hdr) = @_;
…		…
105	Whether to recurse requests or not, e.g. on redirects, authentication	159	Whether to recurse requests or not, e.g. on redirects, authentication
106	retries and so on, and how often to do so.	160	retries and so on, and how often to do so.
107		161
108	=item headers => hashref	162	=item headers => hashref
109		163
110	The request headers to use.	164	The request headers to use. Currently, C<http_request> may provide its own
		165	C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and
		166	will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:>
		167	(this can be suppressed by using C<undef> for these headers in which case
		168	they won't be sent at all).
		169
		170	You really should provide your own C<User-Agent:> header value that is
		171	appropriate for your program - I wouldn't be surprised if the default
		172	AnyEvent string gets blocked by webservers sooner or later.
		173
		174	Also, make sure that your headers names and values do not contain any
		175	embedded newlines.
111		176
112	=item timeout => $seconds	177	=item timeout => $seconds
113		178
114	The time-out to use for various stages - each connect attempt will reset	179	The time-out to use for various stages - each connect attempt will reset
115	the timeout, as will read or write activity. Default timeout is 5 minutes.	180	the timeout, as will read or write activity, i.e. this is not an overall
		181	timeout.
		182
		183	Default timeout is 5 minutes.
116		184
117	=item proxy => [$host, $port[, $scheme]] or undef	185	=item proxy => [$host, $port[, $scheme]] or undef
118		186
119	Use the given http proxy for all requests. If not specified, then the	187	Use the given http proxy for all requests, or no proxy if C<undef> is
120	default proxy (as specified by C<$ENV{http_proxy}>) is used.	188	used.
121		189
122	C<$scheme> must be either missing or C<http> for HTTP, or C<https> for	190	C<$scheme> must be either missing or must be C<http> for HTTP.
123	HTTPS.	191
		192	If not specified, then the default proxy is used (see
		193	C<AnyEvent::HTTP::set_proxy>).
124		194
125	=item body => $string	195	=item body => $string
126		196
127	The request body, usually empty. Will be-sent as-is (future versions of	197	The request body, usually empty. Will be sent as-is (future versions of
128	this module might offer more options).	198	this module might offer more options).
129		199
130	=item cookie_jar => $hash_ref	200	=item cookie_jar => $hash_ref
131		201
132	Passing this parameter enables (simplified) cookie-processing, loosely	202	Passing this parameter enables (simplified) cookie-processing, loosely
133	based on the original netscape specification.	203	based on the original netscape specification.
134		204
135	The C<$hash_ref> must be an (initially empty) hash reference which will	205	The C<$hash_ref> must be an (initially empty) hash reference which
136	get updated automatically. It is possible to save the cookie_jar to	206	will get updated automatically. It is possible to save the cookie jar
137	persistent storage with something like JSON or Storable, but this is not	207	to persistent storage with something like JSON or Storable - see the
138	recommended, as expire times are currently being ignored.	208	C<AnyEvent::HTTP::cookie_jar_expire> function if you wish to remove
		209	expired or session-only cookies, and also for documentation on the format
		210	of the cookie jar.
139		211
140	Note that this cookie implementation is not of very high quality, nor	212	Note that this cookie implementation is not meant to be complete. If
141	meant to be complete. If you want complete cookie management you have to	213	you want complete cookie management you have to do that on your
142	do that on your own. C<cookie_jar> is meant as a quick fix to get some	214	own. C<cookie_jar> is meant as a quick fix to get most cookie-using sites
143	cookie-using sites working. Cookies are a privacy disaster, do not use	215	working. Cookies are a privacy disaster, do not use them unless required
144	them unless required to.	216	to.
		217
		218	When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:>
		219	headers will be set and handled by this module, otherwise they will be
		220	left untouched.
		221
		222	=item tls_ctx => $scheme | $tls_ctx
		223
		224	Specifies the AnyEvent::TLS context to be used for https connections. This
		225	parameter follows the same rules as the C<tls_ctx> parameter to
		226	L<AnyEvent::Handle>, but additionally, the two strings C<low> or
		227	C<high> can be specified, which give you a predefined low-security (no
		228	verification, highest compatibility) and high-security (CA and common-name
		229	verification) TLS context.
		230
		231	The default for this option is C<low>, which could be interpreted as "give
		232	me the page, no matter what".
		233
		234	See also the C<sessionid> parameter.
		235
		236	=item session => $string
		237
		238	The module might reuse connections to the same host internally. Sometimes
		239	(e.g. when using TLS), you do not want to reuse connections from other
		240	sessions. This can be achieved by setting this parameter to some unique
		241	ID (such as the address of an object storing your state data, or the TLS
		242	context) - only connections using the same unique ID will be reused.
		243
		244	=item on_prepare => $callback->($fh)
		245
		246	In rare cases you need to "tune" the socket before it is used to
		247	connect (for exmaple, to bind it on a given IP address). This parameter
		248	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>
		249	and behaves exactly the same way (e.g. it has to provide a
		250	timeout). See the description for the C<$prepare_cb> argument of
		251	C<AnyEvent::Socket::tcp_connect> for details.
		252
		253	=item tcp_connect => $callback->($host, $service, $connect_cb, $prepare_cb)
		254
		255	In even rarer cases you want total control over how AnyEvent::HTTP
		256	establishes connections. Normally it uses L<AnyEvent::Socket::tcp_connect>
		257	to do this, but you can provide your own C<tcp_connect> function -
		258	obviously, it has to follow the same calling conventions, except that it
		259	may always return a connection guard object.
		260
		261	There are probably lots of weird uses for this function, starting from
		262	tracing the hosts C<http_request> actually tries to connect, to (inexact
		263	but fast) host => IP address caching or even socks protocol support.
		264
		265	=item on_header => $callback->($headers)
		266
		267	When specified, this callback will be called with the header hash as soon
		268	as headers have been successfully received from the remote server (not on
		269	locally-generated errors).
		270
		271	It has to return either true (in which case AnyEvent::HTTP will continue),
		272	or false, in which case AnyEvent::HTTP will cancel the download (and call
		273	the finish callback with an error code of C<598>).
		274
		275	This callback is useful, among other things, to quickly reject unwanted
		276	content, which, if it is supposed to be rare, can be faster than first
		277	doing a C<HEAD> request.
		278
		279	The downside is that cancelling the request makes it impossible to re-use
		280	the connection. Also, the C<on_header> callback will not receive any
		281	trailer (headers sent after the response body).
		282
		283	Example: cancel the request unless the content-type is "text/html".
		284
		285	on_header => sub {
		286	$_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/
		287	},
		288
		289	=item on_body => $callback->($partial_body, $headers)
		290
		291	When specified, all body data will be passed to this callback instead of
		292	to the completion callback. The completion callback will get the empty
		293	string instead of the body data.
		294
		295	It has to return either true (in which case AnyEvent::HTTP will continue),
		296	or false, in which case AnyEvent::HTTP will cancel the download (and call
		297	the completion callback with an error code of C<598>).
		298
		299	The downside to cancelling the request is that it makes it impossible to
		300	re-use the connection.
		301
		302	This callback is useful when the data is too large to be held in memory
		303	(so the callback writes it to a file) or when only some information should
		304	be extracted, or when the body should be processed incrementally.
		305
		306	It is usually preferred over doing your own body handling via
		307	C<want_body_handle>, but in case of streaming APIs, where HTTP is
		308	only used to create a connection, C<want_body_handle> is the better
		309	alternative, as it allows you to install your own event handler, reducing
		310	resource usage.
		311
		312	=item want_body_handle => $enable
		313
		314	When enabled (default is disabled), the behaviour of AnyEvent::HTTP
		315	changes considerably: after parsing the headers, and instead of
		316	downloading the body (if any), the completion callback will be
		317	called. Instead of the C<$body> argument containing the body data, the
		318	callback will receive the L<AnyEvent::Handle> object associated with the
		319	connection. In error cases, C<undef> will be passed. When there is no body
		320	(e.g. status C<304>), the empty string will be passed.
		321
		322	The handle object might or might not be in TLS mode, might be connected
		323	to a proxy, be a persistent connection, use chunked transfer encoding
		324	etc., and configured in unspecified ways. The user is responsible for this
		325	handle (it will not be used by this module anymore).
		326
		327	This is useful with some push-type services, where, after the initial
		328	headers, an interactive protocol is used (typical example would be the
		329	push-style twitter API which starts a JSON/XML stream).
		330
		331	If you think you need this, first have a look at C<on_body>, to see if
		332	that doesn't solve your problem in a better way.
		333
		334	=item persistent => $boolean
		335
		336	Try to create/reuse a persistent connection. When this flag is set
		337	(default: true for idempotent requests, false for all others), then
		338	C<http_request> tries to re-use an existing (previously-created)
		339	persistent connection to the host and, failing that, tries to create a new
		340	one.
		341
		342	Requests failing in certain ways will be automatically retried once, which
		343	is dangerous for non-idempotent requests, which is why it defaults to off
		344	for them. The reason for this is because the bozos who designed HTTP/1.1
		345	made it impossible to distinguish between a fatal error and a normal
		346	connection timeout, so you never know whether there was a problem with
		347	your request or not.
		348
		349	When reusing an existent connection, many parameters (such as TLS context)
		350	will be ignored. See the C<session> parameter for a workaround.
		351
		352	=item keepalive => $boolean
		353
		354	Only used when C<persistent> is also true. This parameter decides whether
		355	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		356	(as opposed to only a HTTP/1.1 persistent connection).
		357
		358	The default is true, except when using a proxy, in which case it defaults
		359	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		360
		361	=item handle_params => { key => value ... }
		362
		363	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		364	constructor that is called - not all requests will create a handle, and
		365	sometimes more than one is created, so this parameter is only good for
		366	setting hints.
		367
		368	Example: set the maximum read size to 4096, to potentially conserve memory
		369	at the cost of speed.
		370
		371	handle_params => {
		372	max_read_size => 4096,
		373	},
145		374
146	=back	375	=back
147		376
148	Example: make a simple HTTP GET request for http://www.nethype.de/	377	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
		378	the response body.
149		379
150	http_request GET => "http://www.nethype.de/", sub {	380	http_request GET => "http://www.nethype.de/", sub {
151	my ($body, $hdr) = @_;	381	my ($body, $hdr) = @_;
152	print "$body\n";	382	print "$body\n";
153	};	383	};
154		384
155	Example: make a HTTP HEAD request on https://www.google.com/, use a	385	Example: do a HTTP HEAD request on https://www.google.com/, use a
156	timeout of 30 seconds.	386	timeout of 30 seconds.
157		387
158	http_request	388	http_request
159	GET => "https://www.google.com",	389	HEAD => "https://www.google.com",
		390	headers => { "user-agent" => "MySearchClient 1.0" },
160	timeout => 30,	391	timeout => 30,
161	sub {	392	sub {
162	my ($body, $hdr) = @_;	393	my ($body, $hdr) = @_;
163	use Data::Dumper;	394	use Data::Dumper;
164	print Dumper $hdr;	395	print Dumper $hdr;
165	}	396	}
166	;	397	;
167		398
		399	Example: do another simple HTTP GET request, but immediately try to
		400	cancel it.
		401
		402	my $request = http_request GET => "http://www.nethype.de/", sub {
		403	my ($body, $hdr) = @_;
		404	print "$body\n";
		405	};
		406
		407	undef $request;
		408
168	=cut	409	=cut
169		410
		411	#############################################################################
		412	# wait queue/slots
		413
		414	sub _slot_schedule;
		415	sub _slot_schedule($) {
		416	my $host = shift;
		417
		418	while ($CO_SLOT{$host}[0] < $MAX_PER_HOST) {
		419	if (my $cb = shift @{ $CO_SLOT{$host}[1] }) {
		420	# somebody wants that slot
		421	++$CO_SLOT{$host}[0];
		422	++$ACTIVE;
		423
		424	$cb->(AnyEvent::Util::guard {
		425	--$ACTIVE;
		426	--$CO_SLOT{$host}[0];
		427	_slot_schedule $host;
		428	});
		429	} else {
		430	# nobody wants the slot, maybe we can forget about it
		431	delete $CO_SLOT{$host} unless $CO_SLOT{$host}[0];
		432	last;
		433	}
		434	}
		435	}
		436
		437	# wait for a free slot on host, call callback
		438	sub _get_slot($$) {
		439	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];
		440
		441	_slot_schedule $_[0];
		442	}
		443
		444	#############################################################################
		445	# cookie handling
		446
		447	# expire cookies
		448	sub cookie_jar_expire($;$) {
		449	my ($jar, $session_end) = @_;
		450
		451	%$jar = () if $jar->{version} != 1;
		452
		453	my $anow = AE::now;
		454
		455	while (my ($chost, $paths) = each %$jar) {
		456	next unless ref $paths;
		457
		458	while (my ($cpath, $cookies) = each %$paths) {
		459	while (my ($cookie, $kv) = each %$cookies) {
		460	if (exists $kv->{_expires}) {
		461	delete $cookies->{$cookie}
		462	if $anow > $kv->{_expires};
		463	} elsif ($session_end) {
		464	delete $cookies->{$cookie};
		465	}
		466	}
		467
		468	delete $paths->{$cpath}
		469	unless %$cookies;
		470	}
		471
		472	delete $jar->{$chost}
		473	unless %$paths;
		474	}
		475	}
		476
		477	# extract cookies from jar
		478	sub cookie_jar_extract($$$$) {
		479	my ($jar, $scheme, $host, $path) = @_;
		480
		481	%$jar = () if $jar->{version} != 1;
		482
		483	my @cookies;
		484
		485	while (my ($chost, $paths) = each %$jar) {
		486	next unless ref $paths;
		487
		488	if ($chost =~ /^\./) {
		489	next unless $chost eq substr $host, -length $chost;
		490	} elsif ($chost =~ /\./) {
		491	next unless $chost eq $host;
		492	} else {
		493	next;
		494	}
		495
		496	while (my ($cpath, $cookies) = each %$paths) {
		497	next unless $cpath eq substr $path, 0, length $cpath;
		498
		499	while (my ($cookie, $kv) = each %$cookies) {
		500	next if $scheme ne "https" && exists $kv->{secure};
		501
		502	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
		503	delete $cookies->{$cookie};
		504	next;
		505	}
		506
		507	my $value = $kv->{value};
		508
		509	if ($value =~ /[=;,[:space:]]/) {
		510	$value =~ s/([\\"])/\\$1/g;
		511	$value = "\"$value\"";
		512	}
		513
		514	push @cookies, "$cookie=$value";
		515	}
		516	}
		517	}
		518
		519	\@cookies
		520	}
		521
		522	# parse set_cookie header into jar
		523	sub cookie_jar_set_cookie($$$$) {
		524	my ($jar, $set_cookie, $host, $date) = @_;
		525
		526	my $anow = int AE::now;
		527	my $snow; # server-now
		528
		529	for ($set_cookie) {
		530	# parse NAME=VALUE
		531	my @kv;
		532
		533	# expires is not http-compliant in the original cookie-spec,
		534	# we support the official date format and some extensions
		535	while (
		536	m{
		537	\G\s*
		538	(?:
		539	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
		540	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
		541	)
		542	}gcxsi
		543	) {
		544	my $name = $2;
		545	my $value = $4;
		546
		547	if (defined $1) {
		548	# expires
		549	$name = "expires";
		550	$value = $1;
		551	} elsif (defined $3) {
		552	# quoted
		553	$value = $3;
		554	$value =~ s/\\(.)/$1/gs;
		555	}
		556
		557	push @kv, @kv ? lc $name : $name, $value;
		558
		559	last unless /\G\s*;/gc;
		560	}
		561
		562	last unless @kv;
		563
		564	my $name = shift @kv;
		565	my %kv = (value => shift @kv, @kv);
		566
		567	if (exists $kv{"max-age"}) {
		568	$kv{_expires} = $anow + delete $kv{"max-age"};
		569	} elsif (exists $kv{expires}) {
		570	$snow ||= parse_date ($date) || $anow;
		571	$kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow);
		572	} else {
		573	delete $kv{_expires};
		574	}
		575
		576	my $cdom;
		577	my $cpath = (delete $kv{path}) || "/";
		578
		579	if (exists $kv{domain}) {
		580	$cdom = delete $kv{domain};
		581
		582	$cdom =~ s/^\.?/./; # make sure it starts with a "."
		583
		584	next if $cdom =~ /\.$/;
		585
		586	# this is not rfc-like and not netscape-like. go figure.
		587	my $ndots = $cdom =~ y/.//;
		588	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
		589	} else {
		590	$cdom = $host;
		591	}
		592
		593	# store it
		594	$jar->{version} = 1;
		595	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
		596
		597	redo if /\G\s*,/gc;
		598	}
		599	}
		600
		601	#############################################################################
		602	# keepalive/persistent connection cache
		603
		604	# fetch a connection from the keepalive cache
		605	sub ka_fetch($) {
		606	my $ka_key = shift;
		607
		608	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		609	delete $KA_CACHE{$ka_key}
		610	unless @{ $KA_CACHE{$ka_key} };
		611
		612	$hdl
		613	}
		614
		615	sub ka_store($$) {
		616	my ($ka_key, $hdl) = @_;
		617
		618	my $kaa = $KA_CACHE{$ka_key} ||= [];
		619
		620	my $destroy = sub {
		621	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		622
		623	$hdl->destroy;
		624
		625	@ka
		626	? $KA_CACHE{$ka_key} = \@ka
		627	: delete $KA_CACHE{$ka_key};
		628	};
		629
		630	# on error etc., destroy
		631	$hdl->on_error ($destroy);
		632	$hdl->on_eof ($destroy);
		633	$hdl->on_read ($destroy);
		634	$hdl->timeout ($PERSISTENT_TIMEOUT);
		635
		636	push @$kaa, $hdl;
		637	shift @$kaa while @$kaa > $MAX_PER_HOST;
		638	}
		639
		640	#############################################################################
		641	# utilities
		642
		643	# continue to parse $_ for headers and place them into the arg
		644	sub _parse_hdr() {
		645	my %hdr;
		646
		647	# things seen, not parsed:
		648	# p3pP="NON CUR OTPi OUR NOR UNI"
		649
		650	$hdr{lc $1} .= ",$2"
		651	while /\G
		652	([^:\000-\037]*):
		653	[\011\040]*
		654	((?: [^\012]+ | \012[\011\040] )*)
		655	\012
		656	/gxc;
		657
		658	/\G$/
		659	or return;
		660
		661	# remove the "," prefix we added to all headers above
		662	substr $_, 0, 1, ""
		663	for values %hdr;
		664
		665	\%hdr
		666	}
		667
		668	#############################################################################
		669	# http_get
		670
		671	our $qr_nlnl = qr{(?<![^\012])\015?\012};
		672
		673	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
		674	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };
		675
		676	# maybe it should just become a normal object :/
		677
		678	sub _destroy_state(\%) {
		679	my ($state) = @_;
		680
		681	$state->{handle}->destroy if $state->{handle};
		682	%$state = ();
		683	}
		684
		685	sub _error(\%$$) {
		686	my ($state, $cb, $hdr) = @_;
		687
		688	&_destroy_state ($state);
		689
		690	$cb->(undef, $hdr);
		691	()
		692	}
		693
170	sub http_request($$$;@) {	694	sub http_request($$@) {
171	my $cb = pop;	695	my $cb = pop;
172	my ($method, $url, %arg) = @_;	696	my ($method, $url, %arg) = @_;
173		697
174	my %hdr;	698	my %hdr;
		699
		700	$arg{tls_ctx} = $TLS_CTX_LOW if $arg{tls_ctx} eq "low" || !exists $arg{tls_ctx};
		701	$arg{tls_ctx} = $TLS_CTX_HIGH if $arg{tls_ctx} eq "high";
175		702
176	$method = uc $method;	703	$method = uc $method;
177		704
178	if (my $hdr = $arg{headers}) {	705	if (my $hdr = $arg{headers}) {
179	while (my ($k, $v) = each %$hdr) {	706	while (my ($k, $v) = each %$hdr) {
180	$hdr{lc $k} = $v;	707	$hdr{lc $k} = $v;
181	}	708	}
182	}	709	}
183		710
		711	# pseudo headers for all subsequent responses
		712	my @pseudo = (URL => $url);
		713	push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect};
		714
184	my $recurse = exists $arg{recurse} ? $arg{recurse} : $MAX_RECURSE;	715	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
185		716
186	return $cb->(undef, { Status => 599, Reason => "recursion limit reached" })	717	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
187	if $recurse < 0;	718	if $recurse < 0;
188		719
189	my $proxy = $arg{proxy} || $PROXY;	720	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
190	my $timeout = $arg{timeout} || $TIMEOUT;	721	my $timeout = $arg{timeout} || $TIMEOUT;
191		722
192	$hdr{"user-agent"} ||= $USERAGENT;
193
194	my ($scheme, $authority, $upath, $query, $fragment) =	723	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
195	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|;	724	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
196		725
197	$scheme = lc $scheme;	726	$uscheme = lc $uscheme;
198		727
199	my $uport = $scheme eq "http" ? 80	728	my $uport = $uscheme eq "http" ? 80
200	: $scheme eq "https" ? 443	729	: $uscheme eq "https" ? 443
201	: return $cb->(undef, { Status => 599, Reason => "only http and https URL schemes supported" });	730	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
202		731
203	$authority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	732	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x
204	or return $cb->(undef, { Status => 599, Reason => "unparsable URL" });	733	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
205		734
206	my $uhost = $1;	735	my $uhost = lc $1;
207	$uport = $2 if defined $2;	736	$uport = $2 if defined $2;
208		737
		738	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
		739	unless exists $hdr{host};
		740
209	$uhost =~ s/^\[(.*)\]$/$1/;	741	$uhost =~ s/^\[(.*)\]$/$1/;
210	$upath .= "?$query" if length $query;	742	$upath .= $query if length $query;
211		743
212	$upath =~ s%^/?%/%;	744	$upath =~ s%^/?%/%;
213		745
214	# cookie processing	746	# cookie processing
215	if (my $jar = $arg{cookie_jar}) {	747	if (my $jar = $arg{cookie_jar}) {
216	%$jar = () if $jar->{version} < 1;	748	my $cookies = cookie_jar_extract $jar, $uscheme, $uhost, $upath;
217		749
218	my @cookie;
219
220	while (my ($chost, $v) = each %$jar) {
221	next unless $chost eq substr $uhost, -length $chost;
222	next unless $chost =~ /^\./;
223
224	while (my ($cpath, $v) = each %$v) {
225	next unless $cpath eq substr $upath, 0, length $cpath;
226
227	while (my ($k, $v) = each %$v) {
228	next if $scheme ne "https" && exists $v->{secure};
229	push @cookie, "$k=$v->{value}";
230	}
231	}
232	}
233
234	$hdr{cookie} = join "; ", @cookie	750	$hdr{cookie} = join "; ", @$cookies
235	if @cookie;	751	if @$cookies;
236	}	752	}
237		753
238	my ($rhost, $rport, $rpath); # request host, port, path	754	my ($rhost, $rport, $rscheme, $rpath); # request host, port, path
239		755
240	if ($proxy) {	756	if ($proxy) {
241	($rhost, $rport, $scheme) = @$proxy;	757	($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy);
242	$rpath = $url;	758
		759	$rscheme = "http" unless defined $rscheme;
		760
		761	# don't support https requests over https-proxy transport,
		762	# can't be done with tls as spec'ed, unless you double-encrypt.
		763	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";
		764
		765	$rhost = lc $rhost;
		766	$rscheme = lc $rscheme;
243	} else {	767	} else {
244	($rhost, $rport, $rpath) = ($uhost, $uport, $upath);	768	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
245	$hdr{host} = $uhost;
246	}	769	}
247		770
		771	# leave out fragment and query string, just a heuristic
		772	$hdr{referer} = "$uscheme://$uauthority$upath" unless exists $hdr{referer};
		773	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
		774
248	$hdr{"content-length"} = length $arg{body};	775	$hdr{"content-length"} = length $arg{body}
		776	if length $arg{body} || $method ne "GET";
249		777
250	my %state;	778	my $idempotent = $method =~ /^(?:GET|HEAD|PUT|DELETE|OPTIONS|TRACE)$/;
251		779
252	$state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub {	780	# default value for keepalive is true iff the request is for an idempotent method
253	$state{fh} = shift	781	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
254	or return $cb->(undef, { Status => 599, Reason => "$!" });	782	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		783	my $was_persistent; # true if this is actually a recycled connection
255		784
256	delete $state{connect_guard}; # reduce memory usage, save a tree	785	# the key to use in the keepalive cache
		786	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
257		787
258	# get handle	788	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive " : "" : "close ") . "Te"; #1.1
259	$state{handle} = new AnyEvent::Handle	789	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
260	fh => $state{fh},
261	($scheme eq "https" ? (tls => "connect") : ());
262		790
263	# limit the number of persistent connections	791	my %state = (connect_guard => 1);
264	if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
265	++$KA_COUNT{$_[1]};
266	$state{handle}{ka_count_guard} = AnyEvent::Util::guard { --$KA_COUNT{$_[1]} };
267	$hdr{connection} = "keep-alive";
268	delete $hdr{connection}; # keep-alive not yet supported
269	} else {
270	delete $hdr{connection};
271	}
272		792
273	# (re-)configure handle	793	my $ae_error = 595; # connecting
274	$state{handle}->timeout ($timeout);	794
275	$state{handle}->on_error (sub {	795	# handle actual, non-tunneled, request
276	%state = ();	796	my $handle_actual_request = sub {
277	$cb->(undef, { Status => 599, Reason => "$!" });	797	$ae_error = 596; # request phase
278	});	798
279	$state{handle}->on_eof (sub {	799	my $hdl = $state{handle};
280	%state = ();	800
281	$cb->(undef, { Status => 599, Reason => "unexpected end-of-file" });	801	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
282	});
283		802
284	# send request	803	# send request
285	$state{handle}->push_write (	804	$hdl->push_write (
286	"$method $rpath HTTP/1.0\015\012"	805	"$method $rpath HTTP/1.1\015\012"
287	. (join "", map "$_: $hdr{$_}\015\012", keys %hdr)	806	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
288	. "\015\012"	807	. "\015\012"
289	. (delete $arg{body})	808	. (delete $arg{body})
290	);	809	);
291		810
292	%hdr = (); # reduce memory usage, save a kitten	811	# return if error occured during push_write()
		812	return unless %state;
293		813
		814	# reduce memory usage, save a kitten, also re-use it for the response headers.
		815	%hdr = ();
		816
294	# status line	817	# status line and headers
295	$state{handle}->push_read (line => qr/\015?\012/, sub {	818	$state{read_response} = sub {
296	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) \s+ ([^\015\012]+)/ix	819	return unless %state;
297	or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid server response ($_[1])" }));
298		820
299	my %hdr = ( # response headers
300	HTTPVersion => "\x00$1",
301	Status => "\x00$2",
302	Reason => "\x00$3",
303	);
304
305	# headers, could be optimized a bit
306	$state{handle}->unshift_read (line => qr/\015?\012\015?\012/, sub {
307	for ("$_[1]\012") {	821	for ("$_[1]") {
308	# we support spaces in field names, as lotus domino	822	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
309	# creates them.
310	$hdr{lc $1} .= "\x00$2"
311	while /\G
312	([^:\000-\037]+):
313	[\011\040]*
314	((?: [^\015\012]+ | \015?\012[\011\040] )*)
315	\015?\012
316	/gxc;
317		823
318	/\G$/	824	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
		825	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
		826
		827	# 100 Continue handling
		828	# should not happen as we don't send expect: 100-continue,
		829	# but we handle it just in case.
		830	# since we send the request body regardless, if we get an error
		831	# we are out of-sync, which we currently do NOT handle correctly.
		832	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
		833	if $2 eq 100;
		834
		835	push @pseudo,
		836	HTTPVersion => $1,
		837	Status => $2,
		838	Reason => $3,
		839	;
		840
		841	my $hdr = _parse_hdr
319	or return (%state = (), $cb->(undef, { Status => 599, Reason => "garbled response headers" }));	842	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
		843
		844	%hdr = (%$hdr, @pseudo);
		845	}
		846
		847	# redirect handling
		848	# microsoft and other shitheads don't give a shit for following standards,
		849	# try to support some common forms of broken Location headers.
		850	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
		851	$hdr{location} =~ s/^\.\/+//;
		852
		853	my $url = "$rscheme://$uhost:$uport";
		854
		855	unless ($hdr{location} =~ s/^\///) {
		856	$url .= $upath;
		857	$url =~ s/\/[^\/]*$//;
320	}	858	}
321		859
322	substr $_, 0, 1, ""	860	$hdr{location} = "$url/$hdr{location}";
323	for values %hdr;	861	}
324		862
325	my $finish = sub {	863	my $redirect;
326	%state = ();
327		864
328	# set-cookie processing	865	if ($recurse) {
329	if ($arg{cookie_jar} && exists $hdr{"set-cookie"}) {	866	my $status = $hdr{Status};
330	for (split /\x00/, $hdr{"set-cookie"}) {	867
331	my ($cookie, @arg) = split /;\s*/;	868	# industry standard is to redirect POST as GET for
332	my ($name, $value) = split /=/, $cookie, 2;	869	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
333	my %kv = (value => $value, map { split /=/, $_, 2 } @arg);	870	# also, the UA should ask the user for 301 and 307 and POST,
334		871	# industry standard seems to be to simply follow.
335	my $cdom = (delete $kv{domain}) || $uhost;	872	# we go with the industry standard.
336	my $cpath = (delete $kv{path}) || "/";	873	if ($status == 301 or $status == 302 or $status == 303) {
337		874	# HTTP/1.1 is unclear on how to mutate the method
338	$cdom =~ s/^.?/./; # make sure it starts with a "."	875	$method = "GET" unless $method eq "HEAD";
339		876	$redirect = 1;
340	my $ndots = $cdom =~ y/.//;	877	} elsif ($status == 307) {
341	next if $ndots < ($cdom =~ /[^.]{3}$/ ? 2 : 3);	878	$redirect = 1;
342
343	# store it
344	$arg{cookie_jar}{version} = 1;
345	$arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv;
346	}	879	}
		880	}
		881
		882	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
		883	if ($state{handle}) {
		884	# handle keepalive
		885	if (
		886	$persistent
		887	&& $_[3]
		888	&& ($hdr{HTTPVersion} < 1.1
		889	? $hdr{connection} =~ /\bkeep-?alive\b/i
		890	: $hdr{connection} !~ /\bclose\b/i)
		891	) {
		892	ka_store $ka_key, delete $state{handle};
		893	} else {
		894	# no keepalive, destroy the handle
		895	$state{handle}->destroy;
347	}	896	}
		897	}
348		898
349	if ($_[1]{Status} =~ /^x30[12]$/ && $recurse) {	899	%state = ();
350	# microsoft and other assholes don't give a shit for following standards,
351	# try to support a common form of broken Location header.
352	$_[1]{location} =~ s%^/%$scheme://$uhost:$uport/%;
353		900
354	http_request ($method, $_[1]{location}, %arg, recurse => $recurse - 1, $cb);	901	if (defined $_[1]) {
		902	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		903	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		904	}
		905
		906	# set-cookie processing
		907	if ($arg{cookie_jar}) {
		908	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		909	}
		910
		911	if ($redirect && exists $hdr{location}) {
		912	# we ignore any errors, as it is very common to receive
		913	# Content-Length != 0 but no actual body
		914	# we also access %hdr, as $_[1] might be an erro
		915	$state{recurse} =
		916	http_request (
		917	$method => $hdr{location},
		918	%arg,
		919	recurse => $recurse - 1,
		920	Redirect => [$_[0], \%hdr],
		921	sub {
		922	%state = ();
		923	&$cb
		924	},
		925	);
		926	} else {
		927	$cb->($_[0], \%hdr);
		928	}
		929	};
		930
		931	$ae_error = 597; # body phase
		932
		933	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		934
		935	my $len = $chunked ? undef : $hdr{"content-length"};
		936
		937	# body handling, many different code paths
		938	# - no body expected
		939	# - want_body_handle
		940	# - te chunked
		941	# - 2x length known (with or without on_body)
		942	# - 2x length not known (with or without on_body)
		943	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		944	$finish->(undef, 598 => "Request cancelled by on_header");
		945	} elsif (
		946	$hdr{Status} =~ /^(?:1..|204|205|304)$/
		947	or $method eq "HEAD"
		948	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		949	) {
		950	# no body
		951	$finish->("", undef, undef, 1);
		952
		953	} elsif (!$redirect && $arg{want_body_handle}) {
		954	$_[0]->on_eof (undef);
		955	$_[0]->on_error (undef);
		956	$_[0]->on_read (undef);
		957
		958	$finish->(delete $state{handle});
		959
		960	} elsif ($chunked) {
		961	my $cl = 0;
		962	my $body = "";
		963	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
		964
		965	$state{read_chunk} = sub {
		966	$_[1] =~ /^([0-9a-fA-F]+)/
		967	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		968
		969	my $len = hex $1;
		970
		971	if ($len) {
		972	$cl += $len;
		973
		974	$_[0]->push_read (chunk => $len, sub {
		975	$on_body->($_[1], \%hdr)
		976	or return $finish->(undef, 598 => "Request cancelled by on_body");
		977
		978	$_[0]->push_read (line => sub {
		979	length $_[1]
		980	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		981	$_[0]->push_read (line => $state{read_chunk});
		982	});
		983	});
355	} else {	984	} else {
356	$cb->($_[0], $_[1]);	985	$hdr{"content-length"} ||= $cl;
		986
		987	$_[0]->push_read (line => $qr_nlnl, sub {
		988	if (length $_[1]) {
		989	for ("$_[1]") {
		990	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		991
		992	my $hdr = _parse_hdr
		993	or return $finish->(undef, $ae_error => "Garbled response trailers");
		994
		995	%hdr = (%hdr, %$hdr);
		996	}
		997	}
		998
		999	$finish->($body, undef, undef, 1);
		1000	});
357	}	1001	}
358	};	1002	};
359		1003
360	if ($hdr{Status} =~ /^(?:1..|204|304)$/ or $method eq "HEAD") {	1004	$_[0]->push_read (line => $state{read_chunk});
361	$finish->(undef, \%hdr);	1005
		1006	} elsif ($arg{on_body}) {
		1007	if (defined $len) {
		1008	$_[0]->on_read (sub {
		1009	$len -= length $_[0]{rbuf};
		1010
		1011	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1012	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1013
		1014	$len > 0
		1015	or $finish->("", undef, undef, 1);
		1016	});
362	} else {	1017	} else {
363	if (exists $hdr{"content-length"}) {	1018	$_[0]->on_eof (sub {
364	$_[0]->unshift_read (chunk => $hdr{"content-length"}, sub {
365	# could cache persistent connection now
366	if ($hdr{connection} =~ /\bkeep-alive\b/i) {
367	# but we don't, due to misdesigns, this is annoyingly complex
368	};
369
370	$finish->($_[1], \%hdr);	1019	$finish->("");
371	});	1020	});
		1021	$_[0]->on_read (sub {
		1022	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1023	or $finish->(undef, 598 => "Request cancelled by on_body");
		1024	});
		1025	}
		1026	} else {
		1027	$_[0]->on_eof (undef);
		1028
		1029	if (defined $len) {
		1030	$_[0]->on_read (sub {
		1031	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1032	if $len <= length $_[0]{rbuf};
		1033	});
372	} else {	1034	} else {
373	# too bad, need to read until we get an error or EOF,
374	# no way to detect winged data.
375	$_[0]->on_error (sub {	1035	$_[0]->on_error (sub {
		1036	($! == Errno::EPIPE || !$!)
376	$finish->($_[0]{rbuf}, \%hdr);	1037	? $finish->(delete $_[0]{rbuf})
		1038	: $finish->(undef, $ae_error => $_[2]);
377	});	1039	});
378	$_[0]->on_eof (undef);
379	$_[0]->on_read (sub { });	1040	$_[0]->on_read (sub { });
		1041	}
		1042	}
		1043	};
		1044
		1045	# if keepalive is enabled, then the server closing the connection
		1046	# before a response can happen legally - we retry on idempotent methods.
		1047	if ($was_persistent && $idempotent) {
		1048	my $old_eof = $hdl->{on_eof};
		1049	$hdl->{on_eof} = sub {
		1050	_destroy_state %state;
		1051
		1052	%state = ();
		1053	$state{recurse} =
		1054	http_request (
		1055	$method => $url,
		1056	%arg,
		1057	keepalive => 0,
		1058	sub {
		1059	%state = ();
		1060	&$cb
380	}	1061	}
		1062	);
		1063	};
		1064	$hdl->on_read (sub {
		1065	return unless %state;
		1066
		1067	# as soon as we receive something, a connection close
		1068	# once more becomes a hard error
		1069	$hdl->{on_eof} = $old_eof;
		1070	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1071	});
		1072	} else {
		1073	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1074	}
		1075	};
		1076
		1077	my $prepare_handle = sub {
		1078	my ($hdl) = $state{handle};
		1079
		1080	$hdl->on_error (sub {
		1081	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1082	});
		1083	$hdl->on_eof (sub {
		1084	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1085	});
		1086	$hdl->timeout_reset;
		1087	$hdl->timeout ($timeout);
		1088	};
		1089
		1090	# connected to proxy (or origin server)
		1091	my $connect_cb = sub {
		1092	my $fh = shift
		1093	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1094
		1095	return unless delete $state{connect_guard};
		1096
		1097	# get handle
		1098	$state{handle} = new AnyEvent::Handle
		1099	%{ $arg{handle_params} },
		1100	fh => $fh,
		1101	peername => $uhost,
		1102	tls_ctx => $arg{tls_ctx},
		1103	;
		1104
		1105	$prepare_handle->();
		1106
		1107	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1108
		1109	# now handle proxy-CONNECT method
		1110	if ($proxy && $uscheme eq "https") {
		1111	# oh dear, we have to wrap it into a connect request
		1112
		1113	# maybe re-use $uauthority with patched port?
		1114	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012");
		1115	$state{handle}->push_read (line => $qr_nlnl, sub {
		1116	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1117	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1118
		1119	if ($2 == 200) {
		1120	$rpath = $upath;
		1121	$handle_actual_request->();
		1122	} else {
		1123	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
381	}	1124	}
382	});	1125	});
		1126	} else {
		1127	$handle_actual_request->();
383	});	1128	}
384	}, sub {
385	$timeout
386	};	1129	};
387		1130
		1131	_get_slot $uhost, sub {
		1132	$state{slot_guard} = shift;
		1133
		1134	return unless $state{connect_guard};
		1135
		1136	# try to use an existing keepalive connection, but only if we, ourselves, plan
		1137	# on a keepalive request (in theory, this should be a separate config option).
		1138	if ($persistent && $KA_CACHE{$ka_key}) {
		1139	$was_persistent = 1;
		1140
		1141	$state{handle} = ka_fetch $ka_key;
		1142	$state{handle}->destroyed
		1143	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
		1144	$prepare_handle->();
		1145	$state{handle}->destroyed
		1146	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
		1147	$handle_actual_request->();
		1148
		1149	} else {
		1150	my $tcp_connect = $arg{tcp_connect}
		1151	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
		1152
		1153	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });
		1154	}
		1155	};
		1156
388	defined wantarray && AnyEvent::Util::guard { %state = () }	1157	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
389	}	1158	}
390		1159
391	sub http_get($$;@) {	1160	sub http_get($@) {
392	unshift @_, "GET";	1161	unshift @_, "GET";
393	&http_request	1162	&http_request
394	}	1163	}
395		1164
396	sub http_head($$;@) {	1165	sub http_head($@) {
397	unshift @_, "HEAD";	1166	unshift @_, "HEAD";
398	&http_request	1167	&http_request
399	}	1168	}
400		1169
401	sub http_post($$$;@) {	1170	sub http_post($$@) {
		1171	my $url = shift;
402	unshift @_, "POST", "body";	1172	unshift @_, "POST", $url, "body";
403	&http_request	1173	&http_request
404	}	1174	}
405		1175
406	=back	1176	=back
407		1177
		1178	=head2 DNS CACHING
		1179
		1180	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
		1181	the actual connection, which in turn uses AnyEvent::DNS to resolve
		1182	hostnames. The latter is a simple stub resolver and does no caching
		1183	on its own. If you want DNS caching, you currently have to provide
		1184	your own default resolver (by storing a suitable resolver object in
		1185	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
		1186
408	=head2 GLOBAL FUNCTIONS AND VARIABLES	1187	=head2 GLOBAL FUNCTIONS AND VARIABLES
409		1188
410	=over 4	1189	=over 4
411		1190
412	=item AnyEvent::HTTP::set_proxy "proxy-url"	1191	=item AnyEvent::HTTP::set_proxy "proxy-url"
413		1192
414	Sets the default proxy server to use. The proxy-url must begin with a	1193	Sets the default proxy server to use. The proxy-url must begin with a
415	string of the form C<http://host:port> (optionally C<https:...>).	1194	string of the form C<http://host:port>, croaks otherwise.
		1195
		1196	To clear an already-set proxy, use C<undef>.
		1197
		1198	When AnyEvent::HTTP is laoded for the first time it will query the
		1199	default proxy from the operating system, currently by looking at
		1200	C<$ENV{http_proxy>}.
		1201
		1202	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
		1203
		1204	Remove all cookies from the cookie jar that have been expired. If
		1205	C<$session_end> is given and true, then additionally remove all session
		1206	cookies.
		1207
		1208	You should call this function (with a true C<$session_end>) before you
		1209	save cookies to disk, and you should call this function after loading them
		1210	again. If you have a long-running program you can additonally call this
		1211	function from time to time.
		1212
		1213	A cookie jar is initially an empty hash-reference that is managed by this
		1214	module. It's format is subject to change, but currently it is like this:
		1215
		1216	The key C<version> has to contain C<1>, otherwise the hash gets
		1217	emptied. All other keys are hostnames or IP addresses pointing to
		1218	hash-references. The key for these inner hash references is the
		1219	server path for which this cookie is meant, and the values are again
		1220	hash-references. The keys of those hash-references is the cookie name, and
		1221	the value, you guessed it, is another hash-reference, this time with the
		1222	key-value pairs from the cookie, except for C<expires> and C<max-age>,
		1223	which have been replaced by a C<_expires> key that contains the cookie
		1224	expiry timestamp.
		1225
		1226	Here is an example of a cookie jar with a single cookie, so you have a
		1227	chance of understanding the above paragraph:
		1228
		1229	{
		1230	version => 1,
		1231	"10.0.0.1" => {
		1232	"/" => {
		1233	"mythweb_id" => {
		1234	_expires => 1293917923,
		1235	value => "ooRung9dThee3ooyXooM1Ohm",
		1236	},
		1237	},
		1238	},
		1239	}
		1240
		1241	=item $date = AnyEvent::HTTP::format_date $timestamp
		1242
		1243	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP
		1244	Date (RFC 2616).
		1245
		1246	=item $timestamp = AnyEvent::HTTP::parse_date $date
		1247
		1248	Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) or a
		1249	bunch of minor variations of those, and returns the corresponding POSIX
		1250	timestamp, or C<undef> if the date cannot be parsed.
416		1251
417	=item $AnyEvent::HTTP::MAX_RECURSE	1252	=item $AnyEvent::HTTP::MAX_RECURSE
418		1253
419	The default value for the C<recurse> request parameter (default: C<10>).	1254	The default value for the C<recurse> request parameter (default: C<10>).
420		1255
		1256	=item $AnyEvent::HTTP::TIMEOUT
		1257
		1258	The default timeout for connection operations (default: C<300>).
		1259
421	=item $AnyEvent::HTTP::USERAGENT	1260	=item $AnyEvent::HTTP::USERAGENT
422		1261
423	The default value for the C<User-Agent> header (the default is	1262	The default value for the C<User-Agent> header (the default is
424	C<Mozilla/5.0 (compatible; AnyEvent::HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1263	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
425		1264
426	=item $AnyEvent::HTTP::MAX_PERSISTENT	1265	=item $AnyEvent::HTTP::MAX_PER_HOST
427		1266
428	The maximum number of persistent connections to keep open (default: 8).	1267	The maximum number of concurrent connections to the same host (identified
		1268	by the hostname). If the limit is exceeded, then the additional requests
		1269	are queued until previous connections are closed. Both persistent and
		1270	non-persistent connections are counted in this limit.
429		1271
430	Not implemented currently.	1272	The default value for this is C<4>, and it is highly advisable to not
		1273	increase it much.
		1274
		1275	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1276	connections, older browsers used 2, newers (such as firefox 3) typically
		1277	use 6, and Opera uses 8 because like, they have the fastest browser and
		1278	give a shit for everybody else on the planet.
431		1279
432	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT	1280	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
433		1281
434	The maximum time to cache a persistent connection, in seconds (default: 2).	1282	The time after which idle persistent conenctions get closed by
		1283	AnyEvent::HTTP (default: C<3>).
435		1284
436	Not implemented currently.	1285	=item $AnyEvent::HTTP::ACTIVE
		1286
		1287	The number of active connections. This is not the number of currently
		1288	running requests, but the number of currently open and non-idle TCP
		1289	connections. This number can be useful for load-leveling.
437		1290
438	=back	1291	=back
439		1292
440	=cut	1293	=cut
441		1294
		1295	our @month = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
		1296	our @weekday = qw(Sun Mon Tue Wed Thu Fri Sat);
		1297
		1298	sub format_date($) {
		1299	my ($time) = @_;
		1300
		1301	# RFC 822/1123 format
		1302	my ($S, $M, $H, $mday, $mon, $year, $wday, $yday, undef) = gmtime $time;
		1303
		1304	sprintf "%s, %02d %s %04d %02d:%02d:%02d GMT",
		1305	$weekday[$wday], $mday, $month[$mon], $year + 1900,
		1306	$H, $M, $S;
		1307	}
		1308
		1309	sub parse_date($) {
		1310	my ($date) = @_;
		1311
		1312	my ($d, $m, $y, $H, $M, $S);
		1313
		1314	if ($date =~ /^[A-Z][a-z][a-z]+, ([0-9][0-9]?)[\- ]([A-Z][a-z][a-z])[\- ]([0-9][0-9][0-9][0-9]) ([0-9][0-9]?):([0-9][0-9]?):([0-9][0-9]?) GMT$/) {
		1315	# RFC 822/1123, required by RFC 2616 (with " ")
		1316	# cookie dates (with "-")
		1317
		1318	($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6);
		1319
		1320	} elsif ($date =~ /^[A-Z][a-z][a-z]+, ([0-9][0-9]?)-([A-Z][a-z][a-z])-([0-9][0-9]) ([0-9][0-9]?):([0-9][0-9]?):([0-9][0-9]?) GMT$/) {
		1321	# RFC 850
		1322	($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6);
		1323
		1324	} elsif ($date =~ /^[A-Z][a-z][a-z]+ ([A-Z][a-z][a-z]) ([0-9 ]?[0-9]) ([0-9][0-9]?):([0-9][0-9]?):([0-9][0-9]?) ([0-9][0-9][0-9][0-9])$/) {
		1325	# ISO C's asctime
		1326	($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5);
		1327	}
		1328	# other formats fail in the loop below
		1329
		1330	for (0..11) {
		1331	if ($m eq $month[$_]) {
		1332	require Time::Local;
		1333	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
		1334	}
		1335	}
		1336
		1337	undef
		1338	}
		1339
442	sub set_proxy($) {	1340	sub set_proxy($) {
443	$PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix;	1341	if (length $_[0]) {
		1342	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
		1343	or Carp::croak "$_[0]: invalid proxy URL";
		1344	$PROXY = [$2, $3 || 3128, $1]
		1345	} else {
		1346	undef $PROXY;
		1347	}
444	}	1348	}
445		1349
446	# initialise proxy from environment	1350	# initialise proxy from environment
		1351	eval {
447	set_proxy $ENV{http_proxy};	1352	set_proxy $ENV{http_proxy};
		1353	};
		1354
		1355	=head2 SHOWCASE
		1356
		1357	This section contaisn some more elaborate "real-world" examples or code
		1358	snippets.
		1359
		1360	=head2 HTTP/1.1 FILE DOWNLOAD
		1361
		1362	Downloading files with HTTP can be quite tricky, especially when something
		1363	goes wrong and you want to resume.
		1364
		1365	Here is a function that initiates and resumes a download. It uses the
		1366	last modified time to check for file content changes, and works with many
		1367	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1368	on older servers.
		1369
		1370	It calls the completion callback with either C<undef>, which means a
		1371	nonretryable error occured, C<0> when the download was partial and should
		1372	be retried, and C<1> if it was successful.
		1373
		1374	use AnyEvent::HTTP;
		1375
		1376	sub download($$$) {
		1377	my ($url, $file, $cb) = @_;
		1378
		1379	open my $fh, "+<", $file
		1380	or die "$file: $!";
		1381
		1382	my %hdr;
		1383	my $ofs = 0;
		1384
		1385	warn stat $fh;
		1386	warn -s _;
		1387	if (stat $fh and -s _) {
		1388	$ofs = -s _;
		1389	warn "-s is ", $ofs;
		1390	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1391	$hdr{"range"} = "bytes=$ofs-";
		1392	}
		1393
		1394	http_get $url,
		1395	headers => \%hdr,
		1396	on_header => sub {
		1397	my ($hdr) = @_;
		1398
		1399	if ($hdr->{Status} == 200 && $ofs) {
		1400	# resume failed
		1401	truncate $fh, $ofs = 0;
		1402	}
		1403
		1404	sysseek $fh, $ofs, 0;
		1405
		1406	1
		1407	},
		1408	on_body => sub {
		1409	my ($data, $hdr) = @_;
		1410
		1411	if ($hdr->{Status} =~ /^2/) {
		1412	length $data == syswrite $fh, $data
		1413	or return; # abort on write errors
		1414	}
		1415
		1416	1
		1417	},
		1418	sub {
		1419	my (undef, $hdr) = @_;
		1420
		1421	my $status = $hdr->{Status};
		1422
		1423	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1424	utime $fh, $time, $time;
		1425	}
		1426
		1427	if ($status == 200 || $status == 206 || $status == 416) {
		1428	# download ok || resume ok || file already fully downloaded
		1429	$cb->(1, $hdr);
		1430
		1431	} elsif ($status == 412) {
		1432	# file has changed while resuming, delete and retry
		1433	unlink $file;
		1434	$cb->(0, $hdr);
		1435
		1436	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1437	# retry later
		1438	$cb->(0, $hdr);
		1439
		1440	} else {
		1441	$cb->(undef, $hdr);
		1442	}
		1443	}
		1444	;
		1445	}
		1446
		1447	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1448	if ($_[0]) {
		1449	print "OK!\n";
		1450	} elsif (defined $_[0]) {
		1451	print "please retry later\n";
		1452	} else {
		1453	print "ERROR\n";
		1454	}
		1455	};
		1456
		1457	=head3 SOCKS PROXIES
		1458
		1459	Socks proxies are not directly supported by AnyEvent::HTTP. You can
		1460	compile your perl to support socks, or use an external program such as
		1461	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
		1462	transparently.
		1463
		1464	Alternatively, for AnyEvent::HTTP only, you can use your own
		1465	C<tcp_connect> function that does the proxy handshake - here is an example
		1466	that works with socks4a proxies:
		1467
		1468	use Errno;
		1469	use AnyEvent::Util;
		1470	use AnyEvent::Socket;
		1471	use AnyEvent::Handle;
		1472
		1473	# host, port and username of/for your socks4a proxy
		1474	my $socks_host = "10.0.0.23";
		1475	my $socks_port = 9050;
		1476	my $socks_user = "";
		1477
		1478	sub socks4a_connect {
		1479	my ($host, $port, $connect_cb, $prepare_cb) = @_;
		1480
		1481	my $hdl = new AnyEvent::Handle
		1482	connect => [$socks_host, $socks_port],
		1483	on_prepare => sub { $prepare_cb->($_[0]{fh}) },
		1484	on_error => sub { $connect_cb->() },
		1485	;
		1486
		1487	$hdl->push_write (pack "CCnNZ*Z*", 4, 1, $port, 1, $socks_user, $host);
		1488
		1489	$hdl->push_read (chunk => 8, sub {
		1490	my ($hdl, $chunk) = @_;
		1491	my ($status, $port, $ipn) = unpack "xCna4", $chunk;
		1492
		1493	if ($status == 0x5a) {
		1494	$connect_cb->($hdl->{fh}, (format_address $ipn) . ":$port");
		1495	} else {
		1496	$! = Errno::ENXIO; $connect_cb->();
		1497	}
		1498	});
		1499
		1500	$hdl
		1501	}
		1502
		1503	Use C<socks4a_connect> instead of C<tcp_connect> when doing C<http_request>s,
		1504	possibly after switching off other proxy types:
		1505
		1506	AnyEvent::HTTP::set_proxy undef; # usually you do not want other proxies
		1507
		1508	http_get 'http://www.google.com', tcp_connect => \&socks4a_connect, sub {
		1509	my ($data, $headers) = @_;
		1510	...
		1511	};
448		1512
449	=head1 SEE ALSO	1513	=head1 SEE ALSO
450		1514
451	L<AnyEvent>.	1515	L<AnyEvent>.
452		1516
453	=head1 AUTHOR	1517	=head1 AUTHOR
454		1518
455	Marc Lehmann <schmorp@schmorp.de>	1519	Marc Lehmann <schmorp@schmorp.de>
456	http://home.schmorp.de/	1520	http://home.schmorp.de/
		1521
		1522	With many thanks to Дмитрий Шалашов, who provided countless
		1523	testcases and bugreports.
457		1524
458	=cut	1525	=cut
459		1526
460	1	1527	1
461		1528

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines