ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.73 by root, Fri Dec 31 21:00:36 2010 UTC vs.
Revision 1.108 by root, Thu Jun 30 09:12:39 2011 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 = '2.12';
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, $headers)
73		68
74	Executes an HTTP-GET request. See the http_request function for details on	69	Executes an HTTP-GET request. See the http_request function for details on
…		…
122		117
123	If the server sends a header multiple times, then their contents will be	118	If the server sends a header multiple times, then their contents will be
124	joined together with a comma (C<,>), as per the HTTP spec.	119	joined together with a comma (C<,>), as per the HTTP spec.
125		120
126	If an internal error occurs, such as not being able to resolve a hostname,	121	If an internal error occurs, such as not being able to resolve a hostname,
127	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x>	122	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be
128	(usually C<599>) and the C<Reason> pseudo-header will contain an error	123	C<590>-C<599> and the C<Reason> pseudo-header will contain an error
129	message.	124	message. Currently the following status codes are used:
		125
		126	=over 4
		127
		128	=item 595 - errors during connection etsbalishment, proxy handshake.
		129
		130	=item 596 - errors during TLS negotiation, request sending and header processing.
		131
		132	=item 597 - errors during body receiving or processing.
		133
		134	=item 598 - user aborted request via C<on_header> or C<on_body>.
		135
		136	=item 599 - other, usually nonretryable, errors (garbled URL etc.).
		137
		138	=back
130		139
131	A typical callback might look like this:	140	A typical callback might look like this:
132		141
133	sub {	142	sub {
134	my ($body, $hdr) = @_;	143	my ($body, $hdr) = @_;
…		…
156	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
157	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:>
158	(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
159	they won't be sent at all).	168	they won't be sent at all).
160		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
161	=item timeout => $seconds	177	=item timeout => $seconds
162		178
163	The time-out to use for various stages - each connect attempt will reset	179	The time-out to use for various stages - each connect attempt will reset
164	the timeout, as will read or write activity, i.e. this is not an overall	180	the timeout, as will read or write activity, i.e. this is not an overall
165	timeout.	181	timeout.
166		182
167	Default timeout is 5 minutes.	183	Default timeout is 5 minutes.
168		184
169	=item proxy => [$host, $port[, $scheme]] or undef	185	=item proxy => [$host, $port[, $scheme]] or undef
170		186
171	Use the given http proxy for all requests. If not specified, then the	187	Use the given http proxy for all requests, or no proxy if C<undef> is
172	default proxy (as specified by C<$ENV{http_proxy}>) is used.	188	used.
173		189
174	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.
175	HTTPS.	191
		192	If not specified, then the default proxy is used (see
		193	C<AnyEvent::HTTP::set_proxy>).
176		194
177	=item body => $string	195	=item body => $string
178		196
179	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
180	this module might offer more options).	198	this module might offer more options).
…		…
182	=item cookie_jar => $hash_ref	200	=item cookie_jar => $hash_ref
183		201
184	Passing this parameter enables (simplified) cookie-processing, loosely	202	Passing this parameter enables (simplified) cookie-processing, loosely
185	based on the original netscape specification.	203	based on the original netscape specification.
186		204
187	The C<$hash_ref> must be an (initially empty) hash reference which will	205	The C<$hash_ref> must be an (initially empty) hash reference which
188	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
189	persistent storage with something like JSON or Storable, but this is not	207	to persistent storage with something like JSON or Storable - see the
190	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.
191		211
192	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
193	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
194	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
195	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
196	to.	216	to.
197		217
198	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:>
199	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
…		…
208	verification, highest compatibility) and high-security (CA and common-name	228	verification, highest compatibility) and high-security (CA and common-name
209	verification) TLS context.	229	verification) TLS context.
210		230
211	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
212	me the page, no matter what".	232	me the page, no matter what".
		233
		234	See also the C<sessionid> parameter.
		235
		236	=item session => $string
		237
		238	The module might reuse connections to the same host internally. Sometimes
		239	(e.g. when using TLS), you do not want to reuse connections from other
		240	sessions. This can be achieved by setting this parameter to some unique
		241	ID (such as the address of an object storing your state data, or the TLS
		242	context) - only connections using the same unique ID will be reused.
213		243
214	=item on_prepare => $callback->($fh)	244	=item on_prepare => $callback->($fh)
215		245
216	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
217	connect (for exmaple, to bind it on a given IP address). This parameter	247	connect (for exmaple, to bind it on a given IP address). This parameter
…		…
287	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
288	callback will receive the L<AnyEvent::Handle> object associated with the	318	callback will receive the L<AnyEvent::Handle> object associated with the
289	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
290	(e.g. status C<304>), the empty string will be passed.	320	(e.g. status C<304>), the empty string will be passed.
291		321
292	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
293	a proxy, be a persistent connection etc., and configured in unspecified	323	to a proxy, be a persistent connection, use chunked transfer encoding
294	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
295	module anymore).	325	handle (it will not be used by this module anymore).
296		326
297	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
298	headers, an interactive protocol is used (typical example would be the	328	headers, an interactive protocol is used (typical example would be the
299	push-style twitter API which starts a JSON/XML stream).	329	push-style twitter API which starts a JSON/XML stream).
300		330
301	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
302	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 => $boolean
		353
		354	Only used when C<persistent> is also true. This parameter decides whether
		355	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		356	(as opposed to only a HTTP/1.1 persistent connection).
		357
		358	The default is true, except when using a proxy, in which case it defaults
		359	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		360
		361	=item handle_params => { key => value ... }
		362
		363	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		364	constructor that is called - not all requests will create a handle, and
		365	sometimes more than one is created, so this parameter is only good for
		366	setting hints.
		367
		368	Example: set the maximum read size to 4096, to potentially conserve memory
		369	at the cost of speed.
		370
		371	handle_params => {
		372	max_read_size => 4096,
		373	},
303		374
304	=back	375	=back
305		376
306	Example: do a simple HTTP GET request for http://www.nethype.de/ and print	377	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
307	the response body.	378	the response body.
…		…
314	Example: do a HTTP HEAD request on https://www.google.com/, use a	385	Example: do a HTTP HEAD request on https://www.google.com/, use a
315	timeout of 30 seconds.	386	timeout of 30 seconds.
316		387
317	http_request	388	http_request
318	GET => "https://www.google.com",	389	GET => "https://www.google.com",
		390	headers => { "user-agent" => "MySearchClient 1.0" },
319	timeout => 30,	391	timeout => 30,
320	sub {	392	sub {
321	my ($body, $hdr) = @_;	393	my ($body, $hdr) = @_;
322	use Data::Dumper;	394	use Data::Dumper;
323	print Dumper $hdr;	395	print Dumper $hdr;
…		…
333	};	405	};
334		406
335	undef $request;	407	undef $request;
336		408
337	=cut	409	=cut
		410
		411	#############################################################################
		412	# wait queue/slots
338		413
339	sub _slot_schedule;	414	sub _slot_schedule;
340	sub _slot_schedule($) {	415	sub _slot_schedule($) {
341	my $host = shift;	416	my $host = shift;
342		417
…		…
364	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];	439	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];
365		440
366	_slot_schedule $_[0];	441	_slot_schedule $_[0];
367	}	442	}
368		443
		444	#############################################################################
		445	# cookie handling
		446
		447	# expire cookies
		448	sub cookie_jar_expire($;$) {
		449	my ($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
369	# extract cookies from jar	477	# extract cookies from jar
370	sub cookie_jar_extract($$$$) {	478	sub cookie_jar_extract($$$$) {
371	my ($jar, $uscheme, $uhost, $upath) = @_;	479	my ($jar, $scheme, $host, $path) = @_;
372		480
373	%$jar = () if $jar->{version} != 1;	481	%$jar = () if $jar->{version} != 1;
374		482
375	my @cookies;	483	my @cookies;
376		484
377	while (my ($chost, $paths) = each %$jar) {	485	while (my ($chost, $paths) = each %$jar) {
378	next unless ref $paths;	486	next unless ref $paths;
379		487
380	if ($chost =~ /^\./) {	488	if ($chost =~ /^\./) {
381	next unless $chost eq substr $uhost, -length $chost;	489	next unless $chost eq substr $host, -length $chost;
382	} elsif ($chost =~ /\./) {	490	} elsif ($chost =~ /\./) {
383	next unless $chost eq $uhost;	491	next unless $chost eq $host;
384	} else {	492	} else {
385	next;	493	next;
386	}	494	}
387		495
388	while (my ($cpath, $cookies) = each %$paths) {	496	while (my ($cpath, $cookies) = each %$paths) {
389	next unless $cpath eq substr $upath, 0, length $cpath;	497	next unless $cpath eq substr $path, 0, length $cpath;
390		498
391	while (my ($cookie, $kv) = each %$cookies) {	499	while (my ($cookie, $kv) = each %$cookies) {
392	next if $uscheme ne "https" && exists $kv->{secure};	500	next if $scheme ne "https" && exists $kv->{secure};
393		501
394	if (exists $kv->{expires}) {	502	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
395	if (AE::now > parse_date ($kv->{expires})) {
396	delete $cookies->{$cookie};	503	delete $cookies->{$cookie};
397	next;	504	next;
398	}
399	}	505	}
400		506
401	my $value = $kv->{value};	507	my $value = $kv->{value};
402		508
403	if ($value =~ /[=;,[:space:]]/) {	509	if ($value =~ /[=;,[:space:]]/) {
…		…
412		518
413	\@cookies	519	\@cookies
414	}	520	}
415		521
416	# parse set_cookie header into jar	522	# parse set_cookie header into jar
417	sub cookie_jar_set_cookie($$$) {	523	sub cookie_jar_set_cookie($$$$) {
418	my ($jar, $set_cookie, $uhost) = @_;	524	my ($jar, $set_cookie, $host, $date) = @_;
		525
		526	my $anow = int AE::now;
		527	my $snow; # server-now
419		528
420	for ($set_cookie) {	529	for ($set_cookie) {
421	# parse NAME=VALUE	530	# parse NAME=VALUE
422	my @kv;	531	my @kv;
423		532
		533	# expires is not http-compliant in the original cookie-spec,
		534	# we support the official date format and some extensions
424	while (	535	while (
425	m{	536	m{
426	\G\s*	537	\G\s*
427	(?:	538	(?:
428	expires \s*=\s* ([A-Z][a-z][a-z],\ [^,;]+)	539	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
429	| ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )	540	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
430	)	541	)
431	}gcxsi	542	}gcxsi
432	) {	543	) {
433	my $name = $2;	544	my $name = $2;
434	my $value = $4;	545	my $value = $4;
435		546
436	unless (defined $name) {	547	if (defined $1) {
437	# expires	548	# expires
438	$name = "expires";	549	$name = "expires";
439	$value = $1;	550	$value = $1;
440	} elsif (!defined $value) {	551	} elsif (defined $3) {
441	# quoted	552	# quoted
442	$value = $3;	553	$value = $3;
443	$value =~ s/\\(.)/$1/gs;	554	$value =~ s/\\(.)/$1/gs;
444	}	555	}
445		556
446	push @kv, lc $name, $value;	557	push @kv, @kv ? lc $name : $name, $value;
447		558
448	last unless /\G\s*;/gc;	559	last unless /\G\s*;/gc;
449	}	560	}
450		561
451	last unless @kv;	562	last unless @kv;
452		563
453	my $name = shift @kv;	564	my $name = shift @kv;
454	my %kv = (value => shift @kv, @kv);	565	my %kv = (value => shift @kv, @kv);
455		566
456	$kv{expires} ||= format_date (AE::now + $kv{"max-age"})
457	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	}
458		575
459	my $cdom;	576	my $cdom;
460	my $cpath = (delete $kv{path}) || "/";	577	my $cpath = (delete $kv{path}) || "/";
461		578
462	if (exists $kv{domain}) {	579	if (exists $kv{domain}) {
…		…
468		585
469	# this is not rfc-like and not netscape-like. go figure.	586	# this is not rfc-like and not netscape-like. go figure.
470	my $ndots = $cdom =~ y/.//;	587	my $ndots = $cdom =~ y/.//;
471	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);	588	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
472	} else {	589	} else {
473	$cdom = $uhost;	590	$cdom = $host;
474	}	591	}
475		592
476	# store it	593	# store it
477	$jar->{version} = 1;	594	$jar->{version} = 1;
478	$jar->{$cdom}{$cpath}{$name} = \%kv;	595	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
479		596
480	redo if /\G\s*,/gc;	597	redo if /\G\s*,/gc;
481	}	598	}
482	}	599	}
483		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
484	# continue to parse $_ for headers and place them into the arg	643	# continue to parse $_ for headers and place them into the arg
485	sub parse_hdr() {	644	sub _parse_hdr() {
486	my %hdr;	645	my %hdr;
487		646
488	# things seen, not parsed:	647	# things seen, not parsed:
489	# p3pP="NON CUR OTPi OUR NOR UNI"	648	# p3pP="NON CUR OTPi OUR NOR UNI"
490		649
…		…
504	for values %hdr;	663	for values %hdr;
505		664
506	\%hdr	665	\%hdr
507	}	666	}
508		667
		668	#############################################################################
		669	# http_get
		670
509	our $qr_nlnl = qr{(?<![^\012])\015?\012};	671	our $qr_nlnl = qr{(?<![^\012])\015?\012};
510		672
511	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	673	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
512	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(\%$$) {
		686	my ($state, $cb, $hdr) = @_;
		687
		688	&_destroy_state ($state);
		689
		690	$cb->(undef, $hdr);
		691	()
		692	}
513		693
514	sub http_request($$@) {	694	sub http_request($$@) {
515	my $cb = pop;	695	my $cb = pop;
516	my ($method, $url, %arg) = @_;	696	my ($method, $url, %arg) = @_;
517		697
…		…
535	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	715	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
536		716
537	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })	717	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
538	if $recurse < 0;	718	if $recurse < 0;
539		719
540	my $proxy = $arg{proxy} || $PROXY;	720	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
541	my $timeout = $arg{timeout} || $TIMEOUT;	721	my $timeout = $arg{timeout} || $TIMEOUT;
542		722
543	my ($uscheme, $uauthority, $upath, $query, $fragment) =	723	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
544	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|;	724	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
545		725
546	$uscheme = lc $uscheme;	726	$uscheme = lc $uscheme;
547		727
548	my $uport = $uscheme eq "http" ? 80	728	my $uport = $uscheme eq "http" ? 80
549	: $uscheme eq "https" ? 443	729	: $uscheme eq "https" ? 443
550	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });	730	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
551		731
552	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	732	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x
553	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });	733	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
554		734
555	my $uhost = $1;	735	my $uhost = lc $1;
556	$uport = $2 if defined $2;	736	$uport = $2 if defined $2;
557		737
558	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"	738	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
559	unless exists $hdr{host};	739	unless exists $hdr{host};
560		740
…		…
579	$rscheme = "http" unless defined $rscheme;	759	$rscheme = "http" unless defined $rscheme;
580		760
581	# don't support https requests over https-proxy transport,	761	# don't support https requests over https-proxy transport,
582	# can't be done with tls as spec'ed, unless you double-encrypt.	762	# can't be done with tls as spec'ed, unless you double-encrypt.
583	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";	763	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";
		764
		765	$rhost = lc $rhost;
		766	$rscheme = lc $rscheme;
584	} else {	767	} else {
585	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);	768	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
586	}	769	}
587		770
588	# leave out fragment and query string, just a heuristic	771	# leave out fragment and query string, just a heuristic
…		…
590	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};	773	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
591		774
592	$hdr{"content-length"} = length $arg{body}	775	$hdr{"content-length"} = length $arg{body}
593	if length $arg{body} || $method ne "GET";	776	if length $arg{body} || $method ne "GET";
594		777
595	$hdr{connection} = "close TE"; #1.1	778	my $idempotent = $method =~ /^(?:GET|HEAD|PUT|DELETE|OPTIONS|TRACE)$/;
		779
		780	# default value for keepalive is true iff the request is for an idempotent method
		781	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
		782	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		783	my $was_persistent; # true if this is actually a recycled connection
		784
		785	# the key to use in the keepalive cache
		786	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
		787
		788	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive " : "" : "close ") . "Te"; #1.1
596	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1	789	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
597		790
598	my %state = (connect_guard => 1);	791	my %state = (connect_guard => 1);
		792
		793	my $ae_error = 595; # connecting
		794
		795	# handle actual, non-tunneled, request
		796	my $handle_actual_request = sub {
		797	$ae_error = 596; # request phase
		798
		799	my $hdl = $state{handle};
		800
		801	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
		802
		803	# send request
		804	$hdl->push_write (
		805	"$method $rpath HTTP/1.1\015\012"
		806	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
		807	. "\015\012"
		808	. (delete $arg{body})
		809	);
		810
		811	# return if error occured during push_write()
		812	return unless %state;
		813
		814	# reduce memory usage, save a kitten, also re-use it for the response headers.
		815	%hdr = ();
		816
		817	# status line and headers
		818	$state{read_response} = sub {
		819	return unless %state;
		820
		821	for ("$_[1]") {
		822	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		823
		824	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
		825	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
		826
		827	# 100 Continue handling
		828	# should not happen as we don't send expect: 100-continue,
		829	# but we handle it just in case.
		830	# since we send the request body regardless, if we get an error
		831	# we are out of-sync, which we currently do NOT handle correctly.
		832	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
		833	if $2 eq 100;
		834
		835	push @pseudo,
		836	HTTPVersion => $1,
		837	Status => $2,
		838	Reason => $3,
		839	;
		840
		841	my $hdr = _parse_hdr
		842	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
		843
		844	%hdr = (%$hdr, @pseudo);
		845	}
		846
		847	# redirect handling
		848	# microsoft and other shitheads don't give a shit for following standards,
		849	# try to support some common forms of broken Location headers.
		850	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
		851	$hdr{location} =~ s/^\.\/+//;
		852
		853	my $url = "$rscheme://$uhost:$uport";
		854
		855	unless ($hdr{location} =~ s/^\///) {
		856	$url .= $upath;
		857	$url =~ s/\/[^\/]*$//;
		858	}
		859
		860	$hdr{location} = "$url/$hdr{location}";
		861	}
		862
		863	my $redirect;
		864
		865	if ($recurse) {
		866	my $status = $hdr{Status};
		867
		868	# industry standard is to redirect POST as GET for
		869	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
		870	# also, the UA should ask the user for 301 and 307 and POST,
		871	# industry standard seems to be to simply follow.
		872	# we go with the industry standard.
		873	if ($status == 301 or $status == 302 or $status == 303) {
		874	# HTTP/1.1 is unclear on how to mutate the method
		875	$method = "GET" unless $method eq "HEAD";
		876	$redirect = 1;
		877	} elsif ($status == 307) {
		878	$redirect = 1;
		879	}
		880	}
		881
		882	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
		883	if ($state{handle}) {
		884	# handle keepalive
		885	if (
		886	$persistent
		887	&& $_[3]
		888	&& ($hdr{HTTPVersion} < 1.1
		889	? $hdr{connection} =~ /\bkeep-?alive\b/i
		890	: $hdr{connection} !~ /\bclose\b/i)
		891	) {
		892	ka_store $ka_key, delete $state{handle};
		893	} else {
		894	# no keepalive, destroy the handle
		895	$state{handle}->destroy;
		896	}
		897	}
		898
		899	%state = ();
		900
		901	if (defined $_[1]) {
		902	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		903	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		904	}
		905
		906	# set-cookie processing
		907	if ($arg{cookie_jar}) {
		908	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		909	}
		910
		911	if ($redirect && exists $hdr{location}) {
		912	# we ignore any errors, as it is very common to receive
		913	# Content-Length != 0 but no actual body
		914	# we also access %hdr, as $_[1] might be an erro
		915	$state{recurse} =
		916	http_request (
		917	$method => $hdr{location},
		918	%arg,
		919	recurse => $recurse - 1,
		920	Redirect => [$_[0], \%hdr],
		921	sub {
		922	%state = ();
		923	&$cb
		924	},
		925	);
		926	} else {
		927	$cb->($_[0], \%hdr);
		928	}
		929	};
		930
		931	$ae_error = 597; # body phase
		932
		933	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		934
		935	my $len = $chunked ? undef : $hdr{"content-length"};
		936
		937	# body handling, many different code paths
		938	# - no body expected
		939	# - want_body_handle
		940	# - te chunked
		941	# - 2x length known (with or without on_body)
		942	# - 2x length not known (with or without on_body)
		943	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		944	$finish->(undef, 598 => "Request cancelled by on_header");
		945	} elsif (
		946	$hdr{Status} =~ /^(?:1..|204|205|304)$/
		947	or $method eq "HEAD"
		948	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		949	) {
		950	# no body
		951	$finish->("", undef, undef, 1);
		952
		953	} elsif (!$redirect && $arg{want_body_handle}) {
		954	$_[0]->on_eof (undef);
		955	$_[0]->on_error (undef);
		956	$_[0]->on_read (undef);
		957
		958	$finish->(delete $state{handle});
		959
		960	} elsif ($chunked) {
		961	my $cl = 0;
		962	my $body = "";
		963	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
		964
		965	$state{read_chunk} = sub {
		966	$_[1] =~ /^([0-9a-fA-F]+)/
		967	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		968
		969	my $len = hex $1;
		970
		971	if ($len) {
		972	$cl += $len;
		973
		974	$_[0]->push_read (chunk => $len, sub {
		975	$on_body->($_[1], \%hdr)
		976	or return $finish->(undef, 598 => "Request cancelled by on_body");
		977
		978	$_[0]->push_read (line => sub {
		979	length $_[1]
		980	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		981	$_[0]->push_read (line => $state{read_chunk});
		982	});
		983	});
		984	} else {
		985	$hdr{"content-length"} ||= $cl;
		986
		987	$_[0]->push_read (line => $qr_nlnl, sub {
		988	if (length $_[1]) {
		989	for ("$_[1]") {
		990	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		991
		992	my $hdr = _parse_hdr
		993	or return $finish->(undef, $ae_error => "Garbled response trailers");
		994
		995	%hdr = (%hdr, %$hdr);
		996	}
		997	}
		998
		999	$finish->($body, undef, undef, 1);
		1000	});
		1001	}
		1002	};
		1003
		1004	$_[0]->push_read (line => $state{read_chunk});
		1005
		1006	} elsif ($arg{on_body}) {
		1007	if (defined $len) {
		1008	$_[0]->on_read (sub {
		1009	$len -= length $_[0]{rbuf};
		1010
		1011	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1012	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1013
		1014	$len > 0
		1015	or $finish->("", undef, undef, 1);
		1016	});
		1017	} else {
		1018	$_[0]->on_eof (sub {
		1019	$finish->("");
		1020	});
		1021	$_[0]->on_read (sub {
		1022	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1023	or $finish->(undef, 598 => "Request cancelled by on_body");
		1024	});
		1025	}
		1026	} else {
		1027	$_[0]->on_eof (undef);
		1028
		1029	if (defined $len) {
		1030	$_[0]->on_read (sub {
		1031	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1032	if $len <= length $_[0]{rbuf};
		1033	});
		1034	} else {
		1035	$_[0]->on_error (sub {
		1036	($! == Errno::EPIPE || !$!)
		1037	? $finish->(delete $_[0]{rbuf})
		1038	: $finish->(undef, $ae_error => $_[2]);
		1039	});
		1040	$_[0]->on_read (sub { });
		1041	}
		1042	}
		1043	};
		1044
		1045	# if keepalive is enabled, then the server closing the connection
		1046	# before a response can happen legally - we retry on idempotent methods.
		1047	if ($was_persistent && $idempotent) {
		1048	my $old_eof = $hdl->{on_eof};
		1049	$hdl->{on_eof} = sub {
		1050	_destroy_state %state;
		1051
		1052	%state = ();
		1053	$state{recurse} =
		1054	http_request (
		1055	$method => $url,
		1056	%arg,
		1057	keepalive => 0,
		1058	sub {
		1059	%state = ();
		1060	&$cb
		1061	}
		1062	);
		1063	};
		1064	$hdl->on_read (sub {
		1065	return unless %state;
		1066
		1067	# as soon as we receive something, a connection close
		1068	# once more becomes a hard error
		1069	$hdl->{on_eof} = $old_eof;
		1070	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1071	});
		1072	} else {
		1073	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1074	}
		1075	};
		1076
		1077	my $prepare_handle = sub {
		1078	my ($hdl) = $state{handle};
		1079
		1080	$hdl->on_error (sub {
		1081	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1082	});
		1083	$hdl->on_eof (sub {
		1084	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1085	});
		1086	$hdl->timeout_reset;
		1087	$hdl->timeout ($timeout);
		1088	};
		1089
		1090	# connected to proxy (or origin server)
		1091	my $connect_cb = sub {
		1092	my $fh = shift
		1093	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1094
		1095	return unless delete $state{connect_guard};
		1096
		1097	# get handle
		1098	$state{handle} = new AnyEvent::Handle
		1099	%{ $arg{handle_params} },
		1100	fh => $fh,
		1101	peername => $uhost,
		1102	tls_ctx => $arg{tls_ctx},
		1103	;
		1104
		1105	$prepare_handle->();
		1106
		1107	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1108
		1109	# now handle proxy-CONNECT method
		1110	if ($proxy && $uscheme eq "https") {
		1111	# oh dear, we have to wrap it into a connect request
		1112
		1113	# maybe re-use $uauthority with patched port?
		1114	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012");
		1115	$state{handle}->push_read (line => $qr_nlnl, sub {
		1116	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1117	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1118
		1119	if ($2 == 200) {
		1120	$rpath = $upath;
		1121	$handle_actual_request->();
		1122	} else {
		1123	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
		1124	}
		1125	});
		1126	} else {
		1127	$handle_actual_request->();
		1128	}
		1129	};
599		1130
600	_get_slot $uhost, sub {	1131	_get_slot $uhost, sub {
601	$state{slot_guard} = shift;	1132	$state{slot_guard} = shift;
602		1133
603	return unless $state{connect_guard};	1134	return unless $state{connect_guard};
604		1135
605	my $connect_cb = sub {	1136	# try to use an existing keepalive connection, but only if we, ourselves, plan
606	$state{fh} = shift	1137	# on a keepalive request (in theory, this should be a separate config option).
607	or do {	1138	if ($persistent && $KA_CACHE{$ka_key}) {
608	my $err = "$!";	1139	$was_persistent = 1;
609	%state = ();
610	return $cb->(undef, { @pseudo, Status => 599, Reason => $err });
611	};
612		1140
613	pop; # free memory, save a tree	1141	$state{handle} = ka_fetch $ka_key;
614
615	return unless delete $state{connect_guard};
616
617	# get handle
618	$state{handle} = new AnyEvent::Handle
619	fh => $state{fh},
620	peername => $rhost,
621	tls_ctx => $arg{tls_ctx},
622	# these need to be reconfigured on keepalive handles
623	timeout => $timeout,
624	on_error => sub {
625	%state = ();
626	$cb->(undef, { @pseudo, Status => 599, Reason => $_[2] });
627	},
628	on_eof => sub {
629	%state = ();
630	$cb->(undef, { @pseudo, Status => 599, Reason => "Unexpected end-of-file" });
631	},
632	;
633
634	# limit the number of persistent connections
635	# keepalive not yet supported
636	# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
637	# ++$KA_COUNT{$_[1]};
638	# $state{handle}{ka_count_guard} = AnyEvent::Util::guard {
639	# --$KA_COUNT{$_[1]}
640	# };
641	# $hdr{connection} = "keep-alive";
642	# }
643
644	$state{handle}->starttls ("connect") if $rscheme eq "https";
645
646	# handle actual, non-tunneled, request
647	my $handle_actual_request = sub {
648	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};
649
650	# send request
651	$state{handle}->push_write (	1142	$state{handle}->destroyed
652	"$method $rpath HTTP/1.1\015\012"	1143	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
653	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)	1144	$prepare_handle->();
654	. "\015\012"	1145	$state{handle}->destroyed
655	. (delete $arg{body})	1146	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
656	);
657
658	# return if error occured during push_write()
659	return unless %state;
660
661	%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use
662
663	# status line and headers
664	$state{read_response} = sub {
665	for ("$_[1]") {
666	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
667
668	/^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/igxc
669	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" }));
670
671	# 100 Continue handling
672	# should not happen as we don't send expect: 100-continue,
673	# but we handle it just in case.
674	# since we send the request body regardless, if we get an error
675	# we are out of-sync, which we currently do NOT handle correctly.
676	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
677	if $2 eq 100;
678
679	push @pseudo,
680	HTTPVersion => $1,
681	Status => $2,
682	Reason => $3,
683	;
684
685	my $hdr = parse_hdr
686	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" }));
687
688	%hdr = (%$hdr, @pseudo);
689	}
690
691	# redirect handling
692	# microsoft and other shitheads don't give a shit for following standards,
693	# try to support some common forms of broken Location headers.
694	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
695	$hdr{location} =~ s/^\.\/+//;
696
697	my $url = "$rscheme://$uhost:$uport";
698
699	unless ($hdr{location} =~ s/^\///) {
700	$url .= $upath;
701	$url =~ s/\/[^\/]*$//;
702	}
703
704	$hdr{location} = "$url/$hdr{location}";
705	}
706
707	my $redirect;
708
709	if ($recurse) {
710	my $status = $hdr{Status};
711
712	# industry standard is to redirect POST as GET for
713	# 301, 302 and 303, in contrast to http/1.0 and 1.1.
714	# also, the UA should ask the user for 301 and 307 and POST,
715	# industry standard seems to be to simply follow.
716	# we go with the industry standard.
717	if ($status == 301 or $status == 302 or $status == 303) {
718	# HTTP/1.1 is unclear on how to mutate the method
719	$method = "GET" unless $method eq "HEAD";
720	$redirect = 1;
721	} elsif ($status == 307) {
722	$redirect = 1;
723	}
724	}
725
726	my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive])
727	my $keepalive = pop;
728
729	$state{handle}->destroy if $state{handle};
730	%state = ();
731
732	if (defined $_[1]) {
733	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
734	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
735	}
736
737	# set-cookie processing
738	if ($arg{cookie_jar}) {
739	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost;
740	}
741
742	if ($redirect && exists $hdr{location}) {
743	# we ignore any errors, as it is very common to receive
744	# Content-Length != 0 but no actual body
745	# we also access %hdr, as $_[1] might be an erro
746	http_request (
747	$method => $hdr{location},
748	%arg,
749	recurse => $recurse - 1,
750	Redirect => [$_[0], \%hdr],
751	$cb);
752	} else {
753	$cb->($_[0], \%hdr);
754	}
755	};
756
757	my $len = $hdr{"content-length"};
758
759	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
760	$finish->(undef, 598 => "Request cancelled by on_header");
761	} elsif (
762	$hdr{Status} =~ /^(?:1..|204|205|304)$/
763	or $method eq "HEAD"
764	or (defined $len && !$len)
765	) {
766	# no body
767	$finish->("", undef, undef, 1);
768	} else {
769	# body handling, many different code paths
770	# - no body expected
771	# - want_body_handle
772	# - te chunked
773	# - 2x length known (with or without on_body)
774	# - 2x length not known (with or without on_body)
775	if (!$redirect && $arg{want_body_handle}) {
776	$_[0]->on_eof (undef);
777	$_[0]->on_error (undef);
778	$_[0]->on_read (undef);
779
780	$finish->(delete $state{handle});
781
782	} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) {
783	my $cl = 0;
784	my $body = undef;
785	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
786
787	$_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) });
788
789	my $read_chunk; $read_chunk = sub {
790	$_[1] =~ /^([0-9a-fA-F]+)/
791	or $finish->(undef, 599 => "Garbled chunked transfer encoding");
792
793	my $len = hex $1;
794
795	if ($len) {
796	$cl += $len;
797
798	$_[0]->push_read (chunk => $len, sub {
799	$on_body->($_[1], \%hdr)
800	or return $finish->(undef, 598 => "Request cancelled by on_body");
801
802	$_[0]->push_read (line => sub {
803	length $_[1]
804	and return $finish->(undef, 599 => "Garbled chunked transfer encoding");
805	$_[0]->push_read (line => $read_chunk);
806	});
807	});
808	} else {
809	$hdr{"content-length"} ||= $cl;
810
811	$_[0]->push_read (line => $qr_nlnl, sub {
812	if (length $_[1]) {
813	for ("$_[1]") {
814	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
815
816	my $hdr = parse_hdr
817	or return $finish->(undef, 599 => "Garbled response trailers");
818
819	%hdr = (%hdr, %$hdr);
820	}
821	}
822
823	$finish->($body, undef, undef, 1);
824	});
825	}
826	};
827
828	$_[0]->push_read (line => $read_chunk);
829
830	} elsif ($arg{on_body}) {
831	$_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) });
832
833	if ($len) {
834	$_[0]->on_read (sub {
835	$len -= length $_[0]{rbuf};
836
837	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
838	or return $finish->(undef, 598 => "Request cancelled by on_body");
839
840	$len > 0
841	or $finish->("", undef, undef, 1);
842	});
843	} else {
844	$_[0]->on_eof (sub {
845	$finish->("");
846	});
847	$_[0]->on_read (sub {
848	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
849	or $finish->(undef, 598 => "Request cancelled by on_body");
850	});
851	}
852	} else {
853	$_[0]->on_eof (undef);
854
855	if ($len) {
856	$_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) });
857	$_[0]->on_read (sub {
858	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
859	if $len <= length $_[0]{rbuf};
860	});
861	} else {
862	$_[0]->on_error (sub {
863	($! == Errno::EPIPE || !$!)
864	? $finish->(delete $_[0]{rbuf})
865	: $finish->(undef, 599 => $_[2]);
866	});
867	$_[0]->on_read (sub { });
868	}
869	}
870	}
871	};
872
873	$state{handle}->push_read (line => $qr_nlnl, $state{read_response});
874	};
875
876	# now handle proxy-CONNECT method
877	if ($proxy && $uscheme eq "https") {
878	# oh dear, we have to wrap it into a connect request
879
880	# maybe re-use $uauthority with patched port?
881	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012");
882	$state{handle}->push_read (line => $qr_nlnl, sub {
883	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
884	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }));
885
886	if ($2 == 200) {
887	$rpath = $upath;
888	&$handle_actual_request;	1147	$handle_actual_request->();
889	} else {	1148
890	%state = ();
891	$cb->(undef, { @pseudo, Status => $2, Reason => $3 });
892	}
893	});
894	} else {	1149	} else {
895	&$handle_actual_request;
896	}
897	};
898
899	my $tcp_connect = $arg{tcp_connect}	1150	my $tcp_connect = $arg{tcp_connect}
900	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };	1151	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
901		1152
902	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });	1153	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });
903		1154	}
904	};	1155	};
905		1156
906	defined wantarray && AnyEvent::Util::guard { %state = () }	1157	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
907	}	1158	}
908		1159
909	sub http_get($@) {	1160	sub http_get($@) {
910	unshift @_, "GET";	1161	unshift @_, "GET";
911	&http_request	1162	&http_request
…		…
929	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for	1180	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
930	the actual connection, which in turn uses AnyEvent::DNS to resolve	1181	the actual connection, which in turn uses AnyEvent::DNS to resolve
931	hostnames. The latter is a simple stub resolver and does no caching	1182	hostnames. The latter is a simple stub resolver and does no caching
932	on its own. If you want DNS caching, you currently have to provide	1183	on its own. If you want DNS caching, you currently have to provide
933	your own default resolver (by storing a suitable resolver object in	1184	your own default resolver (by storing a suitable resolver object in
934	C<$AnyEvent::DNS::RESOLVER>).	1185	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
935		1186
936	=head2 GLOBAL FUNCTIONS AND VARIABLES	1187	=head2 GLOBAL FUNCTIONS AND VARIABLES
937		1188
938	=over 4	1189	=over 4
939		1190
940	=item AnyEvent::HTTP::set_proxy "proxy-url"	1191	=item AnyEvent::HTTP::set_proxy "proxy-url"
941		1192
942	Sets the default proxy server to use. The proxy-url must begin with a	1193	Sets the default proxy server to use. The proxy-url must begin with a
943	string of the form C<http://host:port> (optionally C<https:...>), croaks	1194	string of the form C<http://host:port>, croaks otherwise.
944	otherwise.
945		1195
946	To clear an already-set proxy, use C<undef>.	1196	To clear an already-set proxy, use C<undef>.
		1197
		1198	When AnyEvent::HTTP is laoded for the first time it will query the
		1199	default proxy from the operating system, currently by looking at
		1200	C<$ENV{http_proxy>}.
		1201
		1202	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
		1203
		1204	Remove all cookies from the cookie jar that have been expired. If
		1205	C<$session_end> is given and true, then additionally remove all session
		1206	cookies.
		1207
		1208	You should call this function (with a true C<$session_end>) before you
		1209	save cookies to disk, and you should call this function after loading them
		1210	again. If you have a long-running program you can additonally call this
		1211	function from time to time.
		1212
		1213	A cookie jar is initially an empty hash-reference that is managed by this
		1214	module. It's format is subject to change, but currently it is like this:
		1215
		1216	The key C<version> has to contain C<1>, otherwise the hash gets
		1217	emptied. All other keys are hostnames or IP addresses pointing to
		1218	hash-references. The key for these inner hash references is the
		1219	server path for which this cookie is meant, and the values are again
		1220	hash-references. The keys of those hash-references is the cookie name, and
		1221	the value, you guessed it, is another hash-reference, this time with the
		1222	key-value pairs from the cookie, except for C<expires> and C<max-age>,
		1223	which have been replaced by a C<_expires> key that contains the cookie
		1224	expiry timestamp.
		1225
		1226	Here is an example of a cookie jar with a single cookie, so you have a
		1227	chance of understanding the above paragraph:
		1228
		1229	{
		1230	version => 1,
		1231	"10.0.0.1" => {
		1232	"/" => {
		1233	"mythweb_id" => {
		1234	_expires => 1293917923,
		1235	value => "ooRung9dThee3ooyXooM1Ohm",
		1236	},
		1237	},
		1238	},
		1239	}
947		1240
948	=item $date = AnyEvent::HTTP::format_date $timestamp	1241	=item $date = AnyEvent::HTTP::format_date $timestamp
949		1242
950	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP	1243	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP
951	Date (RFC 2616).	1244	Date (RFC 2616).
952		1245
953	=item $timestamp = AnyEvent::HTTP::parse_date $date	1246	=item $timestamp = AnyEvent::HTTP::parse_date $date
954		1247
955	Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) and	1248	Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) or a
956	returns the corresponding POSIX timestamp, or C<undef> if the date cannot	1249	bunch of minor variations of those, and returns the corresponding POSIX
957	be parsed.	1250	timestamp, or C<undef> if the date cannot be parsed.
958		1251
959	=item $AnyEvent::HTTP::MAX_RECURSE	1252	=item $AnyEvent::HTTP::MAX_RECURSE
960		1253
961	The default value for the C<recurse> request parameter (default: C<10>).	1254	The default value for the C<recurse> request parameter (default: C<10>).
		1255
		1256	=item $AnyEvent::HTTP::TIMEOUT
		1257
		1258	The default timeout for conenction operations (default: C<300>).
962		1259
963	=item $AnyEvent::HTTP::USERAGENT	1260	=item $AnyEvent::HTTP::USERAGENT
964		1261
965	The default value for the C<User-Agent> header (the default is	1262	The default value for the C<User-Agent> header (the default is
966	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1263	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
967		1264
968	=item $AnyEvent::HTTP::MAX_PER_HOST	1265	=item $AnyEvent::HTTP::MAX_PER_HOST
969		1266
970	The maximum number of concurrent connections to the same host (identified	1267	The maximum number of concurrent connections to the same host (identified
971	by the hostname). If the limit is exceeded, then the additional requests	1268	by the hostname). If the limit is exceeded, then the additional requests
972	are queued until previous connections are closed.	1269	are queued until previous connections are closed. Both persistent and
		1270	non-persistent connections are counted in this limit.
973		1271
974	The default value for this is C<4>, and it is highly advisable to not	1272	The default value for this is C<4>, and it is highly advisable to not
975	increase it.	1273	increase it much.
		1274
		1275	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1276	connections, older browsers used 2, newers (such as firefox 3) typically
		1277	use 6, and Opera uses 8 because like, they have the fastest browser and
		1278	give a shit for everybody else on the planet.
		1279
		1280	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
		1281
		1282	The time after which idle persistent conenctions get closed by
		1283	AnyEvent::HTTP (default: C<3>).
976		1284
977	=item $AnyEvent::HTTP::ACTIVE	1285	=item $AnyEvent::HTTP::ACTIVE
978		1286
979	The number of active connections. This is not the number of currently	1287	The number of active connections. This is not the number of currently
980	running requests, but the number of currently open and non-idle TCP	1288	running requests, but the number of currently open and non-idle TCP
981	connections. This number of can be useful for load-leveling.	1289	connections. This number can be useful for load-leveling.
982		1290
983	=back	1291	=back
984		1292
985	=cut	1293	=cut
986		1294
…		…
1001	sub parse_date($) {	1309	sub parse_date($) {
1002	my ($date) = @_;	1310	my ($date) = @_;
1003		1311
1004	my ($d, $m, $y, $H, $M, $S);	1312	my ($d, $m, $y, $H, $M, $S);
1005		1313
1006	if ($date =~ /^[A-Z][a-z][a-z], ([0-9][0-9])[\- ]([A-Z][a-z][a-z])[\- ]([0-9][0-9][0-9][0-9]) ([0-9][0-9]):([0-9][0-9]):([0-9][0-9]) GMT$/) {	1314	if ($date =~ /^[A-Z][a-z][a-z]+, ([0-9][0-9]?)[\- ]([A-Z][a-z][a-z])[\- ]([0-9][0-9][0-9][0-9]) ([0-9][0-9]?):([0-9][0-9]?):([0-9][0-9]?) GMT$/) {
1007	# RFC 822/1123, required by RFC 2616 (with " ")	1315	# RFC 822/1123, required by RFC 2616 (with " ")
1008	# cookie dates (with "-")	1316	# cookie dates (with "-")
1009		1317
1010	($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6);	1318	($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6);
1011		1319
1012	} elsif ($date =~ /^[A-Z][a-z]+, ([0-9][0-9])-([A-Z][a-z][a-z])-([0-9][0-9]) ([0-9][0-9]):([0-9][0-9]):([0-9][0-9]) GMT$/) {	1320	} elsif ($date =~ /^[A-Z][a-z][a-z]+, ([0-9][0-9]?)-([A-Z][a-z][a-z])-([0-9][0-9]) ([0-9][0-9]?):([0-9][0-9]?):([0-9][0-9]?) GMT$/) {
1013	# RFC 850	1321	# RFC 850
1014	($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6);	1322	($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6);
1015		1323
1016	} elsif ($date =~ /^[A-Z][a-z][a-z] ([A-Z][a-z][a-z]) ([0-9 ][0-9]) ([0-9][0-9]):([0-9][0-9]):([0-9][0-9]) ([0-9][0-9][0-9][0-9])$/) {	1324	} elsif ($date =~ /^[A-Z][a-z][a-z]+ ([A-Z][a-z][a-z]) ([0-9 ]?[0-9]) ([0-9][0-9]?):([0-9][0-9]?):([0-9][0-9]?) ([0-9][0-9][0-9][0-9])$/) {
1017	# ISO C's asctime	1325	# ISO C's asctime
1018	($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5);	1326	($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5);
1019	}	1327	}
1020	# other formats fail in the loop below	1328	# other formats fail in the loop below
1021		1329
…		…
1029	undef	1337	undef
1030	}	1338	}
1031		1339
1032	sub set_proxy($) {	1340	sub set_proxy($) {
1033	if (length $_[0]) {	1341	if (length $_[0]) {
1034	$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix	1342	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
1035	or Carp::croak "$_[0]: invalid proxy URL";	1343	or Carp::croak "$_[0]: invalid proxy URL";
1036	$PROXY = [$2, $3 || 3128, $1]	1344	$PROXY = [$2, $3 || 3128, $1]
1037	} else {	1345	} else {
1038	undef $PROXY;	1346	undef $PROXY;
1039	}	1347	}
…		…
1042	# initialise proxy from environment	1350	# initialise proxy from environment
1043	eval {	1351	eval {
1044	set_proxy $ENV{http_proxy};	1352	set_proxy $ENV{http_proxy};
1045	};	1353	};
1046		1354
		1355	=head2 SHOWCASE
		1356
		1357	This section contaisn some more elaborate "real-world" examples or code
		1358	snippets.
		1359
		1360	=head2 HTTP/1.1 FILE DOWNLOAD
		1361
		1362	Downloading files with HTTP can be quite tricky, especially when something
		1363	goes wrong and you want to resume.
		1364
		1365	Here is a function that initiates and resumes a download. It uses the
		1366	last modified time to check for file content changes, and works with many
		1367	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1368	on older servers.
		1369
		1370	It calls the completion callback with either C<undef>, which means a
		1371	nonretryable error occured, C<0> when the download was partial and should
		1372	be retried, and C<1> if it was successful.
		1373
		1374	use AnyEvent::HTTP;
		1375
		1376	sub download($$$) {
		1377	my ($url, $file, $cb) = @_;
		1378
		1379	open my $fh, "+<", $file
		1380	or die "$file: $!";
		1381
		1382	my %hdr;
		1383	my $ofs = 0;
		1384
		1385	warn stat $fh;
		1386	warn -s _;
		1387	if (stat $fh and -s _) {
		1388	$ofs = -s _;
		1389	warn "-s is ", $ofs;
		1390	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1391	$hdr{"range"} = "bytes=$ofs-";
		1392	}
		1393
		1394	http_get $url,
		1395	headers => \%hdr,
		1396	on_header => sub {
		1397	my ($hdr) = @_;
		1398
		1399	if ($hdr->{Status} == 200 && $ofs) {
		1400	# resume failed
		1401	truncate $fh, $ofs = 0;
		1402	}
		1403
		1404	sysseek $fh, $ofs, 0;
		1405
		1406	1
		1407	},
		1408	on_body => sub {
		1409	my ($data, $hdr) = @_;
		1410
		1411	if ($hdr->{Status} =~ /^2/) {
		1412	length $data == syswrite $fh, $data
		1413	or return; # abort on write errors
		1414	}
		1415
		1416	1
		1417	},
		1418	sub {
		1419	my (undef, $hdr) = @_;
		1420
		1421	my $status = $hdr->{Status};
		1422
		1423	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1424	utime $fh, $time, $time;
		1425	}
		1426
		1427	if ($status == 200 || $status == 206 || $status == 416) {
		1428	# download ok || resume ok || file already fully downloaded
		1429	$cb->(1, $hdr);
		1430
		1431	} elsif ($status == 412) {
		1432	# file has changed while resuming, delete and retry
		1433	unlink $file;
		1434	$cb->(0, $hdr);
		1435
		1436	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1437	# retry later
		1438	$cb->(0, $hdr);
		1439
		1440	} else {
		1441	$cb->(undef, $hdr);
		1442	}
		1443	}
		1444	;
		1445	}
		1446
		1447	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1448	if ($_[0]) {
		1449	print "OK!\n";
		1450	} elsif (defined $_[0]) {
		1451	print "please retry later\n";
		1452	} else {
		1453	print "ERROR\n";
		1454	}
		1455	};
		1456
1047	=head2 SOCKS PROXIES	1457	=head3 SOCKS PROXIES
1048		1458
1049	Socks proxies are not directly supported by AnyEvent::HTTP. You can	1459	Socks proxies are not directly supported by AnyEvent::HTTP. You can
1050	compile your perl to support socks, or use an external program such as	1460	compile your perl to support socks, or use an external program such as
1051	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy	1461	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
1052	transparently.	1462	transparently.

Diff Legend

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