[ViewVC] Diff of: cvs/AnyEvent-HTTP/HTTP.pm

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.87 by root, Sun Jan 2 08:51:53 2011 UTC vs.
Revision 1.124 by root, Fri May 8 21:26:16 2015 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
…		…
46	use AnyEvent::Util ();	46	use AnyEvent::Util ();
47	use AnyEvent::Handle ();	47	use AnyEvent::Handle ();
48		48
49	use base Exporter::;	49	use base Exporter::;
50		50
51	our $VERSION = '1.5';	51	our $VERSION = 2.21;
52		52
53	our @EXPORT = qw(http_get http_post http_head http_request);	53	our @EXPORT = qw(http_get http_post http_head http_request);
54		54
55	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)";
56	our $MAX_RECURSE = 10;	56	our $MAX_RECURSE = 10;
57	our $MAX_PERSISTENT = 8;
58	our $PERSISTENT_TIMEOUT = 2;	57	our $PERSISTENT_TIMEOUT = 3;
59	our $TIMEOUT = 300;	58	our $TIMEOUT = 300;
60		59	our $MAX_PER_HOST = 4; # changing this is evil
61	# changing these is evil
62	our $MAX_PERSISTENT_PER_HOST = 0;
63	our $MAX_PER_HOST = 4;
64		60
65	our $PROXY;	61	our $PROXY;
66	our $ACTIVE = 0;	62	our $ACTIVE = 0;
67		63
68	my %KA_COUNT; # number of open keep-alive connections per host	64	my %KA_CACHE; # indexed by uhost currently, points to [$handle...] array
69	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
70		66
71	=item http_get $url, key => value..., $cb->($data, $headers)	67	=item http_get $url, key => value..., $cb->($data, $headers)
72		68
73	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
…		…
93	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
94	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
95	destroyed before the callback is called, the request will be cancelled.	91	destroyed before the callback is called, the request will be cancelled.
96		92
97	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
98	(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
99	(and trailers) as second argument.	95	(and trailers) as second argument.
100		96
101	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
102	headers, the "pseudo-headers" (uppercase to avoid clashing with possible	98	headers, the "pseudo-headers" (uppercase to avoid clashing with possible
103	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the	99	response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the
…		…
127	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
128	message. Currently the following status codes are used:	124	message. Currently the following status codes are used:
129		125
130	=over 4	126	=over 4
131		127
132	=item 595 - errors during connection etsbalishment, proxy handshake.	128	=item 595 - errors during connection establishment, proxy handshake.
133		129
134	=item 596 - errors during TLS negotiation, request sending and header processing.	130	=item 596 - errors during TLS negotiation, request sending and header processing.
135		131
136	=item 597 - errors during body receiving or processing.	132	=item 597 - errors during body receiving or processing.
137		133
…		…
158		154
159	=over 4	155	=over 4
160		156
161	=item recurse => $count (default: $MAX_RECURSE)	157	=item recurse => $count (default: $MAX_RECURSE)
162		158
163	Whether to recurse requests or not, e.g. on redirects, authentication	159	Whether to recurse requests or not, e.g. on redirects, authentication and
164	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.
165		166
166	=item headers => hashref	167	=item headers => hashref
167		168
168	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
169	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
170	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:>
171	(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
172	they won't be sent at all).	173	they won't be sent at all).
173		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
174	=item timeout => $seconds	182	=item timeout => $seconds
175		183
176	The time-out to use for various stages - each connect attempt will reset	184	The time-out to use for various stages - each connect attempt will reset
177	the timeout, as will read or write activity, i.e. this is not an overall	185	the timeout, as will read or write activity, i.e. this is not an overall
178	timeout.	186	timeout.
179		187
180	Default timeout is 5 minutes.	188	Default timeout is 5 minutes.
181		189
182	=item proxy => [$host, $port[, $scheme]] or undef	190	=item proxy => [$host, $port[, $scheme]] or undef
183		191
184	Use the given http proxy for all requests. If not specified, then the	192	Use the given http proxy for all requests, or no proxy if C<undef> is
185	default proxy (as specified by C<$ENV{http_proxy}>) is used.	193	used.
186		194
187	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.
188	HTTPS.	196
		197	If not specified, then the default proxy is used (see
		198	C<AnyEvent::HTTP::set_proxy>).
		199
		200	Currently, if your proxy requires authorization, you have to specify an
		201	appropriate "Proxy-Authorization" header in every request.
189		202
190	=item body => $string	203	=item body => $string
191		204
192	The request body, usually empty. Will be sent as-is (future versions of	205	The request body, usually empty. Will be sent as-is (future versions of
193	this module might offer more options).	206	this module might offer more options).
…		…
224	verification) TLS context.	237	verification) TLS context.
225		238
226	The default for this option is C<low>, which could be interpreted as "give	239	The default for this option is C<low>, which could be interpreted as "give
227	me the page, no matter what".	240	me the page, no matter what".
228		241
		242	See also the C<sessionid> parameter.
		243
		244	=item session => $string
		245
		246	The module might reuse connections to the same host internally. Sometimes
		247	(e.g. when using TLS), you do not want to reuse connections from other
		248	sessions. This can be achieved by setting this parameter to some unique
		249	ID (such as the address of an object storing your state data, or the TLS
		250	context) - only connections using the same unique ID will be reused.
		251
229	=item on_prepare => $callback->($fh)	252	=item on_prepare => $callback->($fh)
230		253
231	In rare cases you need to "tune" the socket before it is used to	254	In rare cases you need to "tune" the socket before it is used to
232	connect (for exmaple, to bind it on a given IP address). This parameter	255	connect (for example, to bind it on a given IP address). This parameter
233	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>	256	overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect>
234	and behaves exactly the same way (e.g. it has to provide a	257	and behaves exactly the same way (e.g. it has to provide a
235	timeout). See the description for the C<$prepare_cb> argument of	258	timeout). See the description for the C<$prepare_cb> argument of
236	C<AnyEvent::Socket::tcp_connect> for details.	259	C<AnyEvent::Socket::tcp_connect> for details.
237		260
…		…
302	called. Instead of the C<$body> argument containing the body data, the	325	called. Instead of the C<$body> argument containing the body data, the
303	callback will receive the L<AnyEvent::Handle> object associated with the	326	callback will receive the L<AnyEvent::Handle> object associated with the
304	connection. In error cases, C<undef> will be passed. When there is no body	327	connection. In error cases, C<undef> will be passed. When there is no body
305	(e.g. status C<304>), the empty string will be passed.	328	(e.g. status C<304>), the empty string will be passed.
306		329
307	The handle object might or might not be in TLS mode, might be connected to	330	The handle object might or might not be in TLS mode, might be connected
308	a proxy, be a persistent connection etc., and configured in unspecified	331	to a proxy, be a persistent connection, use chunked transfer encoding
309	ways. The user is responsible for this handle (it will not be used by this	332	etc., and configured in unspecified ways. The user is responsible for this
310	module anymore).	333	handle (it will not be used by this module anymore).
311		334
312	This is useful with some push-type services, where, after the initial	335	This is useful with some push-type services, where, after the initial
313	headers, an interactive protocol is used (typical example would be the	336	headers, an interactive protocol is used (typical example would be the
314	push-style twitter API which starts a JSON/XML stream).	337	push-style twitter API which starts a JSON/XML stream).
315		338
316	If you think you need this, first have a look at C<on_body>, to see if	339	If you think you need this, first have a look at C<on_body>, to see if
317	that doesn't solve your problem in a better way.	340	that doesn't solve your problem in a better way.
		341
		342	=item persistent => $boolean
		343
		344	Try to create/reuse a persistent connection. When this flag is set
		345	(default: true for idempotent requests, false for all others), then
		346	C<http_request> tries to re-use an existing (previously-created)
		347	persistent connection to the host and, failing that, tries to create a new
		348	one.
		349
		350	Requests failing in certain ways will be automatically retried once, which
		351	is dangerous for non-idempotent requests, which is why it defaults to off
		352	for them. The reason for this is because the bozos who designed HTTP/1.1
		353	made it impossible to distinguish between a fatal error and a normal
		354	connection timeout, so you never know whether there was a problem with
		355	your request or not.
		356
		357	When reusing an existent connection, many parameters (such as TLS context)
		358	will be ignored. See the C<session> parameter for a workaround.
		359
		360	=item keepalive => $boolean
		361
		362	Only used when C<persistent> is also true. This parameter decides whether
		363	C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection
		364	(as opposed to only a HTTP/1.1 persistent connection).
		365
		366	The default is true, except when using a proxy, in which case it defaults
		367	to false, as HTTP/1.0 proxies cannot support this in a meaningful way.
		368
		369	=item handle_params => { key => value ... }
		370
		371	The key-value pairs in this hash will be passed to any L<AnyEvent::Handle>
		372	constructor that is called - not all requests will create a handle, and
		373	sometimes more than one is created, so this parameter is only good for
		374	setting hints.
		375
		376	Example: set the maximum read size to 4096, to potentially conserve memory
		377	at the cost of speed.
		378
		379	handle_params => {
		380	max_read_size => 4096,
		381	},
318		382
319	=back	383	=back
320		384
321	Example: do a simple HTTP GET request for http://www.nethype.de/ and print	385	Example: do a simple HTTP GET request for http://www.nethype.de/ and print
322	the response body.	386	the response body.
…		…
328		392
329	Example: do a HTTP HEAD request on https://www.google.com/, use a	393	Example: do a HTTP HEAD request on https://www.google.com/, use a
330	timeout of 30 seconds.	394	timeout of 30 seconds.
331		395
332	http_request	396	http_request
333	GET => "https://www.google.com",	397	HEAD => "https://www.google.com",
		398	headers => { "user-agent" => "MySearchClient 1.0" },
334	timeout => 30,	399	timeout => 30,
335	sub {	400	sub {
336	my ($body, $hdr) = @_;	401	my ($body, $hdr) = @_;
337	use Data::Dumper;	402	use Data::Dumper;
338	print Dumper $hdr;	403	print Dumper $hdr;
…		…
348	};	413	};
349		414
350	undef $request;	415	undef $request;
351		416
352	=cut	417	=cut
		418
		419	#############################################################################
		420	# wait queue/slots
353		421
354	sub _slot_schedule;	422	sub _slot_schedule;
355	sub _slot_schedule($) {	423	sub _slot_schedule($) {
356	my $host = shift;	424	my $host = shift;
357		425
…		…
380		448
381	_slot_schedule $_[0];	449	_slot_schedule $_[0];
382	}	450	}
383		451
384	#############################################################################	452	#############################################################################
		453	# cookie handling
385		454
386	# expire cookies	455	# expire cookies
387	sub cookie_jar_expire($;$) {	456	sub cookie_jar_expire($;$) {
388	my ($jar, $session_end) = @_;	457	my ($jar, $session_end) = @_;
389		458
…		…
413	}	482	}
414	}	483	}
415		484
416	# extract cookies from jar	485	# extract cookies from jar
417	sub cookie_jar_extract($$$$) {	486	sub cookie_jar_extract($$$$) {
418	my ($jar, $uscheme, $uhost, $upath) = @_;	487	my ($jar, $scheme, $host, $path) = @_;
419		488
420	%$jar = () if $jar->{version} != 1;	489	%$jar = () if $jar->{version} != 1;
421		490
422	my @cookies;	491	my @cookies;
423		492
424	while (my ($chost, $paths) = each %$jar) {	493	while (my ($chost, $paths) = each %$jar) {
425	next unless ref $paths;	494	next unless ref $paths;
426		495
427	if ($chost =~ /^\./) {	496	if ($chost =~ /^\./) {
428	next unless $chost eq substr $uhost, -length $chost;	497	next unless $chost eq substr $host, -length $chost;
429	} elsif ($chost =~ /\./) {	498	} elsif ($chost =~ /\./) {
430	next unless $chost eq $uhost;	499	next unless $chost eq $host;
431	} else {	500	} else {
432	next;	501	next;
433	}	502	}
434		503
435	while (my ($cpath, $cookies) = each %$paths) {	504	while (my ($cpath, $cookies) = each %$paths) {
436	next unless $cpath eq substr $upath, 0, length $cpath;	505	next unless $cpath eq substr $path, 0, length $cpath;
437		506
438	while (my ($cookie, $kv) = each %$cookies) {	507	while (my ($cookie, $kv) = each %$cookies) {
439	next if $uscheme ne "https" && exists $kv->{secure};	508	next if $scheme ne "https" && exists $kv->{secure};
440		509
441	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {	510	if (exists $kv->{_expires} and AE::now > $kv->{_expires}) {
442	delete $cookies->{$cookie};	511	delete $cookies->{$cookie};
443	next;	512	next;
444	}	513	}
…		…
458	\@cookies	527	\@cookies
459	}	528	}
460		529
461	# parse set_cookie header into jar	530	# parse set_cookie header into jar
462	sub cookie_jar_set_cookie($$$$) {	531	sub cookie_jar_set_cookie($$$$) {
463	my ($jar, $set_cookie, $uhost, $date) = @_;	532	my ($jar, $set_cookie, $host, $date) = @_;
464		533
465	my $anow = int AE::now;	534	my $anow = int AE::now;
466	my $snow; # server-now	535	my $snow; # server-now
467		536
468	for ($set_cookie) {	537	for ($set_cookie) {
…		…
474	while (	543	while (
475	m{	544	m{
476	\G\s*	545	\G\s*
477	(?:	546	(?:
478	expires \s=\s ([A-Z][a-z][a-z]+,\ [^,;]+)	547	expires \s=\s ([A-Z][a-z][a-z]+,\ [^,;]+)
479	\| ([^=;,[:space:]]+) (?: \s=\s (?: "((?:[^\\"]+\|\\.))" \| ([^=;,[:space:]]) ) )?	548	\| ([^=;,[:space:]]+) (?: \s=\s (?: "((?:[^\\"]+\|\\.))" \| ([^;,[:space:]]) ) )?
480	)	549	)
481	}gcxsi	550	}gcxsi
482	) {	551	) {
483	my $name = $2;	552	my $name = $2;
484	my $value = $4;	553	my $value = $4;
…		…
491	# quoted	560	# quoted
492	$value = $3;	561	$value = $3;
493	$value =~ s/\\(.)/$1/gs;	562	$value =~ s/\\(.)/$1/gs;
494	}	563	}
495		564
496	push @kv, lc $name, $value;	565	push @kv, @kv ? lc $name : $name, $value;
497		566
498	last unless /\G\s*;/gc;	567	last unless /\G\s*;/gc;
499	}	568	}
500		569
501	last unless @kv;	570	last unless @kv;
…		…
524		593
525	# this is not rfc-like and not netscape-like. go figure.	594	# this is not rfc-like and not netscape-like. go figure.
526	my $ndots = $cdom =~ y/.//;	595	my $ndots = $cdom =~ y/.//;
527	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);	596	next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2);
528	} else {	597	} else {
529	$cdom = $uhost;	598	$cdom = $host;
530	}	599	}
531		600
532	# store it	601	# store it
533	$jar->{version} = 1;	602	$jar->{version} = 1;
534	$jar->{lc $cdom}{$cpath}{$name} = \%kv;	603	$jar->{lc $cdom}{$cpath}{$name} = \%kv;
535		604
536	redo if /\G\s*,/gc;	605	redo if /\G\s*,/gc;
537	}	606	}
538	}	607	}
539		608
		609	#############################################################################
		610	# keepalive/persistent connection cache
		611
		612	# fetch a connection from the keepalive cache
		613	sub ka_fetch($) {
		614	my $ka_key = shift;
		615
		616	my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection
		617	delete $KA_CACHE{$ka_key}
		618	unless @{ $KA_CACHE{$ka_key} };
		619
		620	$hdl
		621	}
		622
		623	sub ka_store($$) {
		624	my ($ka_key, $hdl) = @_;
		625
		626	my $kaa = $KA_CACHE{$ka_key} \|\|= [];
		627
		628	my $destroy = sub {
		629	my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} };
		630
		631	$hdl->destroy;
		632
		633	@ka
		634	? $KA_CACHE{$ka_key} = \@ka
		635	: delete $KA_CACHE{$ka_key};
		636	};
		637
		638	# on error etc., destroy
		639	$hdl->on_error ($destroy);
		640	$hdl->on_eof ($destroy);
		641	$hdl->on_read ($destroy);
		642	$hdl->timeout ($PERSISTENT_TIMEOUT);
		643
		644	push @$kaa, $hdl;
		645	shift @$kaa while @$kaa > $MAX_PER_HOST;
		646	}
		647
		648	#############################################################################
		649	# utilities
		650
540	# continue to parse $_ for headers and place them into the arg	651	# continue to parse $_ for headers and place them into the arg
541	sub parse_hdr() {	652	sub _parse_hdr() {
542	my %hdr;	653	my %hdr;
543		654
544	# things seen, not parsed:	655	# things seen, not parsed:
545	# p3pP="NON CUR OTPi OUR NOR UNI"	656	# p3pP="NON CUR OTPi OUR NOR UNI"
546		657
…		…
560	for values %hdr;	671	for values %hdr;
561		672
562	\%hdr	673	\%hdr
563	}	674	}
564		675
		676	#############################################################################
		677	# http_get
		678
565	our $qr_nlnl = qr{(?<![^\012])\015?\012};	679	our $qr_nlnl = qr{(?<![^\012])\015?\012};
566		680
567	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };	681	our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 };
568	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };	682	our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" };
		683
		684	# maybe it should just become a normal object :/
		685
		686	sub _destroy_state(\%) {
		687	my ($state) = @_;
		688
		689	$state->{handle}->destroy if $state->{handle};
		690	%$state = ();
		691	}
		692
		693	sub _error(\%$$) {
		694	my ($state, $cb, $hdr) = @_;
		695
		696	&_destroy_state ($state);
		697
		698	$cb->(undef, $hdr);
		699	()
		700	}
		701
		702	our %IDEMPOTENT = (
		703	DELETE => 1,
		704	GET => 1,
		705	HEAD => 1,
		706	OPTIONS => 1,
		707	PUT => 1,
		708	TRACE => 1,
		709
		710	ACL => 1,
		711	"BASELINE-CONTROL" => 1,
		712	BIND => 1,
		713	CHECKIN => 1,
		714	CHECKOUT => 1,
		715	COPY => 1,
		716	LABEL => 1,
		717	LINK => 1,
		718	MERGE => 1,
		719	MKACTIVITY => 1,
		720	MKCALENDAR => 1,
		721	MKCOL => 1,
		722	MKREDIRECTREF => 1,
		723	MKWORKSPACE => 1,
		724	MOVE => 1,
		725	ORDERPATCH => 1,
		726	PROPFIND => 1,
		727	PROPPATCH => 1,
		728	REBIND => 1,
		729	REPORT => 1,
		730	SEARCH => 1,
		731	UNBIND => 1,
		732	UNCHECKOUT => 1,
		733	UNLINK => 1,
		734	UNLOCK => 1,
		735	UPDATE => 1,
		736	UPDATEREDIRECTREF => 1,
		737	"VERSION-CONTROL" => 1,
		738	);
569		739
570	sub http_request($$@) {	740	sub http_request($$@) {
571	my $cb = pop;	741	my $cb = pop;
572	my ($method, $url, %arg) = @_;	742	my ($method, $url, %arg) = @_;
573		743
…		…
591	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;	761	my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE;
592		762
593	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })	763	return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" })
594	if $recurse < 0;	764	if $recurse < 0;
595		765
596	my $proxy = $arg{proxy} \|\| $PROXY;	766	my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY;
597	my $timeout = $arg{timeout} \|\| $TIMEOUT;	767	my $timeout = $arg{timeout} \|\| $TIMEOUT;
598		768
599	my ($uscheme, $uauthority, $upath, $query, $fragment) =	769	my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment
600	$url =~ m\|(?:([^:/?#]+):)?(?://([^/?#]))?([^?#])(?:(\?[^#]))?(?:#(.))?\|;	770	$url =~ m\|^([^:]+):(?://([^/?#]))?([^?#])(?:(\?[^#]))?(?:#(.))?$\|;
601		771
602	$uscheme = lc $uscheme;	772	$uscheme = lc $uscheme;
603		773
604	my $uport = $uscheme eq "http" ? 80	774	my $uport = $uscheme eq "http" ? 80
605	: $uscheme eq "https" ? 443	775	: $uscheme eq "https" ? 443
606	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });	776	: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" });
607		777
608	$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x	778	$uauthority =~ /^(?: .*\@ )? ([^\@]+?) (?: : (\d+) )?$/x
609	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });	779	or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" });
610		780
611	my $uhost = lc $1;	781	my $uhost = lc $1;
612	$uport = $2 if defined $2;	782	$uport = $2 if defined $2;
613		783
…		…
649	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};	819	$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"};
650		820
651	$hdr{"content-length"} = length $arg{body}	821	$hdr{"content-length"} = length $arg{body}
652	if length $arg{body} \|\| $method ne "GET";	822	if length $arg{body} \|\| $method ne "GET";
653		823
654	$hdr{connection} = "close Te"; #1.1	824	my $idempotent = $IDEMPOTENT{$method};
		825
		826	# default value for keepalive is true iff the request is for an idempotent method
		827	my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent;
		828	my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy;
		829	my $was_persistent; # true if this is actually a recycled connection
		830
		831	# the key to use in the keepalive cache
		832	my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}";
		833
		834	$hdr{connection} = ($persistent ? $keepalive ? "keep-alive, " : "" : "close, ") . "Te"; #1.1
655	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1	835	$hdr{te} = "trailers" unless exists $hdr{te}; #1.1
656		836
657	my %state = (connect_guard => 1);	837	my %state = (connect_guard => 1);
658		838
659	_get_slot $uhost, sub {
660	$state{slot_guard} = shift;
661
662	return unless $state{connect_guard};
663
664	my $ae_error = 595; # connecting	839	my $ae_error = 595; # connecting
665		840
666	# handle actual, non-tunneled, request	841	# handle actual, non-tunneled, request
667	my $handle_actual_request = sub {	842	my $handle_actual_request = sub {
668	$ae_error = 596; # request phase	843	$ae_error = 596; # request phase
669		844
		845	my $hdl = $state{handle};
		846
670	$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls};	847	$hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls};
671		848
672	# send request	849	# send request
673	$state{handle}->push_write (	850	$hdl->push_write (
674	"$method $rpath HTTP/1.1\015\012"	851	"$method $rpath HTTP/1.1\015\012"
675	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)	852	. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr)
676	. "\015\012"	853	. "\015\012"
677	. (delete $arg{body})	854	. $arg{body}
678	);	855	);
679		856
680	# return if error occured during push_write()	857	# return if error occurred during push_write()
		858	return unless %state;
		859
		860	# reduce memory usage, save a kitten, also re-use it for the response headers.
		861	%hdr = ();
		862
		863	# status line and headers
		864	$state{read_response} = sub {
681	return unless %state;	865	return unless %state;
682		866
683	%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use
684
685	# status line and headers
686	$state{read_response} = sub {
687	for ("$_[1]") {	867	for ("$_[1]") {
688	y/\015//d; # weed out any \015, as they show up in the weirdest of places.	868	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
689		869
690	/^HTTP\/0([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]) )? \012/gxci	870	/^HTTP\/0([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]) )? \012/gxci
691	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" }));	871	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" };
692		872
693	# 100 Continue handling	873	# 100 Continue handling
694	# should not happen as we don't send expect: 100-continue,	874	# should not happen as we don't send expect: 100-continue,
695	# but we handle it just in case.	875	# but we handle it just in case.
696	# since we send the request body regardless, if we get an error	876	# since we send the request body regardless, if we get an error
697	# we are out of-sync, which we currently do NOT handle correctly.	877	# we are out of-sync, which we currently do NOT handle correctly.
698	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})	878	return $state{handle}->push_read (line => $qr_nlnl, $state{read_response})
699	if $2 eq 100;	879	if $2 eq 100;
700		880
701	push @pseudo,	881	push @pseudo,
702	HTTPVersion => $1,	882	HTTPVersion => $1,
703	Status => $2,	883	Status => $2,
704	Reason => $3,	884	Reason => $3,
705	;	885	;
706		886
707	my $hdr = parse_hdr	887	my $hdr = _parse_hdr
708	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" }));	888	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" };
709		889
710	%hdr = (%$hdr, @pseudo);	890	%hdr = (%$hdr, @pseudo);
711	}	891	}
712		892
713	# redirect handling	893	# redirect handling
714	# microsoft and other shitheads don't give a shit for following standards,	894	# relative uri handling forced by microsoft and other shitheads.
715	# try to support some common forms of broken Location headers.	895	# we give our best and fall back to URI if available.
		896	if (exists $hdr{location}) {
		897	my $loc = $hdr{location};
		898
		899	if ($loc =~ m%^//%) { # //
		900	$loc = "$rscheme:$loc";
		901
		902	} elsif ($loc eq "") {
		903	$loc = $url;
		904
716	if ($hdr{location} !~ /^(?: $ \| [^:\/?\#]+ : )/x) {	905	} elsif ($loc !~ /^(?: $ \| [^:\/?\#]+ : )/x) { # anything "simple"
717	$hdr{location} =~ s/^\.\/+//;	906	$loc =~ s/^\.\/+//;
718		907
		908	if ($loc !~ m%^[.?#]%) {
719	my $url = "$rscheme://$uhost:$uport";	909	my $prefix = "$rscheme://$uhost:$uport";
720		910
721	unless ($hdr{location} =~ s/^\///) {	911	unless ($loc =~ s/^\///) {
722	$url .= $upath;	912	$prefix .= $upath;
723	$url =~ s/\/[^\/]*$//;	913	$prefix =~ s/\/[^\/]*$//;
724	}	914	}
725		915
726	$hdr{location} = "$url/$hdr{location}";	916	$loc = "$prefix/$loc";
727	}
728		917
729	my $redirect;	918	} elsif (eval { require URI }) { # uri
		919	$loc = URI->new_abs ($loc, $url)->as_string;
730		920
731	if ($recurse) {	921	} else {
732	my $status = $hdr{Status};	922	return _error %state, $cb, { @pseudo, Status => 599, Reason => "Cannot parse Location (URI module missing)" };
733		923	#$hdr{Status} = 599;
734	# industry standard is to redirect POST as GET for	924	#$hdr{Reason} = "Unparsable Redirect (URI module missing)";
735	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
736	# also, the UA should ask the user for 301 and 307 and POST,
737	# industry standard seems to be to simply follow.
738	# we go with the industry standard.
739	if ($status == 301 or $status == 302 or $status == 303) {
740	# HTTP/1.1 is unclear on how to mutate the method
741	$method = "GET" unless $method eq "HEAD";
742	$redirect = 1;	925	#$recurse = 0;
743	} elsif ($status == 307) {
744	$redirect = 1;
745	}	926	}
746	}	927	}
747		928
748	my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive])	929	$hdr{location} = $loc;
749	my $may_keep_alive = $_[3];	930	}
750		931
751	$state{handle}->destroy if $state{handle};	932	my $redirect;
752	%state = ();
753		933
754	if (defined $_[1]) {	934	if ($recurse) {
755	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];	935	my $status = $hdr{Status};
756	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];	936
		937	# industry standard is to redirect POST as GET for
		938	# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1.
		939	# also, the UA should ask the user for 301 and 307 and POST,
		940	# industry standard seems to be to simply follow.
		941	# we go with the industry standard. 308 is defined
		942	# by rfc7538
		943	if ($status == 301 or $status == 302 or $status == 303) {
		944	$redirect = 1;
		945	# HTTP/1.1 is unclear on how to mutate the method
		946	unless ($method eq "HEAD") {
		947	$method = "GET";
		948	delete $arg{body};
757	}	949	}
		950	} elsif ($status == 307 or $status == 308) {
		951	$redirect = 1;
		952	}
		953	}
758		954
759	# set-cookie processing	955	my $finish = sub { # ($data, $err_status, $err_reason[, $persistent])
760	if ($arg{cookie_jar}) {	956	if ($state{handle}) {
761	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};	957	# handle keepalive
		958	if (
		959	$persistent
		960	&& $_[3]
		961	&& ($hdr{HTTPVersion} < 1.1
		962	? $hdr{connection} =~ /\bkeep-?alive\b/i
		963	: $hdr{connection} !~ /\bclose\b/i)
		964	) {
		965	ka_store $ka_key, delete $state{handle};
		966	} else {
		967	# no keepalive, destroy the handle
		968	$state{handle}->destroy;
762	}	969	}
		970	}
763		971
		972	%state = ();
		973
		974	if (defined $_[1]) {
		975	$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1];
		976	$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2];
		977	}
		978
		979	# set-cookie processing
		980	if ($arg{cookie_jar}) {
		981	cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date};
		982	}
		983
764	if ($redirect && exists $hdr{location}) {	984	if ($redirect && exists $hdr{location}) {
765	# we ignore any errors, as it is very common to receive	985	# we ignore any errors, as it is very common to receive
766	# Content-Length != 0 but no actual body	986	# Content-Length != 0 but no actual body
767	# we also access %hdr, as $_[1] might be an erro	987	# we also access %hdr, as $_[1] might be an erro
		988	$state{recurse} =
768	http_request (	989	http_request (
769	$method => $hdr{location},	990	$method => $hdr{location},
770	%arg,	991	%arg,
771	recurse => $recurse - 1,	992	recurse => $recurse - 1,
772	Redirect => [$_[0], \%hdr],	993	Redirect => [$_[0], \%hdr],
		994	sub {
		995	%state = ();
		996	&$cb
		997	},
		998	);
		999	} else {
		1000	$cb->($_[0], \%hdr);
		1001	}
		1002	};
		1003
		1004	$ae_error = 597; # body phase
		1005
		1006	my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct...
		1007
		1008	my $len = $chunked ? undef : $hdr{"content-length"};
		1009
		1010	# body handling, many different code paths
		1011	# - no body expected
		1012	# - want_body_handle
		1013	# - te chunked
		1014	# - 2x length known (with or without on_body)
		1015	# - 2x length not known (with or without on_body)
		1016	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
		1017	$finish->(undef, 598 => "Request cancelled by on_header");
		1018	} elsif (
		1019	$hdr{Status} =~ /^(?:1..\|204\|205\|304)$/
		1020	or $method eq "HEAD"
		1021	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
		1022	) {
		1023	# no body
		1024	$finish->("", undef, undef, 1);
		1025
		1026	} elsif (!$redirect && $arg{want_body_handle}) {
		1027	$_[0]->on_eof (undef);
		1028	$_[0]->on_error (undef);
		1029	$_[0]->on_read (undef);
		1030
		1031	$finish->(delete $state{handle});
		1032
		1033	} elsif ($chunked) {
		1034	my $cl = 0;
		1035	my $body = "";
		1036	my $on_body = $arg{on_body} \|\| sub { $body .= shift; 1 };
		1037
		1038	$state{read_chunk} = sub {
		1039	$_[1] =~ /^([0-9a-fA-F]+)/
		1040	or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1041
		1042	my $len = hex $1;
		1043
		1044	if ($len) {
		1045	$cl += $len;
		1046
		1047	$_[0]->push_read (chunk => $len, sub {
		1048	$on_body->($_[1], \%hdr)
		1049	or return $finish->(undef, 598 => "Request cancelled by on_body");
		1050
		1051	$_[0]->push_read (line => sub {
		1052	length $_[1]
		1053	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
		1054	$_[0]->push_read (line => $state{read_chunk});
773	$cb);	1055	});
		1056	});
774	} else {	1057	} else {
775	$cb->($_[0], \%hdr);	1058	$hdr{"content-length"} \|\|= $cl;
		1059
		1060	$_[0]->push_read (line => $qr_nlnl, sub {
		1061	if (length $_[1]) {
		1062	for ("$_[1]") {
		1063	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
		1064
		1065	my $hdr = _parse_hdr
		1066	or return $finish->(undef, $ae_error => "Garbled response trailers");
		1067
		1068	%hdr = (%hdr, %$hdr);
		1069	}
		1070	}
		1071
		1072	$finish->($body, undef, undef, 1);
		1073	});
776	}	1074	}
777	};	1075	};
778		1076
779	$ae_error = 597; # body phase
780
781	my $len = $hdr{"content-length"};
782
783	# body handling, many different code paths
784	# - no body expected
785	# - want_body_handle
786	# - te chunked
787	# - 2x length known (with or without on_body)
788	# - 2x length not known (with or without on_body)
789	if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) {
790	$finish->(undef, 598 => "Request cancelled by on_header");
791	} elsif (
792	$hdr{Status} =~ /^(?:1..\|204\|205\|304)$/
793	or $method eq "HEAD"
794	or (defined $len && $len == 0) # == 0, not !, because "0 " is true
795	) {
796	# no body
797	$finish->("", undef, undef, 1);
798
799	} elsif (!$redirect && $arg{want_body_handle}) {
800	$_[0]->on_eof (undef);
801	$_[0]->on_error (undef);
802	$_[0]->on_read (undef);
803
804	$finish->(delete $state{handle});
805
806	} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) {
807	my $cl = 0;
808	my $body = undef;
809	my $on_body = $arg{on_body} \|\| sub { $body .= shift; 1 };
810
811	$state{read_chunk} = sub {
812	$_[1] =~ /^([0-9a-fA-F]+)/
813	or $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
814
815	my $len = hex $1;
816
817	if ($len) {
818	$cl += $len;
819
820	$_[0]->push_read (chunk => $len, sub {
821	$on_body->($_[1], \%hdr)
822	or return $finish->(undef, 598 => "Request cancelled by on_body");
823
824	$_[0]->push_read (line => sub {
825	length $_[1]
826	and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding");
827	$_[0]->push_read (line => $state{read_chunk});
828	});
829	});
830	} else {
831	$hdr{"content-length"} \|\|= $cl;
832
833	$_[0]->push_read (line => $qr_nlnl, sub {
834	if (length $_[1]) {
835	for ("$_[1]") {
836	y/\015//d; # weed out any \015, as they show up in the weirdest of places.
837
838	my $hdr = parse_hdr
839	or return $finish->(undef, $ae_error => "Garbled response trailers");
840
841	%hdr = (%hdr, %$hdr);
842	}
843	}
844
845	$finish->($body, undef, undef, 1);
846	});
847	}
848	};
849
850	$_[0]->push_read (line => $state{read_chunk});	1077	$_[0]->push_read (line => $state{read_chunk});
851		1078
852	} elsif ($arg{on_body}) {	1079	} elsif ($arg{on_body}) {
853	if (defined $len) {	1080	if (defined $len) {
854	$_[0]->on_read (sub {	1081	$_[0]->on_read (sub {
855	$len -= length $_[0]{rbuf};	1082	$len -= length $_[0]{rbuf};
856		1083
857	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)	1084	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
858	or return $finish->(undef, 598 => "Request cancelled by on_body");	1085	or return $finish->(undef, 598 => "Request cancelled by on_body");
859		1086
860	$len > 0	1087	$len > 0
861	or $finish->("", undef, undef, 1);	1088	or $finish->("", undef, undef, 1);
862	});
863	} else {
864	$_[0]->on_eof (sub {
865	$finish->("");
866	});
867	$_[0]->on_read (sub {
868	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
869	or $finish->(undef, 598 => "Request cancelled by on_body");
870	});
871	}	1089	});
872	} else {	1090	} else {
873	$_[0]->on_eof (undef);	1091	$_[0]->on_eof (sub {
874		1092	$finish->("");
875	if (defined $len) {	1093	});
876	$_[0]->on_read (sub {	1094	$_[0]->on_read (sub {
877	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)	1095	$arg{on_body}(delete $_[0]{rbuf}, \%hdr)
878	if $len <= length $_[0]{rbuf};	1096	or $finish->(undef, 598 => "Request cancelled by on_body");
879	});
880	} else {
881	$_[0]->on_error (sub {
882	($! == Errno::EPIPE \|\| !$!)
883	? $finish->(delete $_[0]{rbuf})
884	: $finish->(undef, $ae_error => $_[2]);
885	});
886	$_[0]->on_read (sub { });
887	}	1097	});
888	}	1098	}
889	};
890
891	$state{handle}->push_read (line => $qr_nlnl, $state{read_response});
892	};
893
894	my $connect_cb = sub {
895	$state{fh} = shift
896	or do {
897	my $err = "$!";
898	%state = ();
899	return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err });
900	};
901
902	return unless delete $state{connect_guard};
903
904	# get handle
905	$state{handle} = new AnyEvent::Handle
906	fh => $state{fh},
907	peername => $rhost,
908	tls_ctx => $arg{tls_ctx},
909	# these need to be reconfigured on keepalive handles
910	timeout => $timeout,
911	on_error => sub {
912	%state = ();
913	$cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] });
914	},
915	on_eof => sub {
916	%state = ();
917	$cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" });
918	},
919	;
920
921	# limit the number of persistent connections
922	# keepalive not yet supported
923	# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) {
924	# ++$KA_COUNT{$_[1]};
925	# $state{handle}{ka_count_guard} = AnyEvent::Util::guard {
926	# --$KA_COUNT{$_[1]}
927	# };
928	# $hdr{connection} = "keep-alive";
929	# }
930
931	$state{handle}->starttls ("connect") if $rscheme eq "https";
932
933	# now handle proxy-CONNECT method
934	if ($proxy && $uscheme eq "https") {
935	# oh dear, we have to wrap it into a connect request
936
937	# maybe re-use $uauthority with patched port?
938	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012");
939	$state{handle}->push_read (line => $qr_nlnl, sub {
940	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
941	or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }));
942
943	if ($2 == 200) {
944	$rpath = $upath;
945	$handle_actual_request->();
946	} else {
947	%state = ();
948	$cb->(undef, { @pseudo, Status => $2, Reason => $3 });
949	}
950	});
951	} else {	1099	} else {
952	$handle_actual_request->();	1100	$_[0]->on_eof (undef);
		1101
		1102	if (defined $len) {
		1103	$_[0]->on_read (sub {
		1104	$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1)
		1105	if $len <= length $_[0]{rbuf};
		1106	});
		1107	} else {
		1108	$_[0]->on_error (sub {
		1109	($! == Errno::EPIPE \|\| !$!)
		1110	? $finish->(delete $_[0]{rbuf})
		1111	: $finish->(undef, $ae_error => $_[2]);
		1112	});
		1113	$_[0]->on_read (sub { });
		1114	}
953	}	1115	}
954	};	1116	};
955		1117
956	my $tcp_connect = $arg{tcp_connect}	1118	# if keepalive is enabled, then the server closing the connection
957	\|\| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };	1119	# before a response can happen legally - we retry on idempotent methods.
		1120	if ($was_persistent && $idempotent) {
		1121	my $old_eof = $hdl->{on_eof};
		1122	$hdl->{on_eof} = sub {
		1123	_destroy_state %state;
958		1124
959	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} \|\| sub { $timeout });	1125	%state = ();
		1126	$state{recurse} =
		1127	http_request (
		1128	$method => $url,
		1129	%arg,
		1130	recurse => $recurse - 1,
		1131	persistent => 0,
		1132	sub {
		1133	%state = ();
		1134	&$cb
		1135	}
		1136	);
		1137	};
		1138	$hdl->on_read (sub {
		1139	return unless %state;
		1140
		1141	# as soon as we receive something, a connection close
		1142	# once more becomes a hard error
		1143	$hdl->{on_eof} = $old_eof;
		1144	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1145	});
		1146	} else {
		1147	$hdl->push_read (line => $qr_nlnl, $state{read_response});
		1148	}
960	};	1149	};
961		1150
		1151	my $prepare_handle = sub {
		1152	my ($hdl) = $state{handle};
		1153
		1154	$hdl->on_error (sub {
		1155	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] };
		1156	});
		1157	$hdl->on_eof (sub {
		1158	_error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" };
		1159	});
		1160	$hdl->timeout_reset;
		1161	$hdl->timeout ($timeout);
		1162	};
		1163
		1164	# connected to proxy (or origin server)
		1165	my $connect_cb = sub {
		1166	my $fh = shift
		1167	or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" };
		1168
		1169	return unless delete $state{connect_guard};
		1170
		1171	# get handle
		1172	$state{handle} = new AnyEvent::Handle
		1173	%{ $arg{handle_params} },
		1174	fh => $fh,
		1175	peername => $uhost,
		1176	tls_ctx => $arg{tls_ctx},
		1177	;
		1178
		1179	$prepare_handle->();
		1180
		1181	#$state{handle}->starttls ("connect") if $rscheme eq "https";
		1182
		1183	# now handle proxy-CONNECT method
		1184	if ($proxy && $uscheme eq "https") {
		1185	# oh dear, we have to wrap it into a connect request
		1186
		1187	my $auth = exists $hdr{"proxy-authorization"}
		1188	? "proxy-authorization: " . (delete $hdr{"proxy-authorization"}) . "\015\012"
		1189	: "";
		1190
		1191	# maybe re-use $uauthority with patched port?
		1192	$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012$auth\015\012");
		1193	$state{handle}->push_read (line => $qr_nlnl, sub {
		1194	$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix
		1195	or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" };
		1196
		1197	if ($2 == 200) {
		1198	$rpath = $upath;
		1199	$handle_actual_request->();
		1200	} else {
		1201	_error %state, $cb, { @pseudo, Status => $2, Reason => $3 };
		1202	}
		1203	});
		1204	} else {
		1205	delete $hdr{"proxy-authorization"} unless $proxy;
		1206
		1207	$handle_actual_request->();
		1208	}
		1209	};
		1210
		1211	_get_slot $uhost, sub {
		1212	$state{slot_guard} = shift;
		1213
		1214	return unless $state{connect_guard};
		1215
		1216	# try to use an existing keepalive connection, but only if we, ourselves, plan
		1217	# on a keepalive request (in theory, this should be a separate config option).
		1218	if ($persistent && $KA_CACHE{$ka_key}) {
		1219	$was_persistent = 1;
		1220
		1221	$state{handle} = ka_fetch $ka_key;
		1222	$state{handle}->destroyed
		1223	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d#
		1224	$prepare_handle->();
		1225	$state{handle}->destroyed
		1226	and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d#
		1227	$handle_actual_request->();
		1228
		1229	} else {
		1230	my $tcp_connect = $arg{tcp_connect}
		1231	\|\| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect };
		1232
		1233	$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} \|\| sub { $timeout });
		1234	}
		1235	};
		1236
962	defined wantarray && AnyEvent::Util::guard { %state = () }	1237	defined wantarray && AnyEvent::Util::guard { _destroy_state %state }
963	}	1238	}
964		1239
965	sub http_get($@) {	1240	sub http_get($@) {
966	unshift @_, "GET";	1241	unshift @_, "GET";
967	&http_request	1242	&http_request
…		…
985	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for	1260	AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for
986	the actual connection, which in turn uses AnyEvent::DNS to resolve	1261	the actual connection, which in turn uses AnyEvent::DNS to resolve
987	hostnames. The latter is a simple stub resolver and does no caching	1262	hostnames. The latter is a simple stub resolver and does no caching
988	on its own. If you want DNS caching, you currently have to provide	1263	on its own. If you want DNS caching, you currently have to provide
989	your own default resolver (by storing a suitable resolver object in	1264	your own default resolver (by storing a suitable resolver object in
990	C<$AnyEvent::DNS::RESOLVER>).	1265	C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback.
991		1266
992	=head2 GLOBAL FUNCTIONS AND VARIABLES	1267	=head2 GLOBAL FUNCTIONS AND VARIABLES
993		1268
994	=over 4	1269	=over 4
995		1270
996	=item AnyEvent::HTTP::set_proxy "proxy-url"	1271	=item AnyEvent::HTTP::set_proxy "proxy-url"
997		1272
998	Sets the default proxy server to use. The proxy-url must begin with a	1273	Sets the default proxy server to use. The proxy-url must begin with a
999	string of the form C<http://host:port> (optionally C<https:...>), croaks	1274	string of the form C<http://host:port>, croaks otherwise.
1000	otherwise.
1001		1275
1002	To clear an already-set proxy, use C<undef>.	1276	To clear an already-set proxy, use C<undef>.
		1277
		1278	When AnyEvent::HTTP is loaded for the first time it will query the
		1279	default proxy from the operating system, currently by looking at
		1280	C<$ENV{http_proxy>}.
1003		1281
1004	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]	1282	=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end]
1005		1283
1006	Remove all cookies from the cookie jar that have been expired. If	1284	Remove all cookies from the cookie jar that have been expired. If
1007	C<$session_end> is given and true, then additionally remove all session	1285	C<$session_end> is given and true, then additionally remove all session
1008	cookies.	1286	cookies.
1009		1287
1010	You should call this function (with a true C<$session_end>) before you	1288	You should call this function (with a true C<$session_end>) before you
1011	save cookies to disk, and you should call this function after loading them	1289	save cookies to disk, and you should call this function after loading them
1012	again. If you have a long-running program you can additonally call this	1290	again. If you have a long-running program you can additionally call this
1013	function from time to time.	1291	function from time to time.
1014		1292
1015	A cookie jar is initially an empty hash-reference that is managed by this	1293	A cookie jar is initially an empty hash-reference that is managed by this
1016	module. It's format is subject to change, but currently it is like this:	1294	module. Its format is subject to change, but currently it is as follows:
1017		1295
1018	The key C<version> has to contain C<1>, otherwise the hash gets	1296	The key C<version> has to contain C<1>, otherwise the hash gets
1019	emptied. All other keys are hostnames or IP addresses pointing to	1297	emptied. All other keys are hostnames or IP addresses pointing to
1020	hash-references. The key for these inner hash references is the	1298	hash-references. The key for these inner hash references is the
1021	server path for which this cookie is meant, and the values are again	1299	server path for which this cookie is meant, and the values are again
1022	hash-references. The keys of those hash-references is the cookie name, and	1300	hash-references. Each key of those hash-references is a cookie name, and
1023	the value, you guessed it, is another hash-reference, this time with the	1301	the value, you guessed it, is another hash-reference, this time with the
1024	key-value pairs from the cookie, except for C<expires> and C<max-age>,	1302	key-value pairs from the cookie, except for C<expires> and C<max-age>,
1025	which have been replaced by a C<_expires> key that contains the cookie	1303	which have been replaced by a C<_expires> key that contains the cookie
1026	expiry timestamp.	1304	expiry timestamp. Session cookies are indicated by not having an
		1305	C<_expires> key.
1027		1306
1028	Here is an example of a cookie jar with a single cookie, so you have a	1307	Here is an example of a cookie jar with a single cookie, so you have a
1029	chance of understanding the above paragraph:	1308	chance of understanding the above paragraph:
1030		1309
1031	{	1310	{
…		…
1053		1332
1054	=item $AnyEvent::HTTP::MAX_RECURSE	1333	=item $AnyEvent::HTTP::MAX_RECURSE
1055		1334
1056	The default value for the C<recurse> request parameter (default: C<10>).	1335	The default value for the C<recurse> request parameter (default: C<10>).
1057		1336
		1337	=item $AnyEvent::HTTP::TIMEOUT
		1338
		1339	The default timeout for connection operations (default: C<300>).
		1340
1058	=item $AnyEvent::HTTP::USERAGENT	1341	=item $AnyEvent::HTTP::USERAGENT
1059		1342
1060	The default value for the C<User-Agent> header (the default is	1343	The default value for the C<User-Agent> header (the default is
1061	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).	1344	C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>).
1062		1345
1063	=item $AnyEvent::HTTP::MAX_PER_HOST	1346	=item $AnyEvent::HTTP::MAX_PER_HOST
1064		1347
1065	The maximum number of concurrent connections to the same host (identified	1348	The maximum number of concurrent connections to the same host (identified
1066	by the hostname). If the limit is exceeded, then the additional requests	1349	by the hostname). If the limit is exceeded, then additional requests
1067	are queued until previous connections are closed.	1350	are queued until previous connections are closed. Both persistent and
		1351	non-persistent connections are counted in this limit.
1068		1352
1069	The default value for this is C<4>, and it is highly advisable to not	1353	The default value for this is C<4>, and it is highly advisable to not
1070	increase it.	1354	increase it much.
		1355
		1356	For comparison: the RFC's recommend 4 non-persistent or 2 persistent
		1357	connections, older browsers used 2, newer ones (such as firefox 3)
		1358	typically use 6, and Opera uses 8 because like, they have the fastest
		1359	browser and give a shit for everybody else on the planet.
		1360
		1361	=item $AnyEvent::HTTP::PERSISTENT_TIMEOUT
		1362
		1363	The time after which idle persistent connections get closed by
		1364	AnyEvent::HTTP (default: C<3>).
1071		1365
1072	=item $AnyEvent::HTTP::ACTIVE	1366	=item $AnyEvent::HTTP::ACTIVE
1073		1367
1074	The number of active connections. This is not the number of currently	1368	The number of active connections. This is not the number of currently
1075	running requests, but the number of currently open and non-idle TCP	1369	running requests, but the number of currently open and non-idle TCP
1076	connections. This number of can be useful for load-leveling.	1370	connections. This number can be useful for load-leveling.
1077		1371
1078	=back	1372	=back
1079		1373
1080	=cut	1374	=cut
1081		1375
…		…
1115	# other formats fail in the loop below	1409	# other formats fail in the loop below
1116		1410
1117	for (0..11) {	1411	for (0..11) {
1118	if ($m eq $month[$_]) {	1412	if ($m eq $month[$_]) {
1119	require Time::Local;	1413	require Time::Local;
1120	return Time::Local::timegm ($S, $M, $H, $d, $_, $y);	1414	return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) };
1121	}	1415	}
1122	}	1416	}
1123		1417
1124	undef	1418	undef
1125	}	1419	}
1126		1420
1127	sub set_proxy($) {	1421	sub set_proxy($) {
1128	if (length $_[0]) {	1422	if (length $_[0]) {
1129	$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix	1423	$_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix
1130	or Carp::croak "$_[0]: invalid proxy URL";	1424	or Carp::croak "$_[0]: invalid proxy URL";
1131	$PROXY = [$2, $3 \|\| 3128, $1]	1425	$PROXY = [$2, $3 \|\| 3128, $1]
1132	} else {	1426	} else {
1133	undef $PROXY;	1427	undef $PROXY;
1134	}	1428	}
…		…
1137	# initialise proxy from environment	1431	# initialise proxy from environment
1138	eval {	1432	eval {
1139	set_proxy $ENV{http_proxy};	1433	set_proxy $ENV{http_proxy};
1140	};	1434	};
1141		1435
		1436	=head2 SHOWCASE
		1437
		1438	This section contains some more elaborate "real-world" examples or code
		1439	snippets.
		1440
		1441	=head2 HTTP/1.1 FILE DOWNLOAD
		1442
		1443	Downloading files with HTTP can be quite tricky, especially when something
		1444	goes wrong and you want to resume.
		1445
		1446	Here is a function that initiates and resumes a download. It uses the
		1447	last modified time to check for file content changes, and works with many
		1448	HTTP/1.0 servers as well, and usually falls back to a complete re-download
		1449	on older servers.
		1450
		1451	It calls the completion callback with either C<undef>, which means a
		1452	nonretryable error occurred, C<0> when the download was partial and should
		1453	be retried, and C<1> if it was successful.
		1454
		1455	use AnyEvent::HTTP;
		1456
		1457	sub download($$$) {
		1458	my ($url, $file, $cb) = @_;
		1459
		1460	open my $fh, "+<", $file
		1461	or die "$file: $!";
		1462
		1463	my %hdr;
		1464	my $ofs = 0;
		1465
		1466	warn stat $fh;
		1467	warn -s _;
		1468	if (stat $fh and -s _) {
		1469	$ofs = -s _;
		1470	warn "-s is ", $ofs;
		1471	$hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9];
		1472	$hdr{"range"} = "bytes=$ofs-";
		1473	}
		1474
		1475	http_get $url,
		1476	headers => \%hdr,
		1477	on_header => sub {
		1478	my ($hdr) = @_;
		1479
		1480	if ($hdr->{Status} == 200 && $ofs) {
		1481	# resume failed
		1482	truncate $fh, $ofs = 0;
		1483	}
		1484
		1485	sysseek $fh, $ofs, 0;
		1486
		1487	1
		1488	},
		1489	on_body => sub {
		1490	my ($data, $hdr) = @_;
		1491
		1492	if ($hdr->{Status} =~ /^2/) {
		1493	length $data == syswrite $fh, $data
		1494	or return; # abort on write errors
		1495	}
		1496
		1497	1
		1498	},
		1499	sub {
		1500	my (undef, $hdr) = @_;
		1501
		1502	my $status = $hdr->{Status};
		1503
		1504	if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) {
		1505	utime $fh, $time, $time;
		1506	}
		1507
		1508	if ($status == 200 \|\| $status == 206 \|\| $status == 416) {
		1509	# download ok \|\| resume ok \|\| file already fully downloaded
		1510	$cb->(1, $hdr);
		1511
		1512	} elsif ($status == 412) {
		1513	# file has changed while resuming, delete and retry
		1514	unlink $file;
		1515	$cb->(0, $hdr);
		1516
		1517	} elsif ($status == 500 or $status == 503 or $status =~ /^59/) {
		1518	# retry later
		1519	$cb->(0, $hdr);
		1520
		1521	} else {
		1522	$cb->(undef, $hdr);
		1523	}
		1524	}
		1525	;
		1526	}
		1527
		1528	download "http://server/somelargefile", "/tmp/somelargefile", sub {
		1529	if ($_[0]) {
		1530	print "OK!\n";
		1531	} elsif (defined $_[0]) {
		1532	print "please retry later\n";
		1533	} else {
		1534	print "ERROR\n";
		1535	}
		1536	};
		1537
1142	=head2 SOCKS PROXIES	1538	=head3 SOCKS PROXIES
1143		1539
1144	Socks proxies are not directly supported by AnyEvent::HTTP. You can	1540	Socks proxies are not directly supported by AnyEvent::HTTP. You can
1145	compile your perl to support socks, or use an external program such as	1541	compile your perl to support socks, or use an external program such as
1146	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy	1542	F<socksify> (dante) or F<tsocks> to make your program use a socks proxy
1147	transparently.	1543	transparently.

Diff Legend

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

Comparing AnyEvent-HTTP/HTTP.pm (file contents): Revision 1.87 by root, Sun Jan 2 08:51:53 2011 UTC vs. Revision 1.124 by root, Fri May 8 21:26:16 2015 UTC

Diff Legend

Comparing AnyEvent-HTTP/HTTP.pm (file contents):
Revision 1.87 by root, Sun Jan 2 08:51:53 2011 UTC vs.
Revision 1.124 by root, Fri May 8 21:26:16 2015 UTC