ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/lib/AnyEvent.pm
(Generate patch)

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.210 by root, Wed May 13 15:19:43 2009 UTC vs.
Revision 1.217 by root, Tue Jun 23 23:37:32 2009 UTC

1=head1 NAME 1=head1 NAME
2 2
3AnyEvent - provide framework for multiple event loops 3AnyEvent - provide framework for multiple event loops
4 4
5EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event loops 5EV, Event, Glib, Tk, Perl, Event::Lib, Qt and POE are various supported
6event loops.
6 7
7=head1 SYNOPSIS 8=head1 SYNOPSIS
8 9
9 use AnyEvent; 10 use AnyEvent;
10 11
930no warnings; 931no warnings;
931use strict qw(vars subs); 932use strict qw(vars subs);
932 933
933use Carp; 934use Carp;
934 935
935our $VERSION = 4.41; 936our $VERSION = 4.412;
936our $MODEL; 937our $MODEL;
937 938
938our $AUTOLOAD; 939our $AUTOLOAD;
939our @ISA; 940our @ISA;
940 941
941our @REGISTRY; 942our @REGISTRY;
942 943
943our $WIN32; 944our $WIN32;
944 945
945BEGIN { 946BEGIN {
946 my $win32 = ! ! ($^O =~ /mswin32/i); 947 eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }";
947 eval "sub WIN32(){ $win32 }"; 948 eval "sub TAINT(){ " . (${^TAINT}*1) . " }";
949
950 delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
951 if ${^TAINT};
948} 952}
949 953
950our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; 954our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
951 955
952our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred 956our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
1141 AnyEvent::Util::fh_nonblocking ($SIGPIPE_W) if $SIGPIPE_W; # just in case 1145 AnyEvent::Util::fh_nonblocking ($SIGPIPE_W) if $SIGPIPE_W; # just in case
1142 } else { 1146 } else {
1143 pipe $SIGPIPE_R, $SIGPIPE_W; 1147 pipe $SIGPIPE_R, $SIGPIPE_W;
1144 fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R; 1148 fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
1145 fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case 1149 fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
1150
1151 # not strictly required, as $^F is normally 2, but let's make sure...
1152 fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1153 fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1146 } 1154 }
1147 1155
1148 $SIGPIPE_R 1156 $SIGPIPE_R
1149 or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n"; 1157 or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
1150
1151 # not strictly required, as $^F is normally 2, but let's make sure...
1152 fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1153 fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1154 1158
1155 $SIG_IO = AnyEvent->io (fh => $SIGPIPE_R, poll => "r", cb => \&_signal_exec); 1159 $SIG_IO = AnyEvent->io (fh => $SIGPIPE_R, poll => "r", cb => \&_signal_exec);
1156 } 1160 }
1157 1161
1158 my $signal = uc $arg{signal} 1162 my $signal = uc $arg{signal}
1336so on. 1340so on.
1337 1341
1338=head1 ENVIRONMENT VARIABLES 1342=head1 ENVIRONMENT VARIABLES
1339 1343
1340The following environment variables are used by this module or its 1344The following environment variables are used by this module or its
1341submodules: 1345submodules.
1346
1347Note that AnyEvent will remove I<all> environment variables starting with
1348C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
1349enabled.
1342 1350
1343=over 4 1351=over 4
1344 1352
1345=item C<PERL_ANYEVENT_VERBOSE> 1353=item C<PERL_ANYEVENT_VERBOSE>
1346 1354
1886=item * C-based event loops perform very well with small number of 1894=item * C-based event loops perform very well with small number of
1887watchers, as the management overhead dominates. 1895watchers, as the management overhead dominates.
1888 1896
1889=back 1897=back
1890 1898
1899=head2 THE IO::Lambda BENCHMARK
1900
1901Recently I was told about the benchmark in the IO::Lambda manpage, which
1902could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
1903simply compares IO::Lambda with POE, and IO::Lambda looks better (which
1904shouldn't come as a surprise to anybody). As such, the benchmark is
1905fine, and shows that the AnyEvent backend from IO::Lambda isn't very
1906optimal. But how would AnyEvent compare when used without the extra
1907baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
1908
1909The benchmark itself creates an echo-server, and then, for 500 times,
1910connects to the echo server, sends a line, waits for the reply, and then
1911creates the next connection. This is a rather bad benchmark, as it doesn't
1912test the efficiency of the framework, but it is a benchmark nevertheless.
1913
1914 name runtime
1915 Lambda/select 0.330 sec
1916 + optimized 0.122 sec
1917 Lambda/AnyEvent 0.327 sec
1918 + optimized 0.138 sec
1919 Raw sockets/select 0.077 sec
1920 POE/select, components 0.662 sec
1921 POE/select, raw sockets 0.226 sec
1922 POE/select, optimized 0.404 sec
1923
1924 AnyEvent/select/nb 0.085 sec
1925 AnyEvent/EV/nb 0.068 sec
1926 +state machine 0.134 sec
1927
1928The benchmark is also a bit unfair (my fault) - the IO::Lambda
1929benchmarks actually make blocking connects and use 100% blocking I/O,
1930defeating the purpose of an event-based solution. All of the newly
1931written AnyEvent benchmarks use 100% non-blocking connects (using
1932AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
1933resolver), so AnyEvent is at a disadvantage here as non-blocking connects
1934generally require a lot more bookkeeping and event handling than blocking
1935connects (which involve a single syscall only).
1936
1937The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
1938offers similar expressive power as POE and IO::Lambda (using conventional
1939Perl syntax), which means both the echo server and the client are 100%
1940non-blocking w.r.t. I/O, further placing it at a disadvantage.
1941
1942As you can see, AnyEvent + EV even beats the hand-optimised "raw sockets
1943benchmark", while AnyEvent + its pure perl backend easily beats
1944IO::Lambda and POE.
1945
1946And even the 100% non-blocking version written using the high-level (and
1947slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda,
1948even thought it does all of DNS, tcp-connect and socket I/O in a
1949non-blocking way.
1950
1951The two AnyEvent benchmarks can be found as F<eg/ae0.pl> and F<eg/ae2.pl>
1952in the AnyEvent distribution, the remaining benchmarks are part of the
1953IO::lambda distribution and were used without any changes.
1954
1891 1955
1892=head1 SIGNALS 1956=head1 SIGNALS
1893 1957
1894AnyEvent currently installs handlers for these signals: 1958AnyEvent currently installs handlers for these signals:
1895 1959
1952 use AnyEvent; 2016 use AnyEvent;
1953 2017
1954Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can 2018Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
1955be used to probe what backend is used and gain other information (which is 2019be used to probe what backend is used and gain other information (which is
1956probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and 2020probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
1957$ENV{PERL_ANYEGENT_STRICT}. 2021$ENV{PERL_ANYEVENT_STRICT}.
1958 2022
1959 2023
1960=head1 BUGS 2024=head1 BUGS
1961 2025
1962Perl 5.8 has numerous memleaks that sometimes hit this module and are hard 2026Perl 5.8 has numerous memleaks that sometimes hit this module and are hard

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines