hi,
LWP fetches normally by 4k chunks, if you want to time them
you can use the code1 below.
On the other hand for some reason code2 doesn't work.
The output should look something like specified below.
If you wrap "LWP::Protocol::collect", you will time the whole fetch, not every chunk.
One more thing which may be interesting to know is, that the timeout you specify (LWP,Mechanize...) is the timeout for every chunk not the whole request (default:180sec).
So the request timeout could go far above 3 min.
hth
-----code1-----
use LWP::Debug qw(+);
use Time::HiRes qw(gettimeofday tv_interval);
my $ts;
use LWP::Protocol::http;
use Hook::LexWrap;
wrap 'LWP::Protocol::http::SocketMethods::sysread',
pre => sub { $ts = [gettimeofday] },
post => sub { printf ">>%.5f\n" , tv_interval($ts) };
-----code2-----
use LWP::Debug qw(+);
use Aspect;
my $pcut = call qr/LWP::Protocol::http::SocketMethods::sysread/;
before { $ts = [gettimeofday] } $pcut;
after { printf ">>%.5f\n" , tv_interval($ts) } $pcut;
====output=====
LWP::UserAgent::_need_proxy: Not proxied
LWP::Protocol::http::request: ()
>>0.00536
LWP::Protocol::collect: read 796 bytes
>>0.00007
LWP::Protocol::collect: read 3356 bytes
>>0.00071
LWP::Protocol::collect: read 1460 bytes
>>0.00103
LWP::Protocol::collect: read 1460 bytes
>>0.00005
LWP::Protocol::collect: read 997 bytes
LWP::UserAgent::request: Simple response: OK