ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.71 by root, Fri Dec 31 20:50:58 2010 UTC vs.
Revision 1.116 by root, Fri May 17 07:19:23 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, $headers)
73		68
74	Executes an HTTP-GET request. See the http_request function for details on	69	Executes an HTTP-GET request. See the http_request function for details on
…		…
122		117
123	If the server sends a header multiple times, then their contents will be	118	If the server sends a header multiple times, then their contents will be
124	joined together with a comma (C<,>), as per the HTTP spec.	119	joined together with a comma (C<,>), as per the HTTP spec.
125		120
126	If an internal error occurs, such as not being able to resolve a hostname,	121	If an internal error occurs, such as not being able to resolve a hostname,
127	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x>	122	then C<$data> will be C<undef>, C<< $headers->{Status} >> will be
128	(usually C<599>) and the C<Reason> pseudo-header will contain an error	123	C<590>-C<599> and the C<Reason> pseudo-header will contain an error
129	message.	124	message. Currently the following status codes are used:
		125
		126	=over 4
		127
		128	=item 595 - errors during connection establishment, proxy handshake.
		129
		130	=item 596 - errors during TLS negotiation, request sending and header processing.
		131
		132	=item 597 - errors during body receiving or processing.
		133
		134	=item 598 - user aborted request via C<on_header> or C<on_body>.
		135
		136	=item 599 - other, usually nonretryable, errors (garbled URL etc.).
		137
		138	=back
130		139
131	A typical callback might look like this:	140	A typical callback might look like this:
132		141
133	sub {	142	sub {
134	my ($body, $hdr) = @_;	143	my ($body, $hdr) = @_;
…		…
145		154
146	=over 4	155	=over 4
147		156
148	=item recurse => $count (default: $MAX_RECURSE)	157	=item recurse => $count (default: $MAX_RECURSE)
149		158
150	Whether to recurse requests or not, e.g. on redirects, authentication	159	Whether to recurse requests or not, e.g. on redirects, authentication and
151	retries and so on, and how often to do so.	160	other retries and so on, and how often to do so.
152		161
153	=item headers => hashref	162	=item headers => hashref
154		163
155	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
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 => $boolean
		353
		354	Only used when C<persistent> is also true. This parameter decides whether
		355	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		356	(as opposed to only a HTTP/1.1 persistent connection).
		357
		358	The default is true, except when using a proxy, in which case it defaults
		359	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		360
		361	=item handle_params => { key => value ... }
		362
		363	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		364	constructor that is called - not all requests will create a handle, and
		365	sometimes more than one is created, so this parameter is only good for
		366	setting hints.
		367
		368	Example: set the maximum read size to 4096, to potentially conserve memory
		369	at the cost of speed.
		370
		371	handle_params => {
		372	max_read_size => 4096,
		373	},
303		374
304	=back	375	=back
305		376
306	Example: do a simple HTTP GET request for http://www.nethype.de/ and print	377	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
307	the response body.	378	the response body.
…		…
313		384
314	Example: do a HTTP HEAD request on https://www.google.com/, use a	385	Example: do a HTTP HEAD request on https://www.google.com/, use a
315	timeout of 30 seconds.	386	timeout of 30 seconds.
316		387
317	http_request	388	http_request
318	GET => "https://www.google.com",	389	HEAD => "https://www.google.com",
		390	headers => { "user-agent" => "MySearchClient 1.0" },
319	timeout => 30,	391	timeout => 30,
320	sub {	392	sub {
321	my ($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
369	sub cookie_jar_extract($$$$) {	448	sub cookie_jar_expire($;$) {
370	my ($jar, $uscheme, $uhost, $upath) = @_;	449	my ($jar, $session_end) = @_;
371		450
372	%$jar = () if $jar->{version} != 1;	451	%$jar = () if $jar->{version} != 1;
373		452
374	my @cookies;	453	my $anow = AE::now;
375		454
376	while (my ($chost, $paths) = each %$jar) {	455	while (my ($chost, $paths) = each %$jar) {
377	next unless ref $paths;	456	next unless ref $paths;
378		457
		458	while (my ($cpath, $cookies) = each %$paths) {
		459	while (my ($cookie, $kv) = each %$cookies) {
		460	if (exists $kv->{_expires}) {
		461	delete $cookies->{$cookie}
		462	if $anow > $kv->{_expires};
		463	} elsif ($session_end) {
		464	delete $cookies->{$cookie};
		465	}
		466	}
		467
		468	delete $paths->{$cpath}
		469	unless %$cookies;
		470	}
		471
		472	delete $jar->{$chost}
		473	unless %$paths;
		474	}
		475	}
		476
		477	# extract cookies from jar
		478	sub cookie_jar_extract($$$$) {
		479	my ($jar, $scheme, $host, $path) = @_;
		480
		481	%$jar = () if $jar->{version} != 1;
		482
		483	my @cookies;
		484
		485	while (my ($chost, $paths) = each %$jar) {
		486	next unless ref $paths;
		487
379	if ($chost =~ /^\./) {	488	if ($chost =~ /^\./) {
380	next unless $chost eq substr $uhost, -length $chost;	489	next unless $chost eq substr $host, -length $chost;
381	} elsif ($chost =~ /\./) {	490	} elsif ($chost =~ /\./) {
382	next unless $chost eq $uhost;	491	next unless $chost eq $host;
383	} else {	492	} else {
384	next;	493	next;
385	}	494	}
386		495
387	while (my ($cpath, $cookies) = each %$paths) {	496	while (my ($cpath, $cookies) = each %$paths) {
388	next unless $cpath eq substr $upath, 0, length $cpath;	497	next unless $cpath eq substr $path, 0, length $cpath;
389		498
390	while (my ($cookie, $kv) = each %$cookies) {	499	while (my ($cookie, $kv) = each %$cookies) {
391	next if $uscheme ne "https" && exists $kv->{secure};	500	next if $scheme ne "https" && exists $kv->{secure};
392		501
393	if (exists $kv->{expires}) {	502	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
394	if (AE::now > parse_date ($kv->{expires})) {
395	delete $cookies->{$cookie};	503	delete $cookies->{$cookie};
396	next;	504	next;
397	}
398	}	505	}
399		506
400	my $value = $kv->{value};	507	my $value = $kv->{value};
401		508
402	if ($value =~ /[=;,[:space:]]/) {	509	if ($value =~ /[=;,[:space:]]/) {
…		…
410	}	517	}
411		518
412	\@cookies	519	\@cookies
413	}	520	}
414		521
		522	# parse set_cookie header into jar
		523	sub cookie_jar_set_cookie($$$$) {
		524	my ($jar, $set_cookie, $host, $date) = @_;
		525
		526	my $anow = int AE::now;
		527	my $snow; # server-now
		528
		529	for ($set_cookie) {
		530	# parse NAME=VALUE
		531	my @kv;
		532
		533	# expires is not http-compliant in the original cookie-spec,
		534	# we support the official date format and some extensions
		535	while (
		536	m{
		537	\G\s*
		538	(?:
		539	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
		540	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
		541	)
		542	}gcxsi
		543	) {
		544	my $name = $2;
		545	my $value = $4;
		546
		547	if (defined $1) {
		548	# expires
		549	$name = "expires";
		550	$value = $1;
		551	} elsif (defined $3) {
		552	# quoted
		553	$value = $3;
		554	$value =~ s/\\(.)/$1/gs;
		555	}
		556
		557	push @kv, @kv ? lc $name : $name, $value;
		558
		559	last unless /\G\s*;/gc;
		560	}
		561
		562	last unless @kv;
		563
		564	my $name = shift @kv;
		565	my %kv = (value => shift @kv, @kv);
		566
		567	if (exists $kv{"max-age"}) {
		568	$kv{_expires} = $anow + delete $kv{"max-age"};
		569	} elsif (exists $kv{expires}) {
		570	$snow ||= parse_date ($date) || $anow;
		571	$kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow);
		572	} else {
		573	delete $kv{_expires};
		574	}
		575
		576	my $cdom;
		577	my $cpath = (delete $kv{path}) || "/";
		578
		579	if (exists $kv{domain}) {
		580	$cdom = delete $kv{domain};
		581
		582	$cdom =~ s/^\.?/./; # make sure it starts with a "."
		583
		584	next if $cdom =~ /\.$/;
		585
		586	# this is not rfc-like and not netscape-like. go figure.
		587	my $ndots = $cdom =~ y/.//;
		588	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
		589	} else {
		590	$cdom = $host;
		591	}
		592
		593	# store it
		594	$jar->{version} = 1;
		595	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
		596
		597	redo if /\G\s*,/gc;
		598	}
		599	}
		600
		601	#############################################################################
		602	# keepalive/persistent connection cache
		603
		604	# fetch a connection from the keepalive cache
		605	sub ka_fetch($) {
		606	my $ka_key = shift;
		607
		608	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		609	delete $KA_CACHE{$ka_key}
		610	unless @{ $KA_CACHE{$ka_key} };
		611
		612	$hdl
		613	}
		614
		615	sub ka_store($$) {
		616	my ($ka_key, $hdl) = @_;
		617
		618	my $kaa = $KA_CACHE{$ka_key} ||= [];
		619
		620	my $destroy = sub {
		621	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		622
		623	$hdl->destroy;
		624
		625	@ka
		626	? $KA_CACHE{$ka_key} = \@ka
		627	: delete $KA_CACHE{$ka_key};
		628	};
		629
		630	# on error etc., destroy
		631	$hdl->on_error ($destroy);
		632	$hdl->on_eof ($destroy);
		633	$hdl->on_read ($destroy);
		634	$hdl->timeout ($PERSISTENT_TIMEOUT);
		635
		636	push @$kaa, $hdl;
		637	shift @$kaa while @$kaa > $MAX_PER_HOST;
		638	}
		639
		640	#############################################################################
		641	# utilities
		642
415	# continue to parse $_ for headers and place them into the arg	643	# continue to parse $_ for headers and place them into the arg
416	sub parse_hdr() {	644	sub _parse_hdr() {
417	my %hdr;	645	my %hdr;
418		646
419	# things seen, not parsed:	647	# things seen, not parsed:
420	# p3pP="NON CUR OTPi OUR NOR UNI"	648	# p3pP="NON CUR OTPi OUR NOR UNI"
421		649
…		…
435	for values %hdr;	663	for values %hdr;
436		664
437	\%hdr	665	\%hdr
438	}	666	}
439		667
		668	#############################################################################
		669	# http_get
		670
440	our $qr_nlnl = qr{(?<![^\012])\015?\012};	671	our $qr_nlnl = qr{(?<![^\012])\015?\012};
441		672
442	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	673	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
443	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(\%$$) {
		686	my ($state, $cb, $hdr) = @_;
		687
		688	&_destroy_state ($state);
		689
		690	$cb->(undef, $hdr);
		691	()
		692	}
		693
		694	our %IDEMPOTENT = (
		695	DELETE => 1,
		696	GET => 1,
		697	HEAD => 1,
		698	OPTIONS => 1,
		699	PUT => 1,
		700	TRACE => 1,
		701
		702	ACL => 1,
		703	"BASELINE-CONTROL" => 1,
		704	BIND => 1,
		705	CHECKIN => 1,
		706	CHECKOUT => 1,
		707	COPY => 1,
		708	LABEL => 1,
		709	LINK => 1,
		710	MERGE => 1,
		711	MKACTIVITY => 1,
		712	MKCALENDAR => 1,
		713	MKCOL => 1,
		714	MKREDIRECTREF => 1,
		715	MKWORKSPACE => 1,
		716	MOVE => 1,
		717	ORDERPATCH => 1,
		718	PROPFIND => 1,
		719	PROPPATCH => 1,
		720	REBIND => 1,
		721	REPORT => 1,
		722	SEARCH => 1,
		723	UNBIND => 1,
		724	UNCHECKOUT => 1,
		725	UNLINK => 1,
		726	UNLOCK => 1,
		727	UPDATE => 1,
		728	UPDATEREDIRECTREF => 1,
		729	"VERSION-CONTROL" => 1,
		730	);
444		731
445	sub http_request($$@) {	732	sub http_request($$@) {
446	my $cb = pop;	733	my $cb = pop;
447	my ($method, $url, %arg) = @_;	734	my ($method, $url, %arg) = @_;
448		735
…		…
466	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	753	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
467		754
468	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })	755	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
469	if $recurse < 0;	756	if $recurse < 0;
470		757
471	my $proxy = $arg{proxy} || $PROXY;	758	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
472	my $timeout = $arg{timeout} || $TIMEOUT;	759	my $timeout = $arg{timeout} || $TIMEOUT;
473		760
474	my ($uscheme, $uauthority, $upath, $query, $fragment) =	761	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
475	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|;	762	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
476		763
477	$uscheme = lc $uscheme;	764	$uscheme = lc $uscheme;
478		765
479	my $uport = $uscheme eq "http" ? 80	766	my $uport = $uscheme eq "http" ? 80
480	: $uscheme eq "https" ? 443	767	: $uscheme eq "https" ? 443
481	: 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" });
482		769
483	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	770	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x
484	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });	771	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
485		772
486	my $uhost = $1;	773	my $uhost = lc $1;
487	$uport = $2 if defined $2;	774	$uport = $2 if defined $2;
488		775
489	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"	776	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
490	unless exists $hdr{host};	777	unless exists $hdr{host};
491		778
…		…
510	$rscheme = "http" unless defined $rscheme;	797	$rscheme = "http" unless defined $rscheme;
511		798
512	# don't support https requests over https-proxy transport,	799	# don't support https requests over https-proxy transport,
513	# 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.
514	$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;
515	} else {	805	} else {
516	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);	806	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
517	}	807	}
518		808
519	# leave out fragment and query string, just a heuristic	809	# leave out fragment and query string, just a heuristic
…		…
521	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};	811	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
522		812
523	$hdr{"content-length"} = length $arg{body}	813	$hdr{"content-length"} = length $arg{body}
524	if length $arg{body} || $method ne "GET";	814	if length $arg{body} || $method ne "GET";
525		815
526	$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
527	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1	827	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
528		828
529	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 occured 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	};
530		1169
531	_get_slot $uhost, sub {	1170	_get_slot $uhost, sub {
532	$state{slot_guard} = shift;	1171	$state{slot_guard} = shift;
533		1172
534	return unless $state{connect_guard};	1173	return unless $state{connect_guard};
535		1174
536	my $connect_cb = sub {	1175	# try to use an existing keepalive connection, but only if we, ourselves, plan
537	$state{fh} = shift	1176	# on a keepalive request (in theory, this should be a separate config option).
538	or do {	1177	if ($persistent && $KA_CACHE{$ka_key}) {
539	my $err = "$!";	1178	$was_persistent = 1;
540	%state = ();
541	return $cb->(undef, { @pseudo, Status => 599, Reason => $err });
542	};
543		1179
544	pop; # free memory, save a tree	1180	$state{handle} = ka_fetch $ka_key;
545
546	return unless delete $state{connect_guard};
547
548	# get handle
549	$state{handle} = new AnyEvent::Handle
550	fh => $state{fh},
551	peername => $rhost,
552	tls_ctx => $arg{tls_ctx},
553	# these need to be reconfigured on keepalive handles
554	timeout => $timeout,
555	on_error => sub {
556	%state = ();
557	$cb->(undef, { @pseudo, Status => 599, Reason => $_[2] });
558	},
559	on_eof => sub {
560	%state = ();
561	$cb->(undef, { @pseudo, Status => 599, Reason => "Unexpected end-of-file" });
562	},
563	;
564
565	# limit the number of persistent connections
566	# keepalive not yet supported
567	# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
568	# ++$KA_COUNT{$_[1]};
569	# $state{handle}{ka_count_guard} = AnyEvent::Util::guard {
570	# --$KA_COUNT{$_[1]}
571	# };
572	# $hdr{connection} = "keep-alive";
573	# }
574
575	$state{handle}->starttls ("connect") if $rscheme eq "https";
576
577	# handle actual, non-tunneled, request
578	my $handle_actual_request = sub {
579	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};
580
581	# send request
582	$state{handle}->push_write (	1181	$state{handle}->destroyed
583	"$method $rpath HTTP/1.1\015\012"	1182	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
584	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)	1183	$prepare_handle->();
585	. "\015\012"	1184	$state{handle}->destroyed
586	. (delete $arg{body})	1185	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
587	);
588
589	# return if error occured during push_write()
590	return unless %state;
591
592	%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use
593
594	# status line and headers
595	$state{read_response} = sub {
596	for ("$_[1]") {
597	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
598
599	/^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/igxc
600	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" }));
601
602	# 100 Continue handling
603	# should not happen as we don't send expect: 100-continue,
604	# but we handle it just in case.
605	# since we send the request body regardless, if we get an error
606	# we are out of-sync, which we currently do NOT handle correctly.
607	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
608	if $2 eq 100;
609
610	push @pseudo,
611	HTTPVersion => $1,
612	Status => $2,
613	Reason => $3,
614	;
615
616	my $hdr = parse_hdr
617	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" }));
618
619	%hdr = (%$hdr, @pseudo);
620	}
621
622	# redirect handling
623	# microsoft and other shitheads don't give a shit for following standards,
624	# try to support some common forms of broken Location headers.
625	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
626	$hdr{location} =~ s/^\.\/+//;
627
628	my $url = "$rscheme://$uhost:$uport";
629
630	unless ($hdr{location} =~ s/^\///) {
631	$url .= $upath;
632	$url =~ s/\/[^\/]*$//;
633	}
634
635	$hdr{location} = "$url/$hdr{location}";
636	}
637
638	my $redirect;
639
640	if ($recurse) {
641	my $status = $hdr{Status};
642
643	# industry standard is to redirect POST as GET for
644	# 301, 302 and 303, in contrast to http/1.0 and 1.1.
645	# also, the UA should ask the user for 301 and 307 and POST,
646	# industry standard seems to be to simply follow.
647	# we go with the industry standard.
648	if ($status == 301 or $status == 302 or $status == 303) {
649	# HTTP/1.1 is unclear on how to mutate the method
650	$method = "GET" unless $method eq "HEAD";
651	$redirect = 1;
652	} elsif ($status == 307) {
653	$redirect = 1;
654	}
655	}
656
657	my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive])
658	my $keepalive = pop;
659
660	$state{handle}->destroy if $state{handle};
661	%state = ();
662
663	if (defined $_[1]) {
664	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
665	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
666	}
667
668	# set-cookie processing
669	if ($arg{cookie_jar}) {
670	for ($hdr{"set-cookie"}) {
671	# parse NAME=VALUE
672	my @kv;
673
674	while (
675	m{
676	\G\s*
677	(?:
678	expires \s*=\s* ([A-Z][a-z][a-z],\ [^,;]+)
679	| ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )
680	)
681	}gcxsi
682	) {
683	my $name = $2;
684	my $value = $4;
685
686	unless (defined $name) {
687	# expires
688	$name = "expires";
689	$value = $1;
690	} elsif (!defined $value) {
691	# quoted
692	$value = $3;
693	$value =~ s/\\(.)/$1/gs;
694	}
695
696	push @kv, lc $name, $value;
697
698	last unless /\G\s*;/gc;
699	}
700
701	last unless @kv;
702
703	my $name = shift @kv;
704	my %kv = (value => shift @kv, @kv);
705
706	$kv{expires} ||= format_date (AE::now + $kv{"max-age"})
707	if exists $kv{"max-age"};
708
709	my $cdom;
710	my $cpath = (delete $kv{path}) || "/";
711
712	if (exists $kv{domain}) {
713	$cdom = delete $kv{domain};
714
715	$cdom =~ s/^\.?/./; # make sure it starts with a "."
716
717	next if $cdom =~ /\.$/;
718
719	# this is not rfc-like and not netscape-like. go figure.
720	my $ndots = $cdom =~ y/.//;
721	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
722	} else {
723	$cdom = $uhost;
724	}
725
726	# store it
727	$arg{cookie_jar}{version} = 1;
728	$arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv;
729
730	redo if /\G\s*,/gc;
731	}
732	}
733
734	if ($redirect && exists $hdr{location}) {
735	# we ignore any errors, as it is very common to receive
736	# Content-Length != 0 but no actual body
737	# we also access %hdr, as $_[1] might be an erro
738	http_request (
739	$method => $hdr{location},
740	%arg,
741	recurse => $recurse - 1,
742	Redirect => [$_[0], \%hdr],
743	$cb);
744	} else {
745	$cb->($_[0], \%hdr);
746	}
747	};
748
749	my $len = $hdr{"content-length"};
750
751	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
752	$finish->(undef, 598 => "Request cancelled by on_header");
753	} elsif (
754	$hdr{Status} =~ /^(?:1..|204|205|304)$/
755	or $method eq "HEAD"
756	or (defined $len && !$len)
757	) {
758	# no body
759	$finish->("", undef, undef, 1);
760	} else {
761	# body handling, many different code paths
762	# - no body expected
763	# - want_body_handle
764	# - te chunked
765	# - 2x length known (with or without on_body)
766	# - 2x length not known (with or without on_body)
767	if (!$redirect && $arg{want_body_handle}) {
768	$_[0]->on_eof (undef);
769	$_[0]->on_error (undef);
770	$_[0]->on_read (undef);
771
772	$finish->(delete $state{handle});
773
774	} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) {
775	my $cl = 0;
776	my $body = undef;
777	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
778
779	$_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) });
780
781	my $read_chunk; $read_chunk = sub {
782	$_[1] =~ /^([0-9a-fA-F]+)/
783	or $finish->(undef, 599 => "Garbled chunked transfer encoding");
784
785	my $len = hex $1;
786
787	if ($len) {
788	$cl += $len;
789
790	$_[0]->push_read (chunk => $len, sub {
791	$on_body->($_[1], \%hdr)
792	or return $finish->(undef, 598 => "Request cancelled by on_body");
793
794	$_[0]->push_read (line => sub {
795	length $_[1]
796	and return $finish->(undef, 599 => "Garbled chunked transfer encoding");
797	$_[0]->push_read (line => $read_chunk);
798	});
799	});
800	} else {
801	$hdr{"content-length"} ||= $cl;
802
803	$_[0]->push_read (line => $qr_nlnl, sub {
804	if (length $_[1]) {
805	for ("$_[1]") {
806	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
807
808	my $hdr = parse_hdr
809	or return $finish->(undef, 599 => "Garbled response trailers");
810
811	%hdr = (%hdr, %$hdr);
812	}
813	}
814
815	$finish->($body, undef, undef, 1);
816	});
817	}
818	};
819
820	$_[0]->push_read (line => $read_chunk);
821
822	} elsif ($arg{on_body}) {
823	$_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) });
824
825	if ($len) {
826	$_[0]->on_read (sub {
827	$len -= length $_[0]{rbuf};
828
829	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
830	or return $finish->(undef, 598 => "Request cancelled by on_body");
831
832	$len > 0
833	or $finish->("", undef, undef, 1);
834	});
835	} else {
836	$_[0]->on_eof (sub {
837	$finish->("");
838	});
839	$_[0]->on_read (sub {
840	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
841	or $finish->(undef, 598 => "Request cancelled by on_body");
842	});
843	}
844	} else {
845	$_[0]->on_eof (undef);
846
847	if ($len) {
848	$_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) });
849	$_[0]->on_read (sub {
850	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
851	if $len <= length $_[0]{rbuf};
852	});
853	} else {
854	$_[0]->on_error (sub {
855	($! == Errno::EPIPE || !$!)
856	? $finish->(delete $_[0]{rbuf})
857	: $finish->(undef, 599 => $_[2]);
858	});
859	$_[0]->on_read (sub { });
860	}
861	}
862	}
863	};
864
865	$state{handle}->push_read (line => $qr_nlnl, $state{read_response});
866	};
867
868	# now handle proxy-CONNECT method
869	if ($proxy && $uscheme eq "https") {
870	# oh dear, we have to wrap it into a connect request
871
872	# maybe re-use $uauthority with patched port?
873	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012");
874	$state{handle}->push_read (line => $qr_nlnl, sub {
875	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
876	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }));
877
878	if ($2 == 200) {
879	$rpath = $upath;
880	&$handle_actual_request;	1186	$handle_actual_request->();
881	} else {	1187
882	%state = ();
883	$cb->(undef, { @pseudo, Status => $2, Reason => $3 });
884	}
885	});
886	} else {	1188	} else {
887	&$handle_actual_request;
888	}
889	};
890
891	my $tcp_connect = $arg{tcp_connect}	1189	my $tcp_connect = $arg{tcp_connect}
892	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };	1190	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
893		1191
894	$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 });
895		1193	}
896	};	1194	};
897		1195
898	defined wantarray && AnyEvent::Util::guard { %state = () }	1196	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
899	}	1197	}
900		1198
901	sub http_get($@) {	1199	sub http_get($@) {
902	unshift @_, "GET";	1200	unshift @_, "GET";
903	&http_request	1201	&http_request
…		…
921	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for	1219	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
922	the actual connection, which in turn uses AnyEvent::DNS to resolve	1220	the actual connection, which in turn uses AnyEvent::DNS to resolve
923	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
924	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
925	your own default resolver (by storing a suitable resolver object in	1223	your own default resolver (by storing a suitable resolver object in
926	C<$AnyEvent::DNS::RESOLVER>).	1224	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
927		1225
928	=head2 GLOBAL FUNCTIONS AND VARIABLES	1226	=head2 GLOBAL FUNCTIONS AND VARIABLES
929		1227
930	=over 4	1228	=over 4
931		1229
932	=item AnyEvent::HTTP::set_proxy "proxy-url"	1230	=item AnyEvent::HTTP::set_proxy "proxy-url"
933		1231
934	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
935	string of the form C<http://host:port> (optionally C<https:...>), croaks	1233	string of the form C<http://host:port>, croaks otherwise.
936	otherwise.
937		1234
938	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 additonally 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	}
939		1280
940	=item $date = AnyEvent::HTTP::format_date $timestamp	1281	=item $date = AnyEvent::HTTP::format_date $timestamp
941		1282
942	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
943	Date (RFC 2616).	1284	Date (RFC 2616).
944		1285
945	=item $timestamp = AnyEvent::HTTP::parse_date $date	1286	=item $timestamp = AnyEvent::HTTP::parse_date $date
946		1287
947	Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) and	1288	Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) or a
948	returns the corresponding POSIX timestamp, or C<undef> if the date cannot	1289	bunch of minor variations of those, and returns the corresponding POSIX
949	be parsed.	1290	timestamp, or C<undef> if the date cannot be parsed.
950		1291
951	=item $AnyEvent::HTTP::MAX_RECURSE	1292	=item $AnyEvent::HTTP::MAX_RECURSE
952		1293
953	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>).
		1295
		1296	=item $AnyEvent::HTTP::TIMEOUT
		1297
		1298	The default timeout for connection operations (default: C<300>).
954		1299
955	=item $AnyEvent::HTTP::USERAGENT	1300	=item $AnyEvent::HTTP::USERAGENT
956		1301
957	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
958	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)>).
959		1304
960	=item $AnyEvent::HTTP::MAX_PER_HOST	1305	=item $AnyEvent::HTTP::MAX_PER_HOST
961		1306
962	The maximum number of concurrent connections to the same host (identified	1307	The maximum number of concurrent connections to the same host (identified
963	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
964	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.
965		1311
966	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
967	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, newers (such as firefox 3) typically
		1317	use 6, and Opera uses 8 because like, they have the fastest browser and
		1318	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>).
968		1324
969	=item $AnyEvent::HTTP::ACTIVE	1325	=item $AnyEvent::HTTP::ACTIVE
970		1326
971	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
972	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
973	connections. This number of can be useful for load-leveling.	1329	connections. This number can be useful for load-leveling.
974		1330
975	=back	1331	=back
976		1332
977	=cut	1333	=cut
978		1334
…		…
993	sub parse_date($) {	1349	sub parse_date($) {
994	my ($date) = @_;	1350	my ($date) = @_;
995		1351
996	my ($d, $m, $y, $H, $M, $S);	1352	my ($d, $m, $y, $H, $M, $S);
997		1353
998	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$/) {	1354	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$/) {
999	# RFC 822/1123, required by RFC 2616 (with " ")	1355	# RFC 822/1123, required by RFC 2616 (with " ")
1000	# cookie dates (with "-")	1356	# cookie dates (with "-")
1001		1357
1002	($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6);	1358	($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6);
1003		1359
1004	} 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$/) {	1360	} 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$/) {
1005	# RFC 850	1361	# RFC 850
1006	($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6);	1362	($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6);
1007		1363
1008	} 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])$/) {	1364	} 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])$/) {
1009	# ISO C's asctime	1365	# ISO C's asctime
1010	($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5);	1366	($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5);
1011	}	1367	}
1012	# other formats fail in the loop below	1368	# other formats fail in the loop below
1013		1369
1014	for (0..11) {	1370	for (0..11) {
1015	if ($m eq $month[$_]) {	1371	if ($m eq $month[$_]) {
1016	require Time::Local;	1372	require Time::Local;
1017	return Time::Local::timegm ($S, $M, $H, $d, $_, $y);	1373	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
1018	}	1374	}
1019	}	1375	}
1020		1376
1021	undef	1377	undef
1022	}	1378	}
1023		1379
1024	sub set_proxy($) {	1380	sub set_proxy($) {
1025	if (length $_[0]) {	1381	if (length $_[0]) {
1026	$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix	1382	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
1027	or Carp::croak "$_[0]: invalid proxy URL";	1383	or Carp::croak "$_[0]: invalid proxy URL";
1028	$PROXY = [$2, $3 || 3128, $1]	1384	$PROXY = [$2, $3 || 3128, $1]
1029	} else {	1385	} else {
1030	undef $PROXY;	1386	undef $PROXY;
1031	}	1387	}
…		…
1034	# initialise proxy from environment	1390	# initialise proxy from environment
1035	eval {	1391	eval {
1036	set_proxy $ENV{http_proxy};	1392	set_proxy $ENV{http_proxy};
1037	};	1393	};
1038		1394
		1395	=head2 SHOWCASE
		1396
		1397	This section contaisn 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 occured, 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
1039	=head2 SOCKS PROXIES	1497	=head3 SOCKS PROXIES
1040		1498
1041	Socks proxies are not directly supported by AnyEvent::HTTP. You can	1499	Socks proxies are not directly supported by AnyEvent::HTTP. You can
1042	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
1043	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
1044	transparently.	1502	transparently.

Diff Legend

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