… | |
… | |
114 | =over 4 |
114 | =over 4 |
115 | |
115 | |
116 | =item on_prepare => $cb->($handle) |
116 | =item on_prepare => $cb->($handle) |
117 | |
117 | |
118 | This (rarely used) callback is called before a new connection is |
118 | This (rarely used) callback is called before a new connection is |
119 | attempted, but after the file handle has been created. It could be used to |
119 | attempted, but after the file handle has been created (you can access that |
|
|
120 | file handle via C<< $handle->{fh} >>). It could be used to prepare the |
120 | prepare the file handle with parameters required for the actual connect |
121 | file handle with parameters required for the actual connect (as opposed to |
121 | (as opposed to settings that can be changed when the connection is already |
122 | settings that can be changed when the connection is already established). |
122 | established). |
|
|
123 | |
123 | |
124 | The return value of this callback should be the connect timeout value in |
124 | The return value of this callback should be the connect timeout value in |
125 | seconds (or C<0>, or C<undef>, or the empty list, to indicate that the |
125 | seconds (or C<0>, or C<undef>, or the empty list, to indicate that the |
126 | default timeout is to be used). |
126 | default timeout is to be used). |
127 | |
127 | |
… | |
… | |
247 | many seconds pass without a successful read or write on the underlying |
247 | many seconds pass without a successful read or write on the underlying |
248 | file handle (or a call to C<timeout_reset>), the C<on_timeout> callback |
248 | file handle (or a call to C<timeout_reset>), the C<on_timeout> callback |
249 | will be invoked (and if that one is missing, a non-fatal C<ETIMEDOUT> |
249 | will be invoked (and if that one is missing, a non-fatal C<ETIMEDOUT> |
250 | error will be raised). |
250 | error will be raised). |
251 | |
251 | |
252 | There are three variants of the timeouts that work independently |
252 | There are three variants of the timeouts that work independently of each |
253 | of each other, for both read and write, just read, and just write: |
253 | other, for both read and write (triggered when nothing was read I<OR> |
|
|
254 | written), just read (triggered when nothing was read), and just write: |
254 | C<timeout>, C<rtimeout> and C<wtimeout>, with corresponding callbacks |
255 | C<timeout>, C<rtimeout> and C<wtimeout>, with corresponding callbacks |
255 | C<on_timeout>, C<on_rtimeout> and C<on_wtimeout>, and reset functions |
256 | C<on_timeout>, C<on_rtimeout> and C<on_wtimeout>, and reset functions |
256 | C<timeout_reset>, C<rtimeout_reset>, and C<wtimeout_reset>. |
257 | C<timeout_reset>, C<rtimeout_reset>, and C<wtimeout_reset>. |
257 | |
258 | |
258 | Note that timeout processing is active even when you do not have |
259 | Note that timeout processing is active even when you do not have any |
259 | any outstanding read or write requests: If you plan to keep the connection |
260 | outstanding read or write requests: If you plan to keep the connection |
260 | idle then you should disable the timeout temporarily or ignore the timeout |
261 | idle then you should disable the timeout temporarily or ignore the |
261 | in the C<on_timeout> callback, in which case AnyEvent::Handle will simply |
262 | timeout in the corresponding C<on_timeout> callback, in which case |
262 | restart the timeout. |
263 | AnyEvent::Handle will simply restart the timeout. |
263 | |
264 | |
264 | Zero (the default) disables this timeout. |
265 | Zero (the default) disables the corresponding timeout. |
265 | |
266 | |
266 | =item on_timeout => $cb->($handle) |
267 | =item on_timeout => $cb->($handle) |
|
|
268 | |
|
|
269 | =item on_rtimeout => $cb->($handle) |
|
|
270 | |
|
|
271 | =item on_wtimeout => $cb->($handle) |
267 | |
272 | |
268 | Called whenever the inactivity timeout passes. If you return from this |
273 | Called whenever the inactivity timeout passes. If you return from this |
269 | callback, then the timeout will be reset as if some activity had happened, |
274 | callback, then the timeout will be reset as if some activity had happened, |
270 | so this condition is not fatal in any way. |
275 | so this condition is not fatal in any way. |
271 | |
276 | |
… | |
… | |
278 | For example, a server accepting connections from untrusted sources should |
283 | For example, a server accepting connections from untrusted sources should |
279 | be configured to accept only so-and-so much data that it cannot act on |
284 | be configured to accept only so-and-so much data that it cannot act on |
280 | (for example, when expecting a line, an attacker could send an unlimited |
285 | (for example, when expecting a line, an attacker could send an unlimited |
281 | amount of data without a callback ever being called as long as the line |
286 | amount of data without a callback ever being called as long as the line |
282 | isn't finished). |
287 | isn't finished). |
|
|
288 | |
|
|
289 | =item wbuf_max => <bytes> |
|
|
290 | |
|
|
291 | If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) |
|
|
292 | when the write buffer ever (strictly) exceeds this size. This is useful to |
|
|
293 | avoid some forms of denial-of-service attacks. |
|
|
294 | |
|
|
295 | Although the units of this parameter is bytes, this is the I<raw> number |
|
|
296 | of bytes not yet accepted by the kernel. This can make a difference when |
|
|
297 | you e.g. use TLS, as TLS typically makes your write data larger (but it |
|
|
298 | can also make it smaller due to compression). |
|
|
299 | |
|
|
300 | As an example of when this limit is useful, take a chat server that sends |
|
|
301 | chat messages to a client. If the client does not read those in a timely |
|
|
302 | manner then the send buffer in the server would grow unbounded. |
283 | |
303 | |
284 | =item autocork => <boolean> |
304 | =item autocork => <boolean> |
285 | |
305 | |
286 | When disabled (the default), C<push_write> will try to immediately |
306 | When disabled (the default), C<push_write> will try to immediately |
287 | write the data to the handle if possible. This avoids having to register |
307 | write the data to the handle if possible. This avoids having to register |
… | |
… | |
422 | Use the C<< ->starttls >> method if you need to start TLS negotiation later. |
442 | Use the C<< ->starttls >> method if you need to start TLS negotiation later. |
423 | |
443 | |
424 | =item tls_ctx => $anyevent_tls |
444 | =item tls_ctx => $anyevent_tls |
425 | |
445 | |
426 | Use the given C<AnyEvent::TLS> object to create the new TLS connection |
446 | Use the given C<AnyEvent::TLS> object to create the new TLS connection |
427 | (unless a connection object was specified directly). If this parameter is |
447 | (unless a connection object was specified directly). If this |
428 | missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. |
448 | parameter is missing (or C<undef>), then AnyEvent::Handle will use |
|
|
449 | C<AnyEvent::Handle::TLS_CTX>. |
429 | |
450 | |
430 | Instead of an object, you can also specify a hash reference with C<< key |
451 | Instead of an object, you can also specify a hash reference with C<< key |
431 | => value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a |
452 | => value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a |
432 | new TLS context object. |
453 | new TLS context object. |
433 | |
454 | |
… | |
… | |
502 | $self->{connect}[0], |
523 | $self->{connect}[0], |
503 | $self->{connect}[1], |
524 | $self->{connect}[1], |
504 | sub { |
525 | sub { |
505 | my ($fh, $host, $port, $retry) = @_; |
526 | my ($fh, $host, $port, $retry) = @_; |
506 | |
527 | |
507 | delete $self->{_connect}; |
528 | delete $self->{_connect}; # no longer needed |
508 | |
529 | |
509 | if ($fh) { |
530 | if ($fh) { |
510 | $self->{fh} = $fh; |
531 | $self->{fh} = $fh; |
511 | |
532 | |
512 | delete $self->{_skip_drain_rbuf}; |
533 | delete $self->{_skip_drain_rbuf}; |
… | |
… | |
520 | }); |
541 | }); |
521 | |
542 | |
522 | } else { |
543 | } else { |
523 | if ($self->{on_connect_error}) { |
544 | if ($self->{on_connect_error}) { |
524 | $self->{on_connect_error}($self, "$!"); |
545 | $self->{on_connect_error}($self, "$!"); |
525 | $self->destroy; |
546 | $self->destroy if $self; |
526 | } else { |
547 | } else { |
527 | $self->_error ($!, 1); |
548 | $self->_error ($!, 1); |
528 | } |
549 | } |
529 | } |
550 | } |
530 | }, |
551 | }, |
531 | sub { |
552 | sub { |
532 | local $self->{fh} = $_[0]; |
553 | local $self->{fh} = $_[0]; |
533 | |
554 | |
534 | $self->{on_prepare} |
555 | $self->{on_prepare} |
535 | ? $self->{on_prepare}->($self) |
556 | ? $self->{on_prepare}->($self) |
536 | : () |
557 | : () |
537 | } |
558 | } |
538 | ); |
559 | ); |
539 | } |
560 | } |
540 | |
561 | |
… | |
… | |
739 | |
760 | |
740 | =item $handle->rbuf_max ($max_octets) |
761 | =item $handle->rbuf_max ($max_octets) |
741 | |
762 | |
742 | Configures the C<rbuf_max> setting (C<undef> disables it). |
763 | Configures the C<rbuf_max> setting (C<undef> disables it). |
743 | |
764 | |
|
|
765 | =item $handle->wbuf_max ($max_octets) |
|
|
766 | |
|
|
767 | Configures the C<wbuf_max> setting (C<undef> disables it). |
|
|
768 | |
744 | =cut |
769 | =cut |
745 | |
770 | |
746 | sub rbuf_max { |
771 | sub rbuf_max { |
747 | $_[0]{rbuf_max} = $_[1]; |
772 | $_[0]{rbuf_max} = $_[1]; |
748 | } |
773 | } |
749 | |
774 | |
|
|
775 | sub wbuf_max { |
|
|
776 | $_[0]{wbuf_max} = $_[1]; |
|
|
777 | } |
|
|
778 | |
750 | ############################################################################# |
779 | ############################################################################# |
751 | |
780 | |
752 | =item $handle->timeout ($seconds) |
781 | =item $handle->timeout ($seconds) |
753 | |
782 | |
754 | =item $handle->rtimeout ($seconds) |
783 | =item $handle->rtimeout ($seconds) |
755 | |
784 | |
756 | =item $handle->wtimeout ($seconds) |
785 | =item $handle->wtimeout ($seconds) |
757 | |
786 | |
758 | Configures (or disables) the inactivity timeout. |
787 | Configures (or disables) the inactivity timeout. |
|
|
788 | |
|
|
789 | The timeout will be checked instantly, so this method might destroy the |
|
|
790 | handle before it returns. |
759 | |
791 | |
760 | =item $handle->timeout_reset |
792 | =item $handle->timeout_reset |
761 | |
793 | |
762 | =item $handle->rtimeout_reset |
794 | =item $handle->rtimeout_reset |
763 | |
795 | |
… | |
… | |
872 | if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf}); |
904 | if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf}); |
873 | } |
905 | } |
874 | |
906 | |
875 | =item $handle->push_write ($data) |
907 | =item $handle->push_write ($data) |
876 | |
908 | |
877 | Queues the given scalar to be written. You can push as much data as you |
909 | Queues the given scalar to be written. You can push as much data as |
878 | want (only limited by the available memory), as C<AnyEvent::Handle> |
910 | you want (only limited by the available memory and C<wbuf_max>), as |
879 | buffers it independently of the kernel. |
911 | C<AnyEvent::Handle> buffers it independently of the kernel. |
880 | |
912 | |
881 | This method may invoke callbacks (and therefore the handle might be |
913 | This method may invoke callbacks (and therefore the handle might be |
882 | destroyed after it returns). |
914 | destroyed after it returns). |
883 | |
915 | |
884 | =cut |
916 | =cut |
… | |
… | |
912 | $cb->() unless $self->{autocork}; |
944 | $cb->() unless $self->{autocork}; |
913 | |
945 | |
914 | # if still data left in wbuf, we need to poll |
946 | # if still data left in wbuf, we need to poll |
915 | $self->{_ww} = AE::io $self->{fh}, 1, $cb |
947 | $self->{_ww} = AE::io $self->{fh}, 1, $cb |
916 | if length $self->{wbuf}; |
948 | if length $self->{wbuf}; |
|
|
949 | |
|
|
950 | if ( |
|
|
951 | defined $self->{wbuf_max} |
|
|
952 | && $self->{wbuf_max} < length $self->{wbuf} |
|
|
953 | ) { |
|
|
954 | $self->_error (Errno::ENOSPC, 1), return; |
|
|
955 | } |
917 | }; |
956 | }; |
918 | } |
957 | } |
919 | |
958 | |
920 | our %WH; |
959 | our %WH; |
921 | |
960 | |
… | |
… | |
1056 | before it was actually written. One way to do that is to replace your |
1095 | before it was actually written. One way to do that is to replace your |
1057 | C<on_drain> handler by a callback that shuts down the socket (and set |
1096 | C<on_drain> handler by a callback that shuts down the socket (and set |
1058 | C<low_water_mark> to C<0>). This method is a shorthand for just that, and |
1097 | C<low_water_mark> to C<0>). This method is a shorthand for just that, and |
1059 | replaces the C<on_drain> callback with: |
1098 | replaces the C<on_drain> callback with: |
1060 | |
1099 | |
1061 | sub { shutdown $_[0]{fh}, 1 } # for push_shutdown |
1100 | sub { shutdown $_[0]{fh}, 1 } |
1062 | |
1101 | |
1063 | This simply shuts down the write side and signals an EOF condition to the |
1102 | This simply shuts down the write side and signals an EOF condition to the |
1064 | the peer. |
1103 | the peer. |
1065 | |
1104 | |
1066 | You can rely on the normal read queue and C<on_eof> handling |
1105 | You can rely on the normal read queue and C<on_eof> handling |
… | |
… | |
1504 | |
1543 | |
1505 | sub { |
1544 | sub { |
1506 | # accept |
1545 | # accept |
1507 | if ($$rbuf =~ $accept) { |
1546 | if ($$rbuf =~ $accept) { |
1508 | $data .= substr $$rbuf, 0, $+[0], ""; |
1547 | $data .= substr $$rbuf, 0, $+[0], ""; |
1509 | $cb->($self, $data); |
1548 | $cb->($_[0], $data); |
1510 | return 1; |
1549 | return 1; |
1511 | } |
1550 | } |
1512 | |
1551 | |
1513 | # reject |
1552 | # reject |
1514 | if ($reject && $$rbuf =~ $reject) { |
1553 | if ($reject && $$rbuf =~ $reject) { |
1515 | $self->_error (Errno::EBADMSG); |
1554 | $_[0]->_error (Errno::EBADMSG); |
1516 | } |
1555 | } |
1517 | |
1556 | |
1518 | # skip |
1557 | # skip |
1519 | if ($skip && $$rbuf =~ $skip) { |
1558 | if ($skip && $$rbuf =~ $skip) { |
1520 | $data .= substr $$rbuf, 0, $+[0], ""; |
1559 | $data .= substr $$rbuf, 0, $+[0], ""; |
… | |
… | |
1536 | my ($self, $cb) = @_; |
1575 | my ($self, $cb) = @_; |
1537 | |
1576 | |
1538 | sub { |
1577 | sub { |
1539 | unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) { |
1578 | unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) { |
1540 | if ($_[0]{rbuf} =~ /[^0-9]/) { |
1579 | if ($_[0]{rbuf} =~ /[^0-9]/) { |
1541 | $self->_error (Errno::EBADMSG); |
1580 | $_[0]->_error (Errno::EBADMSG); |
1542 | } |
1581 | } |
1543 | return; |
1582 | return; |
1544 | } |
1583 | } |
1545 | |
1584 | |
1546 | my $len = $1; |
1585 | my $len = $1; |
1547 | |
1586 | |
1548 | $self->unshift_read (chunk => $len, sub { |
1587 | $_[0]->unshift_read (chunk => $len, sub { |
1549 | my $string = $_[1]; |
1588 | my $string = $_[1]; |
1550 | $_[0]->unshift_read (chunk => 1, sub { |
1589 | $_[0]->unshift_read (chunk => 1, sub { |
1551 | if ($_[1] eq ",") { |
1590 | if ($_[1] eq ",") { |
1552 | $cb->($_[0], $string); |
1591 | $cb->($_[0], $string); |
1553 | } else { |
1592 | } else { |
1554 | $self->_error (Errno::EBADMSG); |
1593 | $_[0]->_error (Errno::EBADMSG); |
1555 | } |
1594 | } |
1556 | }); |
1595 | }); |
1557 | }); |
1596 | }); |
1558 | |
1597 | |
1559 | 1 |
1598 | 1 |
… | |
… | |
1632 | |
1671 | |
1633 | my $data; |
1672 | my $data; |
1634 | my $rbuf = \$self->{rbuf}; |
1673 | my $rbuf = \$self->{rbuf}; |
1635 | |
1674 | |
1636 | sub { |
1675 | sub { |
1637 | my $ref = eval { $json->incr_parse ($self->{rbuf}) }; |
1676 | my $ref = eval { $json->incr_parse ($_[0]{rbuf}) }; |
1638 | |
1677 | |
1639 | if ($ref) { |
1678 | if ($ref) { |
1640 | $self->{rbuf} = $json->incr_text; |
1679 | $_[0]{rbuf} = $json->incr_text; |
1641 | $json->incr_text = ""; |
1680 | $json->incr_text = ""; |
1642 | $cb->($self, $ref); |
1681 | $cb->($_[0], $ref); |
1643 | |
1682 | |
1644 | 1 |
1683 | 1 |
1645 | } elsif ($@) { |
1684 | } elsif ($@) { |
1646 | # error case |
1685 | # error case |
1647 | $json->incr_skip; |
1686 | $json->incr_skip; |
1648 | |
1687 | |
1649 | $self->{rbuf} = $json->incr_text; |
1688 | $_[0]{rbuf} = $json->incr_text; |
1650 | $json->incr_text = ""; |
1689 | $json->incr_text = ""; |
1651 | |
1690 | |
1652 | $self->_error (Errno::EBADMSG); |
1691 | $_[0]->_error (Errno::EBADMSG); |
1653 | |
1692 | |
1654 | () |
1693 | () |
1655 | } else { |
1694 | } else { |
1656 | $self->{rbuf} = ""; |
1695 | $_[0]{rbuf} = ""; |
1657 | |
1696 | |
1658 | () |
1697 | () |
1659 | } |
1698 | } |
1660 | } |
1699 | } |
1661 | }; |
1700 | }; |
… | |
… | |
1694 | # read remaining chunk |
1733 | # read remaining chunk |
1695 | $_[0]->unshift_read (chunk => $len, sub { |
1734 | $_[0]->unshift_read (chunk => $len, sub { |
1696 | if (my $ref = eval { Storable::thaw ($_[1]) }) { |
1735 | if (my $ref = eval { Storable::thaw ($_[1]) }) { |
1697 | $cb->($_[0], $ref); |
1736 | $cb->($_[0], $ref); |
1698 | } else { |
1737 | } else { |
1699 | $self->_error (Errno::EBADMSG); |
1738 | $_[0]->_error (Errno::EBADMSG); |
1700 | } |
1739 | } |
1701 | }); |
1740 | }); |
1702 | } |
1741 | } |
1703 | |
1742 | |
1704 | 1 |
1743 | 1 |
… | |
… | |
1742 | Note that AnyEvent::Handle will automatically C<start_read> for you when |
1781 | Note that AnyEvent::Handle will automatically C<start_read> for you when |
1743 | you change the C<on_read> callback or push/unshift a read callback, and it |
1782 | you change the C<on_read> callback or push/unshift a read callback, and it |
1744 | will automatically C<stop_read> for you when neither C<on_read> is set nor |
1783 | will automatically C<stop_read> for you when neither C<on_read> is set nor |
1745 | there are any read requests in the queue. |
1784 | there are any read requests in the queue. |
1746 | |
1785 | |
1747 | These methods will have no effect when in TLS mode (as TLS doesn't support |
1786 | In older versions of this module (<= 5.3), these methods had no effect, |
1748 | half-duplex connections). |
1787 | as TLS does not support half-duplex connections. In current versions they |
|
|
1788 | work as expected, as this behaviour is required to avoid certain resource |
|
|
1789 | attacks, where the program would be forced to read (and buffer) arbitrary |
|
|
1790 | amounts of data before being able to send some data. The drawback is that |
|
|
1791 | some readings of the the SSL/TLS specifications basically require this |
|
|
1792 | attack to be working, as SSL/TLS implementations might stall sending data |
|
|
1793 | during a rehandshake. |
|
|
1794 | |
|
|
1795 | As a guideline, during the initial handshake, you should not stop reading, |
|
|
1796 | and as a client, it might cause problems, depending on your applciation. |
1749 | |
1797 | |
1750 | =cut |
1798 | =cut |
1751 | |
1799 | |
1752 | sub stop_read { |
1800 | sub stop_read { |
1753 | my ($self) = @_; |
1801 | my ($self) = @_; |
1754 | |
1802 | |
1755 | delete $self->{_rw} unless $self->{tls}; |
1803 | delete $self->{_rw}; |
1756 | } |
1804 | } |
1757 | |
1805 | |
1758 | sub start_read { |
1806 | sub start_read { |
1759 | my ($self) = @_; |
1807 | my ($self) = @_; |
1760 | |
1808 | |
… | |
… | |
1962 | Net::SSLeay::CTX_set_mode ($tls, 1|2); |
2010 | Net::SSLeay::CTX_set_mode ($tls, 1|2); |
1963 | |
2011 | |
1964 | $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
2012 | $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
1965 | $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
2013 | $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
1966 | |
2014 | |
1967 | Net::SSLeay::BIO_write ($self->{_rbio}, delete $self->{rbuf}); |
2015 | Net::SSLeay::BIO_write ($self->{_rbio}, $self->{rbuf}); |
|
|
2016 | $self->{rbuf} = ""; |
1968 | |
2017 | |
1969 | Net::SSLeay::set_bio ($tls, $self->{_rbio}, $self->{_wbio}); |
2018 | Net::SSLeay::set_bio ($tls, $self->{_rbio}, $self->{_wbio}); |
1970 | |
2019 | |
1971 | $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) } |
2020 | $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) } |
1972 | if $self->{on_starttls}; |
2021 | if $self->{on_starttls}; |
… | |
… | |
2009 | $self->{tls_ctx}->_put_session (delete $self->{tls}) |
2058 | $self->{tls_ctx}->_put_session (delete $self->{tls}) |
2010 | if $self->{tls} > 0; |
2059 | if $self->{tls} > 0; |
2011 | |
2060 | |
2012 | delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)}; |
2061 | delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)}; |
2013 | } |
2062 | } |
|
|
2063 | |
|
|
2064 | =item $handle->resettls |
|
|
2065 | |
|
|
2066 | This rarely-used method simply resets and TLS state on the handle, usually |
|
|
2067 | causing data loss. |
|
|
2068 | |
|
|
2069 | One case where it may be useful is when you want to skip over the data in |
|
|
2070 | the stream but you are not interested in interpreting it, so data loss is |
|
|
2071 | no concern. |
|
|
2072 | |
|
|
2073 | =cut |
|
|
2074 | |
|
|
2075 | *resettls = \&_freetls; |
2014 | |
2076 | |
2015 | sub DESTROY { |
2077 | sub DESTROY { |
2016 | my ($self) = @_; |
2078 | my ($self) = @_; |
2017 | |
2079 | |
2018 | &_freetls; |
2080 | &_freetls; |
… | |
… | |
2134 | |
2196 | |
2135 | It is only safe to "forget" the reference inside EOF or error callbacks, |
2197 | It is only safe to "forget" the reference inside EOF or error callbacks, |
2136 | from within all other callbacks, you need to explicitly call the C<< |
2198 | from within all other callbacks, you need to explicitly call the C<< |
2137 | ->destroy >> method. |
2199 | ->destroy >> method. |
2138 | |
2200 | |
|
|
2201 | =item Why is my C<on_eof> callback never called? |
|
|
2202 | |
|
|
2203 | Probably because your C<on_error> callback is being called instead: When |
|
|
2204 | you have outstanding requests in your read queue, then an EOF is |
|
|
2205 | considered an error as you clearly expected some data. |
|
|
2206 | |
|
|
2207 | To avoid this, make sure you have an empty read queue whenever your handle |
|
|
2208 | is supposed to be "idle" (i.e. connection closes are O.K.). You cna set |
|
|
2209 | an C<on_read> handler that simply pushes the first read requests in the |
|
|
2210 | queue. |
|
|
2211 | |
|
|
2212 | See also the next question, which explains this in a bit more detail. |
|
|
2213 | |
|
|
2214 | =item How can I serve requests in a loop? |
|
|
2215 | |
|
|
2216 | Most protocols consist of some setup phase (authentication for example) |
|
|
2217 | followed by a request handling phase, where the server waits for requests |
|
|
2218 | and handles them, in a loop. |
|
|
2219 | |
|
|
2220 | There are two important variants: The first (traditional, better) variant |
|
|
2221 | handles requests until the server gets some QUIT command, causing it to |
|
|
2222 | close the connection first (highly desirable for a busy TCP server). A |
|
|
2223 | client dropping the connection is an error, which means this variant can |
|
|
2224 | detect an unexpected detection close. |
|
|
2225 | |
|
|
2226 | To handle this case, always make sure you have a on-empty read queue, by |
|
|
2227 | pushing the "read request start" handler on it: |
|
|
2228 | |
|
|
2229 | # we assume a request starts with a single line |
|
|
2230 | my @start_request; @start_request = (line => sub { |
|
|
2231 | my ($hdl, $line) = @_; |
|
|
2232 | |
|
|
2233 | ... handle request |
|
|
2234 | |
|
|
2235 | # push next request read, possibly from a nested callback |
|
|
2236 | $hdl->push_read (@start_request); |
|
|
2237 | }); |
|
|
2238 | |
|
|
2239 | # auth done, now go into request handling loop |
|
|
2240 | # now push the first @start_request |
|
|
2241 | $hdl->push_read (@start_request); |
|
|
2242 | |
|
|
2243 | By always having an outstanding C<push_read>, the handle always expects |
|
|
2244 | some data and raises the C<EPIPE> error when the connction is dropped |
|
|
2245 | unexpectedly. |
|
|
2246 | |
|
|
2247 | The second variant is a protocol where the client can drop the connection |
|
|
2248 | at any time. For TCP, this means that the server machine may run out of |
|
|
2249 | sockets easier, and in general, it means you cnanot distinguish a protocl |
|
|
2250 | failure/client crash from a normal connection close. Nevertheless, these |
|
|
2251 | kinds of protocols are common (and sometimes even the best solution to the |
|
|
2252 | problem). |
|
|
2253 | |
|
|
2254 | Having an outstanding read request at all times is possible if you ignore |
|
|
2255 | C<EPIPE> errors, but this doesn't help with when the client drops the |
|
|
2256 | connection during a request, which would still be an error. |
|
|
2257 | |
|
|
2258 | A better solution is to push the initial request read in an C<on_read> |
|
|
2259 | callback. This avoids an error, as when the server doesn't expect data |
|
|
2260 | (i.e. is idly waiting for the next request, an EOF will not raise an |
|
|
2261 | error, but simply result in an C<on_eof> callback. It is also a bit slower |
|
|
2262 | and simpler: |
|
|
2263 | |
|
|
2264 | # auth done, now go into request handling loop |
|
|
2265 | $hdl->on_read (sub { |
|
|
2266 | my ($hdl) = @_; |
|
|
2267 | |
|
|
2268 | # called each time we receive data but the read queue is empty |
|
|
2269 | # simply start read the request |
|
|
2270 | |
|
|
2271 | $hdl->push_read (line => sub { |
|
|
2272 | my ($hdl, $line) = @_; |
|
|
2273 | |
|
|
2274 | ... handle request |
|
|
2275 | |
|
|
2276 | # do nothing special when the request has been handled, just |
|
|
2277 | # let the request queue go empty. |
|
|
2278 | }); |
|
|
2279 | }); |
|
|
2280 | |
2139 | =item I get different callback invocations in TLS mode/Why can't I pause |
2281 | =item I get different callback invocations in TLS mode/Why can't I pause |
2140 | reading? |
2282 | reading? |
2141 | |
2283 | |
2142 | Unlike, say, TCP, TLS connections do not consist of two independent |
2284 | Unlike, say, TCP, TLS connections do not consist of two independent |
2143 | communication channels, one for each direction. Or put differently, the |
2285 | communication channels, one for each direction. Or put differently, the |
… | |
… | |
2163 | $handle->on_read (sub { }); |
2305 | $handle->on_read (sub { }); |
2164 | $handle->on_eof (undef); |
2306 | $handle->on_eof (undef); |
2165 | $handle->on_error (sub { |
2307 | $handle->on_error (sub { |
2166 | my $data = delete $_[0]{rbuf}; |
2308 | my $data = delete $_[0]{rbuf}; |
2167 | }); |
2309 | }); |
|
|
2310 | |
|
|
2311 | Note that this example removes the C<rbuf> member from the handle object, |
|
|
2312 | which is not normally allowed by the API. It is expressly permitted in |
|
|
2313 | this case only, as the handle object needs to be destroyed afterwards. |
2168 | |
2314 | |
2169 | The reason to use C<on_error> is that TCP connections, due to latencies |
2315 | The reason to use C<on_error> is that TCP connections, due to latencies |
2170 | and packets loss, might get closed quite violently with an error, when in |
2316 | and packets loss, might get closed quite violently with an error, when in |
2171 | fact all data has been received. |
2317 | fact all data has been received. |
2172 | |
2318 | |