ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.79 by root, Sat Jan 1 20:01:07 2011 UTC vs.
Revision 1.119 by root, Sun Jun 8 23:33:28 2014 UTC

…		…
15	This module is an L<AnyEvent> user, you need to make sure that you use and	15	This module is an L<AnyEvent> user, you need to make sure that you use and
16	run a supported event loop.	16	run a supported event loop.
17		17
18	This module implements a simple, stateless and non-blocking HTTP	18	This module implements a simple, stateless and non-blocking HTTP
19	client. It supports GET, POST and other request methods, cookies and more,	19	client. It supports GET, POST and other request methods, cookies and more,
20	all on a very low level. It can follow redirects supports proxies and	20	all on a very low level. It can follow redirects, supports proxies, and
21	automatically limits the number of connections to the values specified in	21	automatically limits the number of connections to the values specified in
22	the RFC.	22	the RFC.
23		23
24	It should generally be a "good client" that is enough for most HTTP	24	It should generally be a "good client" that is enough for most HTTP
25	tasks. Simple tasks should be simple, but complex tasks should still be	25	tasks. Simple tasks should be simple, but complex tasks should still be
…		…
36		36
37	=cut	37	=cut
38		38
39	package AnyEvent::HTTP;	39	package AnyEvent::HTTP;
40		40
41	use strict;	41	use common::sense;
42	no warnings;
43		42
44	use Errno ();	43	use Errno ();
45		44
46	use AnyEvent 5.0 ();	45	use AnyEvent 5.0 ();
47	use AnyEvent::Util ();	46	use AnyEvent::Util ();
48	use AnyEvent::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 = 2.2;
53		52
54	our @EXPORT = qw(http_get http_post http_head http_request);	53	our @EXPORT = qw(http_get http_post http_head http_request);
55		54
56	our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)";	55	our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)";
57	our $MAX_RECURSE = 10;	56	our $MAX_RECURSE = 10;
58	our $MAX_PERSISTENT = 8;
59	our $PERSISTENT_TIMEOUT = 2;	57	our $PERSISTENT_TIMEOUT = 3;
60	our $TIMEOUT = 300;	58	our $TIMEOUT = 300;
61		59	our $MAX_PER_HOST = 4; # changing this is evil
62	# changing these is evil
63	our $MAX_PERSISTENT_PER_HOST = 0;
64	our $MAX_PER_HOST = 4;
65		60
66	our $PROXY;	61	our $PROXY;
67	our $ACTIVE = 0;	62	our $ACTIVE = 0;
68		63
69	my %KA_COUNT; # number of open keep-alive connections per host	64	my %KA_CACHE; # indexed by uhost currently, points to [$handle...] array
70	my %CO_SLOT; # number of open connections, and wait queue, per host	65	my %CO_SLOT; # number of open connections, and wait queue, per host
71		66
72	=item http_get $url, key => value..., $cb->($data, $headers)	67	=item http_get $url, key => value..., $cb->($data, $headers)
73		68
74	Executes an HTTP-GET request. See the http_request function for details on	69	Executes an HTTP-GET request. See the http_request function for details on
…		…
94	C<http_request> returns a "cancellation guard" - you have to keep the	89	C<http_request> returns a "cancellation guard" - you have to keep the
95	object at least alive until the callback get called. If the object gets	90	object at least alive until the callback get called. If the object gets
96	destroyed before the callback is called, the request will be cancelled.	91	destroyed before the callback is called, the request will be cancelled.
97		92
98	The callback will be called with the response body data as first argument	93	The callback will be called with the response body data as first argument
99	(or C<undef> if an error occured), and a hash-ref with response headers	94	(or C<undef> if an error occurred), and a hash-ref with response headers
100	(and trailers) as second argument.	95	(and trailers) as second argument.
101		96
102	All the headers in that hash are lowercased. In addition to the response	97	All the headers in that hash are lowercased. In addition to the response
103	headers, the "pseudo-headers" (uppercase to avoid clashing with possible	98	headers, the "pseudo-headers" (uppercase to avoid clashing with possible
104	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the	99	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the
…		…
128	C<590>-C<599> and the C<Reason> pseudo-header will contain an error	123	C<590>-C<599> and the C<Reason> pseudo-header will contain an error
129	message. Currently the following status codes are used:	124	message. Currently the following status codes are used:
130		125
131	=over 4	126	=over 4
132		127
133	=item 595 - errors during connection etsbalishment, proxy handshake.	128	=item 595 - errors during connection establishment, proxy handshake.
134		129
135	=item 596 - errors during TLS negotiation, request sending and header processing.	130	=item 596 - errors during TLS negotiation, request sending and header processing.
136		131
137	=item 597 - errors during body receiving or processing.	132	=item 597 - errors during body receiving or processing.
138		133
…		…
159		154
160	=over 4	155	=over 4
161		156
162	=item recurse => $count (default: $MAX_RECURSE)	157	=item recurse => $count (default: $MAX_RECURSE)
163		158
164	Whether to recurse requests or not, e.g. on redirects, authentication	159	Whether to recurse requests or not, e.g. on redirects, authentication and
165	retries and so on, and how often to do so.	160	other retries and so on, and how often to do so.
		161
		162	Only redirects to http and https URLs are supported. While most common
		163	redirection forms are handled entirely within this module, some require
		164	the use of the optional L<URI> module. If it is required but missing, then
		165	the request will fail with an error.
166		166
167	=item headers => hashref	167	=item headers => hashref
168		168
169	The request headers to use. Currently, C<http_request> may provide its own	169	The request headers to use. Currently, C<http_request> may provide its own
170	C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and	170	C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and
171	will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:>	171	will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:>
172	(this can be suppressed by using C<undef> for these headers in which case	172	(this can be suppressed by using C<undef> for these headers in which case
173	they won't be sent at all).	173	they won't be sent at all).
174		174
		175	You really should provide your own C<User-Agent:> header value that is
		176	appropriate for your program - I wouldn't be surprised if the default
		177	AnyEvent string gets blocked by webservers sooner or later.
		178
		179	Also, make sure that your headers names and values do not contain any
		180	embedded newlines.
		181
175	=item timeout => $seconds	182	=item timeout => $seconds
176		183
177	The time-out to use for various stages - each connect attempt will reset	184	The time-out to use for various stages - each connect attempt will reset
178	the timeout, as will read or write activity, i.e. this is not an overall	185	the timeout, as will read or write activity, i.e. this is not an overall
179	timeout.	186	timeout.
180		187
181	Default timeout is 5 minutes.	188	Default timeout is 5 minutes.
182		189
183	=item proxy => [$host, $port[, $scheme]] or undef	190	=item proxy => [$host, $port[, $scheme]] or undef
184		191
185	Use the given http proxy for all requests. If not specified, then the	192	Use the given http proxy for all requests, or no proxy if C<undef> is
186	default proxy (as specified by C<$ENV{http_proxy}>) is used.	193	used.
187		194
188	C<$scheme> must be either missing, C<http> for HTTP or C<https> for	195	C<$scheme> must be either missing or must be C<http> for HTTP.
189	HTTPS.	196
		197	If not specified, then the default proxy is used (see
		198	C<AnyEvent::HTTP::set_proxy>).
190		199
191	=item body => $string	200	=item body => $string
192		201
193	The request body, usually empty. Will be sent as-is (future versions of	202	The request body, usually empty. Will be sent as-is (future versions of
194	this module might offer more options).	203	this module might offer more options).
…		…
196	=item cookie_jar => $hash_ref	205	=item cookie_jar => $hash_ref
197		206
198	Passing this parameter enables (simplified) cookie-processing, loosely	207	Passing this parameter enables (simplified) cookie-processing, loosely
199	based on the original netscape specification.	208	based on the original netscape specification.
200		209
201	The C<$hash_ref> must be an (initially empty) hash reference which will	210	The C<$hash_ref> must be an (initially empty) hash reference which
202	get updated automatically. It is possible to save the cookie jar to	211	will get updated automatically. It is possible to save the cookie jar
203	persistent storage with something like JSON or Storable, but this is not	212	to persistent storage with something like JSON or Storable - see the
204	recommended, as session-only cookies might survive longer than expected.	213	C<AnyEvent::HTTP::cookie_jar_expire> function if you wish to remove
		214	expired or session-only cookies, and also for documentation on the format
		215	of the cookie jar.
205		216
206	Note that this cookie implementation is not meant to be complete. If	217	Note that this cookie implementation is not meant to be complete. If
207	you want complete cookie management you have to do that on your	218	you want complete cookie management you have to do that on your
208	own. C<cookie_jar> is meant as a quick fix to get some cookie-using sites	219	own. C<cookie_jar> is meant as a quick fix to get most cookie-using sites
209	working. Cookies are a privacy disaster, do not use them unless required	220	working. Cookies are a privacy disaster, do not use them unless required
210	to.	221	to.
211		222
212	When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:>	223	When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:>
213	headers will be set and handled by this module, otherwise they will be	224	headers will be set and handled by this module, otherwise they will be
…		…
223	verification) TLS context.	234	verification) TLS context.
224		235
225	The default for this option is C<low>, which could be interpreted as "give	236	The default for this option is C<low>, which could be interpreted as "give
226	me the page, no matter what".	237	me the page, no matter what".
227		238
		239	See also the C<sessionid> parameter.
		240
		241	=item session => $string
		242
		243	The module might reuse connections to the same host internally. Sometimes
		244	(e.g. when using TLS), you do not want to reuse connections from other
		245	sessions. This can be achieved by setting this parameter to some unique
		246	ID (such as the address of an object storing your state data, or the TLS
		247	context) - only connections using the same unique ID will be reused.
		248
228	=item on_prepare => $callback->($fh)	249	=item on_prepare => $callback->($fh)
229		250
230	In rare cases you need to "tune" the socket before it is used to	251	In rare cases you need to "tune" the socket before it is used to
231	connect (for exmaple, to bind it on a given IP address). This parameter	252	connect (for example, to bind it on a given IP address). This parameter
232	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>	253	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>
233	and behaves exactly the same way (e.g. it has to provide a	254	and behaves exactly the same way (e.g. it has to provide a
234	timeout). See the description for the C<$prepare_cb> argument of	255	timeout). See the description for the C<$prepare_cb> argument of
235	C<AnyEvent::Socket::tcp_connect> for details.	256	C<AnyEvent::Socket::tcp_connect> for details.
236		257
…		…
301	called. Instead of the C<$body> argument containing the body data, the	322	called. Instead of the C<$body> argument containing the body data, the
302	callback will receive the L<AnyEvent::Handle> object associated with the	323	callback will receive the L<AnyEvent::Handle> object associated with the
303	connection. In error cases, C<undef> will be passed. When there is no body	324	connection. In error cases, C<undef> will be passed. When there is no body
304	(e.g. status C<304>), the empty string will be passed.	325	(e.g. status C<304>), the empty string will be passed.
305		326
306	The handle object might or might not be in TLS mode, might be connected to	327	The handle object might or might not be in TLS mode, might be connected
307	a proxy, be a persistent connection etc., and configured in unspecified	328	to a proxy, be a persistent connection, use chunked transfer encoding
308	ways. The user is responsible for this handle (it will not be used by this	329	etc., and configured in unspecified ways. The user is responsible for this
309	module anymore).	330	handle (it will not be used by this module anymore).
310		331
311	This is useful with some push-type services, where, after the initial	332	This is useful with some push-type services, where, after the initial
312	headers, an interactive protocol is used (typical example would be the	333	headers, an interactive protocol is used (typical example would be the
313	push-style twitter API which starts a JSON/XML stream).	334	push-style twitter API which starts a JSON/XML stream).
314		335
315	If you think you need this, first have a look at C<on_body>, to see if	336	If you think you need this, first have a look at C<on_body>, to see if
316	that doesn't solve your problem in a better way.	337	that doesn't solve your problem in a better way.
		338
		339	=item persistent => $boolean
		340
		341	Try to create/reuse a persistent connection. When this flag is set
		342	(default: true for idempotent requests, false for all others), then
		343	C<http_request> tries to re-use an existing (previously-created)
		344	persistent connection to the host and, failing that, tries to create a new
		345	one.
		346
		347	Requests failing in certain ways will be automatically retried once, which
		348	is dangerous for non-idempotent requests, which is why it defaults to off
		349	for them. The reason for this is because the bozos who designed HTTP/1.1
		350	made it impossible to distinguish between a fatal error and a normal
		351	connection timeout, so you never know whether there was a problem with
		352	your request or not.
		353
		354	When reusing an existent connection, many parameters (such as TLS context)
		355	will be ignored. See the C<session> parameter for a workaround.
		356
		357	=item keepalive => $boolean
		358
		359	Only used when C<persistent> is also true. This parameter decides whether
		360	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		361	(as opposed to only a HTTP/1.1 persistent connection).
		362
		363	The default is true, except when using a proxy, in which case it defaults
		364	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		365
		366	=item handle_params => { key => value ... }
		367
		368	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		369	constructor that is called - not all requests will create a handle, and
		370	sometimes more than one is created, so this parameter is only good for
		371	setting hints.
		372
		373	Example: set the maximum read size to 4096, to potentially conserve memory
		374	at the cost of speed.
		375
		376	handle_params => {
		377	max_read_size => 4096,
		378	},
317		379
318	=back	380	=back
319		381
320	Example: do a simple HTTP GET request for http://www.nethype.de/ and print	382	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
321	the response body.	383	the response body.
…		…
327		389
328	Example: do a HTTP HEAD request on https://www.google.com/, use a	390	Example: do a HTTP HEAD request on https://www.google.com/, use a
329	timeout of 30 seconds.	391	timeout of 30 seconds.
330		392
331	http_request	393	http_request
332	GET => "https://www.google.com",	394	HEAD => "https://www.google.com",
		395	headers => { "user-agent" => "MySearchClient 1.0" },
333	timeout => 30,	396	timeout => 30,
334	sub {	397	sub {
335	my ($body, $hdr) = @_;	398	my ($body, $hdr) = @_;
336	use Data::Dumper;	399	use Data::Dumper;
337	print Dumper $hdr;	400	print Dumper $hdr;
…		…
347	};	410	};
348		411
349	undef $request;	412	undef $request;
350		413
351	=cut	414	=cut
		415
		416	#############################################################################
		417	# wait queue/slots
352		418
353	sub _slot_schedule;	419	sub _slot_schedule;
354	sub _slot_schedule($) {	420	sub _slot_schedule($) {
355	my $host = shift;	421	my $host = shift;
356		422
…		…
378	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];	444	push @{ $CO_SLOT{$_[0]}[1] }, $_[1];
379		445
380	_slot_schedule $_[0];	446	_slot_schedule $_[0];
381	}	447	}
382		448
		449	#############################################################################
		450	# cookie handling
		451
		452	# expire cookies
		453	sub cookie_jar_expire($;$) {
		454	my ($jar, $session_end) = @_;
		455
		456	%$jar = () if $jar->{version} != 1;
		457
		458	my $anow = AE::now;
		459
		460	while (my ($chost, $paths) = each %$jar) {
		461	next unless ref $paths;
		462
		463	while (my ($cpath, $cookies) = each %$paths) {
		464	while (my ($cookie, $kv) = each %$cookies) {
		465	if (exists $kv->{_expires}) {
		466	delete $cookies->{$cookie}
		467	if $anow > $kv->{_expires};
		468	} elsif ($session_end) {
		469	delete $cookies->{$cookie};
		470	}
		471	}
		472
		473	delete $paths->{$cpath}
		474	unless %$cookies;
		475	}
		476
		477	delete $jar->{$chost}
		478	unless %$paths;
		479	}
		480	}
		481
383	# extract cookies from jar	482	# extract cookies from jar
384	sub cookie_jar_extract($$$$) {	483	sub cookie_jar_extract($$$$) {
385	my ($jar, $uscheme, $uhost, $upath) = @_;	484	my ($jar, $scheme, $host, $path) = @_;
386		485
387	%$jar = () if $jar->{version} != 1;	486	%$jar = () if $jar->{version} != 1;
388		487
389	my @cookies;	488	my @cookies;
390		489
391	while (my ($chost, $paths) = each %$jar) {	490	while (my ($chost, $paths) = each %$jar) {
392	next unless ref $paths;	491	next unless ref $paths;
393		492
394	if ($chost =~ /^\./) {	493	if ($chost =~ /^\./) {
395	next unless $chost eq substr $uhost, -length $chost;	494	next unless $chost eq substr $host, -length $chost;
396	} elsif ($chost =~ /\./) {	495	} elsif ($chost =~ /\./) {
397	next unless $chost eq $uhost;	496	next unless $chost eq $host;
398	} else {	497	} else {
399	next;	498	next;
400	}	499	}
401		500
402	while (my ($cpath, $cookies) = each %$paths) {	501	while (my ($cpath, $cookies) = each %$paths) {
403	next unless $cpath eq substr $upath, 0, length $cpath;	502	next unless $cpath eq substr $path, 0, length $cpath;
404		503
405	while (my ($cookie, $kv) = each %$cookies) {	504	while (my ($cookie, $kv) = each %$cookies) {
406	next if $uscheme ne "https" && exists $kv->{secure};	505	next if $scheme ne "https" && exists $kv->{secure};
407		506
408	if (exists $kv->{expires}) {	507	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
409	if (AE::now > parse_date ($kv->{expires})) {
410	delete $cookies->{$cookie};	508	delete $cookies->{$cookie};
411	next;	509	next;
412	}
413	}	510	}
414		511
415	my $value = $kv->{value};	512	my $value = $kv->{value};
416		513
417	if ($value =~ /[=;,[:space:]]/) {	514	if ($value =~ /[=;,[:space:]]/) {
…		…
426		523
427	\@cookies	524	\@cookies
428	}	525	}
429		526
430	# parse set_cookie header into jar	527	# parse set_cookie header into jar
431	sub cookie_jar_set_cookie($$$) {	528	sub cookie_jar_set_cookie($$$$) {
432	my ($jar, $set_cookie, $uhost) = @_;	529	my ($jar, $set_cookie, $host, $date) = @_;
		530
		531	my $anow = int AE::now;
		532	my $snow; # server-now
433		533
434	for ($set_cookie) {	534	for ($set_cookie) {
435	# parse NAME=VALUE	535	# parse NAME=VALUE
436	my @kv;	536	my @kv;
437		537
…		…
440	while (	540	while (
441	m{	541	m{
442	\G\s*	542	\G\s*
443	(?:	543	(?:
444	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)	544	expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+)
445	| ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )	545	| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )?
446	)	546	)
447	}gcxsi	547	}gcxsi
448	) {	548	) {
449	my $name = $2;	549	my $name = $2;
450	my $value = $4;	550	my $value = $4;
451		551
452	unless (defined $name) {	552	if (defined $1) {
453	# expires	553	# expires
454	$name = "expires";	554	$name = "expires";
455	$value = $1;	555	$value = $1;
456	} elsif (!defined $value) {	556	} elsif (defined $3) {
457	# quoted	557	# quoted
458	$value = $3;	558	$value = $3;
459	$value =~ s/\\(.)/$1/gs;	559	$value =~ s/\\(.)/$1/gs;
460	}	560	}
461		561
462	push @kv, lc $name, $value;	562	push @kv, @kv ? lc $name : $name, $value;
463		563
464	last unless /\G\s*;/gc;	564	last unless /\G\s*;/gc;
465	}	565	}
466		566
467	last unless @kv;	567	last unless @kv;
468		568
469	my $name = shift @kv;	569	my $name = shift @kv;
470	my %kv = (value => shift @kv, @kv);	570	my %kv = (value => shift @kv, @kv);
471		571
472	$kv{expires} ||= format_date (AE::now + $kv{"max-age"})
473	if exists $kv{"max-age"};	572	if (exists $kv{"max-age"}) {
		573	$kv{_expires} = $anow + delete $kv{"max-age"};
		574	} elsif (exists $kv{expires}) {
		575	$snow ||= parse_date ($date) || $anow;
		576	$kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow);
		577	} else {
		578	delete $kv{_expires};
		579	}
474		580
475	my $cdom;	581	my $cdom;
476	my $cpath = (delete $kv{path}) || "/";	582	my $cpath = (delete $kv{path}) || "/";
477		583
478	if (exists $kv{domain}) {	584	if (exists $kv{domain}) {
…		…
484		590
485	# this is not rfc-like and not netscape-like. go figure.	591	# this is not rfc-like and not netscape-like. go figure.
486	my $ndots = $cdom =~ y/.//;	592	my $ndots = $cdom =~ y/.//;
487	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);	593	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
488	} else {	594	} else {
489	$cdom = $uhost;	595	$cdom = $host;
490	}	596	}
491		597
492	# store it	598	# store it
493	$jar->{version} = 1;	599	$jar->{version} = 1;
494	$jar->{$cdom}{$cpath}{$name} = \%kv;	600	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
495		601
496	redo if /\G\s*,/gc;	602	redo if /\G\s*,/gc;
497	}	603	}
498	}	604	}
499		605
		606	#############################################################################
		607	# keepalive/persistent connection cache
		608
		609	# fetch a connection from the keepalive cache
		610	sub ka_fetch($) {
		611	my $ka_key = shift;
		612
		613	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		614	delete $KA_CACHE{$ka_key}
		615	unless @{ $KA_CACHE{$ka_key} };
		616
		617	$hdl
		618	}
		619
		620	sub ka_store($$) {
		621	my ($ka_key, $hdl) = @_;
		622
		623	my $kaa = $KA_CACHE{$ka_key} ||= [];
		624
		625	my $destroy = sub {
		626	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		627
		628	$hdl->destroy;
		629
		630	@ka
		631	? $KA_CACHE{$ka_key} = \@ka
		632	: delete $KA_CACHE{$ka_key};
		633	};
		634
		635	# on error etc., destroy
		636	$hdl->on_error ($destroy);
		637	$hdl->on_eof ($destroy);
		638	$hdl->on_read ($destroy);
		639	$hdl->timeout ($PERSISTENT_TIMEOUT);
		640
		641	push @$kaa, $hdl;
		642	shift @$kaa while @$kaa > $MAX_PER_HOST;
		643	}
		644
		645	#############################################################################
		646	# utilities
		647
500	# continue to parse $_ for headers and place them into the arg	648	# continue to parse $_ for headers and place them into the arg
501	sub parse_hdr() {	649	sub _parse_hdr() {
502	my %hdr;	650	my %hdr;
503		651
504	# things seen, not parsed:	652	# things seen, not parsed:
505	# p3pP="NON CUR OTPi OUR NOR UNI"	653	# p3pP="NON CUR OTPi OUR NOR UNI"
506		654
…		…
520	for values %hdr;	668	for values %hdr;
521		669
522	\%hdr	670	\%hdr
523	}	671	}
524		672
		673	#############################################################################
		674	# http_get
		675
525	our $qr_nlnl = qr{(?<![^\012])\015?\012};	676	our $qr_nlnl = qr{(?<![^\012])\015?\012};
526		677
527	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	678	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
528	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };	679	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };
		680
		681	# maybe it should just become a normal object :/
		682
		683	sub _destroy_state(\%) {
		684	my ($state) = @_;
		685
		686	$state->{handle}->destroy if $state->{handle};
		687	%$state = ();
		688	}
		689
		690	sub _error(\%$$) {
		691	my ($state, $cb, $hdr) = @_;
		692
		693	&_destroy_state ($state);
		694
		695	$cb->(undef, $hdr);
		696	()
		697	}
		698
		699	our %IDEMPOTENT = (
		700	DELETE => 1,
		701	GET => 1,
		702	HEAD => 1,
		703	OPTIONS => 1,
		704	PUT => 1,
		705	TRACE => 1,
		706
		707	ACL => 1,
		708	"BASELINE-CONTROL" => 1,
		709	BIND => 1,
		710	CHECKIN => 1,
		711	CHECKOUT => 1,
		712	COPY => 1,
		713	LABEL => 1,
		714	LINK => 1,
		715	MERGE => 1,
		716	MKACTIVITY => 1,
		717	MKCALENDAR => 1,
		718	MKCOL => 1,
		719	MKREDIRECTREF => 1,
		720	MKWORKSPACE => 1,
		721	MOVE => 1,
		722	ORDERPATCH => 1,
		723	PROPFIND => 1,
		724	PROPPATCH => 1,
		725	REBIND => 1,
		726	REPORT => 1,
		727	SEARCH => 1,
		728	UNBIND => 1,
		729	UNCHECKOUT => 1,
		730	UNLINK => 1,
		731	UNLOCK => 1,
		732	UPDATE => 1,
		733	UPDATEREDIRECTREF => 1,
		734	"VERSION-CONTROL" => 1,
		735	);
529		736
530	sub http_request($$@) {	737	sub http_request($$@) {
531	my $cb = pop;	738	my $cb = pop;
532	my ($method, $url, %arg) = @_;	739	my ($method, $url, %arg) = @_;
533		740
…		…
551	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	758	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
552		759
553	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })	760	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
554	if $recurse < 0;	761	if $recurse < 0;
555		762
556	my $proxy = $arg{proxy} || $PROXY;	763	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
557	my $timeout = $arg{timeout} || $TIMEOUT;	764	my $timeout = $arg{timeout} || $TIMEOUT;
558		765
559	my ($uscheme, $uauthority, $upath, $query, $fragment) =	766	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
560	$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|;	767	$url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|;
561		768
562	$uscheme = lc $uscheme;	769	$uscheme = lc $uscheme;
563		770
564	my $uport = $uscheme eq "http" ? 80	771	my $uport = $uscheme eq "http" ? 80
565	: $uscheme eq "https" ? 443	772	: $uscheme eq "https" ? 443
566	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });	773	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
567		774
568	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	775	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x
569	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });	776	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
570		777
571	my $uhost = $1;	778	my $uhost = lc $1;
572	$uport = $2 if defined $2;	779	$uport = $2 if defined $2;
573		780
574	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"	781	$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"
575	unless exists $hdr{host};	782	unless exists $hdr{host};
576		783
…		…
595	$rscheme = "http" unless defined $rscheme;	802	$rscheme = "http" unless defined $rscheme;
596		803
597	# don't support https requests over https-proxy transport,	804	# don't support https requests over https-proxy transport,
598	# can't be done with tls as spec'ed, unless you double-encrypt.	805	# can't be done with tls as spec'ed, unless you double-encrypt.
599	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";	806	$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https";
		807
		808	$rhost = lc $rhost;
		809	$rscheme = lc $rscheme;
600	} else {	810	} else {
601	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);	811	($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath);
602	}	812	}
603		813
604	# leave out fragment and query string, just a heuristic	814	# leave out fragment and query string, just a heuristic
…		…
606	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};	816	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
607		817
608	$hdr{"content-length"} = length $arg{body}	818	$hdr{"content-length"} = length $arg{body}
609	if length $arg{body} || $method ne "GET";	819	if length $arg{body} || $method ne "GET";
610		820
611	$hdr{connection} = "close TE"; #1.1	821	my $idempotent = $IDEMPOTENT{$method};
		822
		823	# default value for keepalive is true iff the request is for an idempotent method
		824	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
		825	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		826	my $was_persistent; # true if this is actually a recycled connection
		827
		828	# the key to use in the keepalive cache
		829	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
		830
		831	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive, " : "" : "close, ") . "Te"; #1.1
612	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1	832	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
613		833
614	my %state = (connect_guard => 1);	834	my %state = (connect_guard => 1);
		835
		836	my $ae_error = 595; # connecting
		837
		838	# handle actual, non-tunneled, request
		839	my $handle_actual_request = sub {
		840	$ae_error = 596; # request phase
		841
		842	my $hdl = $state{handle};
		843
		844	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
		845
		846	# send request
		847	$hdl->push_write (
		848	"$method $rpath HTTP/1.1\015\012"
		849	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
		850	. "\015\012"
		851	. (delete $arg{body})
		852	);
		853
		854	# return if error occurred during push_write()
		855	return unless %state;
		856
		857	# reduce memory usage, save a kitten, also re-use it for the response headers.
		858	%hdr = ();
		859
		860	# status line and headers
		861	$state{read_response} = sub {
		862	return unless %state;
		863
		864	for ("$_[1]") {
		865	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		866
		867	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
		868	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
		869
		870	# 100 Continue handling
		871	# should not happen as we don't send expect: 100-continue,
		872	# but we handle it just in case.
		873	# since we send the request body regardless, if we get an error
		874	# we are out of-sync, which we currently do NOT handle correctly.
		875	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
		876	if $2 eq 100;
		877
		878	push @pseudo,
		879	HTTPVersion => $1,
		880	Status => $2,
		881	Reason => $3,
		882	;
		883
		884	my $hdr = _parse_hdr
		885	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
		886
		887	%hdr = (%$hdr, @pseudo);
		888	}
		889
		890	# redirect handling
		891	# relative uri handling forced by microsoft and other shitheads.
		892	# we give our best and fall back to URI if available.
		893	if (exists $hdr{location}) {
		894	my $loc = $hdr{location};
		895
		896	if ($loc =~ m%^//%) { # //
		897	$loc = "$rscheme:$loc";
		898
		899	} elsif ($loc eq "") {
		900	$loc = $url;
		901
		902	} elsif ($loc !~ /^(?: $ | [^:\/?\#]+ : )/x) { # anything "simple"
		903	$loc =~ s/^\.\/+//;
		904
		905	if ($loc !~ m%^[.?#]%) {
		906	my $prefix = "$rscheme://$uhost:$uport";
		907
		908	unless ($loc =~ s/^\///) {
		909	$prefix .= $upath;
		910	$prefix =~ s/\/[^\/]*$//;
		911	}
		912
		913	$loc = "$prefix/$loc";
		914
		915	} elsif (eval { require URI }) { # uri
		916	$loc = URI->new_abs ($loc, $url)->as_string;
		917
		918	} else {
		919	return _error %state, $cb, { @pseudo, Status => 599, Reason => "Cannot parse Location (URI module missing)" };
		920	#$hdr{Status} = 599;
		921	#$hdr{Reason} = "Unparsable Redirect (URI module missing)";
		922	#$recurse = 0;
		923	}
		924	}
		925
		926	$hdr{location} = $loc;
		927	}
		928
		929	my $redirect;
		930
		931	if ($recurse) {
		932	my $status = $hdr{Status};
		933
		934	# industry standard is to redirect POST as GET for
		935	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
		936	# also, the UA should ask the user for 301 and 307 and POST,
		937	# industry standard seems to be to simply follow.
		938	# we go with the industry standard. 308 is defined
		939	# by rfc7238
		940	if ($status == 301 or $status == 302 or $status == 303) {
		941	# HTTP/1.1 is unclear on how to mutate the method
		942	$method = "GET" unless $method eq "HEAD";
		943	$redirect = 1;
		944	} elsif ($status == 307 or $status == 308) {
		945	$redirect = 1;
		946	}
		947	}
		948
		949	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
		950	if ($state{handle}) {
		951	# handle keepalive
		952	if (
		953	$persistent
		954	&& $_[3]
		955	&& ($hdr{HTTPVersion} < 1.1
		956	? $hdr{connection} =~ /\bkeep-?alive\b/i
		957	: $hdr{connection} !~ /\bclose\b/i)
		958	) {
		959	ka_store $ka_key, delete $state{handle};
		960	} else {
		961	# no keepalive, destroy the handle
		962	$state{handle}->destroy;
		963	}
		964	}
		965
		966	%state = ();
		967
		968	if (defined $_[1]) {
		969	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		970	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		971	}
		972
		973	# set-cookie processing
		974	if ($arg{cookie_jar}) {
		975	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		976	}
		977
		978	if ($redirect && exists $hdr{location}) {
		979	# we ignore any errors, as it is very common to receive
		980	# Content-Length != 0 but no actual body
		981	# we also access %hdr, as $_[1] might be an erro
		982	$state{recurse} =
		983	http_request (
		984	$method => $hdr{location},
		985	%arg,
		986	recurse => $recurse - 1,
		987	Redirect => [$_[0], \%hdr],
		988	sub {
		989	%state = ();
		990	&$cb
		991	},
		992	);
		993	} else {
		994	$cb->($_[0], \%hdr);
		995	}
		996	};
		997
		998	$ae_error = 597; # body phase
		999
		1000	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		1001
		1002	my $len = $chunked ? undef : $hdr{"content-length"};
		1003
		1004	# body handling, many different code paths
		1005	# - no body expected
		1006	# - want_body_handle
		1007	# - te chunked
		1008	# - 2x length known (with or without on_body)
		1009	# - 2x length not known (with or without on_body)
		1010	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		1011	$finish->(undef, 598 => "Request cancelled by on_header");
		1012	} elsif (
		1013	$hdr{Status} =~ /^(?:1..|204|205|304)$/
		1014	or $method eq "HEAD"
		1015	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		1016	) {
		1017	# no body
		1018	$finish->("", undef, undef, 1);
		1019
		1020	} elsif (!$redirect && $arg{want_body_handle}) {
		1021	$_[0]->on_eof (undef);
		1022	$_[0]->on_error (undef);
		1023	$_[0]->on_read (undef);
		1024
		1025	$finish->(delete $state{handle});
		1026
		1027	} elsif ($chunked) {
		1028	my $cl = 0;
		1029	my $body = "";
		1030	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
		1031
		1032	$state{read_chunk} = sub {
		1033	$_[1] =~ /^([0-9a-fA-F]+)/
		1034	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1035
		1036	my $len = hex $1;
		1037
		1038	if ($len) {
		1039	$cl += $len;
		1040
		1041	$_[0]->push_read (chunk => $len, sub {
		1042	$on_body->($_[1], \%hdr)
		1043	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1044
		1045	$_[0]->push_read (line => sub {
		1046	length $_[1]
		1047	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1048	$_[0]->push_read (line => $state{read_chunk});
		1049	});
		1050	});
		1051	} else {
		1052	$hdr{"content-length"} ||= $cl;
		1053
		1054	$_[0]->push_read (line => $qr_nlnl, sub {
		1055	if (length $_[1]) {
		1056	for ("$_[1]") {
		1057	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		1058
		1059	my $hdr = _parse_hdr
		1060	or return $finish->(undef, $ae_error => "Garbled response trailers");
		1061
		1062	%hdr = (%hdr, %$hdr);
		1063	}
		1064	}
		1065
		1066	$finish->($body, undef, undef, 1);
		1067	});
		1068	}
		1069	};
		1070
		1071	$_[0]->push_read (line => $state{read_chunk});
		1072
		1073	} elsif ($arg{on_body}) {
		1074	if (defined $len) {
		1075	$_[0]->on_read (sub {
		1076	$len -= length $_[0]{rbuf};
		1077
		1078	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1079	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1080
		1081	$len > 0
		1082	or $finish->("", undef, undef, 1);
		1083	});
		1084	} else {
		1085	$_[0]->on_eof (sub {
		1086	$finish->("");
		1087	});
		1088	$_[0]->on_read (sub {
		1089	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
		1090	or $finish->(undef, 598 => "Request cancelled by on_body");
		1091	});
		1092	}
		1093	} else {
		1094	$_[0]->on_eof (undef);
		1095
		1096	if (defined $len) {
		1097	$_[0]->on_read (sub {
		1098	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1099	if $len <= length $_[0]{rbuf};
		1100	});
		1101	} else {
		1102	$_[0]->on_error (sub {
		1103	($! == Errno::EPIPE || !$!)
		1104	? $finish->(delete $_[0]{rbuf})
		1105	: $finish->(undef, $ae_error => $_[2]);
		1106	});
		1107	$_[0]->on_read (sub { });
		1108	}
		1109	}
		1110	};
		1111
		1112	# if keepalive is enabled, then the server closing the connection
		1113	# before a response can happen legally - we retry on idempotent methods.
		1114	if ($was_persistent && $idempotent) {
		1115	my $old_eof = $hdl->{on_eof};
		1116	$hdl->{on_eof} = sub {
		1117	_destroy_state %state;
		1118
		1119	%state = ();
		1120	$state{recurse} =
		1121	http_request (
		1122	$method => $url,
		1123	%arg,
		1124	recurse => $recurse - 1,
		1125	keepalive => 0,
		1126	sub {
		1127	%state = ();
		1128	&$cb
		1129	}
		1130	);
		1131	};
		1132	$hdl->on_read (sub {
		1133	return unless %state;
		1134
		1135	# as soon as we receive something, a connection close
		1136	# once more becomes a hard error
		1137	$hdl->{on_eof} = $old_eof;
		1138	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1139	});
		1140	} else {
		1141	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1142	}
		1143	};
		1144
		1145	my $prepare_handle = sub {
		1146	my ($hdl) = $state{handle};
		1147
		1148	$hdl->on_error (sub {
		1149	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1150	});
		1151	$hdl->on_eof (sub {
		1152	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1153	});
		1154	$hdl->timeout_reset;
		1155	$hdl->timeout ($timeout);
		1156	};
		1157
		1158	# connected to proxy (or origin server)
		1159	my $connect_cb = sub {
		1160	my $fh = shift
		1161	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1162
		1163	return unless delete $state{connect_guard};
		1164
		1165	# get handle
		1166	$state{handle} = new AnyEvent::Handle
		1167	%{ $arg{handle_params} },
		1168	fh => $fh,
		1169	peername => $uhost,
		1170	tls_ctx => $arg{tls_ctx},
		1171	;
		1172
		1173	$prepare_handle->();
		1174
		1175	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1176
		1177	# now handle proxy-CONNECT method
		1178	if ($proxy && $uscheme eq "https") {
		1179	# oh dear, we have to wrap it into a connect request
		1180
		1181	# maybe re-use $uauthority with patched port?
		1182	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012");
		1183	$state{handle}->push_read (line => $qr_nlnl, sub {
		1184	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1185	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1186
		1187	if ($2 == 200) {
		1188	$rpath = $upath;
		1189	$handle_actual_request->();
		1190	} else {
		1191	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
		1192	}
		1193	});
		1194	} else {
		1195	$handle_actual_request->();
		1196	}
		1197	};
615		1198
616	_get_slot $uhost, sub {	1199	_get_slot $uhost, sub {
617	$state{slot_guard} = shift;	1200	$state{slot_guard} = shift;
618		1201
619	return unless $state{connect_guard};	1202	return unless $state{connect_guard};
620		1203
621	my $ae_error = 595; # connecting	1204	# try to use an existing keepalive connection, but only if we, ourselves, plan
		1205	# on a keepalive request (in theory, this should be a separate config option).
		1206	if ($persistent && $KA_CACHE{$ka_key}) {
		1207	$was_persistent = 1;
622		1208
623	my $connect_cb = sub {	1209	$state{handle} = ka_fetch $ka_key;
624	$state{fh} = shift
625	or do {
626	my $err = "$!";
627	%state = ();
628	return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err });
629	};
630
631	return unless delete $state{connect_guard};
632
633	# get handle
634	$state{handle} = new AnyEvent::Handle
635	fh => $state{fh},
636	peername => $rhost,
637	tls_ctx => $arg{tls_ctx},
638	# these need to be reconfigured on keepalive handles
639	timeout => $timeout,
640	on_error => sub {
641	%state = ();
642	$cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] });
643	},
644	on_eof => sub {
645	%state = ();
646	$cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" });
647	},
648	;
649
650	# limit the number of persistent connections
651	# keepalive not yet supported
652	# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
653	# ++$KA_COUNT{$_[1]};
654	# $state{handle}{ka_count_guard} = AnyEvent::Util::guard {
655	# --$KA_COUNT{$_[1]}
656	# };
657	# $hdr{connection} = "keep-alive";
658	# }
659
660	$state{handle}->starttls ("connect") if $rscheme eq "https";
661
662	# handle actual, non-tunneled, request
663	my $handle_actual_request = sub {
664	$ae_error = 596; # request phase
665
666	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};
667
668	# send request
669	$state{handle}->push_write (	1210	$state{handle}->destroyed
670	"$method $rpath HTTP/1.1\015\012"	1211	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
671	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)	1212	$prepare_handle->();
672	. "\015\012"	1213	$state{handle}->destroyed
673	. (delete $arg{body})	1214	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
674	);
675
676	# return if error occured during push_write()
677	return unless %state;
678
679	%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use
680
681	# status line and headers
682	$state{read_response} = sub {
683	for ("$_[1]") {
684	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
685
686	/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci
687	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" }));
688
689	# 100 Continue handling
690	# should not happen as we don't send expect: 100-continue,
691	# but we handle it just in case.
692	# since we send the request body regardless, if we get an error
693	# we are out of-sync, which we currently do NOT handle correctly.
694	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
695	if $2 eq 100;
696
697	push @pseudo,
698	HTTPVersion => $1,
699	Status => $2,
700	Reason => $3,
701	;
702
703	my $hdr = parse_hdr
704	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" }));
705
706	%hdr = (%$hdr, @pseudo);
707	}
708
709	# redirect handling
710	# microsoft and other shitheads don't give a shit for following standards,
711	# try to support some common forms of broken Location headers.
712	if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) {
713	$hdr{location} =~ s/^\.\/+//;
714
715	my $url = "$rscheme://$uhost:$uport";
716
717	unless ($hdr{location} =~ s/^\///) {
718	$url .= $upath;
719	$url =~ s/\/[^\/]*$//;
720	}
721
722	$hdr{location} = "$url/$hdr{location}";
723	}
724
725	my $redirect;
726
727	if ($recurse) {
728	my $status = $hdr{Status};
729
730	# industry standard is to redirect POST as GET for
731	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
732	# also, the UA should ask the user for 301 and 307 and POST,
733	# industry standard seems to be to simply follow.
734	# we go with the industry standard.
735	if ($status == 301 or $status == 302 or $status == 303) {
736	# HTTP/1.1 is unclear on how to mutate the method
737	$method = "GET" unless $method eq "HEAD";
738	$redirect = 1;
739	} elsif ($status == 307) {
740	$redirect = 1;
741	}
742	}
743
744	my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive])
745	my $may_keep_alive = $_[3];
746
747	$state{handle}->destroy if $state{handle};
748	%state = ();
749
750	if (defined $_[1]) {
751	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
752	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
753	}
754
755	# set-cookie processing
756	if ($arg{cookie_jar}) {
757	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost;
758	}
759
760	if ($redirect && exists $hdr{location}) {
761	# we ignore any errors, as it is very common to receive
762	# Content-Length != 0 but no actual body
763	# we also access %hdr, as $_[1] might be an erro
764	http_request (
765	$method => $hdr{location},
766	%arg,
767	recurse => $recurse - 1,
768	Redirect => [$_[0], \%hdr],
769	$cb);
770	} else {
771	$cb->($_[0], \%hdr);
772	}
773	};
774
775	$ae_error = 597; # body phase
776
777	my $len = $hdr{"content-length"};
778
779	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
780	$finish->(undef, 598 => "Request cancelled by on_header");
781	} elsif (
782	$hdr{Status} =~ /^(?:1..|204|205|304)$/
783	or $method eq "HEAD"
784	or (defined $len && !$len)
785	) {
786	# no body
787	$finish->("", undef, undef, 1);
788	} else {
789	# body handling, many different code paths
790	# - no body expected
791	# - want_body_handle
792	# - te chunked
793	# - 2x length known (with or without on_body)
794	# - 2x length not known (with or without on_body)
795	if (!$redirect && $arg{want_body_handle}) {
796	$_[0]->on_eof (undef);
797	$_[0]->on_error (undef);
798	$_[0]->on_read (undef);
799
800	$finish->(delete $state{handle});
801
802	} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) {
803	my $cl = 0;
804	my $body = undef;
805	my $on_body = $arg{on_body} || sub { $body .= shift; 1 };
806
807	my $read_chunk; $read_chunk = sub {
808	$_[1] =~ /^([0-9a-fA-F]+)/
809	or $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
810
811	my $len = hex $1;
812
813	if ($len) {
814	$cl += $len;
815
816	$_[0]->push_read (chunk => $len, sub {
817	$on_body->($_[1], \%hdr)
818	or return $finish->(undef, 598 => "Request cancelled by on_body");
819
820	$_[0]->push_read (line => sub {
821	length $_[1]
822	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
823	$_[0]->push_read (line => $read_chunk);
824	});
825	});
826	} else {
827	$hdr{"content-length"} ||= $cl;
828
829	$_[0]->push_read (line => $qr_nlnl, sub {
830	if (length $_[1]) {
831	for ("$_[1]") {
832	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
833
834	my $hdr = parse_hdr
835	or return $finish->(undef, $ae_error => "Garbled response trailers");
836
837	%hdr = (%hdr, %$hdr);
838	}
839	}
840
841	$finish->($body, undef, undef, 1);
842	});
843	}
844	};
845
846	$_[0]->push_read (line => $read_chunk);
847
848	} elsif ($arg{on_body}) {
849	if ($len) {
850	$_[0]->on_read (sub {
851	$len -= length $_[0]{rbuf};
852
853	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
854	or return $finish->(undef, 598 => "Request cancelled by on_body");
855
856	$len > 0
857	or $finish->("", undef, undef, 1);
858	});
859	} else {
860	$_[0]->on_eof (sub {
861	$finish->("");
862	});
863	$_[0]->on_read (sub {
864	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
865	or $finish->(undef, 598 => "Request cancelled by on_body");
866	});
867	}
868	} else {
869	$_[0]->on_eof (undef);
870
871	if ($len) {
872	$_[0]->on_read (sub {
873	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
874	if $len <= length $_[0]{rbuf};
875	});
876	} else {
877	$_[0]->on_error (sub {
878	($! == Errno::EPIPE || !$!)
879	? $finish->(delete $_[0]{rbuf})
880	: $finish->(undef, $ae_error => $_[2]);
881	});
882	$_[0]->on_read (sub { });
883	}
884	}
885	}
886	};
887
888	$state{handle}->push_read (line => $qr_nlnl, $state{read_response});
889	};
890
891	# now handle proxy-CONNECT method
892	if ($proxy && $uscheme eq "https") {
893	# oh dear, we have to wrap it into a connect request
894
895	# maybe re-use $uauthority with patched port?
896	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012");
897	$state{handle}->push_read (line => $qr_nlnl, sub {
898	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
899	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }));
900
901	if ($2 == 200) {
902	$rpath = $upath;
903	&$handle_actual_request;	1215	$handle_actual_request->();
904	} else {	1216
905	%state = ();
906	$cb->(undef, { @pseudo, Status => $2, Reason => $3 });
907	}
908	});
909	} else {	1217	} else {
910	&$handle_actual_request;
911	}
912	};
913
914	my $tcp_connect = $arg{tcp_connect}	1218	my $tcp_connect = $arg{tcp_connect}
915	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };	1219	|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
916		1220
917	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });	1221	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout });
918		1222	}
919	};	1223	};
920		1224
921	defined wantarray && AnyEvent::Util::guard { %state = () }	1225	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
922	}	1226	}
923		1227
924	sub http_get($@) {	1228	sub http_get($@) {
925	unshift @_, "GET";	1229	unshift @_, "GET";
926	&http_request	1230	&http_request
…		…
944	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for	1248	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
945	the actual connection, which in turn uses AnyEvent::DNS to resolve	1249	the actual connection, which in turn uses AnyEvent::DNS to resolve
946	hostnames. The latter is a simple stub resolver and does no caching	1250	hostnames. The latter is a simple stub resolver and does no caching
947	on its own. If you want DNS caching, you currently have to provide	1251	on its own. If you want DNS caching, you currently have to provide
948	your own default resolver (by storing a suitable resolver object in	1252	your own default resolver (by storing a suitable resolver object in
949	C<$AnyEvent::DNS::RESOLVER>).	1253	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
950		1254
951	=head2 GLOBAL FUNCTIONS AND VARIABLES	1255	=head2 GLOBAL FUNCTIONS AND VARIABLES
952		1256
953	=over 4	1257	=over 4
954		1258
955	=item AnyEvent::HTTP::set_proxy "proxy-url"	1259	=item AnyEvent::HTTP::set_proxy "proxy-url"
956		1260
957	Sets the default proxy server to use. The proxy-url must begin with a	1261	Sets the default proxy server to use. The proxy-url must begin with a
958	string of the form C<http://host:port> (optionally C<https:...>), croaks	1262	string of the form C<http://host:port>, croaks otherwise.
959	otherwise.
960		1263
961	To clear an already-set proxy, use C<undef>.	1264	To clear an already-set proxy, use C<undef>.
		1265
		1266	When AnyEvent::HTTP is loaded for the first time it will query the
		1267	default proxy from the operating system, currently by looking at
		1268	C<$ENV{http_proxy>}.
		1269
		1270	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
		1271
		1272	Remove all cookies from the cookie jar that have been expired. If
		1273	C<$session_end> is given and true, then additionally remove all session
		1274	cookies.
		1275
		1276	You should call this function (with a true C<$session_end>) before you
		1277	save cookies to disk, and you should call this function after loading them
		1278	again. If you have a long-running program you can additionally call this
		1279	function from time to time.
		1280
		1281	A cookie jar is initially an empty hash-reference that is managed by this
		1282	module. It's format is subject to change, but currently it is like this:
		1283
		1284	The key C<version> has to contain C<1>, otherwise the hash gets
		1285	emptied. All other keys are hostnames or IP addresses pointing to
		1286	hash-references. The key for these inner hash references is the
		1287	server path for which this cookie is meant, and the values are again
		1288	hash-references. Each key of those hash-references is a cookie name, and
		1289	the value, you guessed it, is another hash-reference, this time with the
		1290	key-value pairs from the cookie, except for C<expires> and C<max-age>,
		1291	which have been replaced by a C<_expires> key that contains the cookie
		1292	expiry timestamp. Session cookies are indicated by not having an
		1293	C<_expires> key.
		1294
		1295	Here is an example of a cookie jar with a single cookie, so you have a
		1296	chance of understanding the above paragraph:
		1297
		1298	{
		1299	version => 1,
		1300	"10.0.0.1" => {
		1301	"/" => {
		1302	"mythweb_id" => {
		1303	_expires => 1293917923,
		1304	value => "ooRung9dThee3ooyXooM1Ohm",
		1305	},
		1306	},
		1307	},
		1308	}
962		1309
963	=item $date = AnyEvent::HTTP::format_date $timestamp	1310	=item $date = AnyEvent::HTTP::format_date $timestamp
964		1311
965	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP	1312	Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP
966	Date (RFC 2616).	1313	Date (RFC 2616).
…		…
973		1320
974	=item $AnyEvent::HTTP::MAX_RECURSE	1321	=item $AnyEvent::HTTP::MAX_RECURSE
975		1322
976	The default value for the C<recurse> request parameter (default: C<10>).	1323	The default value for the C<recurse> request parameter (default: C<10>).
977		1324
		1325	=item $AnyEvent::HTTP::TIMEOUT
		1326
		1327	The default timeout for connection operations (default: C<300>).
		1328
978	=item $AnyEvent::HTTP::USERAGENT	1329	=item $AnyEvent::HTTP::USERAGENT
979		1330
980	The default value for the C<User-Agent> header (the default is	1331	The default value for the C<User-Agent> header (the default is
981	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1332	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
982		1333
983	=item $AnyEvent::HTTP::MAX_PER_HOST	1334	=item $AnyEvent::HTTP::MAX_PER_HOST
984		1335
985	The maximum number of concurrent connections to the same host (identified	1336	The maximum number of concurrent connections to the same host (identified
986	by the hostname). If the limit is exceeded, then the additional requests	1337	by the hostname). If the limit is exceeded, then the additional requests
987	are queued until previous connections are closed.	1338	are queued until previous connections are closed. Both persistent and
		1339	non-persistent connections are counted in this limit.
988		1340
989	The default value for this is C<4>, and it is highly advisable to not	1341	The default value for this is C<4>, and it is highly advisable to not
990	increase it.	1342	increase it much.
		1343
		1344	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1345	connections, older browsers used 2, newer ones (such as firefox 3)
		1346	typically use 6, and Opera uses 8 because like, they have the fastest
		1347	browser and give a shit for everybody else on the planet.
		1348
		1349	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
		1350
		1351	The time after which idle persistent connections get closed by
		1352	AnyEvent::HTTP (default: C<3>).
991		1353
992	=item $AnyEvent::HTTP::ACTIVE	1354	=item $AnyEvent::HTTP::ACTIVE
993		1355
994	The number of active connections. This is not the number of currently	1356	The number of active connections. This is not the number of currently
995	running requests, but the number of currently open and non-idle TCP	1357	running requests, but the number of currently open and non-idle TCP
996	connections. This number of can be useful for load-leveling.	1358	connections. This number can be useful for load-leveling.
997		1359
998	=back	1360	=back
999		1361
1000	=cut	1362	=cut
1001		1363
…		…
1035	# other formats fail in the loop below	1397	# other formats fail in the loop below
1036		1398
1037	for (0..11) {	1399	for (0..11) {
1038	if ($m eq $month[$_]) {	1400	if ($m eq $month[$_]) {
1039	require Time::Local;	1401	require Time::Local;
1040	return Time::Local::timegm ($S, $M, $H, $d, $_, $y);	1402	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
1041	}	1403	}
1042	}	1404	}
1043		1405
1044	undef	1406	undef
1045	}	1407	}
1046		1408
1047	sub set_proxy($) {	1409	sub set_proxy($) {
1048	if (length $_[0]) {	1410	if (length $_[0]) {
1049	$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix	1411	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
1050	or Carp::croak "$_[0]: invalid proxy URL";	1412	or Carp::croak "$_[0]: invalid proxy URL";
1051	$PROXY = [$2, $3 || 3128, $1]	1413	$PROXY = [$2, $3 || 3128, $1]
1052	} else {	1414	} else {
1053	undef $PROXY;	1415	undef $PROXY;
1054	}	1416	}
…		…
1057	# initialise proxy from environment	1419	# initialise proxy from environment
1058	eval {	1420	eval {
1059	set_proxy $ENV{http_proxy};	1421	set_proxy $ENV{http_proxy};
1060	};	1422	};
1061		1423
		1424	=head2 SHOWCASE
		1425
		1426	This section contains some more elaborate "real-world" examples or code
		1427	snippets.
		1428
		1429	=head2 HTTP/1.1 FILE DOWNLOAD
		1430
		1431	Downloading files with HTTP can be quite tricky, especially when something
		1432	goes wrong and you want to resume.
		1433
		1434	Here is a function that initiates and resumes a download. It uses the
		1435	last modified time to check for file content changes, and works with many
		1436	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1437	on older servers.
		1438
		1439	It calls the completion callback with either C<undef>, which means a
		1440	nonretryable error occurred, C<0> when the download was partial and should
		1441	be retried, and C<1> if it was successful.
		1442
		1443	use AnyEvent::HTTP;
		1444
		1445	sub download($$$) {
		1446	my ($url, $file, $cb) = @_;
		1447
		1448	open my $fh, "+<", $file
		1449	or die "$file: $!";
		1450
		1451	my %hdr;
		1452	my $ofs = 0;
		1453
		1454	warn stat $fh;
		1455	warn -s _;
		1456	if (stat $fh and -s _) {
		1457	$ofs = -s _;
		1458	warn "-s is ", $ofs;
		1459	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1460	$hdr{"range"} = "bytes=$ofs-";
		1461	}
		1462
		1463	http_get $url,
		1464	headers => \%hdr,
		1465	on_header => sub {
		1466	my ($hdr) = @_;
		1467
		1468	if ($hdr->{Status} == 200 && $ofs) {
		1469	# resume failed
		1470	truncate $fh, $ofs = 0;
		1471	}
		1472
		1473	sysseek $fh, $ofs, 0;
		1474
		1475	1
		1476	},
		1477	on_body => sub {
		1478	my ($data, $hdr) = @_;
		1479
		1480	if ($hdr->{Status} =~ /^2/) {
		1481	length $data == syswrite $fh, $data
		1482	or return; # abort on write errors
		1483	}
		1484
		1485	1
		1486	},
		1487	sub {
		1488	my (undef, $hdr) = @_;
		1489
		1490	my $status = $hdr->{Status};
		1491
		1492	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1493	utime $fh, $time, $time;
		1494	}
		1495
		1496	if ($status == 200 || $status == 206 || $status == 416) {
		1497	# download ok || resume ok || file already fully downloaded
		1498	$cb->(1, $hdr);
		1499
		1500	} elsif ($status == 412) {
		1501	# file has changed while resuming, delete and retry
		1502	unlink $file;
		1503	$cb->(0, $hdr);
		1504
		1505	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1506	# retry later
		1507	$cb->(0, $hdr);
		1508
		1509	} else {
		1510	$cb->(undef, $hdr);
		1511	}
		1512	}
		1513	;
		1514	}
		1515
		1516	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1517	if ($_[0]) {
		1518	print "OK!\n";
		1519	} elsif (defined $_[0]) {
		1520	print "please retry later\n";
		1521	} else {
		1522	print "ERROR\n";
		1523	}
		1524	};
		1525
1062	=head2 SOCKS PROXIES	1526	=head3 SOCKS PROXIES
1063		1527
1064	Socks proxies are not directly supported by AnyEvent::HTTP. You can	1528	Socks proxies are not directly supported by AnyEvent::HTTP. You can
1065	compile your perl to support socks, or use an external program such as	1529	compile your perl to support socks, or use an external program such as
1066	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy	1530	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
1067	transparently.	1531	transparently.

Diff Legend

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