ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.81 by root, Sun Jan 2 01:20:17 2011 UTC vs.
Revision 1.127 by root, Sun Aug 28 09:31:29 2016 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::Handle ();	47	use AnyEvent::Handle ();
49		48
50	use base Exporter::;	49	use base Exporter::;
51		50
52	our $VERSION = '1.5';	51	our $VERSION = 2.23;
53		52
54	our @EXPORT = qw(http_get http_post http_head http_request);	53	our @EXPORT = qw(http_get http_post http_head http_request);
55		54
56	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)";
57	our $MAX_RECURSE = 10;	56	our $MAX_RECURSE = 10;
58	our $MAX_PERSISTENT = 8;
59	our $PERSISTENT_TIMEOUT = 2;	57	our $PERSISTENT_TIMEOUT = 3;
60	our $TIMEOUT = 300;	58	our $TIMEOUT = 300;
61		59	our $MAX_PER_HOST = 4; # changing this is evil
62	# changing these is evil
63	our $MAX_PERSISTENT_PER_HOST = 0;
64	our $MAX_PER_HOST = 4;
65		60
66	our $PROXY;	61	our $PROXY;
67	our $ACTIVE = 0;	62	our $ACTIVE = 0;
68		63
69	my %KA_COUNT; # number of open keep-alive connections per host	64	my %KA_CACHE; # indexed by uhost currently, points to [$handle...] array
70	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
71		66
72	=item http_get $url, key => value..., $cb->($data, $headers)	67	=item http_get $url, key => value..., $cb->($data, $headers)
73		68
74	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
…		…
94	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
95	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
96	destroyed before the callback is called, the request will be cancelled.	91	destroyed before the callback is called, the request will be cancelled.
97		92
98	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
99	(or C<undef> if an error occured), and a hash-ref with response headers	94	(or C<undef> if an error occurred), and a hash-ref with response headers
100	(and trailers) as second argument.	95	(and trailers) as second argument.
101		96
102	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
103	headers, the "pseudo-headers" (uppercase to avoid clashing with possible	98	headers, the "pseudo-headers" (uppercase to avoid clashing with possible
104	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the	99	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the
…		…
128	C<590>-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
129	message. Currently the following status codes are used:	124	message. Currently the following status codes are used:
130		125
131	=over 4	126	=over 4
132		127
133	=item 595 - errors during connection etsbalishment, proxy handshake.	128	=item 595 - errors during connection establishment, proxy handshake.
134		129
135	=item 596 - errors during TLS negotiation, request sending and header processing.	130	=item 596 - errors during TLS negotiation, request sending and header processing.
136		131
137	=item 597 - errors during body receiving or processing.	132	=item 597 - errors during body receiving or processing.
138		133
…		…
159		154
160	=over 4	155	=over 4
161		156
162	=item recurse => $count (default: $MAX_RECURSE)	157	=item recurse => $count (default: $MAX_RECURSE)
163		158
164	Whether to recurse requests or not, e.g. on redirects, authentication	159	Whether to recurse requests or not, e.g. on redirects, authentication and
165	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.
166		166
167	=item headers => hashref	167	=item headers => hashref
168		168
169	The request headers to use. Currently, C<http_request> may provide its own	169	The request headers to use. Currently, C<http_request> may provide its own
170	C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and	170	C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and
171	will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:>	171	will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:>
172	(this can be suppressed by using C<undef> for these headers in which case	172	(this can be suppressed by using C<undef> for these headers in which case
173	they won't be sent at all).	173	they won't be sent at all).
174		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.
		181
175	=item timeout => $seconds	182	=item timeout => $seconds
176		183
177	The time-out to use for various stages - each connect attempt will reset	184	The time-out to use for various stages - each connect attempt will reset
178	the timeout, as will read or write activity, i.e. this is not an overall	185	the timeout, as will read or write activity, i.e. this is not an overall
179	timeout.	186	timeout.
180		187
181	Default timeout is 5 minutes.	188	Default timeout is 5 minutes.
182		189
183	=item proxy => [$host, $port[, $scheme]] or undef	190	=item proxy => [$host, $port[, $scheme]] or undef
184		191
185	Use the given http proxy for all requests. If not specified, then the	192	Use the given http proxy for all requests, or no proxy if C<undef> is
186	default proxy (as specified by C<$ENV{http_proxy}>) is used.	193	used.
187		194
188	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.
189	HTTPS.	196
		197	If not specified, then the default proxy is used (see
		198	C<AnyEvent::HTTP::set_proxy>).
		199
		200	Currently, if your proxy requires authorization, you have to specify an
		201	appropriate "Proxy-Authorization" header in every request.
190		202
191	=item body => $string	203	=item body => $string
192		204
193	The request body, usually empty. Will be sent as-is (future versions of	205	The request body, usually empty. Will be sent as-is (future versions of
194	this module might offer more options).	206	this module might offer more options).
…		…
225	verification) TLS context.	237	verification) TLS context.
226		238
227	The default for this option is C<low>, which could be interpreted as "give	239	The default for this option is C<low>, which could be interpreted as "give
228	me the page, no matter what".	240	me the page, no matter what".
229		241
		242	See also the C<sessionid> parameter.
		243
		244	=item session => $string
		245
		246	The module might reuse connections to the same host internally. Sometimes
		247	(e.g. when using TLS), you do not want to reuse connections from other
		248	sessions. This can be achieved by setting this parameter to some unique
		249	ID (such as the address of an object storing your state data, or the TLS
		250	context) - only connections using the same unique ID will be reused.
		251
230	=item on_prepare => $callback->($fh)	252	=item on_prepare => $callback->($fh)
231		253
232	In rare cases you need to "tune" the socket before it is used to	254	In rare cases you need to "tune" the socket before it is used to
233	connect (for exmaple, to bind it on a given IP address). This parameter	255	connect (for example, to bind it on a given IP address). This parameter
234	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>	256	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>
235	and behaves exactly the same way (e.g. it has to provide a	257	and behaves exactly the same way (e.g. it has to provide a
236	timeout). See the description for the C<$prepare_cb> argument of	258	timeout). See the description for the C<$prepare_cb> argument of
237	C<AnyEvent::Socket::tcp_connect> for details.	259	C<AnyEvent::Socket::tcp_connect> for details.
238		260
…		…
303	called. Instead of the C<$body> argument containing the body data, the	325	called. Instead of the C<$body> argument containing the body data, the
304	callback will receive the L<AnyEvent::Handle> object associated with the	326	callback will receive the L<AnyEvent::Handle> object associated with the
305	connection. In error cases, C<undef> will be passed. When there is no body	327	connection. In error cases, C<undef> will be passed. When there is no body
306	(e.g. status C<304>), the empty string will be passed.	328	(e.g. status C<304>), the empty string will be passed.
307		329
308	The handle object might or might not be in TLS mode, might be connected to	330	The handle object might or might not be in TLS mode, might be connected
309	a proxy, be a persistent connection etc., and configured in unspecified	331	to a proxy, be a persistent connection, use chunked transfer encoding
310	ways. The user is responsible for this handle (it will not be used by this	332	etc., and configured in unspecified ways. The user is responsible for this
311	module anymore).	333	handle (it will not be used by this module anymore).
312		334
313	This is useful with some push-type services, where, after the initial	335	This is useful with some push-type services, where, after the initial
314	headers, an interactive protocol is used (typical example would be the	336	headers, an interactive protocol is used (typical example would be the
315	push-style twitter API which starts a JSON/XML stream).	337	push-style twitter API which starts a JSON/XML stream).
316		338
317	If you think you need this, first have a look at C<on_body>, to see if	339	If you think you need this, first have a look at C<on_body>, to see if
318	that doesn't solve your problem in a better way.	340	that doesn't solve your problem in a better way.
		341
		342	=item persistent => $boolean
		343
		344	Try to create/reuse a persistent connection. When this flag is set
		345	(default: true for idempotent requests, false for all others), then
		346	C<http_request> tries to re-use an existing (previously-created)
		347	persistent connection to the host and, failing that, tries to create a new
		348	one.
		349
		350	Requests failing in certain ways will be automatically retried once, which
		351	is dangerous for non-idempotent requests, which is why it defaults to off
		352	for them. The reason for this is because the bozos who designed HTTP/1.1
		353	made it impossible to distinguish between a fatal error and a normal
		354	connection timeout, so you never know whether there was a problem with
		355	your request or not.
		356
		357	When reusing an existent connection, many parameters (such as TLS context)
		358	will be ignored. See the C<session> parameter for a workaround.
		359
		360	=item keepalive => $boolean
		361
		362	Only used when C<persistent> is also true. This parameter decides whether
		363	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		364	(as opposed to only a HTTP/1.1 persistent connection).
		365
		366	The default is true, except when using a proxy, in which case it defaults
		367	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		368
		369	=item handle_params => { key => value ... }
		370
		371	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		372	constructor that is called - not all requests will create a handle, and
		373	sometimes more than one is created, so this parameter is only good for
		374	setting hints.
		375
		376	Example: set the maximum read size to 4096, to potentially conserve memory
		377	at the cost of speed.
		378
		379	handle_params => {
		380	max_read_size => 4096,
		381	},
319		382
320	=back	383	=back
321		384
322	Example: do a simple HTTP GET request for http://www.nethype.de/ and print	385	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
323	the response body.	386	the response body.
…		…
329		392
330	Example: do a HTTP HEAD request on https://www.google.com/, use a	393	Example: do a HTTP HEAD request on https://www.google.com/, use a
331	timeout of 30 seconds.	394	timeout of 30 seconds.
332		395
333	http_request	396	http_request
334	GET => "https://www.google.com",	397	HEAD => "https://www.google.com",
		398	headers => { "user-agent" => "MySearchClient 1.0" },
335	timeout => 30,	399	timeout => 30,
336	sub {	400	sub {
337	my ($body, $hdr) = @_;	401	my ($body, $hdr) = @_;
338	use Data::Dumper;	402	use Data::Dumper;
339	print Dumper $hdr;	403	print Dumper $hdr;
…		…
349	};	413	};
350		414
351	undef $request;	415	undef $request;
352		416
353	=cut	417	=cut
		418
		419	#############################################################################
		420	# wait queue/slots
354		421
355	sub _slot_schedule;	422	sub _slot_schedule;
356	sub _slot_schedule($) {	423	sub _slot_schedule($) {
357	my $host = shift;	424	my $host = shift;
358		425
…		…
381		448
382	_slot_schedule $_[0];	449	_slot_schedule $_[0];
383	}	450	}
384		451
385	#############################################################################	452	#############################################################################
		453	# cookie handling
386		454
387	# expire cookies	455	# expire cookies
388	sub cookie_jar_expire($;$) {	456	sub cookie_jar_expire($;$) {
389	my ($jar, $session_end) = @_;	457	my ($jar, $session_end) = @_;
390		458
…		…
414	}	482	}
415	}	483	}
416		484
417	# extract cookies from jar	485	# extract cookies from jar
418	sub cookie_jar_extract($$$$) {	486	sub cookie_jar_extract($$$$) {
419	my ($jar, $uscheme, $uhost, $upath) = @_;	487	my ($jar, $scheme, $host, $path) = @_;
420		488
421	%$jar = () if $jar->{version} != 1;	489	%$jar = () if $jar->{version} != 1;
422		490
423	my @cookies;	491	my @cookies;
424		492
425	while (my ($chost, $paths) = each %$jar) {	493	while (my ($chost, $paths) = each %$jar) {
426	next unless ref $paths;	494	next unless ref $paths;
427		495
428	if ($chost =~ /^\./) {	496	if ($chost =~ /^\./) {
429	next unless $chost eq substr $uhost, -length $chost;	497	next unless $chost eq substr $host, -length $chost;
430	} elsif ($chost =~ /\./) {	498	} elsif ($chost =~ /\./) {
431	next unless $chost eq $uhost;	499	next unless $chost eq $host;
432	} else {	500	} else {
433	next;	501	next;
434	}	502	}
435		503
436	while (my ($cpath, $cookies) = each %$paths) {	504	while (my ($cpath, $cookies) = each %$paths) {
437	next unless $cpath eq substr $upath, 0, length $cpath;	505	next unless $cpath eq substr $path, 0, length $cpath;
438		506
439	while (my ($cookie, $kv) = each %$cookies) {	507	while (my ($cookie, $kv) = each %$cookies) {
440	next if $uscheme ne "https" && exists $kv->{secure};	508	next if $scheme ne "https" && exists $kv->{secure};
441		509
442	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {	510	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
443	delete $cookies->{$cookie};	511	delete $cookies->{$cookie};
444	next;	512	next;
445	}	513	}
…		…
459	\@cookies	527	\@cookies
460	}	528	}
461		529
462	# parse set_cookie header into jar	530	# parse set_cookie header into jar
463	sub cookie_jar_set_cookie($$$$) {	531	sub cookie_jar_set_cookie($$$$) {
464	my ($jar, $set_cookie, $uhost, $date) = @_;	532	my ($jar, $set_cookie, $host, $date) = @_;
465		533
466	my $anow = int AE::now;	534	my $anow = int AE::now;
467	my $snow; # server-now	535	my $snow; # server-now
468		536
469	for ($set_cookie) {	537	for ($set_cookie) {
…		…
475	while (	543	while (
476	m{	544	m{
477	\G\s*	545	\G\s*
478	(?:	546	(?:
479	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)	547	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
480	| ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )	548	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
481	)	549	)
482	}gcxsi	550	}gcxsi
483	) {	551	) {
484	my $name = $2;	552	my $name = $2;
485	my $value = $4;	553	my $value = $4;
486		554
487	unless (defined $name) {	555	if (defined $1) {
488	# expires	556	# expires
489	$name = "expires";	557	$name = "expires";
490	$value = $1;	558	$value = $1;
491	} elsif (!defined $value) {	559	} elsif (defined $3) {
492	# quoted	560	# quoted
493	$value = $3;	561	$value = $3;
494	$value =~ s/\\(.)/$1/gs;	562	$value =~ s/\\(.)/$1/gs;
495	}	563	}
496		564
497	push @kv, lc $name, $value;	565	push @kv, @kv ? lc $name : $name, $value;
498		566
499	last unless /\G\s*;/gc;	567	last unless /\G\s*;/gc;
500	}	568	}
501		569
502	last unless @kv;	570	last unless @kv;
…		…
525		593
526	# this is not rfc-like and not netscape-like. go figure.	594	# this is not rfc-like and not netscape-like. go figure.
527	my $ndots = $cdom =~ y/.//;	595	my $ndots = $cdom =~ y/.//;
528	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);	596	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
529	} else {	597	} else {
530	$cdom = $uhost;	598	$cdom = $host;
531	}	599	}
532		600
533	# store it	601	# store it
534	$jar->{version} = 1;	602	$jar->{version} = 1;
535	$jar->{$cdom}{$cpath}{$name} = \%kv;	603	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
536		604
537	redo if /\G\s*,/gc;	605	redo if /\G\s*,/gc;
538	}	606	}
539	}	607	}
540		608
		609	#############################################################################
		610	# keepalive/persistent connection cache
		611
		612	# fetch a connection from the keepalive cache
		613	sub ka_fetch($) {
		614	my $ka_key = shift;
		615
		616	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		617	delete $KA_CACHE{$ka_key}
		618	unless @{ $KA_CACHE{$ka_key} };
		619
		620	$hdl
		621	}
		622
		623	sub ka_store($$) {
		624	my ($ka_key, $hdl) = @_;
		625
		626	my $kaa = $KA_CACHE{$ka_key} ||= [];
		627
		628	my $destroy = sub {
		629	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		630
		631	$hdl->destroy;
		632
		633	@ka
		634	? $KA_CACHE{$ka_key} = \@ka
		635	: delete $KA_CACHE{$ka_key};
		636	};
		637
		638	# on error etc., destroy
		639	$hdl->on_error ($destroy);
		640	$hdl->on_eof ($destroy);
		641	$hdl->on_read ($destroy);
		642	$hdl->timeout ($PERSISTENT_TIMEOUT);
		643
		644	push @$kaa, $hdl;
		645	shift @$kaa while @$kaa > $MAX_PER_HOST;
		646	}
		647
		648	#############################################################################
		649	# utilities
		650
541	# continue to parse $_ for headers and place them into the arg	651	# continue to parse $_ for headers and place them into the arg
542	sub parse_hdr() {	652	sub _parse_hdr() {
543	my %hdr;	653	my %hdr;
544		654
545	# things seen, not parsed:	655	# things seen, not parsed:
546	# p3pP="NON CUR OTPi OUR NOR UNI"	656	# p3pP="NON CUR OTPi OUR NOR UNI"
547		657
…		…
561	for values %hdr;	671	for values %hdr;
562		672
563	\%hdr	673	\%hdr
564	}	674	}
565		675
		676	#############################################################################
		677	# http_get
		678
566	our $qr_nlnl = qr{(?<![^\012])\015?\012};	679	our $qr_nlnl = qr{(?<![^\012])\015?\012};
567		680
568	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	681	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
569	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };	682	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };
		683
		684	# maybe it should just become a normal object :/
		685
		686	sub _destroy_state(\%) {
		687	my ($state) = @_;
		688
		689	$state->{handle}->destroy if $state->{handle};
		690	%$state = ();
		691	}
		692
		693	sub _error(\%$$) {
		694	my ($state, $cb, $hdr) = @_;
		695
		696	&_destroy_state ($state);
		697
		698	$cb->(undef, $hdr);
		699	()
		700	}
		701
		702	our %IDEMPOTENT = (
		703	DELETE => 1,
		704	GET => 1,
		705	HEAD => 1,
		706	OPTIONS => 1,
		707	PUT => 1,
		708	TRACE => 1,
		709
		710	ACL => 1,
		711	"BASELINE-CONTROL" => 1,
		712	BIND => 1,
		713	CHECKIN => 1,
		714	CHECKOUT => 1,
		715	COPY => 1,
		716	LABEL => 1,
		717	LINK => 1,
		718	MERGE => 1,
		719	MKACTIVITY => 1,
		720	MKCALENDAR => 1,
		721	MKCOL => 1,
		722	MKREDIRECTREF => 1,
		723	MKWORKSPACE => 1,
		724	MOVE => 1,
		725	ORDERPATCH => 1,
		726	PROPFIND => 1,
		727	PROPPATCH => 1,
		728	REBIND => 1,
		729	REPORT => 1,
		730	SEARCH => 1,
		731	UNBIND => 1,
		732	UNCHECKOUT => 1,
		733	UNLINK => 1,
		734	UNLOCK => 1,
		735	UPDATE => 1,
		736	UPDATEREDIRECTREF => 1,
		737	"VERSION-CONTROL" => 1,
		738	);
570		739
571	sub http_request($$@) {	740	sub http_request($$@) {
572	my $cb = pop;	741	my $cb = pop;
573	my ($method, $url, %arg) = @_;	742	my ($method, $url, %arg) = @_;
574		743
…		…
592	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	761	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
593		762
594	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })	763	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
595	if $recurse < 0;	764	if $recurse < 0;
596		765
597	my $proxy = $arg{proxy} || $PROXY;	766	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
598	my $timeout = $arg{timeout} || $TIMEOUT;	767	my $timeout = $arg{timeout} || $TIMEOUT;
599		768
600	my ($uscheme, $uauthority, $upath, $query, $fragment) =	769	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
601	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|;	770	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
602		771
603	$uscheme = lc $uscheme;	772	$uscheme = lc $uscheme;
604		773
605	my $uport = $uscheme eq "http" ? 80	774	my $uport = $uscheme eq "http" ? 80
606	: $uscheme eq "https" ? 443	775	: $uscheme eq "https" ? 443
607	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });	776	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
608		777
609	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	778	$uauthority =~ /^(?: .*\@ )? ([^\@]+?) (?: : (\d+) )?$/x
610	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });	779	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
611		780
612	my $uhost = $1;	781	my $uhost = lc $1;
613	$uport = $2 if defined $2;	782	$uport = $2 if defined $2;
614		783
615	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"	784	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
616	unless exists $hdr{host};	785	unless exists $hdr{host};
617		786
…		…
636	$rscheme = "http" unless defined $rscheme;	805	$rscheme = "http" unless defined $rscheme;
637		806
638	# don't support https requests over https-proxy transport,	807	# don't support https requests over https-proxy transport,
639	# can't be done with tls as spec'ed, unless you double-encrypt.	808	# can't be done with tls as spec'ed, unless you double-encrypt.
640	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";	809	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";
		810
		811	$rhost = lc $rhost;
		812	$rscheme = lc $rscheme;
641	} else {	813	} else {
642	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);	814	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
643	}	815	}
644		816
645	# leave out fragment and query string, just a heuristic	817	# leave out fragment and query string, just a heuristic
…		…
647	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};	819	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
648		820
649	$hdr{"content-length"} = length $arg{body}	821	$hdr{"content-length"} = length $arg{body}
650	if length $arg{body} || $method ne "GET";	822	if length $arg{body} || $method ne "GET";
651		823
652	$hdr{connection} = "close TE"; #1.1	824	my $idempotent = $IDEMPOTENT{$method};
		825
		826	# default value for keepalive is true iff the request is for an idempotent method
		827	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
		828	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		829	my $was_persistent; # true if this is actually a recycled connection
		830
		831	# the key to use in the keepalive cache
		832	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
		833
		834	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive, " : "" : "close, ") . "Te"; #1.1
653	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1	835	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
654		836
655	my %state = (connect_guard => 1);	837	my %state = (connect_guard => 1);
		838
		839	my $ae_error = 595; # connecting
		840
		841	# handle actual, non-tunneled, request
		842	my $handle_actual_request = sub {
		843	$ae_error = 596; # request phase
		844
		845	my $hdl = $state{handle};
		846
		847	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
		848
		849	# send request
		850	$hdl->push_write (
		851	"$method $rpath HTTP/1.1\015\012"
		852	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
		853	. "\015\012"
		854	. $arg{body}
		855	);
		856
		857	# return if error occurred during push_write()
		858	return unless %state;
		859
		860	# reduce memory usage, save a kitten, also re-use it for the response headers.
		861	%hdr = ();
		862
		863	# status line and headers
		864	$state{read_response} = sub {
		865	return unless %state;
		866
		867	for ("$_[1]") {
		868	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		869
		870	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
		871	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
		872
		873	# 100 Continue handling
		874	# should not happen as we don't send expect: 100-continue,
		875	# but we handle it just in case.
		876	# since we send the request body regardless, if we get an error
		877	# we are out of-sync, which we currently do NOT handle correctly.
		878	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
		879	if $2 eq 100;
		880
		881	push @pseudo,
		882	HTTPVersion => $1,
		883	Status => $2,
		884	Reason => $3,
		885	;
		886
		887	my $hdr = _parse_hdr
		888	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
		889
		890	%hdr = (%$hdr, @pseudo);
		891	}
		892
		893	# redirect handling
		894	# relative uri handling forced by microsoft and other shitheads.
		895	# we give our best and fall back to URI if available.
		896	if (exists $hdr{location}) {
		897	my $loc = $hdr{location};
		898
		899	if ($loc =~ m%^//%) { # //
		900	$loc = "$uscheme:$loc";
		901
		902	} elsif ($loc eq "") {
		903	$loc = $url;
		904
		905	} elsif ($loc !~ /^(?: $ | [^:\/?\#]+ : )/x) { # anything "simple"
		906	$loc =~ s/^\.\/+//;
		907
		908	if ($loc !~ m%^[.?#]%) {
		909	my $prefix = "$uscheme://$uhost:$uport";
		910
		911	unless ($loc =~ s/^\///) {
		912	$prefix .= $upath;
		913	$prefix =~ s/\/[^\/]*$//;
		914	}
		915
		916	$loc = "$prefix/$loc";
		917
		918	} elsif (eval { require URI }) { # uri
		919	$loc = URI->new_abs ($loc, $url)->as_string;
		920
		921	} else {
		922	return _error %state, $cb, { @pseudo, Status => 599, Reason => "Cannot parse Location (URI module missing)" };
		923	#$hdr{Status} = 599;
		924	#$hdr{Reason} = "Unparsable Redirect (URI module missing)";
		925	#$recurse = 0;
		926	}
		927	}
		928
		929	$hdr{location} = $loc;
		930	}
		931
		932	my $redirect;
		933
		934	if ($recurse) {
		935	my $status = $hdr{Status};
		936
		937	# industry standard is to redirect POST as GET for
		938	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
		939	# also, the UA should ask the user for 301 and 307 and POST,
		940	# industry standard seems to be to simply follow.
		941	# we go with the industry standard. 308 is defined
		942	# by rfc7538
		943	if ($status == 301 or $status == 302 or $status == 303) {
		944	$redirect = 1;
		945	# HTTP/1.1 is unclear on how to mutate the method
		946	unless ($method eq "HEAD") {
		947	$method = "GET";
		948	delete $arg{body};
		949	}
		950	} elsif ($status == 307 or $status == 308) {
		951	$redirect = 1;
		952	}
		953	}
		954
		955	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
		956	if ($state{handle}) {
		957	# handle keepalive
		958	if (
		959	$persistent
		960	&& $_[3]
		961	&& ($hdr{HTTPVersion} < 1.1
		962	? $hdr{connection} =~ /\bkeep-?alive\b/i
		963	: $hdr{connection} !~ /\bclose\b/i)
		964	) {
		965	ka_store $ka_key, delete $state{handle};
		966	} else {
		967	# no keepalive, destroy the handle
		968	$state{handle}->destroy;
		969	}
		970	}
		971
		972	%state = ();
		973
		974	if (defined $_[1]) {
		975	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		976	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		977	}
		978
		979	# set-cookie processing
		980	if ($arg{cookie_jar}) {
		981	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		982	}
		983
		984	if ($redirect && exists $hdr{location}) {
		985	# we ignore any errors, as it is very common to receive
		986	# Content-Length != 0 but no actual body
		987	# we also access %hdr, as $_[1] might be an erro
		988	$state{recurse} =
		989	http_request (
		990	$method => $hdr{location},
		991	%arg,
		992	recurse => $recurse - 1,
		993	Redirect => [$_[0], \%hdr],
		994	sub {
		995	%state = ();
		996	&$cb
		997	},
		998	);
		999	} else {
		1000	$cb->($_[0], \%hdr);
		1001	}
		1002	};
		1003
		1004	$ae_error = 597; # body phase
		1005
		1006	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		1007
		1008	my $len = $chunked ? undef : $hdr{"content-length"};
		1009
		1010	# body handling, many different code paths
		1011	# - no body expected
		1012	# - want_body_handle
		1013	# - te chunked
		1014	# - 2x length known (with or without on_body)
		1015	# - 2x length not known (with or without on_body)
		1016	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		1017	$finish->(undef, 598 => "Request cancelled by on_header");
		1018	} elsif (
		1019	$hdr{Status} =~ /^(?:1..|204|205|304)$/
		1020	or $method eq "HEAD"
		1021	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		1022	) {
		1023	# no body
		1024	$finish->("", undef, undef, 1);
		1025
		1026	} elsif (!$redirect && $arg{want_body_handle}) {
		1027	$_[0]->on_eof (undef);
		1028	$_[0]->on_error (undef);
		1029	$_[0]->on_read (undef);
		1030
		1031	$finish->(delete $state{handle});
		1032
		1033	} elsif ($chunked) {
		1034	my $cl = 0;
		1035	my $body = "";
		1036	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
		1037
		1038	$state{read_chunk} = sub {
		1039	$_[1] =~ /^([0-9a-fA-F]+)/
		1040	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1041
		1042	my $len = hex $1;
		1043
		1044	if ($len) {
		1045	$cl += $len;
		1046
		1047	$_[0]->push_read (chunk => $len, sub {
		1048	$on_body->($_[1], \%hdr)
		1049	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1050
		1051	$_[0]->push_read (line => sub {
		1052	length $_[1]
		1053	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1054	$_[0]->push_read (line => $state{read_chunk});
		1055	});
		1056	});
		1057	} else {
		1058	$hdr{"content-length"} ||= $cl;
		1059
		1060	$_[0]->push_read (line => $qr_nlnl, sub {
		1061	if (length $_[1]) {
		1062	for ("$_[1]") {
		1063	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		1064
		1065	my $hdr = _parse_hdr
		1066	or return $finish->(undef, $ae_error => "Garbled response trailers");
		1067
		1068	%hdr = (%hdr, %$hdr);
		1069	}
		1070	}
		1071
		1072	$finish->($body, undef, undef, 1);
		1073	});
		1074	}
		1075	};
		1076
		1077	$_[0]->push_read (line => $state{read_chunk});
		1078
		1079	} elsif ($arg{on_body}) {
		1080	if (defined $len) {
		1081	$_[0]->on_read (sub {
		1082	$len -= length $_[0]{rbuf};
		1083
		1084	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1085	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1086
		1087	$len > 0
		1088	or $finish->("", undef, undef, 1);
		1089	});
		1090	} else {
		1091	$_[0]->on_eof (sub {
		1092	$finish->("");
		1093	});
		1094	$_[0]->on_read (sub {
		1095	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1096	or $finish->(undef, 598 => "Request cancelled by on_body");
		1097	});
		1098	}
		1099	} else {
		1100	$_[0]->on_eof (undef);
		1101
		1102	if (defined $len) {
		1103	$_[0]->on_read (sub {
		1104	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1105	if $len <= length $_[0]{rbuf};
		1106	});
		1107	} else {
		1108	$_[0]->on_error (sub {
		1109	($! == Errno::EPIPE || !$!)
		1110	? $finish->(delete $_[0]{rbuf})
		1111	: $finish->(undef, $ae_error => $_[2]);
		1112	});
		1113	$_[0]->on_read (sub { });
		1114	}
		1115	}
		1116	};
		1117
		1118	# if keepalive is enabled, then the server closing the connection
		1119	# before a response can happen legally - we retry on idempotent methods.
		1120	if ($was_persistent && $idempotent) {
		1121	my $old_eof = $hdl->{on_eof};
		1122	$hdl->{on_eof} = sub {
		1123	_destroy_state %state;
		1124
		1125	%state = ();
		1126	$state{recurse} =
		1127	http_request (
		1128	$method => $url,
		1129	%arg,
		1130	recurse => $recurse - 1,
		1131	persistent => 0,
		1132	sub {
		1133	%state = ();
		1134	&$cb
		1135	}
		1136	);
		1137	};
		1138	$hdl->on_read (sub {
		1139	return unless %state;
		1140
		1141	# as soon as we receive something, a connection close
		1142	# once more becomes a hard error
		1143	$hdl->{on_eof} = $old_eof;
		1144	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1145	});
		1146	} else {
		1147	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1148	}
		1149	};
		1150
		1151	my $prepare_handle = sub {
		1152	my ($hdl) = $state{handle};
		1153
		1154	$hdl->on_error (sub {
		1155	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1156	});
		1157	$hdl->on_eof (sub {
		1158	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1159	});
		1160	$hdl->timeout_reset;
		1161	$hdl->timeout ($timeout);
		1162	};
		1163
		1164	# connected to proxy (or origin server)
		1165	my $connect_cb = sub {
		1166	my $fh = shift
		1167	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1168
		1169	return unless delete $state{connect_guard};
		1170
		1171	# get handle
		1172	$state{handle} = new AnyEvent::Handle
		1173	%{ $arg{handle_params} },
		1174	fh => $fh,
		1175	peername => $uhost,
		1176	tls_ctx => $arg{tls_ctx},
		1177	;
		1178
		1179	$prepare_handle->();
		1180
		1181	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1182
		1183	# now handle proxy-CONNECT method
		1184	if ($proxy && $uscheme eq "https") {
		1185	# oh dear, we have to wrap it into a connect request
		1186
		1187	my $auth = exists $hdr{"proxy-authorization"}
		1188	? "proxy-authorization: " . (delete $hdr{"proxy-authorization"}) . "\015\012"
		1189	: "";
		1190
		1191	# maybe re-use $uauthority with patched port?
		1192	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012$auth\015\012");
		1193	$state{handle}->push_read (line => $qr_nlnl, sub {
		1194	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1195	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1196
		1197	if ($2 == 200) {
		1198	$rpath = $upath;
		1199	$handle_actual_request->();
		1200	} else {
		1201	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
		1202	}
		1203	});
		1204	} else {
		1205	delete $hdr{"proxy-authorization"} unless $proxy;
		1206
		1207	$handle_actual_request->();
		1208	}
		1209	};
656		1210
657	_get_slot $uhost, sub {	1211	_get_slot $uhost, sub {
658	$state{slot_guard} = shift;	1212	$state{slot_guard} = shift;
659		1213
660	return unless $state{connect_guard};	1214	return unless $state{connect_guard};
661		1215
662	my $ae_error = 595; # connecting	1216	# try to use an existing keepalive connection, but only if we, ourselves, plan
		1217	# on a keepalive request (in theory, this should be a separate config option).
		1218	if ($persistent && $KA_CACHE{$ka_key}) {
		1219	$was_persistent = 1;
663		1220
664	my $connect_cb = sub {	1221	$state{handle} = ka_fetch $ka_key;
665	$state{fh} = shift
666	or do {
667	my $err = "$!";
668	%state = ();
669	return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err });
670	};
671
672	return unless delete $state{connect_guard};
673
674	# get handle
675	$state{handle} = new AnyEvent::Handle
676	fh => $state{fh},
677	peername => $rhost,
678	tls_ctx => $arg{tls_ctx},
679	# these need to be reconfigured on keepalive handles
680	timeout => $timeout,
681	on_error => sub {
682	%state = ();
683	$cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] });
684	},
685	on_eof => sub {
686	%state = ();
687	$cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" });
688	},
689	;
690
691	# limit the number of persistent connections
692	# keepalive not yet supported
693	# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
694	# ++$KA_COUNT{$_[1]};
695	# $state{handle}{ka_count_guard} = AnyEvent::Util::guard {
696	# --$KA_COUNT{$_[1]}
697	# };
698	# $hdr{connection} = "keep-alive";
699	# }
700
701	$state{handle}->starttls ("connect") if $rscheme eq "https";
702
703	# handle actual, non-tunneled, request
704	my $handle_actual_request = sub {
705	$ae_error = 596; # request phase
706
707	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};
708
709	# send request
710	$state{handle}->push_write (	1222	$state{handle}->destroyed
711	"$method $rpath HTTP/1.1\015\012"	1223	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
712	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)	1224	$prepare_handle->();
713	. "\015\012"	1225	$state{handle}->destroyed
714	. (delete $arg{body})	1226	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
715	);
716
717	# return if error occured during push_write()
718	return unless %state;
719
720	%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use
721
722	# status line and headers
723	$state{read_response} = sub {
724	for ("$_[1]") {
725	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
726
727	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
728	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" }));
729
730	# 100 Continue handling
731	# should not happen as we don't send expect: 100-continue,
732	# but we handle it just in case.
733	# since we send the request body regardless, if we get an error
734	# we are out of-sync, which we currently do NOT handle correctly.
735	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
736	if $2 eq 100;
737
738	push @pseudo,
739	HTTPVersion => $1,
740	Status => $2,
741	Reason => $3,
742	;
743
744	my $hdr = parse_hdr
745	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" }));
746
747	%hdr = (%$hdr, @pseudo);
748	}
749
750	# redirect handling
751	# microsoft and other shitheads don't give a shit for following standards,
752	# try to support some common forms of broken Location headers.
753	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
754	$hdr{location} =~ s/^\.\/+//;
755
756	my $url = "$rscheme://$uhost:$uport";
757
758	unless ($hdr{location} =~ s/^\///) {
759	$url .= $upath;
760	$url =~ s/\/[^\/]*$//;
761	}
762
763	$hdr{location} = "$url/$hdr{location}";
764	}
765
766	my $redirect;
767
768	if ($recurse) {
769	my $status = $hdr{Status};
770
771	# industry standard is to redirect POST as GET for
772	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
773	# also, the UA should ask the user for 301 and 307 and POST,
774	# industry standard seems to be to simply follow.
775	# we go with the industry standard.
776	if ($status == 301 or $status == 302 or $status == 303) {
777	# HTTP/1.1 is unclear on how to mutate the method
778	$method = "GET" unless $method eq "HEAD";
779	$redirect = 1;
780	} elsif ($status == 307) {
781	$redirect = 1;
782	}
783	}
784
785	my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive])
786	my $may_keep_alive = $_[3];
787
788	$state{handle}->destroy if $state{handle};
789	%state = ();
790
791	if (defined $_[1]) {
792	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
793	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
794	}
795
796	# set-cookie processing
797	if ($arg{cookie_jar}) {
798	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
799	}
800
801	if ($redirect && exists $hdr{location}) {
802	# we ignore any errors, as it is very common to receive
803	# Content-Length != 0 but no actual body
804	# we also access %hdr, as $_[1] might be an erro
805	http_request (
806	$method => $hdr{location},
807	%arg,
808	recurse => $recurse - 1,
809	Redirect => [$_[0], \%hdr],
810	$cb);
811	} else {
812	$cb->($_[0], \%hdr);
813	}
814	};
815
816	$ae_error = 597; # body phase
817
818	my $len = $hdr{"content-length"};
819
820	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
821	$finish->(undef, 598 => "Request cancelled by on_header");
822	} elsif (
823	$hdr{Status} =~ /^(?:1..|204|205|304)$/
824	or $method eq "HEAD"
825	or (defined $len && !$len)
826	) {
827	# no body
828	$finish->("", undef, undef, 1);
829	} else {
830	# body handling, many different code paths
831	# - no body expected
832	# - want_body_handle
833	# - te chunked
834	# - 2x length known (with or without on_body)
835	# - 2x length not known (with or without on_body)
836	if (!$redirect && $arg{want_body_handle}) {
837	$_[0]->on_eof (undef);
838	$_[0]->on_error (undef);
839	$_[0]->on_read (undef);
840
841	$finish->(delete $state{handle});
842
843	} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) {
844	my $cl = 0;
845	my $body = undef;
846	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
847
848	$state{read_chunk} = sub {
849	$_[1] =~ /^([0-9a-fA-F]+)/
850	or $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
851
852	my $len = hex $1;
853
854	if ($len) {
855	$cl += $len;
856
857	$_[0]->push_read (chunk => $len, sub {
858	$on_body->($_[1], \%hdr)
859	or return $finish->(undef, 598 => "Request cancelled by on_body");
860
861	$_[0]->push_read (line => sub {
862	length $_[1]
863	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
864	$_[0]->push_read (line => $state{read_chunk});
865	});
866	});
867	} else {
868	$hdr{"content-length"} ||= $cl;
869
870	$_[0]->push_read (line => $qr_nlnl, sub {
871	if (length $_[1]) {
872	for ("$_[1]") {
873	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
874
875	my $hdr = parse_hdr
876	or return $finish->(undef, $ae_error => "Garbled response trailers");
877
878	%hdr = (%hdr, %$hdr);
879	}
880	}
881
882	$finish->($body, undef, undef, 1);
883	});
884	}
885	};
886
887	$_[0]->push_read (line => $state{read_chunk});
888
889	} elsif ($arg{on_body}) {
890	if ($len) {
891	$_[0]->on_read (sub {
892	$len -= length $_[0]{rbuf};
893
894	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
895	or return $finish->(undef, 598 => "Request cancelled by on_body");
896
897	$len > 0
898	or $finish->("", undef, undef, 1);
899	});
900	} else {
901	$_[0]->on_eof (sub {
902	$finish->("");
903	});
904	$_[0]->on_read (sub {
905	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
906	or $finish->(undef, 598 => "Request cancelled by on_body");
907	});
908	}
909	} else {
910	$_[0]->on_eof (undef);
911
912	if ($len) {
913	$_[0]->on_read (sub {
914	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
915	if $len <= length $_[0]{rbuf};
916	});
917	} else {
918	$_[0]->on_error (sub {
919	($! == Errno::EPIPE || !$!)
920	? $finish->(delete $_[0]{rbuf})
921	: $finish->(undef, $ae_error => $_[2]);
922	});
923	$_[0]->on_read (sub { });
924	}
925	}
926	}
927	};
928
929	$state{handle}->push_read (line => $qr_nlnl, $state{read_response});
930	};
931
932	# now handle proxy-CONNECT method
933	if ($proxy && $uscheme eq "https") {
934	# oh dear, we have to wrap it into a connect request
935
936	# maybe re-use $uauthority with patched port?
937	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012");
938	$state{handle}->push_read (line => $qr_nlnl, sub {
939	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
940	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }));
941
942	if ($2 == 200) {
943	$rpath = $upath;
944	$handle_actual_request->();
945	} else {
946	%state = ();
947	$cb->(undef, { @pseudo, Status => $2, Reason => $3 });
948	}
949	});
950	} else {
951	$handle_actual_request->();	1227	$handle_actual_request->();
952	}
953	};
954		1228
		1229	} else {
955	my $tcp_connect = $arg{tcp_connect}	1230	my $tcp_connect = $arg{tcp_connect}
956	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };	1231	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
957		1232
958	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });	1233	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });
959		1234	}
960	};	1235	};
961		1236
962	defined wantarray && AnyEvent::Util::guard { %state = () }	1237	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
963	}	1238	}
964		1239
965	sub http_get($@) {	1240	sub http_get($@) {
966	unshift @_, "GET";	1241	unshift @_, "GET";
967	&http_request	1242	&http_request
…		…
985	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for	1260	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
986	the actual connection, which in turn uses AnyEvent::DNS to resolve	1261	the actual connection, which in turn uses AnyEvent::DNS to resolve
987	hostnames. The latter is a simple stub resolver and does no caching	1262	hostnames. The latter is a simple stub resolver and does no caching
988	on its own. If you want DNS caching, you currently have to provide	1263	on its own. If you want DNS caching, you currently have to provide
989	your own default resolver (by storing a suitable resolver object in	1264	your own default resolver (by storing a suitable resolver object in
990	C<$AnyEvent::DNS::RESOLVER>).	1265	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
991		1266
992	=head2 GLOBAL FUNCTIONS AND VARIABLES	1267	=head2 GLOBAL FUNCTIONS AND VARIABLES
993		1268
994	=over 4	1269	=over 4
995		1270
996	=item AnyEvent::HTTP::set_proxy "proxy-url"	1271	=item AnyEvent::HTTP::set_proxy "proxy-url"
997		1272
998	Sets the default proxy server to use. The proxy-url must begin with a	1273	Sets the default proxy server to use. The proxy-url must begin with a
999	string of the form C<http://host:port> (optionally C<https:...>), croaks	1274	string of the form C<http://host:port>, croaks otherwise.
1000	otherwise.
1001		1275
1002	To clear an already-set proxy, use C<undef>.	1276	To clear an already-set proxy, use C<undef>.
		1277
		1278	When AnyEvent::HTTP is loaded for the first time it will query the
		1279	default proxy from the operating system, currently by looking at
		1280	C<$ENV{http_proxy>}.
1003		1281
1004	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]	1282	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
1005		1283
1006	Remove all cookies from the cookie jar that have been expired. If	1284	Remove all cookies from the cookie jar that have been expired. If
1007	C<$session_end> is given and true, then additionally remove all session	1285	C<$session_end> is given and true, then additionally remove all session
1008	cookies.	1286	cookies.
1009		1287
1010	You should call this function (with a true C<$session_end>) before you	1288	You should call this function (with a true C<$session_end>) before you
1011	save cookies to disk, and you should call this function after loading them	1289	save cookies to disk, and you should call this function after loading them
1012	again. If you have a long-running program you can additonally call this	1290	again. If you have a long-running program you can additionally call this
1013	function from time to time.	1291	function from time to time.
1014		1292
1015	A cookie jar is initially an empty hash-reference that is managed by this	1293	A cookie jar is initially an empty hash-reference that is managed by this
1016	module. It's format is subject to change, but currently it is like this:	1294	module. Its format is subject to change, but currently it is as follows:
1017		1295
1018	The key C<version> has to contain C<1>, otherwise the hash gets	1296	The key C<version> has to contain C<1>, otherwise the hash gets
1019	emptied. All other keys are hostnames or IP addresses pointing to	1297	emptied. All other keys are hostnames or IP addresses pointing to
1020	hash-references. The key for these inner hash references is the	1298	hash-references. The key for these inner hash references is the
1021	server path for which this cookie is meant, and the values are again	1299	server path for which this cookie is meant, and the values are again
1022	hash-references. The keys of those hash-references is the cookie name, and	1300	hash-references. Each key of those hash-references is a cookie name, and
1023	the value, you guessed it, is another hash-reference, this time with the	1301	the value, you guessed it, is another hash-reference, this time with the
1024	key-value pairs from the cookie, except for C<expires> and C<max-age>,	1302	key-value pairs from the cookie, except for C<expires> and C<max-age>,
1025	which have been replaced by a C<_expires> key that contains the cookie	1303	which have been replaced by a C<_expires> key that contains the cookie
1026	expiry timestamp.	1304	expiry timestamp. Session cookies are indicated by not having an
		1305	C<_expires> key.
1027		1306
1028	Here is an example of a cookie jar with a single cookie, so you have a	1307	Here is an example of a cookie jar with a single cookie, so you have a
1029	chance of understanding the above paragraph:	1308	chance of understanding the above paragraph:
1030		1309
1031	{	1310	{
…		…
1053		1332
1054	=item $AnyEvent::HTTP::MAX_RECURSE	1333	=item $AnyEvent::HTTP::MAX_RECURSE
1055		1334
1056	The default value for the C<recurse> request parameter (default: C<10>).	1335	The default value for the C<recurse> request parameter (default: C<10>).
1057		1336
		1337	=item $AnyEvent::HTTP::TIMEOUT
		1338
		1339	The default timeout for connection operations (default: C<300>).
		1340
1058	=item $AnyEvent::HTTP::USERAGENT	1341	=item $AnyEvent::HTTP::USERAGENT
1059		1342
1060	The default value for the C<User-Agent> header (the default is	1343	The default value for the C<User-Agent> header (the default is
1061	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1344	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
1062		1345
1063	=item $AnyEvent::HTTP::MAX_PER_HOST	1346	=item $AnyEvent::HTTP::MAX_PER_HOST
1064		1347
1065	The maximum number of concurrent connections to the same host (identified	1348	The maximum number of concurrent connections to the same host (identified
1066	by the hostname). If the limit is exceeded, then the additional requests	1349	by the hostname). If the limit is exceeded, then additional requests
1067	are queued until previous connections are closed.	1350	are queued until previous connections are closed. Both persistent and
		1351	non-persistent connections are counted in this limit.
1068		1352
1069	The default value for this is C<4>, and it is highly advisable to not	1353	The default value for this is C<4>, and it is highly advisable to not
1070	increase it.	1354	increase it much.
		1355
		1356	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1357	connections, older browsers used 2, newer ones (such as firefox 3)
		1358	typically use 6, and Opera uses 8 because like, they have the fastest
		1359	browser and give a shit for everybody else on the planet.
		1360
		1361	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
		1362
		1363	The time after which idle persistent connections get closed by
		1364	AnyEvent::HTTP (default: C<3>).
1071		1365
1072	=item $AnyEvent::HTTP::ACTIVE	1366	=item $AnyEvent::HTTP::ACTIVE
1073		1367
1074	The number of active connections. This is not the number of currently	1368	The number of active connections. This is not the number of currently
1075	running requests, but the number of currently open and non-idle TCP	1369	running requests, but the number of currently open and non-idle TCP
1076	connections. This number of can be useful for load-leveling.	1370	connections. This number can be useful for load-leveling.
1077		1371
1078	=back	1372	=back
1079		1373
1080	=cut	1374	=cut
1081		1375
…		…
1115	# other formats fail in the loop below	1409	# other formats fail in the loop below
1116		1410
1117	for (0..11) {	1411	for (0..11) {
1118	if ($m eq $month[$_]) {	1412	if ($m eq $month[$_]) {
1119	require Time::Local;	1413	require Time::Local;
1120	return Time::Local::timegm ($S, $M, $H, $d, $_, $y);	1414	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
1121	}	1415	}
1122	}	1416	}
1123		1417
1124	undef	1418	undef
1125	}	1419	}
1126		1420
1127	sub set_proxy($) {	1421	sub set_proxy($) {
1128	if (length $_[0]) {	1422	if (length $_[0]) {
1129	$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix	1423	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
1130	or Carp::croak "$_[0]: invalid proxy URL";	1424	or Carp::croak "$_[0]: invalid proxy URL";
1131	$PROXY = [$2, $3 || 3128, $1]	1425	$PROXY = [$2, $3 || 3128, $1]
1132	} else {	1426	} else {
1133	undef $PROXY;	1427	undef $PROXY;
1134	}	1428	}
…		…
1137	# initialise proxy from environment	1431	# initialise proxy from environment
1138	eval {	1432	eval {
1139	set_proxy $ENV{http_proxy};	1433	set_proxy $ENV{http_proxy};
1140	};	1434	};
1141		1435
		1436	=head2 SHOWCASE
		1437
		1438	This section contains some more elaborate "real-world" examples or code
		1439	snippets.
		1440
		1441	=head2 HTTP/1.1 FILE DOWNLOAD
		1442
		1443	Downloading files with HTTP can be quite tricky, especially when something
		1444	goes wrong and you want to resume.
		1445
		1446	Here is a function that initiates and resumes a download. It uses the
		1447	last modified time to check for file content changes, and works with many
		1448	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1449	on older servers.
		1450
		1451	It calls the completion callback with either C<undef>, which means a
		1452	nonretryable error occurred, C<0> when the download was partial and should
		1453	be retried, and C<1> if it was successful.
		1454
		1455	use AnyEvent::HTTP;
		1456
		1457	sub download($$$) {
		1458	my ($url, $file, $cb) = @_;
		1459
		1460	open my $fh, "+<", $file
		1461	or die "$file: $!";
		1462
		1463	my %hdr;
		1464	my $ofs = 0;
		1465
		1466	if (stat $fh and -s _) {
		1467	$ofs = -s _;
		1468	warn "-s is ", $ofs;
		1469	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1470	$hdr{"range"} = "bytes=$ofs-";
		1471	}
		1472
		1473	http_get $url,
		1474	headers => \%hdr,
		1475	on_header => sub {
		1476	my ($hdr) = @_;
		1477
		1478	if ($hdr->{Status} == 200 && $ofs) {
		1479	# resume failed
		1480	truncate $fh, $ofs = 0;
		1481	}
		1482
		1483	sysseek $fh, $ofs, 0;
		1484
		1485	1
		1486	},
		1487	on_body => sub {
		1488	my ($data, $hdr) = @_;
		1489
		1490	if ($hdr->{Status} =~ /^2/) {
		1491	length $data == syswrite $fh, $data
		1492	or return; # abort on write errors
		1493	}
		1494
		1495	1
		1496	},
		1497	sub {
		1498	my (undef, $hdr) = @_;
		1499
		1500	my $status = $hdr->{Status};
		1501
		1502	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1503	utime $time, $time, $fh;
		1504	}
		1505
		1506	if ($status == 200 || $status == 206 || $status == 416) {
		1507	# download ok || resume ok || file already fully downloaded
		1508	$cb->(1, $hdr);
		1509
		1510	} elsif ($status == 412) {
		1511	# file has changed while resuming, delete and retry
		1512	unlink $file;
		1513	$cb->(0, $hdr);
		1514
		1515	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1516	# retry later
		1517	$cb->(0, $hdr);
		1518
		1519	} else {
		1520	$cb->(undef, $hdr);
		1521	}
		1522	}
		1523	;
		1524	}
		1525
		1526	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1527	if ($_[0]) {
		1528	print "OK!\n";
		1529	} elsif (defined $_[0]) {
		1530	print "please retry later\n";
		1531	} else {
		1532	print "ERROR\n";
		1533	}
		1534	};
		1535
1142	=head2 SOCKS PROXIES	1536	=head3 SOCKS PROXIES
1143		1537
1144	Socks proxies are not directly supported by AnyEvent::HTTP. You can	1538	Socks proxies are not directly supported by AnyEvent::HTTP. You can
1145	compile your perl to support socks, or use an external program such as	1539	compile your perl to support socks, or use an external program such as
1146	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy	1540	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
1147	transparently.	1541	transparently.

Diff Legend

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