The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

IO::Async::Resolver::StupidCache - a trivial caching layer around an IO::Async::Resolver

SYNOPSIS

 use IO::Async::Loop 0.62;
 use IO::Async::Resolver::StupidCache;

 my $loop = IO::Async::Loop->new;

 # Wrap the existing resolver in a cache
 $loop->set_resolver(
    IO::Async::Resolver::StupidCache->new( source => $loop->resolver )
 );

 # $loop->resolve requests will now be cached

DESCRIPTION

This object class provides a wrapper around another IO::Async::Resolver instance, which applies a simple caching layer to avoid making identical lookups. This can be useful, for example, when performing a large number of HTTP requests to the same host or a small set of hosts, or other cases where it is expected that the same few resolver queries will be made over and over.

This is called a "stupid" cache because it is made without awareness of TTL values or other cache-relevant information that may be provided by DNS or other resolve methods. As such, it should not be relied upon to give always-accurate answers.

PARAMETERS

The following named parameters may be passed to new or configure:

source => IO::Async::Resolver

Optional. The source of the cache data. If not supplied, a new IO::Async::Resolver instance will be constructed.

ttl => INT

Optional. Time-to-live of cache entries in seconds. If not supplied a default of 5 minutes will apply.

max_size => INT

Optional. Maximum number of entries to keep in the cache. Entries will be evicted at random over this limit. If not supplied a default of 1000 entries will apply.

METHODS

The following methods documented with a trailing call to ->get return Future instances.

$resolver = $cache->source

Returns the source resolver

@result = $cache->resolve( %args )->get

@addrs = $cache->getaddrinfo( %args )->get

( $host, $service ) = $cache->getnameinfo( %args )->get

These methods perform identically to the base IO::Async::Resolver class, except that the results are cached.

Returned Futures are created with the without_cancel method, so that multiple concurrent waiters are shielded from cancellation by one another.

AUTHOR

Paul Evans <leonerd@leonerd.org.uk>