ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.51 by root, Fri Aug 14 15:21:33 2009 UTC vs.
Revision 1.119 by root, Sun Jun 8 23:33:28 2014 UTC

…		…
15	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
16	run a supported event loop.	16	run a supported event loop.
17		17
18	This module implements a simple, stateless and non-blocking HTTP	18	This module implements a simple, stateless and non-blocking HTTP
19	client. It supports GET, POST and other request methods, cookies and more,	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	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	21	automatically limits the number of connections to the values specified in
22	the RFC.	22	the RFC.
23		23
24	It should generally be a "good client" that is enough for most HTTP	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	25	tasks. Simple tasks should be simple, but complex tasks should still be
…		…
36		36
37	=cut	37	=cut
38		38
39	package AnyEvent::HTTP;	39	package AnyEvent::HTTP;
40		40
41	use strict;	41	use common::sense;
42	no warnings;
43		42
44	use Errno ();	43	use Errno ();
45		44
46	use AnyEvent 5.0 ();	45	use AnyEvent 5.0 ();
47	use AnyEvent::Util ();	46	use AnyEvent::Util ();
48	use AnyEvent::Socket ();
49	use AnyEvent::Handle ();	47	use AnyEvent::Handle ();
50		48
51	use base Exporter::;	49	use base Exporter::;
52		50
53	our $VERSION = '1.43';	51	our $VERSION = 2.2;
54		52
55	our @EXPORT = qw(http_get http_post http_head http_request);	53	our @EXPORT = qw(http_get http_post http_head http_request);
56		54
57	our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)";	55	our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)";
58	our $MAX_RECURSE = 10;	56	our $MAX_RECURSE = 10;
59	our $MAX_PERSISTENT = 8;
60	our $PERSISTENT_TIMEOUT = 2;	57	our $PERSISTENT_TIMEOUT = 3;
61	our $TIMEOUT = 300;	58	our $TIMEOUT = 300;
62		59	our $MAX_PER_HOST = 4; # changing this is evil
63	# changing these is evil
64	our $MAX_PERSISTENT_PER_HOST = 0;
65	our $MAX_PER_HOST = 4;
66		60
67	our $PROXY;	61	our $PROXY;
68	our $ACTIVE = 0;	62	our $ACTIVE = 0;
69		63
70	my %KA_COUNT; # number of open keep-alive connections per host	64	my %KA_CACHE; # indexed by uhost currently, points to [$handle...] array
71	my %CO_SLOT; # number of open connections, and wait queue, per host	65	my %CO_SLOT; # number of open connections, and wait queue, per host
72		66
73	=item http_get $url, key => value..., $cb->($data, $headers)	67	=item http_get $url, key => value..., $cb->($data, $headers)
74		68
75	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
…		…
92	must be an absolute http or https URL.	86	must be an absolute http or https URL.
93		87
94	When called in void context, nothing is returned. In other contexts,	88	When called in void context, nothing is returned. In other contexts,
95	C<http_request> returns a "cancellation guard" - you have to keep the	89	C<http_request> returns a "cancellation guard" - you have to keep the
96	object at least alive until the callback get called. If the object gets	90	object at least alive until the callback get called. If the object gets
97	destroyed before the callbakc is called, the request will be cancelled.	91	destroyed before the callback is called, the request will be cancelled.
98		92
99	The callback will be called with the response body data as first argument	93	The callback will be called with the response body data as first argument
100	(or C<undef> if an error occured), and a hash-ref with response headers as	94	(or C<undef> if an error occurred), and a hash-ref with response headers
101	second argument.	95	(and trailers) as second argument.
102		96
103	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
104	headers, the "pseudo-headers" C<HTTPVersion>, C<Status> and C<Reason>	98	headers, the "pseudo-headers" (uppercase to avoid clashing with possible
		99	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the
105	contain the three parts of the HTTP Status-Line of the same name. 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
106	pseudo-header C<URL> contains the original URL (which can differ from the	105	The pseudo-header C<URL> contains the actual URL (which can differ from
107	requested URL when following redirects).	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	$headers->{Redirect}[1]{Redirect} >> will then contain the original
		116	response, and so on.
108		117
109	If the server sends a header multiple times, then their contents will be	118	If the server sends a header multiple times, then their contents will be
110	joined together with a comma (C<,>), as per the HTTP spec.	119	joined together with a comma (C<,>), as per the HTTP spec.
111		120
112	If an internal error occurs, such as not being able to resolve a hostname,	121	If an internal error occurs, such as not being able to resolve a hostname,
113	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x>	122	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be
114	(usually C<599>) and the C<Reason> pseudo-header will contain an error	123	C<590>-C<599> and the C<Reason> pseudo-header will contain an error
115	message.	124	message. Currently the following status codes are used:
		125
		126	=over 4
		127
		128	=item 595 - errors during connection establishment, proxy handshake.
		129
		130	=item 596 - errors during TLS negotiation, request sending and header processing.
		131
		132	=item 597 - errors during body receiving or processing.
		133
		134	=item 598 - user aborted request via C<on_header> or C<on_body>.
		135
		136	=item 599 - other, usually nonretryable, errors (garbled URL etc.).
		137
		138	=back
116		139
117	A typical callback might look like this:	140	A typical callback might look like this:
118		141
119	sub {	142	sub {
120	my ($body, $hdr) = @_;	143	my ($body, $hdr) = @_;
…		…
131		154
132	=over 4	155	=over 4
133		156
134	=item recurse => $count (default: $MAX_RECURSE)	157	=item recurse => $count (default: $MAX_RECURSE)
135		158
136	Whether to recurse requests or not, e.g. on redirects, authentication	159	Whether to recurse requests or not, e.g. on redirects, authentication and
137	retries and so on, and how often to do so.	160	other retries and so on, and how often to do so.
		161
		162	Only redirects to http and https URLs are supported. While most common
		163	redirection forms are handled entirely within this module, some require
		164	the use of the optional L<URI> module. If it is required but missing, then
		165	the request will fail with an error.
138		166
139	=item headers => hashref	167	=item headers => hashref
140		168
141	The request headers to use. Currently, C<http_request> may provide its	169	The request headers to use. Currently, C<http_request> may provide its own
142	own C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers	170	C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and
143	and will provide defaults for C<User-Agent:> and C<Referer:> (this can be	171	will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:>
144	suppressed by using C<undef> for these headers in which case they won't be	172	(this can be suppressed by using C<undef> for these headers in which case
145	sent at all).	173	they won't be sent at all).
		174
		175	You really should provide your own C<User-Agent:> header value that is
		176	appropriate for your program - I wouldn't be surprised if the default
		177	AnyEvent string gets blocked by webservers sooner or later.
		178
		179	Also, make sure that your headers names and values do not contain any
		180	embedded newlines.
146		181
147	=item timeout => $seconds	182	=item timeout => $seconds
148		183
149	The time-out to use for various stages - each connect attempt will reset	184	The time-out to use for various stages - each connect attempt will reset
150	the timeout, as will read or write activity, i.e. this is not an overall	185	the timeout, as will read or write activity, i.e. this is not an overall
…		…
152		187
153	Default timeout is 5 minutes.	188	Default timeout is 5 minutes.
154		189
155	=item proxy => [$host, $port[, $scheme]] or undef	190	=item proxy => [$host, $port[, $scheme]] or undef
156		191
157	Use the given http proxy for all requests. If not specified, then the	192	Use the given http proxy for all requests, or no proxy if C<undef> is
158	default proxy (as specified by C<$ENV{http_proxy}>) is used.	193	used.
159		194
160	C<$scheme> must be either missing, C<http> for HTTP or C<https> for	195	C<$scheme> must be either missing or must be C<http> for HTTP.
161	HTTPS.	196
		197	If not specified, then the default proxy is used (see
		198	C<AnyEvent::HTTP::set_proxy>).
162		199
163	=item body => $string	200	=item body => $string
164		201
165	The request body, usually empty. Will be-sent as-is (future versions of	202	The request body, usually empty. Will be sent as-is (future versions of
166	this module might offer more options).	203	this module might offer more options).
167		204
168	=item cookie_jar => $hash_ref	205	=item cookie_jar => $hash_ref
169		206
170	Passing this parameter enables (simplified) cookie-processing, loosely	207	Passing this parameter enables (simplified) cookie-processing, loosely
171	based on the original netscape specification.	208	based on the original netscape specification.
172		209
173	The C<$hash_ref> must be an (initially empty) hash reference which will	210	The C<$hash_ref> must be an (initially empty) hash reference which
174	get updated automatically. It is possible to save the cookie_jar to	211	will get updated automatically. It is possible to save the cookie jar
175	persistent storage with something like JSON or Storable, but this is not	212	to persistent storage with something like JSON or Storable - see the
176	recommended, as expiry times are currently being ignored.	213	C<AnyEvent::HTTP::cookie_jar_expire> function if you wish to remove
		214	expired or session-only cookies, and also for documentation on the format
		215	of the cookie jar.
177		216
178	Note that this cookie implementation is not of very high quality, nor	217	Note that this cookie implementation is not meant to be complete. If
179	meant to be complete. If you want complete cookie management you have to	218	you want complete cookie management you have to do that on your
180	do that on your own. C<cookie_jar> is meant as a quick fix to get some	219	own. C<cookie_jar> is meant as a quick fix to get most cookie-using sites
181	cookie-using sites working. Cookies are a privacy disaster, do not use	220	working. Cookies are a privacy disaster, do not use them unless required
182	them unless required to.	221	to.
		222
		223	When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:>
		224	headers will be set and handled by this module, otherwise they will be
		225	left untouched.
183		226
184	=item tls_ctx => $scheme | $tls_ctx	227	=item tls_ctx => $scheme | $tls_ctx
185		228
186	Specifies the AnyEvent::TLS context to be used for https connections. This	229	Specifies the AnyEvent::TLS context to be used for https connections. This
187	parameter follows the same rules as the C<tls_ctx> parameter to	230	parameter follows the same rules as the C<tls_ctx> parameter to
…		…
191	verification) TLS context.	234	verification) TLS context.
192		235
193	The default for this option is C<low>, which could be interpreted as "give	236	The default for this option is C<low>, which could be interpreted as "give
194	me the page, no matter what".	237	me the page, no matter what".
195		238
		239	See also the C<sessionid> parameter.
		240
		241	=item session => $string
		242
		243	The module might reuse connections to the same host internally. Sometimes
		244	(e.g. when using TLS), you do not want to reuse connections from other
		245	sessions. This can be achieved by setting this parameter to some unique
		246	ID (such as the address of an object storing your state data, or the TLS
		247	context) - only connections using the same unique ID will be reused.
		248
196	=item on_prepare => $callback->($fh)	249	=item on_prepare => $callback->($fh)
197		250
198	In rare cases you need to "tune" the socket before it is used to	251	In rare cases you need to "tune" the socket before it is used to
199	connect (for exmaple, to bind it on a given IP address). This parameter	252	connect (for example, to bind it on a given IP address). This parameter
200	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>	253	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>
201	and behaves exactly the same way (e.g. it has to provide a	254	and behaves exactly the same way (e.g. it has to provide a
202	timeout). See the description for the C<$prepare_cb> argument of	255	timeout). See the description for the C<$prepare_cb> argument of
203	C<AnyEvent::Socket::tcp_connect> for details.	256	C<AnyEvent::Socket::tcp_connect> for details.
204		257
		258	=item tcp_connect => $callback->($host, $service, $connect_cb, $prepare_cb)
		259
		260	In even rarer cases you want total control over how AnyEvent::HTTP
		261	establishes connections. Normally it uses L<AnyEvent::Socket::tcp_connect>
		262	to do this, but you can provide your own C<tcp_connect> function -
		263	obviously, it has to follow the same calling conventions, except that it
		264	may always return a connection guard object.
		265
		266	There are probably lots of weird uses for this function, starting from
		267	tracing the hosts C<http_request> actually tries to connect, to (inexact
		268	but fast) host => IP address caching or even socks protocol support.
		269
205	=item on_header => $callback->($headers)	270	=item on_header => $callback->($headers)
206		271
207	When specified, this callback will be called with the header hash as soon	272	When specified, this callback will be called with the header hash as soon
208	as headers have been successfully received from the remote server (not on	273	as headers have been successfully received from the remote server (not on
209	locally-generated errors).	274	locally-generated errors).
…		…
214		279
215	This callback is useful, among other things, to quickly reject unwanted	280	This callback is useful, among other things, to quickly reject unwanted
216	content, which, if it is supposed to be rare, can be faster than first	281	content, which, if it is supposed to be rare, can be faster than first
217	doing a C<HEAD> request.	282	doing a C<HEAD> request.
218		283
		284	The downside is that cancelling the request makes it impossible to re-use
		285	the connection. Also, the C<on_header> callback will not receive any
		286	trailer (headers sent after the response body).
		287
219	Example: cancel the request unless the content-type is "text/html".	288	Example: cancel the request unless the content-type is "text/html".
220		289
221	on_header => sub {	290	on_header => sub {
222	$_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/	291	$_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/
223	},	292	},
…		…
229	string instead of the body data.	298	string instead of the body data.
230		299
231	It has to return either true (in which case AnyEvent::HTTP will continue),	300	It has to return either true (in which case AnyEvent::HTTP will continue),
232	or false, in which case AnyEvent::HTTP will cancel the download (and call	301	or false, in which case AnyEvent::HTTP will cancel the download (and call
233	the completion callback with an error code of C<598>).	302	the completion callback with an error code of C<598>).
		303
		304	The downside to cancelling the request is that it makes it impossible to
		305	re-use the connection.
234		306
235	This callback is useful when the data is too large to be held in memory	307	This callback is useful when the data is too large to be held in memory
236	(so the callback writes it to a file) or when only some information should	308	(so the callback writes it to a file) or when only some information should
237	be extracted, or when the body should be processed incrementally.	309	be extracted, or when the body should be processed incrementally.
238		310
…		…
250	called. Instead of the C<$body> argument containing the body data, the	322	called. Instead of the C<$body> argument containing the body data, the
251	callback will receive the L<AnyEvent::Handle> object associated with the	323	callback will receive the L<AnyEvent::Handle> object associated with the
252	connection. In error cases, C<undef> will be passed. When there is no body	324	connection. In error cases, C<undef> will be passed. When there is no body
253	(e.g. status C<304>), the empty string will be passed.	325	(e.g. status C<304>), the empty string will be passed.
254		326
255	The handle object might or might not be in TLS mode, might be connected to	327	The handle object might or might not be in TLS mode, might be connected
256	a proxy, be a persistent connection etc., and configured in unspecified	328	to a proxy, be a persistent connection, use chunked transfer encoding
257	ways. The user is responsible for this handle (it will not be used by this	329	etc., and configured in unspecified ways. The user is responsible for this
258	module anymore).	330	handle (it will not be used by this module anymore).
259		331
260	This is useful with some push-type services, where, after the initial	332	This is useful with some push-type services, where, after the initial
261	headers, an interactive protocol is used (typical example would be the	333	headers, an interactive protocol is used (typical example would be the
262	push-style twitter API which starts a JSON/XML stream).	334	push-style twitter API which starts a JSON/XML stream).
263		335
264	If you think you need this, first have a look at C<on_body>, to see if	336	If you think you need this, first have a look at C<on_body>, to see if
265	that doesn't solve your problem in a better way.	337	that doesn't solve your problem in a better way.
266		338
		339	=item persistent => $boolean
		340
		341	Try to create/reuse a persistent connection. When this flag is set
		342	(default: true for idempotent requests, false for all others), then
		343	C<http_request> tries to re-use an existing (previously-created)
		344	persistent connection to the host and, failing that, tries to create a new
		345	one.
		346
		347	Requests failing in certain ways will be automatically retried once, which
		348	is dangerous for non-idempotent requests, which is why it defaults to off
		349	for them. The reason for this is because the bozos who designed HTTP/1.1
		350	made it impossible to distinguish between a fatal error and a normal
		351	connection timeout, so you never know whether there was a problem with
		352	your request or not.
		353
		354	When reusing an existent connection, many parameters (such as TLS context)
		355	will be ignored. See the C<session> parameter for a workaround.
		356
		357	=item keepalive => $boolean
		358
		359	Only used when C<persistent> is also true. This parameter decides whether
		360	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		361	(as opposed to only a HTTP/1.1 persistent connection).
		362
		363	The default is true, except when using a proxy, in which case it defaults
		364	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		365
		366	=item handle_params => { key => value ... }
		367
		368	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		369	constructor that is called - not all requests will create a handle, and
		370	sometimes more than one is created, so this parameter is only good for
		371	setting hints.
		372
		373	Example: set the maximum read size to 4096, to potentially conserve memory
		374	at the cost of speed.
		375
		376	handle_params => {
		377	max_read_size => 4096,
		378	},
		379
267	=back	380	=back
268		381
269	Example: make a simple HTTP GET request for http://www.nethype.de/	382	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
		383	the response body.
270		384
271	http_request GET => "http://www.nethype.de/", sub {	385	http_request GET => "http://www.nethype.de/", sub {
272	my ($body, $hdr) = @_;	386	my ($body, $hdr) = @_;
273	print "$body\n";	387	print "$body\n";
274	};	388	};
275		389
276	Example: make a HTTP HEAD request on https://www.google.com/, use a	390	Example: do a HTTP HEAD request on https://www.google.com/, use a
277	timeout of 30 seconds.	391	timeout of 30 seconds.
278		392
279	http_request	393	http_request
280	GET => "https://www.google.com",	394	HEAD => "https://www.google.com",
		395	headers => { "user-agent" => "MySearchClient 1.0" },
281	timeout => 30,	396	timeout => 30,
282	sub {	397	sub {
283	my ($body, $hdr) = @_;	398	my ($body, $hdr) = @_;
284	use Data::Dumper;	399	use Data::Dumper;
285	print Dumper $hdr;	400	print Dumper $hdr;
286	}	401	}
287	;	402	;
288		403
289	Example: make another simple HTTP GET request, but immediately try to	404	Example: do another simple HTTP GET request, but immediately try to
290	cancel it.	405	cancel it.
291		406
292	my $request = http_request GET => "http://www.nethype.de/", sub {	407	my $request = http_request GET => "http://www.nethype.de/", sub {
293	my ($body, $hdr) = @_;	408	my ($body, $hdr) = @_;
294	print "$body\n";	409	print "$body\n";
295	};	410	};
296		411
297	undef $request;	412	undef $request;
298		413
299	=cut	414	=cut
		415
		416	#############################################################################
		417	# wait queue/slots
300		418
301	sub _slot_schedule;	419	sub _slot_schedule;
302	sub _slot_schedule($) {	420	sub _slot_schedule($) {
303	my $host = shift;	421	my $host = shift;
304		422
…		…
326	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];	444	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];
327		445
328	_slot_schedule $_[0];	446	_slot_schedule $_[0];
329	}	447	}
330		448
331	our $qr_nl = qr{\015?\012};	449	#############################################################################
		450	# cookie handling
		451
		452	# expire cookies
		453	sub cookie_jar_expire($;$) {
		454	my ($jar, $session_end) = @_;
		455
		456	%$jar = () if $jar->{version} != 1;
		457
		458	my $anow = AE::now;
		459
		460	while (my ($chost, $paths) = each %$jar) {
		461	next unless ref $paths;
		462
		463	while (my ($cpath, $cookies) = each %$paths) {
		464	while (my ($cookie, $kv) = each %$cookies) {
		465	if (exists $kv->{_expires}) {
		466	delete $cookies->{$cookie}
		467	if $anow > $kv->{_expires};
		468	} elsif ($session_end) {
		469	delete $cookies->{$cookie};
		470	}
		471	}
		472
		473	delete $paths->{$cpath}
		474	unless %$cookies;
		475	}
		476
		477	delete $jar->{$chost}
		478	unless %$paths;
		479	}
		480	}
		481
		482	# extract cookies from jar
		483	sub cookie_jar_extract($$$$) {
		484	my ($jar, $scheme, $host, $path) = @_;
		485
		486	%$jar = () if $jar->{version} != 1;
		487
		488	my @cookies;
		489
		490	while (my ($chost, $paths) = each %$jar) {
		491	next unless ref $paths;
		492
		493	if ($chost =~ /^\./) {
		494	next unless $chost eq substr $host, -length $chost;
		495	} elsif ($chost =~ /\./) {
		496	next unless $chost eq $host;
		497	} else {
		498	next;
		499	}
		500
		501	while (my ($cpath, $cookies) = each %$paths) {
		502	next unless $cpath eq substr $path, 0, length $cpath;
		503
		504	while (my ($cookie, $kv) = each %$cookies) {
		505	next if $scheme ne "https" && exists $kv->{secure};
		506
		507	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
		508	delete $cookies->{$cookie};
		509	next;
		510	}
		511
		512	my $value = $kv->{value};
		513
		514	if ($value =~ /[=;,[:space:]]/) {
		515	$value =~ s/([\\"])/\\$1/g;
		516	$value = "\"$value\"";
		517	}
		518
		519	push @cookies, "$cookie=$value";
		520	}
		521	}
		522	}
		523
		524	\@cookies
		525	}
		526
		527	# parse set_cookie header into jar
		528	sub cookie_jar_set_cookie($$$$) {
		529	my ($jar, $set_cookie, $host, $date) = @_;
		530
		531	my $anow = int AE::now;
		532	my $snow; # server-now
		533
		534	for ($set_cookie) {
		535	# parse NAME=VALUE
		536	my @kv;
		537
		538	# expires is not http-compliant in the original cookie-spec,
		539	# we support the official date format and some extensions
		540	while (
		541	m{
		542	\G\s*
		543	(?:
		544	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
		545	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
		546	)
		547	}gcxsi
		548	) {
		549	my $name = $2;
		550	my $value = $4;
		551
		552	if (defined $1) {
		553	# expires
		554	$name = "expires";
		555	$value = $1;
		556	} elsif (defined $3) {
		557	# quoted
		558	$value = $3;
		559	$value =~ s/\\(.)/$1/gs;
		560	}
		561
		562	push @kv, @kv ? lc $name : $name, $value;
		563
		564	last unless /\G\s*;/gc;
		565	}
		566
		567	last unless @kv;
		568
		569	my $name = shift @kv;
		570	my %kv = (value => shift @kv, @kv);
		571
		572	if (exists $kv{"max-age"}) {
		573	$kv{_expires} = $anow + delete $kv{"max-age"};
		574	} elsif (exists $kv{expires}) {
		575	$snow ||= parse_date ($date) || $anow;
		576	$kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow);
		577	} else {
		578	delete $kv{_expires};
		579	}
		580
		581	my $cdom;
		582	my $cpath = (delete $kv{path}) || "/";
		583
		584	if (exists $kv{domain}) {
		585	$cdom = delete $kv{domain};
		586
		587	$cdom =~ s/^\.?/./; # make sure it starts with a "."
		588
		589	next if $cdom =~ /\.$/;
		590
		591	# this is not rfc-like and not netscape-like. go figure.
		592	my $ndots = $cdom =~ y/.//;
		593	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
		594	} else {
		595	$cdom = $host;
		596	}
		597
		598	# store it
		599	$jar->{version} = 1;
		600	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
		601
		602	redo if /\G\s*,/gc;
		603	}
		604	}
		605
		606	#############################################################################
		607	# keepalive/persistent connection cache
		608
		609	# fetch a connection from the keepalive cache
		610	sub ka_fetch($) {
		611	my $ka_key = shift;
		612
		613	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		614	delete $KA_CACHE{$ka_key}
		615	unless @{ $KA_CACHE{$ka_key} };
		616
		617	$hdl
		618	}
		619
		620	sub ka_store($$) {
		621	my ($ka_key, $hdl) = @_;
		622
		623	my $kaa = $KA_CACHE{$ka_key} ||= [];
		624
		625	my $destroy = sub {
		626	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		627
		628	$hdl->destroy;
		629
		630	@ka
		631	? $KA_CACHE{$ka_key} = \@ka
		632	: delete $KA_CACHE{$ka_key};
		633	};
		634
		635	# on error etc., destroy
		636	$hdl->on_error ($destroy);
		637	$hdl->on_eof ($destroy);
		638	$hdl->on_read ($destroy);
		639	$hdl->timeout ($PERSISTENT_TIMEOUT);
		640
		641	push @$kaa, $hdl;
		642	shift @$kaa while @$kaa > $MAX_PER_HOST;
		643	}
		644
		645	#############################################################################
		646	# utilities
		647
		648	# continue to parse $_ for headers and place them into the arg
		649	sub _parse_hdr() {
		650	my %hdr;
		651
		652	# things seen, not parsed:
		653	# p3pP="NON CUR OTPi OUR NOR UNI"
		654
		655	$hdr{lc $1} .= ",$2"
		656	while /\G
		657	([^:\000-\037]*):
		658	[\011\040]*
		659	((?: [^\012]+ | \012[\011\040] )*)
		660	\012
		661	/gxc;
		662
		663	/\G$/
		664	or return;
		665
		666	# remove the "," prefix we added to all headers above
		667	substr $_, 0, 1, ""
		668	for values %hdr;
		669
		670	\%hdr
		671	}
		672
		673	#############################################################################
		674	# http_get
		675
332	our $qr_nlnl = qr{(?<![^\012])\015?\012};	676	our $qr_nlnl = qr{(?<![^\012])\015?\012};
333		677
334	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	678	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
335	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };	679	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };
		680
		681	# maybe it should just become a normal object :/
		682
		683	sub _destroy_state(\%) {
		684	my ($state) = @_;
		685
		686	$state->{handle}->destroy if $state->{handle};
		687	%$state = ();
		688	}
		689
		690	sub _error(\%$$) {
		691	my ($state, $cb, $hdr) = @_;
		692
		693	&_destroy_state ($state);
		694
		695	$cb->(undef, $hdr);
		696	()
		697	}
		698
		699	our %IDEMPOTENT = (
		700	DELETE => 1,
		701	GET => 1,
		702	HEAD => 1,
		703	OPTIONS => 1,
		704	PUT => 1,
		705	TRACE => 1,
		706
		707	ACL => 1,
		708	"BASELINE-CONTROL" => 1,
		709	BIND => 1,
		710	CHECKIN => 1,
		711	CHECKOUT => 1,
		712	COPY => 1,
		713	LABEL => 1,
		714	LINK => 1,
		715	MERGE => 1,
		716	MKACTIVITY => 1,
		717	MKCALENDAR => 1,
		718	MKCOL => 1,
		719	MKREDIRECTREF => 1,
		720	MKWORKSPACE => 1,
		721	MOVE => 1,
		722	ORDERPATCH => 1,
		723	PROPFIND => 1,
		724	PROPPATCH => 1,
		725	REBIND => 1,
		726	REPORT => 1,
		727	SEARCH => 1,
		728	UNBIND => 1,
		729	UNCHECKOUT => 1,
		730	UNLINK => 1,
		731	UNLOCK => 1,
		732	UPDATE => 1,
		733	UPDATEREDIRECTREF => 1,
		734	"VERSION-CONTROL" => 1,
		735	);
336		736
337	sub http_request($$@) {	737	sub http_request($$@) {
338	my $cb = pop;	738	my $cb = pop;
339	my ($method, $url, %arg) = @_;	739	my ($method, $url, %arg) = @_;
340		740
…		…
349	while (my ($k, $v) = each %$hdr) {	749	while (my ($k, $v) = each %$hdr) {
350	$hdr{lc $k} = $v;	750	$hdr{lc $k} = $v;
351	}	751	}
352	}	752	}
353		753
		754	# pseudo headers for all subsequent responses
		755	my @pseudo = (URL => $url);
		756	push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect};
		757
354	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	758	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
355		759
356	return $cb->(undef, { Status => 599, Reason => "Too many redirections", URL => $url })	760	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
357	if $recurse < 0;	761	if $recurse < 0;
358		762
359	my $proxy = $arg{proxy} || $PROXY;	763	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
360	my $timeout = $arg{timeout} || $TIMEOUT;	764	my $timeout = $arg{timeout} || $TIMEOUT;
361		765
362	my ($uscheme, $uauthority, $upath, $query, $fragment) =	766	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
363	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|;	767	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
364		768
365	$uscheme = lc $uscheme;	769	$uscheme = lc $uscheme;
366		770
367	my $uport = $uscheme eq "http" ? 80	771	my $uport = $uscheme eq "http" ? 80
368	: $uscheme eq "https" ? 443	772	: $uscheme eq "https" ? 443
369	: return $cb->(undef, { Status => 599, Reason => "Only http and https URL schemes supported", URL => $url });	773	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
370		774
371	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	775	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x
372	or return $cb->(undef, { Status => 599, Reason => "Unparsable URL", URL => $url });	776	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
373		777
374	my $uhost = $1;	778	my $uhost = lc $1;
375	$uport = $2 if defined $2;	779	$uport = $2 if defined $2;
376		780
377	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost";	781	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
		782	unless exists $hdr{host};
378		783
379	$uhost =~ s/^\[(.*)\]$/$1/;	784	$uhost =~ s/^\[(.*)\]$/$1/;
380	$upath .= "?$query" if length $query;	785	$upath .= $query if length $query;
381		786
382	$upath =~ s%^/?%/%;	787	$upath =~ s%^/?%/%;
383		788
384	# cookie processing	789	# cookie processing
385	if (my $jar = $arg{cookie_jar}) {	790	if (my $jar = $arg{cookie_jar}) {
386	%$jar = () if $jar->{version} != 1;	791	my $cookies = cookie_jar_extract $jar, $uscheme, $uhost, $upath;
387		792
388	my @cookie;
389
390	while (my ($chost, $v) = each %$jar) {
391	if ($chost =~ /^\./) {
392	next unless $chost eq substr $uhost, -length $chost;
393	} elsif ($chost =~ /\./) {
394	next unless $chost eq $uhost;
395	} else {
396	next;
397	}
398
399	while (my ($cpath, $v) = each %$v) {
400	next unless $cpath eq substr $upath, 0, length $cpath;
401
402	while (my ($k, $v) = each %$v) {
403	next if $uscheme ne "https" && exists $v->{secure};
404	my $value = $v->{value};
405	$value =~ s/([\\"])/\\$1/g;
406	push @cookie, "$k=\"$value\"";
407	}
408	}
409	}
410
411	$hdr{cookie} = join "; ", @cookie	793	$hdr{cookie} = join "; ", @$cookies
412	if @cookie;	794	if @$cookies;
413	}	795	}
414		796
415	my ($rhost, $rport, $rscheme, $rpath); # request host, port, path	797	my ($rhost, $rport, $rscheme, $rpath); # request host, port, path
416		798
417	if ($proxy) {	799	if ($proxy) {
…		…
420	$rscheme = "http" unless defined $rscheme;	802	$rscheme = "http" unless defined $rscheme;
421		803
422	# don't support https requests over https-proxy transport,	804	# don't support https requests over https-proxy transport,
423	# can't be done with tls as spec'ed, unless you double-encrypt.	805	# can't be done with tls as spec'ed, unless you double-encrypt.
424	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";	806	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";
		807
		808	$rhost = lc $rhost;
		809	$rscheme = lc $rscheme;
425	} else {	810	} else {
426	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);	811	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
427	}	812	}
428		813
429	# leave out fragment and query string, just a heuristic	814	# leave out fragment and query string, just a heuristic
430	$hdr{referer} ||= "$uscheme://$uauthority$upath" unless exists $hdr{referer};	815	$hdr{referer} = "$uscheme://$uauthority$upath" unless exists $hdr{referer};
431	$hdr{"user-agent"} ||= $USERAGENT unless exists $hdr{"user-agent"};	816	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
432		817
433	$hdr{"content-length"} = length $arg{body};	818	$hdr{"content-length"} = length $arg{body}
		819	if length $arg{body} || $method ne "GET";
		820
		821	my $idempotent = $IDEMPOTENT{$method};
		822
		823	# default value for keepalive is true iff the request is for an idempotent method
		824	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
		825	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		826	my $was_persistent; # true if this is actually a recycled connection
		827
		828	# the key to use in the keepalive cache
		829	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
		830
		831	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive, " : "" : "close, ") . "Te"; #1.1
		832	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
434		833
435	my %state = (connect_guard => 1);	834	my %state = (connect_guard => 1);
		835
		836	my $ae_error = 595; # connecting
		837
		838	# handle actual, non-tunneled, request
		839	my $handle_actual_request = sub {
		840	$ae_error = 596; # request phase
		841
		842	my $hdl = $state{handle};
		843
		844	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
		845
		846	# send request
		847	$hdl->push_write (
		848	"$method $rpath HTTP/1.1\015\012"
		849	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
		850	. "\015\012"
		851	. (delete $arg{body})
		852	);
		853
		854	# return if error occurred during push_write()
		855	return unless %state;
		856
		857	# reduce memory usage, save a kitten, also re-use it for the response headers.
		858	%hdr = ();
		859
		860	# status line and headers
		861	$state{read_response} = sub {
		862	return unless %state;
		863
		864	for ("$_[1]") {
		865	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		866
		867	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
		868	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
		869
		870	# 100 Continue handling
		871	# should not happen as we don't send expect: 100-continue,
		872	# but we handle it just in case.
		873	# since we send the request body regardless, if we get an error
		874	# we are out of-sync, which we currently do NOT handle correctly.
		875	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
		876	if $2 eq 100;
		877
		878	push @pseudo,
		879	HTTPVersion => $1,
		880	Status => $2,
		881	Reason => $3,
		882	;
		883
		884	my $hdr = _parse_hdr
		885	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
		886
		887	%hdr = (%$hdr, @pseudo);
		888	}
		889
		890	# redirect handling
		891	# relative uri handling forced by microsoft and other shitheads.
		892	# we give our best and fall back to URI if available.
		893	if (exists $hdr{location}) {
		894	my $loc = $hdr{location};
		895
		896	if ($loc =~ m%^//%) { # //
		897	$loc = "$rscheme:$loc";
		898
		899	} elsif ($loc eq "") {
		900	$loc = $url;
		901
		902	} elsif ($loc !~ /^(?: $ | [^:\/?\#]+ : )/x) { # anything "simple"
		903	$loc =~ s/^\.\/+//;
		904
		905	if ($loc !~ m%^[.?#]%) {
		906	my $prefix = "$rscheme://$uhost:$uport";
		907
		908	unless ($loc =~ s/^\///) {
		909	$prefix .= $upath;
		910	$prefix =~ s/\/[^\/]*$//;
		911	}
		912
		913	$loc = "$prefix/$loc";
		914
		915	} elsif (eval { require URI }) { # uri
		916	$loc = URI->new_abs ($loc, $url)->as_string;
		917
		918	} else {
		919	return _error %state, $cb, { @pseudo, Status => 599, Reason => "Cannot parse Location (URI module missing)" };
		920	#$hdr{Status} = 599;
		921	#$hdr{Reason} = "Unparsable Redirect (URI module missing)";
		922	#$recurse = 0;
		923	}
		924	}
		925
		926	$hdr{location} = $loc;
		927	}
		928
		929	my $redirect;
		930
		931	if ($recurse) {
		932	my $status = $hdr{Status};
		933
		934	# industry standard is to redirect POST as GET for
		935	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
		936	# also, the UA should ask the user for 301 and 307 and POST,
		937	# industry standard seems to be to simply follow.
		938	# we go with the industry standard. 308 is defined
		939	# by rfc7238
		940	if ($status == 301 or $status == 302 or $status == 303) {
		941	# HTTP/1.1 is unclear on how to mutate the method
		942	$method = "GET" unless $method eq "HEAD";
		943	$redirect = 1;
		944	} elsif ($status == 307 or $status == 308) {
		945	$redirect = 1;
		946	}
		947	}
		948
		949	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
		950	if ($state{handle}) {
		951	# handle keepalive
		952	if (
		953	$persistent
		954	&& $_[3]
		955	&& ($hdr{HTTPVersion} < 1.1
		956	? $hdr{connection} =~ /\bkeep-?alive\b/i
		957	: $hdr{connection} !~ /\bclose\b/i)
		958	) {
		959	ka_store $ka_key, delete $state{handle};
		960	} else {
		961	# no keepalive, destroy the handle
		962	$state{handle}->destroy;
		963	}
		964	}
		965
		966	%state = ();
		967
		968	if (defined $_[1]) {
		969	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		970	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		971	}
		972
		973	# set-cookie processing
		974	if ($arg{cookie_jar}) {
		975	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		976	}
		977
		978	if ($redirect && exists $hdr{location}) {
		979	# we ignore any errors, as it is very common to receive
		980	# Content-Length != 0 but no actual body
		981	# we also access %hdr, as $_[1] might be an erro
		982	$state{recurse} =
		983	http_request (
		984	$method => $hdr{location},
		985	%arg,
		986	recurse => $recurse - 1,
		987	Redirect => [$_[0], \%hdr],
		988	sub {
		989	%state = ();
		990	&$cb
		991	},
		992	);
		993	} else {
		994	$cb->($_[0], \%hdr);
		995	}
		996	};
		997
		998	$ae_error = 597; # body phase
		999
		1000	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		1001
		1002	my $len = $chunked ? undef : $hdr{"content-length"};
		1003
		1004	# body handling, many different code paths
		1005	# - no body expected
		1006	# - want_body_handle
		1007	# - te chunked
		1008	# - 2x length known (with or without on_body)
		1009	# - 2x length not known (with or without on_body)
		1010	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		1011	$finish->(undef, 598 => "Request cancelled by on_header");
		1012	} elsif (
		1013	$hdr{Status} =~ /^(?:1..|204|205|304)$/
		1014	or $method eq "HEAD"
		1015	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		1016	) {
		1017	# no body
		1018	$finish->("", undef, undef, 1);
		1019
		1020	} elsif (!$redirect && $arg{want_body_handle}) {
		1021	$_[0]->on_eof (undef);
		1022	$_[0]->on_error (undef);
		1023	$_[0]->on_read (undef);
		1024
		1025	$finish->(delete $state{handle});
		1026
		1027	} elsif ($chunked) {
		1028	my $cl = 0;
		1029	my $body = "";
		1030	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
		1031
		1032	$state{read_chunk} = sub {
		1033	$_[1] =~ /^([0-9a-fA-F]+)/
		1034	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1035
		1036	my $len = hex $1;
		1037
		1038	if ($len) {
		1039	$cl += $len;
		1040
		1041	$_[0]->push_read (chunk => $len, sub {
		1042	$on_body->($_[1], \%hdr)
		1043	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1044
		1045	$_[0]->push_read (line => sub {
		1046	length $_[1]
		1047	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1048	$_[0]->push_read (line => $state{read_chunk});
		1049	});
		1050	});
		1051	} else {
		1052	$hdr{"content-length"} ||= $cl;
		1053
		1054	$_[0]->push_read (line => $qr_nlnl, sub {
		1055	if (length $_[1]) {
		1056	for ("$_[1]") {
		1057	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		1058
		1059	my $hdr = _parse_hdr
		1060	or return $finish->(undef, $ae_error => "Garbled response trailers");
		1061
		1062	%hdr = (%hdr, %$hdr);
		1063	}
		1064	}
		1065
		1066	$finish->($body, undef, undef, 1);
		1067	});
		1068	}
		1069	};
		1070
		1071	$_[0]->push_read (line => $state{read_chunk});
		1072
		1073	} elsif ($arg{on_body}) {
		1074	if (defined $len) {
		1075	$_[0]->on_read (sub {
		1076	$len -= length $_[0]{rbuf};
		1077
		1078	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1079	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1080
		1081	$len > 0
		1082	or $finish->("", undef, undef, 1);
		1083	});
		1084	} else {
		1085	$_[0]->on_eof (sub {
		1086	$finish->("");
		1087	});
		1088	$_[0]->on_read (sub {
		1089	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1090	or $finish->(undef, 598 => "Request cancelled by on_body");
		1091	});
		1092	}
		1093	} else {
		1094	$_[0]->on_eof (undef);
		1095
		1096	if (defined $len) {
		1097	$_[0]->on_read (sub {
		1098	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1099	if $len <= length $_[0]{rbuf};
		1100	});
		1101	} else {
		1102	$_[0]->on_error (sub {
		1103	($! == Errno::EPIPE || !$!)
		1104	? $finish->(delete $_[0]{rbuf})
		1105	: $finish->(undef, $ae_error => $_[2]);
		1106	});
		1107	$_[0]->on_read (sub { });
		1108	}
		1109	}
		1110	};
		1111
		1112	# if keepalive is enabled, then the server closing the connection
		1113	# before a response can happen legally - we retry on idempotent methods.
		1114	if ($was_persistent && $idempotent) {
		1115	my $old_eof = $hdl->{on_eof};
		1116	$hdl->{on_eof} = sub {
		1117	_destroy_state %state;
		1118
		1119	%state = ();
		1120	$state{recurse} =
		1121	http_request (
		1122	$method => $url,
		1123	%arg,
		1124	recurse => $recurse - 1,
		1125	keepalive => 0,
		1126	sub {
		1127	%state = ();
		1128	&$cb
		1129	}
		1130	);
		1131	};
		1132	$hdl->on_read (sub {
		1133	return unless %state;
		1134
		1135	# as soon as we receive something, a connection close
		1136	# once more becomes a hard error
		1137	$hdl->{on_eof} = $old_eof;
		1138	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1139	});
		1140	} else {
		1141	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1142	}
		1143	};
		1144
		1145	my $prepare_handle = sub {
		1146	my ($hdl) = $state{handle};
		1147
		1148	$hdl->on_error (sub {
		1149	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1150	});
		1151	$hdl->on_eof (sub {
		1152	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1153	});
		1154	$hdl->timeout_reset;
		1155	$hdl->timeout ($timeout);
		1156	};
		1157
		1158	# connected to proxy (or origin server)
		1159	my $connect_cb = sub {
		1160	my $fh = shift
		1161	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1162
		1163	return unless delete $state{connect_guard};
		1164
		1165	# get handle
		1166	$state{handle} = new AnyEvent::Handle
		1167	%{ $arg{handle_params} },
		1168	fh => $fh,
		1169	peername => $uhost,
		1170	tls_ctx => $arg{tls_ctx},
		1171	;
		1172
		1173	$prepare_handle->();
		1174
		1175	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1176
		1177	# now handle proxy-CONNECT method
		1178	if ($proxy && $uscheme eq "https") {
		1179	# oh dear, we have to wrap it into a connect request
		1180
		1181	# maybe re-use $uauthority with patched port?
		1182	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012");
		1183	$state{handle}->push_read (line => $qr_nlnl, sub {
		1184	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1185	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1186
		1187	if ($2 == 200) {
		1188	$rpath = $upath;
		1189	$handle_actual_request->();
		1190	} else {
		1191	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
		1192	}
		1193	});
		1194	} else {
		1195	$handle_actual_request->();
		1196	}
		1197	};
436		1198
437	_get_slot $uhost, sub {	1199	_get_slot $uhost, sub {
438	$state{slot_guard} = shift;	1200	$state{slot_guard} = shift;
439		1201
440	return unless $state{connect_guard};	1202	return unless $state{connect_guard};
441		1203
442	$state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub {	1204	# try to use an existing keepalive connection, but only if we, ourselves, plan
443	$state{fh} = shift	1205	# on a keepalive request (in theory, this should be a separate config option).
444	or do {	1206	if ($persistent && $KA_CACHE{$ka_key}) {
445	my $err = "$!";	1207	$was_persistent = 1;
446	%state = ();
447	return $cb->(undef, { Status => 599, Reason => $err, URL => $url });
448	};
449		1208
450	pop; # free memory, save a tree	1209	$state{handle} = ka_fetch $ka_key;
		1210	$state{handle}->destroyed
		1211	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
		1212	$prepare_handle->();
		1213	$state{handle}->destroyed
		1214	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
		1215	$handle_actual_request->();
451		1216
452	return unless delete $state{connect_guard};
453
454	# get handle
455	$state{handle} = new AnyEvent::Handle
456	fh => $state{fh},
457	timeout => $timeout,
458	peername => $rhost,
459	tls_ctx => $arg{tls_ctx};
460
461	# limit the number of persistent connections
462	# keepalive not yet supported
463	if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
464	++$KA_COUNT{$_[1]};
465	$state{handle}{ka_count_guard} = AnyEvent::Util::guard {
466	--$KA_COUNT{$_[1]}
467	};
468	$hdr{connection} = "keep-alive";
469	} else {	1217	} else {
470	delete $hdr{connection};	1218	my $tcp_connect = $arg{tcp_connect}
471	}	1219	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
472		1220
473	# (re-)configure handle	1221	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });
474	$state{handle}->on_error (sub {
475	%state = ();
476	$cb->(undef, { Status => 599, Reason => $_[2], URL => $url });
477	});
478	$state{handle}->on_eof (sub {
479	%state = ();
480	$cb->(undef, { Status => 599, Reason => "Unexpected end-of-file", URL => $url });
481	});
482
483	$state{handle}->starttls ("connect") if $rscheme eq "https";
484
485	# handle actual, non-tunneled, request
486	my $handle_actual_request = sub {
487	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};
488
489	# send request
490	$state{handle}->push_write (
491	"$method $rpath HTTP/1.0\015\012"
492	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
493	. "\015\012"
494	. (delete $arg{body})
495	);
496
497	%hdr = (); # reduce memory usage, save a kitten
498
499	# status line
500	$state{handle}->push_read (line => $qr_nl, sub {
501	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
502	or return (%state = (), $cb->(undef, { Status => 599, Reason => "Invalid server response ($_[1])", URL => $url }));
503
504	my %hdr = ( # response headers
505	HTTPVersion => ",$1",
506	Status => ",$2",
507	Reason => ",$3",
508	URL => ",$url"
509	);
510
511	# headers, could be optimized a bit
512	$state{handle}->unshift_read (line => $qr_nlnl, sub {
513	for ("$_[1]") {
514	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
515
516	# things seen, not parsed:
517	# p3pP="NON CUR OTPi OUR NOR UNI"
518
519	$hdr{lc $1} .= ",$2"
520	while /\G
521	([^:\000-\037]*):
522	[\011\040]*
523	((?: [^\012]+ | \012[\011\040] )*)
524	\012
525	/gxc;
526
527	/\G$/
528	or return (%state = (), $cb->(undef, { Status => 599, Reason => "Garbled response headers", URL => $url }));
529	}
530
531	substr $_, 0, 1, ""
532	for values %hdr;
533
534	# redirect handling
535	# microsoft and other shitheads don't give a shit for following standards,
536	# try to support some common forms of broken Location headers.
537	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
538	$hdr{location} =~ s/^\.\/+//;
539
540	my $url = "$rscheme://$uhost:$uport";
541
542	unless ($hdr{location} =~ s/^\///) {
543	$url .= $upath;
544	$url =~ s/\/[^\/]*$//;
545	}
546
547	$hdr{location} = "$url/$hdr{location}";
548	}
549
550	my $redirect;
551
552	if ($recurse) {
553	if ($hdr{Status} =~ /^30[12]$/ && $method ne "POST") {
554	# apparently, mozilla et al. just change POST to GET here
555	# more research is needed before we do the same
556	$redirect = 1;
557	} elsif ($hdr{Status} == 303) {
558	# even http/1.1 is unclear on how to mutate the method
559	$method = "GET" unless $method eq "HEAD";
560	$redirect = 1;
561	} elsif ($hdr{Status} == 307 && $method =~ /^(?:GET|HEAD)$/) {
562	$redirect = 1;
563	}
564	}
565
566	my $finish = sub {
567	$state{handle}->destroy if $state{handle};
568	%state = ();
569
570	# set-cookie processing
571	if ($arg{cookie_jar}) {
572	for ($_[1]{"set-cookie"}) {
573	# parse NAME=VALUE
574	my @kv;
575
576	while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) {
577	my $name = $1;
578	my $value = $3;
579
580	unless ($value) {
581	$value = $2;
582	$value =~ s/\\(.)/$1/gs;
583	}
584
585	push @kv, $name => $value;
586
587	last unless /\G\s*;/gc;
588	}
589
590	last unless @kv;
591
592	my $name = shift @kv;
593	my %kv = (value => shift @kv, @kv);
594
595	my $cdom;
596	my $cpath = (delete $kv{path}) || "/";
597
598	if (exists $kv{domain}) {
599	$cdom = delete $kv{domain};
600		1222	}
601	$cdom =~ s/^\.?/./; # make sure it starts with a "."
602
603	next if $cdom =~ /\.$/;
604
605	# this is not rfc-like and not netscape-like. go figure.
606	my $ndots = $cdom =~ y/.//;
607	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
608	} else {
609	$cdom = $uhost;
610	}
611
612	# store it
613	$arg{cookie_jar}{version} = 1;
614	$arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv;
615
616	redo if /\G\s*,/gc;
617	}
618	}
619
620	if ($redirect && exists $hdr{location}) {
621	# we ignore any errors, as it is very common to receive
622	# Content-Length != 0 but no actual body
623	# we also access %hdr, as $_[1] might be an erro
624	http_request ($method => $hdr{location}, %arg, recurse => $recurse - 1, $cb);
625	} else {
626	$cb->($_[0], $_[1]);
627	}
628	};
629
630	my $len = $hdr{"content-length"};
631
632	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
633	$finish->(undef, { Status => 598, Reason => "Request cancelled by on_header", URL => $url });
634	} elsif (
635	$hdr{Status} =~ /^(?:1..|[23]04)$/
636	or $method eq "HEAD"
637	or (defined $len && !$len)
638	) {
639	# no body
640	$finish->("", \%hdr);
641	} else {
642	# body handling, four different code paths
643	# for want_body_handle, on_body (2x), normal (2x)
644	# we might read too much here, but it does not matter yet (no pers. connections)
645	if (!$redirect && $arg{want_body_handle}) {
646	$_[0]->on_eof (undef);
647	$_[0]->on_error (undef);
648	$_[0]->on_read (undef);
649
650	$finish->(delete $state{handle}, \%hdr);
651
652	} elsif ($arg{on_body}) {
653	$_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }) });
654	if ($len) {
655	$_[0]->on_eof (undef);
656	$_[0]->on_read (sub {
657	$len -= length $_[0]{rbuf};
658
659	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
660	or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", URL => $url });
661
662	$len > 0
663	or $finish->("", \%hdr);
664	});
665	} else {
666	$_[0]->on_eof (sub {
667	$finish->("", \%hdr);
668	});
669	$_[0]->on_read (sub {
670	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
671	or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", URL => $url });
672	});
673	}
674	} else {
675	$_[0]->on_eof (undef);
676
677	if ($len) {
678	$_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }) });
679	$_[0]->on_read (sub {
680	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), \%hdr)
681	if $len <= length $_[0]{rbuf};
682	});
683	} else {
684	$_[0]->on_error (sub {
685	$! == Errno::EPIPE || !$!
686	? $finish->(delete $_[0]{rbuf}, \%hdr)
687	: $finish->(undef, { Status => 599, Reason => $_[2], URL => $url });
688	});
689	$_[0]->on_read (sub { });
690	}
691	}
692	}
693	});
694	});
695	};
696
697	# now handle proxy-CONNECT method
698	if ($proxy && $uscheme eq "https") {
699	# oh dear, we have to wrap it into a connect request
700
701	# maybe re-use $uauthority with patched port?
702	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012");
703	$state{handle}->push_read (line => $qr_nlnl, sub {
704	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
705	or return (%state = (), $cb->(undef, { Status => 599, Reason => "Invalid proxy connect response ($_[1])", URL => $url }));
706
707	if ($2 == 200) {
708	$rpath = $upath;
709	&$handle_actual_request;
710	} else {
711	%state = ();
712	$cb->(undef, { Status => $2, Reason => $3, URL => $url });
713	}
714	});
715	} else {
716	&$handle_actual_request;
717	}
718
719	}, $arg{on_prepare} || sub { $timeout };
720	};	1223	};
721		1224
722	defined wantarray && AnyEvent::Util::guard { %state = () }	1225	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
723	}	1226	}
724		1227
725	sub http_get($@) {	1228	sub http_get($@) {
726	unshift @_, "GET";	1229	unshift @_, "GET";
727	&http_request	1230	&http_request
…		…
738	&http_request	1241	&http_request
739	}	1242	}
740		1243
741	=back	1244	=back
742		1245
		1246	=head2 DNS CACHING
		1247
		1248	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
		1249	the actual connection, which in turn uses AnyEvent::DNS to resolve
		1250	hostnames. The latter is a simple stub resolver and does no caching
		1251	on its own. If you want DNS caching, you currently have to provide
		1252	your own default resolver (by storing a suitable resolver object in
		1253	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
		1254
743	=head2 GLOBAL FUNCTIONS AND VARIABLES	1255	=head2 GLOBAL FUNCTIONS AND VARIABLES
744		1256
745	=over 4	1257	=over 4
746		1258
747	=item AnyEvent::HTTP::set_proxy "proxy-url"	1259	=item AnyEvent::HTTP::set_proxy "proxy-url"
748		1260
749	Sets the default proxy server to use. The proxy-url must begin with a	1261	Sets the default proxy server to use. The proxy-url must begin with a
750	string of the form C<http://host:port> (optionally C<https:...>).	1262	string of the form C<http://host:port>, croaks otherwise.
		1263
		1264	To clear an already-set proxy, use C<undef>.
		1265
		1266	When AnyEvent::HTTP is loaded for the first time it will query the
		1267	default proxy from the operating system, currently by looking at
		1268	C<$ENV{http_proxy>}.
		1269
		1270	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
		1271
		1272	Remove all cookies from the cookie jar that have been expired. If
		1273	C<$session_end> is given and true, then additionally remove all session
		1274	cookies.
		1275
		1276	You should call this function (with a true C<$session_end>) before you
		1277	save cookies to disk, and you should call this function after loading them
		1278	again. If you have a long-running program you can additionally call this
		1279	function from time to time.
		1280
		1281	A cookie jar is initially an empty hash-reference that is managed by this
		1282	module. It's format is subject to change, but currently it is like this:
		1283
		1284	The key C<version> has to contain C<1>, otherwise the hash gets
		1285	emptied. All other keys are hostnames or IP addresses pointing to
		1286	hash-references. The key for these inner hash references is the
		1287	server path for which this cookie is meant, and the values are again
		1288	hash-references. Each key of those hash-references is a cookie name, and
		1289	the value, you guessed it, is another hash-reference, this time with the
		1290	key-value pairs from the cookie, except for C<expires> and C<max-age>,
		1291	which have been replaced by a C<_expires> key that contains the cookie
		1292	expiry timestamp. Session cookies are indicated by not having an
		1293	C<_expires> key.
		1294
		1295	Here is an example of a cookie jar with a single cookie, so you have a
		1296	chance of understanding the above paragraph:
		1297
		1298	{
		1299	version => 1,
		1300	"10.0.0.1" => {
		1301	"/" => {
		1302	"mythweb_id" => {
		1303	_expires => 1293917923,
		1304	value => "ooRung9dThee3ooyXooM1Ohm",
		1305	},
		1306	},
		1307	},
		1308	}
		1309
		1310	=item $date = AnyEvent::HTTP::format_date $timestamp
		1311
		1312	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP
		1313	Date (RFC 2616).
		1314
		1315	=item $timestamp = AnyEvent::HTTP::parse_date $date
		1316
		1317	Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) or a
		1318	bunch of minor variations of those, and returns the corresponding POSIX
		1319	timestamp, or C<undef> if the date cannot be parsed.
751		1320
752	=item $AnyEvent::HTTP::MAX_RECURSE	1321	=item $AnyEvent::HTTP::MAX_RECURSE
753		1322
754	The default value for the C<recurse> request parameter (default: C<10>).	1323	The default value for the C<recurse> request parameter (default: C<10>).
		1324
		1325	=item $AnyEvent::HTTP::TIMEOUT
		1326
		1327	The default timeout for connection operations (default: C<300>).
755		1328
756	=item $AnyEvent::HTTP::USERAGENT	1329	=item $AnyEvent::HTTP::USERAGENT
757		1330
758	The default value for the C<User-Agent> header (the default is	1331	The default value for the C<User-Agent> header (the default is
759	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1332	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
760		1333
761	=item $AnyEvent::HTTP::MAX_PER_HOST	1334	=item $AnyEvent::HTTP::MAX_PER_HOST
762		1335
763	The maximum number of concurrent connections to the same host (identified	1336	The maximum number of concurrent connections to the same host (identified
764	by the hostname). If the limit is exceeded, then the additional requests	1337	by the hostname). If the limit is exceeded, then the additional requests
765	are queued until previous connections are closed.	1338	are queued until previous connections are closed. Both persistent and
		1339	non-persistent connections are counted in this limit.
766		1340
767	The default value for this is C<4>, and it is highly advisable to not	1341	The default value for this is C<4>, and it is highly advisable to not
768	increase it.	1342	increase it much.
		1343
		1344	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1345	connections, older browsers used 2, newer ones (such as firefox 3)
		1346	typically use 6, and Opera uses 8 because like, they have the fastest
		1347	browser and give a shit for everybody else on the planet.
		1348
		1349	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
		1350
		1351	The time after which idle persistent connections get closed by
		1352	AnyEvent::HTTP (default: C<3>).
769		1353
770	=item $AnyEvent::HTTP::ACTIVE	1354	=item $AnyEvent::HTTP::ACTIVE
771		1355
772	The number of active connections. This is not the number of currently	1356	The number of active connections. This is not the number of currently
773	running requests, but the number of currently open and non-idle TCP	1357	running requests, but the number of currently open and non-idle TCP
774	connections. This number of can be useful for load-leveling.	1358	connections. This number can be useful for load-leveling.
775		1359
776	=back	1360	=back
777		1361
778	=cut	1362	=cut
779		1363
		1364	our @month = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
		1365	our @weekday = qw(Sun Mon Tue Wed Thu Fri Sat);
		1366
		1367	sub format_date($) {
		1368	my ($time) = @_;
		1369
		1370	# RFC 822/1123 format
		1371	my ($S, $M, $H, $mday, $mon, $year, $wday, $yday, undef) = gmtime $time;
		1372
		1373	sprintf "%s, %02d %s %04d %02d:%02d:%02d GMT",
		1374	$weekday[$wday], $mday, $month[$mon], $year + 1900,
		1375	$H, $M, $S;
		1376	}
		1377
		1378	sub parse_date($) {
		1379	my ($date) = @_;
		1380
		1381	my ($d, $m, $y, $H, $M, $S);
		1382
		1383	if ($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$/) {
		1384	# RFC 822/1123, required by RFC 2616 (with " ")
		1385	# cookie dates (with "-")
		1386
		1387	($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6);
		1388
		1389	} elsif ($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$/) {
		1390	# RFC 850
		1391	($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6);
		1392
		1393	} elsif ($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])$/) {
		1394	# ISO C's asctime
		1395	($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5);
		1396	}
		1397	# other formats fail in the loop below
		1398
		1399	for (0..11) {
		1400	if ($m eq $month[$_]) {
		1401	require Time::Local;
		1402	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
		1403	}
		1404	}
		1405
		1406	undef
		1407	}
		1408
780	sub set_proxy($) {	1409	sub set_proxy($) {
781	$PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix;	1410	if (length $_[0]) {
		1411	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
		1412	or Carp::croak "$_[0]: invalid proxy URL";
		1413	$PROXY = [$2, $3 || 3128, $1]
		1414	} else {
		1415	undef $PROXY;
		1416	}
782	}	1417	}
783		1418
784	# initialise proxy from environment	1419	# initialise proxy from environment
		1420	eval {
785	set_proxy $ENV{http_proxy};	1421	set_proxy $ENV{http_proxy};
		1422	};
		1423
		1424	=head2 SHOWCASE
		1425
		1426	This section contains some more elaborate "real-world" examples or code
		1427	snippets.
		1428
		1429	=head2 HTTP/1.1 FILE DOWNLOAD
		1430
		1431	Downloading files with HTTP can be quite tricky, especially when something
		1432	goes wrong and you want to resume.
		1433
		1434	Here is a function that initiates and resumes a download. It uses the
		1435	last modified time to check for file content changes, and works with many
		1436	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1437	on older servers.
		1438
		1439	It calls the completion callback with either C<undef>, which means a
		1440	nonretryable error occurred, C<0> when the download was partial and should
		1441	be retried, and C<1> if it was successful.
		1442
		1443	use AnyEvent::HTTP;
		1444
		1445	sub download($$$) {
		1446	my ($url, $file, $cb) = @_;
		1447
		1448	open my $fh, "+<", $file
		1449	or die "$file: $!";
		1450
		1451	my %hdr;
		1452	my $ofs = 0;
		1453
		1454	warn stat $fh;
		1455	warn -s _;
		1456	if (stat $fh and -s _) {
		1457	$ofs = -s _;
		1458	warn "-s is ", $ofs;
		1459	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1460	$hdr{"range"} = "bytes=$ofs-";
		1461	}
		1462
		1463	http_get $url,
		1464	headers => \%hdr,
		1465	on_header => sub {
		1466	my ($hdr) = @_;
		1467
		1468	if ($hdr->{Status} == 200 && $ofs) {
		1469	# resume failed
		1470	truncate $fh, $ofs = 0;
		1471	}
		1472
		1473	sysseek $fh, $ofs, 0;
		1474
		1475	1
		1476	},
		1477	on_body => sub {
		1478	my ($data, $hdr) = @_;
		1479
		1480	if ($hdr->{Status} =~ /^2/) {
		1481	length $data == syswrite $fh, $data
		1482	or return; # abort on write errors
		1483	}
		1484
		1485	1
		1486	},
		1487	sub {
		1488	my (undef, $hdr) = @_;
		1489
		1490	my $status = $hdr->{Status};
		1491
		1492	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1493	utime $fh, $time, $time;
		1494	}
		1495
		1496	if ($status == 200 || $status == 206 || $status == 416) {
		1497	# download ok || resume ok || file already fully downloaded
		1498	$cb->(1, $hdr);
		1499
		1500	} elsif ($status == 412) {
		1501	# file has changed while resuming, delete and retry
		1502	unlink $file;
		1503	$cb->(0, $hdr);
		1504
		1505	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1506	# retry later
		1507	$cb->(0, $hdr);
		1508
		1509	} else {
		1510	$cb->(undef, $hdr);
		1511	}
		1512	}
		1513	;
		1514	}
		1515
		1516	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1517	if ($_[0]) {
		1518	print "OK!\n";
		1519	} elsif (defined $_[0]) {
		1520	print "please retry later\n";
		1521	} else {
		1522	print "ERROR\n";
		1523	}
		1524	};
		1525
		1526	=head3 SOCKS PROXIES
		1527
		1528	Socks proxies are not directly supported by AnyEvent::HTTP. You can
		1529	compile your perl to support socks, or use an external program such as
		1530	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
		1531	transparently.
		1532
		1533	Alternatively, for AnyEvent::HTTP only, you can use your own
		1534	C<tcp_connect> function that does the proxy handshake - here is an example
		1535	that works with socks4a proxies:
		1536
		1537	use Errno;
		1538	use AnyEvent::Util;
		1539	use AnyEvent::Socket;
		1540	use AnyEvent::Handle;
		1541
		1542	# host, port and username of/for your socks4a proxy
		1543	my $socks_host = "10.0.0.23";
		1544	my $socks_port = 9050;
		1545	my $socks_user = "";
		1546
		1547	sub socks4a_connect {
		1548	my ($host, $port, $connect_cb, $prepare_cb) = @_;
		1549
		1550	my $hdl = new AnyEvent::Handle
		1551	connect => [$socks_host, $socks_port],
		1552	on_prepare => sub { $prepare_cb->($_[0]{fh}) },
		1553	on_error => sub { $connect_cb->() },
		1554	;
		1555
		1556	$hdl->push_write (pack "CCnNZ*Z*", 4, 1, $port, 1, $socks_user, $host);
		1557
		1558	$hdl->push_read (chunk => 8, sub {
		1559	my ($hdl, $chunk) = @_;
		1560	my ($status, $port, $ipn) = unpack "xCna4", $chunk;
		1561
		1562	if ($status == 0x5a) {
		1563	$connect_cb->($hdl->{fh}, (format_address $ipn) . ":$port");
		1564	} else {
		1565	$! = Errno::ENXIO; $connect_cb->();
		1566	}
		1567	});
		1568
		1569	$hdl
		1570	}
		1571
		1572	Use C<socks4a_connect> instead of C<tcp_connect> when doing C<http_request>s,
		1573	possibly after switching off other proxy types:
		1574
		1575	AnyEvent::HTTP::set_proxy undef; # usually you do not want other proxies
		1576
		1577	http_get 'http://www.google.com', tcp_connect => \&socks4a_connect, sub {
		1578	my ($data, $headers) = @_;
		1579	...
		1580	};
786		1581
787	=head1 SEE ALSO	1582	=head1 SEE ALSO
788		1583
789	L<AnyEvent>.	1584	L<AnyEvent>.
790		1585

Diff Legend

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