ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.79 by root, Sat Jan 1 20:01:07 2011 UTC vs.
Revision 1.117 by root, Mon Sep 9 21:41:43 2013 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 = '2.15';
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.
166		161
167	=item headers => hashref	162	=item headers => hashref
168		163
169	The request headers to use. Currently, C<http_request> may provide its own	164	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	165	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:>	166	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	167	(this can be suppressed by using C<undef> for these headers in which case
173	they won't be sent at all).	168	they won't be sent at all).
174		169
		170	You really should provide your own C<User-Agent:> header value that is
		171	appropriate for your program - I wouldn't be surprised if the default
		172	AnyEvent string gets blocked by webservers sooner or later.
		173
		174	Also, make sure that your headers names and values do not contain any
		175	embedded newlines.
		176
175	=item timeout => $seconds	177	=item timeout => $seconds
176		178
177	The time-out to use for various stages - each connect attempt will reset	179	The time-out to use for various stages - each connect attempt will reset
178	the timeout, as will read or write activity, i.e. this is not an overall	180	the timeout, as will read or write activity, i.e. this is not an overall
179	timeout.	181	timeout.
180		182
181	Default timeout is 5 minutes.	183	Default timeout is 5 minutes.
182		184
183	=item proxy => [$host, $port[, $scheme]] or undef	185	=item proxy => [$host, $port[, $scheme]] or undef
184		186
185	Use the given http proxy for all requests. If not specified, then the	187	Use the given http proxy for all requests, or no proxy if C<undef> is
186	default proxy (as specified by C<$ENV{http_proxy}>) is used.	188	used.
187		189
188	C<$scheme> must be either missing, C<http> for HTTP or C<https> for	190	C<$scheme> must be either missing or must be C<http> for HTTP.
189	HTTPS.	191
		192	If not specified, then the default proxy is used (see
		193	C<AnyEvent::HTTP::set_proxy>).
190		194
191	=item body => $string	195	=item body => $string
192		196
193	The request body, usually empty. Will be sent as-is (future versions of	197	The request body, usually empty. Will be sent as-is (future versions of
194	this module might offer more options).	198	this module might offer more options).
…		…
196	=item cookie_jar => $hash_ref	200	=item cookie_jar => $hash_ref
197		201
198	Passing this parameter enables (simplified) cookie-processing, loosely	202	Passing this parameter enables (simplified) cookie-processing, loosely
199	based on the original netscape specification.	203	based on the original netscape specification.
200		204
201	The C<$hash_ref> must be an (initially empty) hash reference which will	205	The C<$hash_ref> must be an (initially empty) hash reference which
202	get updated automatically. It is possible to save the cookie jar to	206	will get updated automatically. It is possible to save the cookie jar
203	persistent storage with something like JSON or Storable, but this is not	207	to persistent storage with something like JSON or Storable - see the
204	recommended, as session-only cookies might survive longer than expected.	208	C<AnyEvent::HTTP::cookie_jar_expire> function if you wish to remove
		209	expired or session-only cookies, and also for documentation on the format
		210	of the cookie jar.
205		211
206	Note that this cookie implementation is not meant to be complete. If	212	Note that this cookie implementation is not meant to be complete. If
207	you want complete cookie management you have to do that on your	213	you want complete cookie management you have to do that on your
208	own. C<cookie_jar> is meant as a quick fix to get some cookie-using sites	214	own. C<cookie_jar> is meant as a quick fix to get most cookie-using sites
209	working. Cookies are a privacy disaster, do not use them unless required	215	working. Cookies are a privacy disaster, do not use them unless required
210	to.	216	to.
211		217
212	When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:>	218	When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:>
213	headers will be set and handled by this module, otherwise they will be	219	headers will be set and handled by this module, otherwise they will be
…		…
223	verification) TLS context.	229	verification) TLS context.
224		230
225	The default for this option is C<low>, which could be interpreted as "give	231	The default for this option is C<low>, which could be interpreted as "give
226	me the page, no matter what".	232	me the page, no matter what".
227		233
		234	See also the C<sessionid> parameter.
		235
		236	=item session => $string
		237
		238	The module might reuse connections to the same host internally. Sometimes
		239	(e.g. when using TLS), you do not want to reuse connections from other
		240	sessions. This can be achieved by setting this parameter to some unique
		241	ID (such as the address of an object storing your state data, or the TLS
		242	context) - only connections using the same unique ID will be reused.
		243
228	=item on_prepare => $callback->($fh)	244	=item on_prepare => $callback->($fh)
229		245
230	In rare cases you need to "tune" the socket before it is used to	246	In rare cases you need to "tune" the socket before it is used to
231	connect (for exmaple, to bind it on a given IP address). This parameter	247	connect (for example, to bind it on a given IP address). This parameter
232	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>	248	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>
233	and behaves exactly the same way (e.g. it has to provide a	249	and behaves exactly the same way (e.g. it has to provide a
234	timeout). See the description for the C<$prepare_cb> argument of	250	timeout). See the description for the C<$prepare_cb> argument of
235	C<AnyEvent::Socket::tcp_connect> for details.	251	C<AnyEvent::Socket::tcp_connect> for details.
236		252
…		…
301	called. Instead of the C<$body> argument containing the body data, the	317	called. Instead of the C<$body> argument containing the body data, the
302	callback will receive the L<AnyEvent::Handle> object associated with the	318	callback will receive the L<AnyEvent::Handle> object associated with the
303	connection. In error cases, C<undef> will be passed. When there is no body	319	connection. In error cases, C<undef> will be passed. When there is no body
304	(e.g. status C<304>), the empty string will be passed.	320	(e.g. status C<304>), the empty string will be passed.
305		321
306	The handle object might or might not be in TLS mode, might be connected to	322	The handle object might or might not be in TLS mode, might be connected
307	a proxy, be a persistent connection etc., and configured in unspecified	323	to a proxy, be a persistent connection, use chunked transfer encoding
308	ways. The user is responsible for this handle (it will not be used by this	324	etc., and configured in unspecified ways. The user is responsible for this
309	module anymore).	325	handle (it will not be used by this module anymore).
310		326
311	This is useful with some push-type services, where, after the initial	327	This is useful with some push-type services, where, after the initial
312	headers, an interactive protocol is used (typical example would be the	328	headers, an interactive protocol is used (typical example would be the
313	push-style twitter API which starts a JSON/XML stream).	329	push-style twitter API which starts a JSON/XML stream).
314		330
315	If you think you need this, first have a look at C<on_body>, to see if	331	If you think you need this, first have a look at C<on_body>, to see if
316	that doesn't solve your problem in a better way.	332	that doesn't solve your problem in a better way.
		333
		334	=item persistent => $boolean
		335
		336	Try to create/reuse a persistent connection. When this flag is set
		337	(default: true for idempotent requests, false for all others), then
		338	C<http_request> tries to re-use an existing (previously-created)
		339	persistent connection to the host and, failing that, tries to create a new
		340	one.
		341
		342	Requests failing in certain ways will be automatically retried once, which
		343	is dangerous for non-idempotent requests, which is why it defaults to off
		344	for them. The reason for this is because the bozos who designed HTTP/1.1
		345	made it impossible to distinguish between a fatal error and a normal
		346	connection timeout, so you never know whether there was a problem with
		347	your request or not.
		348
		349	When reusing an existent connection, many parameters (such as TLS context)
		350	will be ignored. See the C<session> parameter for a workaround.
		351
		352	=item keepalive => $boolean
		353
		354	Only used when C<persistent> is also true. This parameter decides whether
		355	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		356	(as opposed to only a HTTP/1.1 persistent connection).
		357
		358	The default is true, except when using a proxy, in which case it defaults
		359	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		360
		361	=item handle_params => { key => value ... }
		362
		363	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		364	constructor that is called - not all requests will create a handle, and
		365	sometimes more than one is created, so this parameter is only good for
		366	setting hints.
		367
		368	Example: set the maximum read size to 4096, to potentially conserve memory
		369	at the cost of speed.
		370
		371	handle_params => {
		372	max_read_size => 4096,
		373	},
317		374
318	=back	375	=back
319		376
320	Example: do a simple HTTP GET request for http://www.nethype.de/ and print	377	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
321	the response body.	378	the response body.
…		…
327		384
328	Example: do a HTTP HEAD request on https://www.google.com/, use a	385	Example: do a HTTP HEAD request on https://www.google.com/, use a
329	timeout of 30 seconds.	386	timeout of 30 seconds.
330		387
331	http_request	388	http_request
332	GET => "https://www.google.com",	389	HEAD => "https://www.google.com",
		390	headers => { "user-agent" => "MySearchClient 1.0" },
333	timeout => 30,	391	timeout => 30,
334	sub {	392	sub {
335	my ($body, $hdr) = @_;	393	my ($body, $hdr) = @_;
336	use Data::Dumper;	394	use Data::Dumper;
337	print Dumper $hdr;	395	print Dumper $hdr;
…		…
347	};	405	};
348		406
349	undef $request;	407	undef $request;
350		408
351	=cut	409	=cut
		410
		411	#############################################################################
		412	# wait queue/slots
352		413
353	sub _slot_schedule;	414	sub _slot_schedule;
354	sub _slot_schedule($) {	415	sub _slot_schedule($) {
355	my $host = shift;	416	my $host = shift;
356		417
…		…
378	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];	439	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];
379		440
380	_slot_schedule $_[0];	441	_slot_schedule $_[0];
381	}	442	}
382		443
		444	#############################################################################
		445	# cookie handling
		446
		447	# expire cookies
		448	sub cookie_jar_expire($;$) {
		449	my ($jar, $session_end) = @_;
		450
		451	%$jar = () if $jar->{version} != 1;
		452
		453	my $anow = AE::now;
		454
		455	while (my ($chost, $paths) = each %$jar) {
		456	next unless ref $paths;
		457
		458	while (my ($cpath, $cookies) = each %$paths) {
		459	while (my ($cookie, $kv) = each %$cookies) {
		460	if (exists $kv->{_expires}) {
		461	delete $cookies->{$cookie}
		462	if $anow > $kv->{_expires};
		463	} elsif ($session_end) {
		464	delete $cookies->{$cookie};
		465	}
		466	}
		467
		468	delete $paths->{$cpath}
		469	unless %$cookies;
		470	}
		471
		472	delete $jar->{$chost}
		473	unless %$paths;
		474	}
		475	}
		476
383	# extract cookies from jar	477	# extract cookies from jar
384	sub cookie_jar_extract($$$$) {	478	sub cookie_jar_extract($$$$) {
385	my ($jar, $uscheme, $uhost, $upath) = @_;	479	my ($jar, $scheme, $host, $path) = @_;
386		480
387	%$jar = () if $jar->{version} != 1;	481	%$jar = () if $jar->{version} != 1;
388		482
389	my @cookies;	483	my @cookies;
390		484
391	while (my ($chost, $paths) = each %$jar) {	485	while (my ($chost, $paths) = each %$jar) {
392	next unless ref $paths;	486	next unless ref $paths;
393		487
394	if ($chost =~ /^\./) {	488	if ($chost =~ /^\./) {
395	next unless $chost eq substr $uhost, -length $chost;	489	next unless $chost eq substr $host, -length $chost;
396	} elsif ($chost =~ /\./) {	490	} elsif ($chost =~ /\./) {
397	next unless $chost eq $uhost;	491	next unless $chost eq $host;
398	} else {	492	} else {
399	next;	493	next;
400	}	494	}
401		495
402	while (my ($cpath, $cookies) = each %$paths) {	496	while (my ($cpath, $cookies) = each %$paths) {
403	next unless $cpath eq substr $upath, 0, length $cpath;	497	next unless $cpath eq substr $path, 0, length $cpath;
404		498
405	while (my ($cookie, $kv) = each %$cookies) {	499	while (my ($cookie, $kv) = each %$cookies) {
406	next if $uscheme ne "https" && exists $kv->{secure};	500	next if $scheme ne "https" && exists $kv->{secure};
407		501
408	if (exists $kv->{expires}) {	502	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
409	if (AE::now > parse_date ($kv->{expires})) {
410	delete $cookies->{$cookie};	503	delete $cookies->{$cookie};
411	next;	504	next;
412	}
413	}	505	}
414		506
415	my $value = $kv->{value};	507	my $value = $kv->{value};
416		508
417	if ($value =~ /[=;,[:space:]]/) {	509	if ($value =~ /[=;,[:space:]]/) {
…		…
426		518
427	\@cookies	519	\@cookies
428	}	520	}
429		521
430	# parse set_cookie header into jar	522	# parse set_cookie header into jar
431	sub cookie_jar_set_cookie($$$) {	523	sub cookie_jar_set_cookie($$$$) {
432	my ($jar, $set_cookie, $uhost) = @_;	524	my ($jar, $set_cookie, $host, $date) = @_;
		525
		526	my $anow = int AE::now;
		527	my $snow; # server-now
433		528
434	for ($set_cookie) {	529	for ($set_cookie) {
435	# parse NAME=VALUE	530	# parse NAME=VALUE
436	my @kv;	531	my @kv;
437		532
…		…
440	while (	535	while (
441	m{	536	m{
442	\G\s*	537	\G\s*
443	(?:	538	(?:
444	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)	539	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
445	| ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )	540	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
446	)	541	)
447	}gcxsi	542	}gcxsi
448	) {	543	) {
449	my $name = $2;	544	my $name = $2;
450	my $value = $4;	545	my $value = $4;
451		546
452	unless (defined $name) {	547	if (defined $1) {
453	# expires	548	# expires
454	$name = "expires";	549	$name = "expires";
455	$value = $1;	550	$value = $1;
456	} elsif (!defined $value) {	551	} elsif (defined $3) {
457	# quoted	552	# quoted
458	$value = $3;	553	$value = $3;
459	$value =~ s/\\(.)/$1/gs;	554	$value =~ s/\\(.)/$1/gs;
460	}	555	}
461		556
462	push @kv, lc $name, $value;	557	push @kv, @kv ? lc $name : $name, $value;
463		558
464	last unless /\G\s*;/gc;	559	last unless /\G\s*;/gc;
465	}	560	}
466		561
467	last unless @kv;	562	last unless @kv;
468		563
469	my $name = shift @kv;	564	my $name = shift @kv;
470	my %kv = (value => shift @kv, @kv);	565	my %kv = (value => shift @kv, @kv);
471		566
472	$kv{expires} ||= format_date (AE::now + $kv{"max-age"})
473	if exists $kv{"max-age"};	567	if (exists $kv{"max-age"}) {
		568	$kv{_expires} = $anow + delete $kv{"max-age"};
		569	} elsif (exists $kv{expires}) {
		570	$snow ||= parse_date ($date) || $anow;
		571	$kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow);
		572	} else {
		573	delete $kv{_expires};
		574	}
474		575
475	my $cdom;	576	my $cdom;
476	my $cpath = (delete $kv{path}) || "/";	577	my $cpath = (delete $kv{path}) || "/";
477		578
478	if (exists $kv{domain}) {	579	if (exists $kv{domain}) {
…		…
484		585
485	# this is not rfc-like and not netscape-like. go figure.	586	# this is not rfc-like and not netscape-like. go figure.
486	my $ndots = $cdom =~ y/.//;	587	my $ndots = $cdom =~ y/.//;
487	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);	588	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
488	} else {	589	} else {
489	$cdom = $uhost;	590	$cdom = $host;
490	}	591	}
491		592
492	# store it	593	# store it
493	$jar->{version} = 1;	594	$jar->{version} = 1;
494	$jar->{$cdom}{$cpath}{$name} = \%kv;	595	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
495		596
496	redo if /\G\s*,/gc;	597	redo if /\G\s*,/gc;
497	}	598	}
498	}	599	}
499		600
		601	#############################################################################
		602	# keepalive/persistent connection cache
		603
		604	# fetch a connection from the keepalive cache
		605	sub ka_fetch($) {
		606	my $ka_key = shift;
		607
		608	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		609	delete $KA_CACHE{$ka_key}
		610	unless @{ $KA_CACHE{$ka_key} };
		611
		612	$hdl
		613	}
		614
		615	sub ka_store($$) {
		616	my ($ka_key, $hdl) = @_;
		617
		618	my $kaa = $KA_CACHE{$ka_key} ||= [];
		619
		620	my $destroy = sub {
		621	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		622
		623	$hdl->destroy;
		624
		625	@ka
		626	? $KA_CACHE{$ka_key} = \@ka
		627	: delete $KA_CACHE{$ka_key};
		628	};
		629
		630	# on error etc., destroy
		631	$hdl->on_error ($destroy);
		632	$hdl->on_eof ($destroy);
		633	$hdl->on_read ($destroy);
		634	$hdl->timeout ($PERSISTENT_TIMEOUT);
		635
		636	push @$kaa, $hdl;
		637	shift @$kaa while @$kaa > $MAX_PER_HOST;
		638	}
		639
		640	#############################################################################
		641	# utilities
		642
500	# continue to parse $_ for headers and place them into the arg	643	# continue to parse $_ for headers and place them into the arg
501	sub parse_hdr() {	644	sub _parse_hdr() {
502	my %hdr;	645	my %hdr;
503		646
504	# things seen, not parsed:	647	# things seen, not parsed:
505	# p3pP="NON CUR OTPi OUR NOR UNI"	648	# p3pP="NON CUR OTPi OUR NOR UNI"
506		649
…		…
520	for values %hdr;	663	for values %hdr;
521		664
522	\%hdr	665	\%hdr
523	}	666	}
524		667
		668	#############################################################################
		669	# http_get
		670
525	our $qr_nlnl = qr{(?<![^\012])\015?\012};	671	our $qr_nlnl = qr{(?<![^\012])\015?\012};
526		672
527	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	673	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
528	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };	674	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };
		675
		676	# maybe it should just become a normal object :/
		677
		678	sub _destroy_state(\%) {
		679	my ($state) = @_;
		680
		681	$state->{handle}->destroy if $state->{handle};
		682	%$state = ();
		683	}
		684
		685	sub _error(\%$$) {
		686	my ($state, $cb, $hdr) = @_;
		687
		688	&_destroy_state ($state);
		689
		690	$cb->(undef, $hdr);
		691	()
		692	}
		693
		694	our %IDEMPOTENT = (
		695	DELETE => 1,
		696	GET => 1,
		697	HEAD => 1,
		698	OPTIONS => 1,
		699	PUT => 1,
		700	TRACE => 1,
		701
		702	ACL => 1,
		703	"BASELINE-CONTROL" => 1,
		704	BIND => 1,
		705	CHECKIN => 1,
		706	CHECKOUT => 1,
		707	COPY => 1,
		708	LABEL => 1,
		709	LINK => 1,
		710	MERGE => 1,
		711	MKACTIVITY => 1,
		712	MKCALENDAR => 1,
		713	MKCOL => 1,
		714	MKREDIRECTREF => 1,
		715	MKWORKSPACE => 1,
		716	MOVE => 1,
		717	ORDERPATCH => 1,
		718	PROPFIND => 1,
		719	PROPPATCH => 1,
		720	REBIND => 1,
		721	REPORT => 1,
		722	SEARCH => 1,
		723	UNBIND => 1,
		724	UNCHECKOUT => 1,
		725	UNLINK => 1,
		726	UNLOCK => 1,
		727	UPDATE => 1,
		728	UPDATEREDIRECTREF => 1,
		729	"VERSION-CONTROL" => 1,
		730	);
529		731
530	sub http_request($$@) {	732	sub http_request($$@) {
531	my $cb = pop;	733	my $cb = pop;
532	my ($method, $url, %arg) = @_;	734	my ($method, $url, %arg) = @_;
533		735
…		…
551	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	753	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
552		754
553	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })	755	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
554	if $recurse < 0;	756	if $recurse < 0;
555		757
556	my $proxy = $arg{proxy} || $PROXY;	758	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
557	my $timeout = $arg{timeout} || $TIMEOUT;	759	my $timeout = $arg{timeout} || $TIMEOUT;
558		760
559	my ($uscheme, $uauthority, $upath, $query, $fragment) =	761	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
560	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|;	762	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
561		763
562	$uscheme = lc $uscheme;	764	$uscheme = lc $uscheme;
563		765
564	my $uport = $uscheme eq "http" ? 80	766	my $uport = $uscheme eq "http" ? 80
565	: $uscheme eq "https" ? 443	767	: $uscheme eq "https" ? 443
566	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });	768	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
567		769
568	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	770	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x
569	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });	771	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
570		772
571	my $uhost = $1;	773	my $uhost = lc $1;
572	$uport = $2 if defined $2;	774	$uport = $2 if defined $2;
573		775
574	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"	776	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
575	unless exists $hdr{host};	777	unless exists $hdr{host};
576		778
…		…
595	$rscheme = "http" unless defined $rscheme;	797	$rscheme = "http" unless defined $rscheme;
596		798
597	# don't support https requests over https-proxy transport,	799	# don't support https requests over https-proxy transport,
598	# can't be done with tls as spec'ed, unless you double-encrypt.	800	# can't be done with tls as spec'ed, unless you double-encrypt.
599	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";	801	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";
		802
		803	$rhost = lc $rhost;
		804	$rscheme = lc $rscheme;
600	} else {	805	} else {
601	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);	806	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
602	}	807	}
603		808
604	# leave out fragment and query string, just a heuristic	809	# leave out fragment and query string, just a heuristic
…		…
606	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};	811	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
607		812
608	$hdr{"content-length"} = length $arg{body}	813	$hdr{"content-length"} = length $arg{body}
609	if length $arg{body} || $method ne "GET";	814	if length $arg{body} || $method ne "GET";
610		815
611	$hdr{connection} = "close TE"; #1.1	816	my $idempotent = $IDEMPOTENT{$method};
		817
		818	# default value for keepalive is true iff the request is for an idempotent method
		819	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
		820	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		821	my $was_persistent; # true if this is actually a recycled connection
		822
		823	# the key to use in the keepalive cache
		824	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
		825
		826	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive " : "" : "close ") . "Te"; #1.1
612	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1	827	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
613		828
614	my %state = (connect_guard => 1);	829	my %state = (connect_guard => 1);
		830
		831	my $ae_error = 595; # connecting
		832
		833	# handle actual, non-tunneled, request
		834	my $handle_actual_request = sub {
		835	$ae_error = 596; # request phase
		836
		837	my $hdl = $state{handle};
		838
		839	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
		840
		841	# send request
		842	$hdl->push_write (
		843	"$method $rpath HTTP/1.1\015\012"
		844	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
		845	. "\015\012"
		846	. (delete $arg{body})
		847	);
		848
		849	# return if error occurred during push_write()
		850	return unless %state;
		851
		852	# reduce memory usage, save a kitten, also re-use it for the response headers.
		853	%hdr = ();
		854
		855	# status line and headers
		856	$state{read_response} = sub {
		857	return unless %state;
		858
		859	for ("$_[1]") {
		860	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		861
		862	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
		863	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
		864
		865	# 100 Continue handling
		866	# should not happen as we don't send expect: 100-continue,
		867	# but we handle it just in case.
		868	# since we send the request body regardless, if we get an error
		869	# we are out of-sync, which we currently do NOT handle correctly.
		870	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
		871	if $2 eq 100;
		872
		873	push @pseudo,
		874	HTTPVersion => $1,
		875	Status => $2,
		876	Reason => $3,
		877	;
		878
		879	my $hdr = _parse_hdr
		880	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
		881
		882	%hdr = (%$hdr, @pseudo);
		883	}
		884
		885	# redirect handling
		886	# microsoft and other shitheads don't give a shit for following standards,
		887	# try to support some common forms of broken Location headers.
		888	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
		889	$hdr{location} =~ s/^\.\/+//;
		890
		891	my $url = "$rscheme://$uhost:$uport";
		892
		893	unless ($hdr{location} =~ s/^\///) {
		894	$url .= $upath;
		895	$url =~ s/\/[^\/]*$//;
		896	}
		897
		898	$hdr{location} = "$url/$hdr{location}";
		899	}
		900
		901	my $redirect;
		902
		903	if ($recurse) {
		904	my $status = $hdr{Status};
		905
		906	# industry standard is to redirect POST as GET for
		907	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
		908	# also, the UA should ask the user for 301 and 307 and POST,
		909	# industry standard seems to be to simply follow.
		910	# we go with the industry standard.
		911	if ($status == 301 or $status == 302 or $status == 303) {
		912	# HTTP/1.1 is unclear on how to mutate the method
		913	$method = "GET" unless $method eq "HEAD";
		914	$redirect = 1;
		915	} elsif ($status == 307) {
		916	$redirect = 1;
		917	}
		918	}
		919
		920	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
		921	if ($state{handle}) {
		922	# handle keepalive
		923	if (
		924	$persistent
		925	&& $_[3]
		926	&& ($hdr{HTTPVersion} < 1.1
		927	? $hdr{connection} =~ /\bkeep-?alive\b/i
		928	: $hdr{connection} !~ /\bclose\b/i)
		929	) {
		930	ka_store $ka_key, delete $state{handle};
		931	} else {
		932	# no keepalive, destroy the handle
		933	$state{handle}->destroy;
		934	}
		935	}
		936
		937	%state = ();
		938
		939	if (defined $_[1]) {
		940	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		941	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		942	}
		943
		944	# set-cookie processing
		945	if ($arg{cookie_jar}) {
		946	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		947	}
		948
		949	if ($redirect && exists $hdr{location}) {
		950	# we ignore any errors, as it is very common to receive
		951	# Content-Length != 0 but no actual body
		952	# we also access %hdr, as $_[1] might be an erro
		953	$state{recurse} =
		954	http_request (
		955	$method => $hdr{location},
		956	%arg,
		957	recurse => $recurse - 1,
		958	Redirect => [$_[0], \%hdr],
		959	sub {
		960	%state = ();
		961	&$cb
		962	},
		963	);
		964	} else {
		965	$cb->($_[0], \%hdr);
		966	}
		967	};
		968
		969	$ae_error = 597; # body phase
		970
		971	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		972
		973	my $len = $chunked ? undef : $hdr{"content-length"};
		974
		975	# body handling, many different code paths
		976	# - no body expected
		977	# - want_body_handle
		978	# - te chunked
		979	# - 2x length known (with or without on_body)
		980	# - 2x length not known (with or without on_body)
		981	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		982	$finish->(undef, 598 => "Request cancelled by on_header");
		983	} elsif (
		984	$hdr{Status} =~ /^(?:1..|204|205|304)$/
		985	or $method eq "HEAD"
		986	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		987	) {
		988	# no body
		989	$finish->("", undef, undef, 1);
		990
		991	} elsif (!$redirect && $arg{want_body_handle}) {
		992	$_[0]->on_eof (undef);
		993	$_[0]->on_error (undef);
		994	$_[0]->on_read (undef);
		995
		996	$finish->(delete $state{handle});
		997
		998	} elsif ($chunked) {
		999	my $cl = 0;
		1000	my $body = "";
		1001	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
		1002
		1003	$state{read_chunk} = sub {
		1004	$_[1] =~ /^([0-9a-fA-F]+)/
		1005	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1006
		1007	my $len = hex $1;
		1008
		1009	if ($len) {
		1010	$cl += $len;
		1011
		1012	$_[0]->push_read (chunk => $len, sub {
		1013	$on_body->($_[1], \%hdr)
		1014	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1015
		1016	$_[0]->push_read (line => sub {
		1017	length $_[1]
		1018	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1019	$_[0]->push_read (line => $state{read_chunk});
		1020	});
		1021	});
		1022	} else {
		1023	$hdr{"content-length"} ||= $cl;
		1024
		1025	$_[0]->push_read (line => $qr_nlnl, sub {
		1026	if (length $_[1]) {
		1027	for ("$_[1]") {
		1028	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		1029
		1030	my $hdr = _parse_hdr
		1031	or return $finish->(undef, $ae_error => "Garbled response trailers");
		1032
		1033	%hdr = (%hdr, %$hdr);
		1034	}
		1035	}
		1036
		1037	$finish->($body, undef, undef, 1);
		1038	});
		1039	}
		1040	};
		1041
		1042	$_[0]->push_read (line => $state{read_chunk});
		1043
		1044	} elsif ($arg{on_body}) {
		1045	if (defined $len) {
		1046	$_[0]->on_read (sub {
		1047	$len -= length $_[0]{rbuf};
		1048
		1049	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1050	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1051
		1052	$len > 0
		1053	or $finish->("", undef, undef, 1);
		1054	});
		1055	} else {
		1056	$_[0]->on_eof (sub {
		1057	$finish->("");
		1058	});
		1059	$_[0]->on_read (sub {
		1060	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1061	or $finish->(undef, 598 => "Request cancelled by on_body");
		1062	});
		1063	}
		1064	} else {
		1065	$_[0]->on_eof (undef);
		1066
		1067	if (defined $len) {
		1068	$_[0]->on_read (sub {
		1069	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1070	if $len <= length $_[0]{rbuf};
		1071	});
		1072	} else {
		1073	$_[0]->on_error (sub {
		1074	($! == Errno::EPIPE || !$!)
		1075	? $finish->(delete $_[0]{rbuf})
		1076	: $finish->(undef, $ae_error => $_[2]);
		1077	});
		1078	$_[0]->on_read (sub { });
		1079	}
		1080	}
		1081	};
		1082
		1083	# if keepalive is enabled, then the server closing the connection
		1084	# before a response can happen legally - we retry on idempotent methods.
		1085	if ($was_persistent && $idempotent) {
		1086	my $old_eof = $hdl->{on_eof};
		1087	$hdl->{on_eof} = sub {
		1088	_destroy_state %state;
		1089
		1090	%state = ();
		1091	$state{recurse} =
		1092	http_request (
		1093	$method => $url,
		1094	%arg,
		1095	recurse => $recurse - 1,
		1096	keepalive => 0,
		1097	sub {
		1098	%state = ();
		1099	&$cb
		1100	}
		1101	);
		1102	};
		1103	$hdl->on_read (sub {
		1104	return unless %state;
		1105
		1106	# as soon as we receive something, a connection close
		1107	# once more becomes a hard error
		1108	$hdl->{on_eof} = $old_eof;
		1109	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1110	});
		1111	} else {
		1112	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1113	}
		1114	};
		1115
		1116	my $prepare_handle = sub {
		1117	my ($hdl) = $state{handle};
		1118
		1119	$hdl->on_error (sub {
		1120	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1121	});
		1122	$hdl->on_eof (sub {
		1123	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1124	});
		1125	$hdl->timeout_reset;
		1126	$hdl->timeout ($timeout);
		1127	};
		1128
		1129	# connected to proxy (or origin server)
		1130	my $connect_cb = sub {
		1131	my $fh = shift
		1132	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1133
		1134	return unless delete $state{connect_guard};
		1135
		1136	# get handle
		1137	$state{handle} = new AnyEvent::Handle
		1138	%{ $arg{handle_params} },
		1139	fh => $fh,
		1140	peername => $uhost,
		1141	tls_ctx => $arg{tls_ctx},
		1142	;
		1143
		1144	$prepare_handle->();
		1145
		1146	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1147
		1148	# now handle proxy-CONNECT method
		1149	if ($proxy && $uscheme eq "https") {
		1150	# oh dear, we have to wrap it into a connect request
		1151
		1152	# maybe re-use $uauthority with patched port?
		1153	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012");
		1154	$state{handle}->push_read (line => $qr_nlnl, sub {
		1155	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1156	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1157
		1158	if ($2 == 200) {
		1159	$rpath = $upath;
		1160	$handle_actual_request->();
		1161	} else {
		1162	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
		1163	}
		1164	});
		1165	} else {
		1166	$handle_actual_request->();
		1167	}
		1168	};
615		1169
616	_get_slot $uhost, sub {	1170	_get_slot $uhost, sub {
617	$state{slot_guard} = shift;	1171	$state{slot_guard} = shift;
618		1172
619	return unless $state{connect_guard};	1173	return unless $state{connect_guard};
620		1174
621	my $ae_error = 595; # connecting	1175	# try to use an existing keepalive connection, but only if we, ourselves, plan
		1176	# on a keepalive request (in theory, this should be a separate config option).
		1177	if ($persistent && $KA_CACHE{$ka_key}) {
		1178	$was_persistent = 1;
622		1179
623	my $connect_cb = sub {	1180	$state{handle} = ka_fetch $ka_key;
624	$state{fh} = shift
625	or do {
626	my $err = "$!";
627	%state = ();
628	return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err });
629	};
630
631	return unless delete $state{connect_guard};
632
633	# get handle
634	$state{handle} = new AnyEvent::Handle
635	fh => $state{fh},
636	peername => $rhost,
637	tls_ctx => $arg{tls_ctx},
638	# these need to be reconfigured on keepalive handles
639	timeout => $timeout,
640	on_error => sub {
641	%state = ();
642	$cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] });
643	},
644	on_eof => sub {
645	%state = ();
646	$cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" });
647	},
648	;
649
650	# limit the number of persistent connections
651	# keepalive not yet supported
652	# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
653	# ++$KA_COUNT{$_[1]};
654	# $state{handle}{ka_count_guard} = AnyEvent::Util::guard {
655	# --$KA_COUNT{$_[1]}
656	# };
657	# $hdr{connection} = "keep-alive";
658	# }
659
660	$state{handle}->starttls ("connect") if $rscheme eq "https";
661
662	# handle actual, non-tunneled, request
663	my $handle_actual_request = sub {
664	$ae_error = 596; # request phase
665
666	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};
667
668	# send request
669	$state{handle}->push_write (	1181	$state{handle}->destroyed
670	"$method $rpath HTTP/1.1\015\012"	1182	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
671	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)	1183	$prepare_handle->();
672	. "\015\012"	1184	$state{handle}->destroyed
673	. (delete $arg{body})	1185	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
674	);
675
676	# return if error occured during push_write()
677	return unless %state;
678
679	%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use
680
681	# status line and headers
682	$state{read_response} = sub {
683	for ("$_[1]") {
684	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
685
686	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
687	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" }));
688
689	# 100 Continue handling
690	# should not happen as we don't send expect: 100-continue,
691	# but we handle it just in case.
692	# since we send the request body regardless, if we get an error
693	# we are out of-sync, which we currently do NOT handle correctly.
694	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
695	if $2 eq 100;
696
697	push @pseudo,
698	HTTPVersion => $1,
699	Status => $2,
700	Reason => $3,
701	;
702
703	my $hdr = parse_hdr
704	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" }));
705
706	%hdr = (%$hdr, @pseudo);
707	}
708
709	# redirect handling
710	# microsoft and other shitheads don't give a shit for following standards,
711	# try to support some common forms of broken Location headers.
712	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
713	$hdr{location} =~ s/^\.\/+//;
714
715	my $url = "$rscheme://$uhost:$uport";
716
717	unless ($hdr{location} =~ s/^\///) {
718	$url .= $upath;
719	$url =~ s/\/[^\/]*$//;
720	}
721
722	$hdr{location} = "$url/$hdr{location}";
723	}
724
725	my $redirect;
726
727	if ($recurse) {
728	my $status = $hdr{Status};
729
730	# industry standard is to redirect POST as GET for
731	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
732	# also, the UA should ask the user for 301 and 307 and POST,
733	# industry standard seems to be to simply follow.
734	# we go with the industry standard.
735	if ($status == 301 or $status == 302 or $status == 303) {
736	# HTTP/1.1 is unclear on how to mutate the method
737	$method = "GET" unless $method eq "HEAD";
738	$redirect = 1;
739	} elsif ($status == 307) {
740	$redirect = 1;
741	}
742	}
743
744	my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive])
745	my $may_keep_alive = $_[3];
746
747	$state{handle}->destroy if $state{handle};
748	%state = ();
749
750	if (defined $_[1]) {
751	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
752	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
753	}
754
755	# set-cookie processing
756	if ($arg{cookie_jar}) {
757	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost;
758	}
759
760	if ($redirect && exists $hdr{location}) {
761	# we ignore any errors, as it is very common to receive
762	# Content-Length != 0 but no actual body
763	# we also access %hdr, as $_[1] might be an erro
764	http_request (
765	$method => $hdr{location},
766	%arg,
767	recurse => $recurse - 1,
768	Redirect => [$_[0], \%hdr],
769	$cb);
770	} else {
771	$cb->($_[0], \%hdr);
772	}
773	};
774
775	$ae_error = 597; # body phase
776
777	my $len = $hdr{"content-length"};
778
779	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
780	$finish->(undef, 598 => "Request cancelled by on_header");
781	} elsif (
782	$hdr{Status} =~ /^(?:1..|204|205|304)$/
783	or $method eq "HEAD"
784	or (defined $len && !$len)
785	) {
786	# no body
787	$finish->("", undef, undef, 1);
788	} else {
789	# body handling, many different code paths
790	# - no body expected
791	# - want_body_handle
792	# - te chunked
793	# - 2x length known (with or without on_body)
794	# - 2x length not known (with or without on_body)
795	if (!$redirect && $arg{want_body_handle}) {
796	$_[0]->on_eof (undef);
797	$_[0]->on_error (undef);
798	$_[0]->on_read (undef);
799
800	$finish->(delete $state{handle});
801
802	} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) {
803	my $cl = 0;
804	my $body = undef;
805	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
806
807	my $read_chunk; $read_chunk = sub {
808	$_[1] =~ /^([0-9a-fA-F]+)/
809	or $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
810
811	my $len = hex $1;
812
813	if ($len) {
814	$cl += $len;
815
816	$_[0]->push_read (chunk => $len, sub {
817	$on_body->($_[1], \%hdr)
818	or return $finish->(undef, 598 => "Request cancelled by on_body");
819
820	$_[0]->push_read (line => sub {
821	length $_[1]
822	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
823	$_[0]->push_read (line => $read_chunk);
824	});
825	});
826	} else {
827	$hdr{"content-length"} ||= $cl;
828
829	$_[0]->push_read (line => $qr_nlnl, sub {
830	if (length $_[1]) {
831	for ("$_[1]") {
832	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
833
834	my $hdr = parse_hdr
835	or return $finish->(undef, $ae_error => "Garbled response trailers");
836
837	%hdr = (%hdr, %$hdr);
838	}
839	}
840
841	$finish->($body, undef, undef, 1);
842	});
843	}
844	};
845
846	$_[0]->push_read (line => $read_chunk);
847
848	} elsif ($arg{on_body}) {
849	if ($len) {
850	$_[0]->on_read (sub {
851	$len -= length $_[0]{rbuf};
852
853	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
854	or return $finish->(undef, 598 => "Request cancelled by on_body");
855
856	$len > 0
857	or $finish->("", undef, undef, 1);
858	});
859	} else {
860	$_[0]->on_eof (sub {
861	$finish->("");
862	});
863	$_[0]->on_read (sub {
864	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
865	or $finish->(undef, 598 => "Request cancelled by on_body");
866	});
867	}
868	} else {
869	$_[0]->on_eof (undef);
870
871	if ($len) {
872	$_[0]->on_read (sub {
873	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
874	if $len <= length $_[0]{rbuf};
875	});
876	} else {
877	$_[0]->on_error (sub {
878	($! == Errno::EPIPE || !$!)
879	? $finish->(delete $_[0]{rbuf})
880	: $finish->(undef, $ae_error => $_[2]);
881	});
882	$_[0]->on_read (sub { });
883	}
884	}
885	}
886	};
887
888	$state{handle}->push_read (line => $qr_nlnl, $state{read_response});
889	};
890
891	# now handle proxy-CONNECT method
892	if ($proxy && $uscheme eq "https") {
893	# oh dear, we have to wrap it into a connect request
894
895	# maybe re-use $uauthority with patched port?
896	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012");
897	$state{handle}->push_read (line => $qr_nlnl, sub {
898	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
899	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }));
900
901	if ($2 == 200) {
902	$rpath = $upath;
903	&$handle_actual_request;	1186	$handle_actual_request->();
904	} else {	1187
905	%state = ();
906	$cb->(undef, { @pseudo, Status => $2, Reason => $3 });
907	}
908	});
909	} else {	1188	} else {
910	&$handle_actual_request;
911	}
912	};
913
914	my $tcp_connect = $arg{tcp_connect}	1189	my $tcp_connect = $arg{tcp_connect}
915	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };	1190	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
916		1191
917	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });	1192	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });
918		1193	}
919	};	1194	};
920		1195
921	defined wantarray && AnyEvent::Util::guard { %state = () }	1196	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
922	}	1197	}
923		1198
924	sub http_get($@) {	1199	sub http_get($@) {
925	unshift @_, "GET";	1200	unshift @_, "GET";
926	&http_request	1201	&http_request
…		…
944	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for	1219	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
945	the actual connection, which in turn uses AnyEvent::DNS to resolve	1220	the actual connection, which in turn uses AnyEvent::DNS to resolve
946	hostnames. The latter is a simple stub resolver and does no caching	1221	hostnames. The latter is a simple stub resolver and does no caching
947	on its own. If you want DNS caching, you currently have to provide	1222	on its own. If you want DNS caching, you currently have to provide
948	your own default resolver (by storing a suitable resolver object in	1223	your own default resolver (by storing a suitable resolver object in
949	C<$AnyEvent::DNS::RESOLVER>).	1224	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
950		1225
951	=head2 GLOBAL FUNCTIONS AND VARIABLES	1226	=head2 GLOBAL FUNCTIONS AND VARIABLES
952		1227
953	=over 4	1228	=over 4
954		1229
955	=item AnyEvent::HTTP::set_proxy "proxy-url"	1230	=item AnyEvent::HTTP::set_proxy "proxy-url"
956		1231
957	Sets the default proxy server to use. The proxy-url must begin with a	1232	Sets the default proxy server to use. The proxy-url must begin with a
958	string of the form C<http://host:port> (optionally C<https:...>), croaks	1233	string of the form C<http://host:port>, croaks otherwise.
959	otherwise.
960		1234
961	To clear an already-set proxy, use C<undef>.	1235	To clear an already-set proxy, use C<undef>.
		1236
		1237	When AnyEvent::HTTP is loaded for the first time it will query the
		1238	default proxy from the operating system, currently by looking at
		1239	C<$ENV{http_proxy>}.
		1240
		1241	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
		1242
		1243	Remove all cookies from the cookie jar that have been expired. If
		1244	C<$session_end> is given and true, then additionally remove all session
		1245	cookies.
		1246
		1247	You should call this function (with a true C<$session_end>) before you
		1248	save cookies to disk, and you should call this function after loading them
		1249	again. If you have a long-running program you can additionally call this
		1250	function from time to time.
		1251
		1252	A cookie jar is initially an empty hash-reference that is managed by this
		1253	module. It's format is subject to change, but currently it is like this:
		1254
		1255	The key C<version> has to contain C<1>, otherwise the hash gets
		1256	emptied. All other keys are hostnames or IP addresses pointing to
		1257	hash-references. The key for these inner hash references is the
		1258	server path for which this cookie is meant, and the values are again
		1259	hash-references. Each key of those hash-references is a cookie name, and
		1260	the value, you guessed it, is another hash-reference, this time with the
		1261	key-value pairs from the cookie, except for C<expires> and C<max-age>,
		1262	which have been replaced by a C<_expires> key that contains the cookie
		1263	expiry timestamp. Session cookies are indicated by not having an
		1264	C<_expires> key.
		1265
		1266	Here is an example of a cookie jar with a single cookie, so you have a
		1267	chance of understanding the above paragraph:
		1268
		1269	{
		1270	version => 1,
		1271	"10.0.0.1" => {
		1272	"/" => {
		1273	"mythweb_id" => {
		1274	_expires => 1293917923,
		1275	value => "ooRung9dThee3ooyXooM1Ohm",
		1276	},
		1277	},
		1278	},
		1279	}
962		1280
963	=item $date = AnyEvent::HTTP::format_date $timestamp	1281	=item $date = AnyEvent::HTTP::format_date $timestamp
964		1282
965	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP	1283	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP
966	Date (RFC 2616).	1284	Date (RFC 2616).
…		…
973		1291
974	=item $AnyEvent::HTTP::MAX_RECURSE	1292	=item $AnyEvent::HTTP::MAX_RECURSE
975		1293
976	The default value for the C<recurse> request parameter (default: C<10>).	1294	The default value for the C<recurse> request parameter (default: C<10>).
977		1295
		1296	=item $AnyEvent::HTTP::TIMEOUT
		1297
		1298	The default timeout for connection operations (default: C<300>).
		1299
978	=item $AnyEvent::HTTP::USERAGENT	1300	=item $AnyEvent::HTTP::USERAGENT
979		1301
980	The default value for the C<User-Agent> header (the default is	1302	The default value for the C<User-Agent> header (the default is
981	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1303	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
982		1304
983	=item $AnyEvent::HTTP::MAX_PER_HOST	1305	=item $AnyEvent::HTTP::MAX_PER_HOST
984		1306
985	The maximum number of concurrent connections to the same host (identified	1307	The maximum number of concurrent connections to the same host (identified
986	by the hostname). If the limit is exceeded, then the additional requests	1308	by the hostname). If the limit is exceeded, then the additional requests
987	are queued until previous connections are closed.	1309	are queued until previous connections are closed. Both persistent and
		1310	non-persistent connections are counted in this limit.
988		1311
989	The default value for this is C<4>, and it is highly advisable to not	1312	The default value for this is C<4>, and it is highly advisable to not
990	increase it.	1313	increase it much.
		1314
		1315	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1316	connections, older browsers used 2, newer ones (such as firefox 3)
		1317	typically use 6, and Opera uses 8 because like, they have the fastest
		1318	browser and give a shit for everybody else on the planet.
		1319
		1320	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
		1321
		1322	The time after which idle persistent connections get closed by
		1323	AnyEvent::HTTP (default: C<3>).
991		1324
992	=item $AnyEvent::HTTP::ACTIVE	1325	=item $AnyEvent::HTTP::ACTIVE
993		1326
994	The number of active connections. This is not the number of currently	1327	The number of active connections. This is not the number of currently
995	running requests, but the number of currently open and non-idle TCP	1328	running requests, but the number of currently open and non-idle TCP
996	connections. This number of can be useful for load-leveling.	1329	connections. This number can be useful for load-leveling.
997		1330
998	=back	1331	=back
999		1332
1000	=cut	1333	=cut
1001		1334
…		…
1035	# other formats fail in the loop below	1368	# other formats fail in the loop below
1036		1369
1037	for (0..11) {	1370	for (0..11) {
1038	if ($m eq $month[$_]) {	1371	if ($m eq $month[$_]) {
1039	require Time::Local;	1372	require Time::Local;
1040	return Time::Local::timegm ($S, $M, $H, $d, $_, $y);	1373	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
1041	}	1374	}
1042	}	1375	}
1043		1376
1044	undef	1377	undef
1045	}	1378	}
1046		1379
1047	sub set_proxy($) {	1380	sub set_proxy($) {
1048	if (length $_[0]) {	1381	if (length $_[0]) {
1049	$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix	1382	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
1050	or Carp::croak "$_[0]: invalid proxy URL";	1383	or Carp::croak "$_[0]: invalid proxy URL";
1051	$PROXY = [$2, $3 || 3128, $1]	1384	$PROXY = [$2, $3 || 3128, $1]
1052	} else {	1385	} else {
1053	undef $PROXY;	1386	undef $PROXY;
1054	}	1387	}
…		…
1057	# initialise proxy from environment	1390	# initialise proxy from environment
1058	eval {	1391	eval {
1059	set_proxy $ENV{http_proxy};	1392	set_proxy $ENV{http_proxy};
1060	};	1393	};
1061		1394
		1395	=head2 SHOWCASE
		1396
		1397	This section contains some more elaborate "real-world" examples or code
		1398	snippets.
		1399
		1400	=head2 HTTP/1.1 FILE DOWNLOAD
		1401
		1402	Downloading files with HTTP can be quite tricky, especially when something
		1403	goes wrong and you want to resume.
		1404
		1405	Here is a function that initiates and resumes a download. It uses the
		1406	last modified time to check for file content changes, and works with many
		1407	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1408	on older servers.
		1409
		1410	It calls the completion callback with either C<undef>, which means a
		1411	nonretryable error occurred, C<0> when the download was partial and should
		1412	be retried, and C<1> if it was successful.
		1413
		1414	use AnyEvent::HTTP;
		1415
		1416	sub download($$$) {
		1417	my ($url, $file, $cb) = @_;
		1418
		1419	open my $fh, "+<", $file
		1420	or die "$file: $!";
		1421
		1422	my %hdr;
		1423	my $ofs = 0;
		1424
		1425	warn stat $fh;
		1426	warn -s _;
		1427	if (stat $fh and -s _) {
		1428	$ofs = -s _;
		1429	warn "-s is ", $ofs;
		1430	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1431	$hdr{"range"} = "bytes=$ofs-";
		1432	}
		1433
		1434	http_get $url,
		1435	headers => \%hdr,
		1436	on_header => sub {
		1437	my ($hdr) = @_;
		1438
		1439	if ($hdr->{Status} == 200 && $ofs) {
		1440	# resume failed
		1441	truncate $fh, $ofs = 0;
		1442	}
		1443
		1444	sysseek $fh, $ofs, 0;
		1445
		1446	1
		1447	},
		1448	on_body => sub {
		1449	my ($data, $hdr) = @_;
		1450
		1451	if ($hdr->{Status} =~ /^2/) {
		1452	length $data == syswrite $fh, $data
		1453	or return; # abort on write errors
		1454	}
		1455
		1456	1
		1457	},
		1458	sub {
		1459	my (undef, $hdr) = @_;
		1460
		1461	my $status = $hdr->{Status};
		1462
		1463	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1464	utime $fh, $time, $time;
		1465	}
		1466
		1467	if ($status == 200 || $status == 206 || $status == 416) {
		1468	# download ok || resume ok || file already fully downloaded
		1469	$cb->(1, $hdr);
		1470
		1471	} elsif ($status == 412) {
		1472	# file has changed while resuming, delete and retry
		1473	unlink $file;
		1474	$cb->(0, $hdr);
		1475
		1476	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1477	# retry later
		1478	$cb->(0, $hdr);
		1479
		1480	} else {
		1481	$cb->(undef, $hdr);
		1482	}
		1483	}
		1484	;
		1485	}
		1486
		1487	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1488	if ($_[0]) {
		1489	print "OK!\n";
		1490	} elsif (defined $_[0]) {
		1491	print "please retry later\n";
		1492	} else {
		1493	print "ERROR\n";
		1494	}
		1495	};
		1496
1062	=head2 SOCKS PROXIES	1497	=head3 SOCKS PROXIES
1063		1498
1064	Socks proxies are not directly supported by AnyEvent::HTTP. You can	1499	Socks proxies are not directly supported by AnyEvent::HTTP. You can
1065	compile your perl to support socks, or use an external program such as	1500	compile your perl to support socks, or use an external program such as
1066	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy	1501	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
1067	transparently.	1502	transparently.

Diff Legend

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