… | |
… | |
122 | |
122 | |
123 | If the server sends a header multiple times, then their contents will be |
123 | If the server sends a header multiple times, then their contents will be |
124 | joined together with a comma (C<,>), as per the HTTP spec. |
124 | joined together with a comma (C<,>), as per the HTTP spec. |
125 | |
125 | |
126 | If an internal error occurs, such as not being able to resolve a hostname, |
126 | If an internal error occurs, such as not being able to resolve a hostname, |
127 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x> |
127 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be |
128 | (usually C<599>) and the C<Reason> pseudo-header will contain an error |
128 | C<590>-C<599> and the C<Reason> pseudo-header will contain an error |
129 | message. |
129 | message. Currently the following status codes are used: |
|
|
130 | |
|
|
131 | =over 4 |
|
|
132 | |
|
|
133 | =item 595 - errors during connection etsbalishment, proxy handshake. |
|
|
134 | |
|
|
135 | =item 596 - errors during TLS negotiation, request sending and header processing. |
|
|
136 | |
|
|
137 | =item 597 - errors during body receive or processing. |
|
|
138 | |
|
|
139 | =item 598 - user aborted request in C<on_header> or C<on_body>. |
|
|
140 | |
|
|
141 | =item 599 - other, usually nonretryable, errors (garbled URL etc.). |
|
|
142 | |
|
|
143 | =back |
130 | |
144 | |
131 | A typical callback might look like this: |
145 | A typical callback might look like this: |
132 | |
146 | |
133 | sub { |
147 | sub { |
134 | my ($body, $hdr) = @_; |
148 | my ($body, $hdr) = @_; |
… | |
… | |
412 | |
426 | |
413 | \@cookies |
427 | \@cookies |
414 | } |
428 | } |
415 | |
429 | |
416 | # parse set_cookie header into jar |
430 | # parse set_cookie header into jar |
417 | sub cookie_jar_set_cookie($$) { |
431 | sub cookie_jar_set_cookie($$$) { |
418 | my ($jar, $set_cookie) = @_; |
432 | my ($jar, $set_cookie, $uhost) = @_; |
419 | |
433 | |
420 | for ($set_cookie) { |
434 | for ($set_cookie) { |
421 | # parse NAME=VALUE |
435 | # parse NAME=VALUE |
422 | my @kv; |
436 | my @kv; |
423 | |
437 | |
… | |
… | |
472 | } else { |
486 | } else { |
473 | $cdom = $uhost; |
487 | $cdom = $uhost; |
474 | } |
488 | } |
475 | |
489 | |
476 | # store it |
490 | # store it |
477 | $arg{cookie_jar}{version} = 1; |
491 | $jar->{version} = 1; |
478 | $arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv; |
492 | $jar->{$cdom}{$cpath}{$name} = \%kv; |
479 | |
493 | |
480 | redo if /\G\s*,/gc; |
494 | redo if /\G\s*,/gc; |
481 | } |
495 | } |
482 | } |
|
|
483 | } |
496 | } |
484 | |
497 | |
485 | # continue to parse $_ for headers and place them into the arg |
498 | # continue to parse $_ for headers and place them into the arg |
486 | sub parse_hdr() { |
499 | sub parse_hdr() { |
487 | my %hdr; |
500 | my %hdr; |
… | |
… | |
601 | _get_slot $uhost, sub { |
614 | _get_slot $uhost, sub { |
602 | $state{slot_guard} = shift; |
615 | $state{slot_guard} = shift; |
603 | |
616 | |
604 | return unless $state{connect_guard}; |
617 | return unless $state{connect_guard}; |
605 | |
618 | |
|
|
619 | my $ae_error = 595; # connecting |
|
|
620 | |
606 | my $connect_cb = sub { |
621 | my $connect_cb = sub { |
607 | $state{fh} = shift |
622 | $state{fh} = shift |
608 | or do { |
623 | or do { |
609 | my $err = "$!"; |
624 | my $err = "$!"; |
610 | %state = (); |
625 | %state = (); |
611 | return $cb->(undef, { @pseudo, Status => 599, Reason => $err }); |
626 | return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err }); |
612 | }; |
627 | }; |
613 | |
|
|
614 | pop; # free memory, save a tree |
|
|
615 | |
628 | |
616 | return unless delete $state{connect_guard}; |
629 | return unless delete $state{connect_guard}; |
617 | |
630 | |
618 | # get handle |
631 | # get handle |
619 | $state{handle} = new AnyEvent::Handle |
632 | $state{handle} = new AnyEvent::Handle |
… | |
… | |
622 | tls_ctx => $arg{tls_ctx}, |
635 | tls_ctx => $arg{tls_ctx}, |
623 | # these need to be reconfigured on keepalive handles |
636 | # these need to be reconfigured on keepalive handles |
624 | timeout => $timeout, |
637 | timeout => $timeout, |
625 | on_error => sub { |
638 | on_error => sub { |
626 | %state = (); |
639 | %state = (); |
627 | $cb->(undef, { @pseudo, Status => 599, Reason => $_[2] }); |
640 | $cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] }); |
628 | }, |
641 | }, |
629 | on_eof => sub { |
642 | on_eof => sub { |
630 | %state = (); |
643 | %state = (); |
631 | $cb->(undef, { @pseudo, Status => 599, Reason => "Unexpected end-of-file" }); |
644 | $cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" }); |
632 | }, |
645 | }, |
633 | ; |
646 | ; |
634 | |
647 | |
635 | # limit the number of persistent connections |
648 | # limit the number of persistent connections |
636 | # keepalive not yet supported |
649 | # keepalive not yet supported |
… | |
… | |
644 | |
657 | |
645 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
658 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
646 | |
659 | |
647 | # handle actual, non-tunneled, request |
660 | # handle actual, non-tunneled, request |
648 | my $handle_actual_request = sub { |
661 | my $handle_actual_request = sub { |
|
|
662 | $ae_error = 596; # request phase |
|
|
663 | |
649 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
664 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
650 | |
665 | |
651 | # send request |
666 | # send request |
652 | $state{handle}->push_write ( |
667 | $state{handle}->push_write ( |
653 | "$method $rpath HTTP/1.1\015\012" |
668 | "$method $rpath HTTP/1.1\015\012" |
… | |
… | |
664 | # status line and headers |
679 | # status line and headers |
665 | $state{read_response} = sub { |
680 | $state{read_response} = sub { |
666 | for ("$_[1]") { |
681 | for ("$_[1]") { |
667 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
682 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
668 | |
683 | |
669 | /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/igxc |
684 | /^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci |
670 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" })); |
685 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" })); |
671 | |
686 | |
672 | # 100 Continue handling |
687 | # 100 Continue handling |
673 | # should not happen as we don't send expect: 100-continue, |
688 | # should not happen as we don't send expect: 100-continue, |
674 | # but we handle it just in case. |
689 | # but we handle it just in case. |
… | |
… | |
709 | |
724 | |
710 | if ($recurse) { |
725 | if ($recurse) { |
711 | my $status = $hdr{Status}; |
726 | my $status = $hdr{Status}; |
712 | |
727 | |
713 | # industry standard is to redirect POST as GET for |
728 | # industry standard is to redirect POST as GET for |
714 | # 301, 302 and 303, in contrast to http/1.0 and 1.1. |
729 | # 301, 302 and 303, in contrast to HTTP/1.0 and 1.1. |
715 | # also, the UA should ask the user for 301 and 307 and POST, |
730 | # also, the UA should ask the user for 301 and 307 and POST, |
716 | # industry standard seems to be to simply follow. |
731 | # industry standard seems to be to simply follow. |
717 | # we go with the industry standard. |
732 | # we go with the industry standard. |
718 | if ($status == 301 or $status == 302 or $status == 303) { |
733 | if ($status == 301 or $status == 302 or $status == 303) { |
719 | # HTTP/1.1 is unclear on how to mutate the method |
734 | # HTTP/1.1 is unclear on how to mutate the method |
… | |
… | |
723 | $redirect = 1; |
738 | $redirect = 1; |
724 | } |
739 | } |
725 | } |
740 | } |
726 | |
741 | |
727 | my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive]) |
742 | my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive]) |
728 | my $keepalive = pop; |
743 | my $may_keep_alive = $_[3]; |
729 | |
744 | |
730 | $state{handle}->destroy if $state{handle}; |
745 | $state{handle}->destroy if $state{handle}; |
731 | %state = (); |
746 | %state = (); |
732 | |
747 | |
733 | if (defined $_[1]) { |
748 | if (defined $_[1]) { |
… | |
… | |
735 | $hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2]; |
750 | $hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2]; |
736 | } |
751 | } |
737 | |
752 | |
738 | # set-cookie processing |
753 | # set-cookie processing |
739 | if ($arg{cookie_jar}) { |
754 | if ($arg{cookie_jar}) { |
740 | cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}; |
755 | cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost; |
|
|
756 | } |
741 | |
757 | |
742 | if ($redirect && exists $hdr{location}) { |
758 | if ($redirect && exists $hdr{location}) { |
743 | # we ignore any errors, as it is very common to receive |
759 | # we ignore any errors, as it is very common to receive |
744 | # Content-Length != 0 but no actual body |
760 | # Content-Length != 0 but no actual body |
745 | # we also access %hdr, as $_[1] might be an erro |
761 | # we also access %hdr, as $_[1] might be an erro |
… | |
… | |
751 | $cb); |
767 | $cb); |
752 | } else { |
768 | } else { |
753 | $cb->($_[0], \%hdr); |
769 | $cb->($_[0], \%hdr); |
754 | } |
770 | } |
755 | }; |
771 | }; |
|
|
772 | |
|
|
773 | $ae_error = 597; # body phase |
756 | |
774 | |
757 | my $len = $hdr{"content-length"}; |
775 | my $len = $hdr{"content-length"}; |
758 | |
776 | |
759 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
777 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
760 | $finish->(undef, 598 => "Request cancelled by on_header"); |
778 | $finish->(undef, 598 => "Request cancelled by on_header"); |
… | |
… | |
782 | } elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) { |
800 | } elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) { |
783 | my $cl = 0; |
801 | my $cl = 0; |
784 | my $body = undef; |
802 | my $body = undef; |
785 | my $on_body = $arg{on_body} || sub { $body .= shift; 1 }; |
803 | my $on_body = $arg{on_body} || sub { $body .= shift; 1 }; |
786 | |
804 | |
787 | $_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) }); |
|
|
788 | |
|
|
789 | my $read_chunk; $read_chunk = sub { |
805 | my $read_chunk; $read_chunk = sub { |
790 | $_[1] =~ /^([0-9a-fA-F]+)/ |
806 | $_[1] =~ /^([0-9a-fA-F]+)/ |
791 | or $finish->(undef, 599 => "Garbled chunked transfer encoding"); |
807 | or $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
792 | |
808 | |
793 | my $len = hex $1; |
809 | my $len = hex $1; |
794 | |
810 | |
795 | if ($len) { |
811 | if ($len) { |
796 | $cl += $len; |
812 | $cl += $len; |
… | |
… | |
799 | $on_body->($_[1], \%hdr) |
815 | $on_body->($_[1], \%hdr) |
800 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
816 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
801 | |
817 | |
802 | $_[0]->push_read (line => sub { |
818 | $_[0]->push_read (line => sub { |
803 | length $_[1] |
819 | length $_[1] |
804 | and return $finish->(undef, 599 => "Garbled chunked transfer encoding"); |
820 | and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
805 | $_[0]->push_read (line => $read_chunk); |
821 | $_[0]->push_read (line => $read_chunk); |
806 | }); |
822 | }); |
807 | }); |
823 | }); |
808 | } else { |
824 | } else { |
809 | $hdr{"content-length"} ||= $cl; |
825 | $hdr{"content-length"} ||= $cl; |
… | |
… | |
812 | if (length $_[1]) { |
828 | if (length $_[1]) { |
813 | for ("$_[1]") { |
829 | for ("$_[1]") { |
814 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
830 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
815 | |
831 | |
816 | my $hdr = parse_hdr |
832 | my $hdr = parse_hdr |
817 | or return $finish->(undef, 599 => "Garbled response trailers"); |
833 | or return $finish->(undef, $ae_error => "Garbled response trailers"); |
818 | |
834 | |
819 | %hdr = (%hdr, %$hdr); |
835 | %hdr = (%hdr, %$hdr); |
820 | } |
836 | } |
821 | } |
837 | } |
822 | |
838 | |
… | |
… | |
826 | }; |
842 | }; |
827 | |
843 | |
828 | $_[0]->push_read (line => $read_chunk); |
844 | $_[0]->push_read (line => $read_chunk); |
829 | |
845 | |
830 | } elsif ($arg{on_body}) { |
846 | } elsif ($arg{on_body}) { |
831 | $_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) }); |
|
|
832 | |
|
|
833 | if ($len) { |
847 | if ($len) { |
834 | $_[0]->on_read (sub { |
848 | $_[0]->on_read (sub { |
835 | $len -= length $_[0]{rbuf}; |
849 | $len -= length $_[0]{rbuf}; |
836 | |
850 | |
837 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
851 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
… | |
… | |
851 | } |
865 | } |
852 | } else { |
866 | } else { |
853 | $_[0]->on_eof (undef); |
867 | $_[0]->on_eof (undef); |
854 | |
868 | |
855 | if ($len) { |
869 | if ($len) { |
856 | $_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) }); |
|
|
857 | $_[0]->on_read (sub { |
870 | $_[0]->on_read (sub { |
858 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1) |
871 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1) |
859 | if $len <= length $_[0]{rbuf}; |
872 | if $len <= length $_[0]{rbuf}; |
860 | }); |
873 | }); |
861 | } else { |
874 | } else { |
862 | $_[0]->on_error (sub { |
875 | $_[0]->on_error (sub { |
863 | ($! == Errno::EPIPE || !$!) |
876 | ($! == Errno::EPIPE || !$!) |
864 | ? $finish->(delete $_[0]{rbuf}) |
877 | ? $finish->(delete $_[0]{rbuf}) |
865 | : $finish->(undef, 599 => $_[2]); |
878 | : $finish->(undef, $ae_error => $_[2]); |
866 | }); |
879 | }); |
867 | $_[0]->on_read (sub { }); |
880 | $_[0]->on_read (sub { }); |
868 | } |
881 | } |
869 | } |
882 | } |
870 | } |
883 | } |