Changelog for AnyEvent-Net-Curl-Queued 0.037 2012-11-26T13:15:04 [Stanislaw Pusep ] - fixed encoding issues revealed by upgrading to Digest::SHA v5.74 - postponing $cv->send to properly destroy ::Multi instance - followlocation-caused header propagation fix v0.036 2012-11-06T15:00:02 [Stanislaw Pusep ] - curly brackets cleanup - initialization exception testing - proper usage of BUILDARGS/FOREIGNBUILDARGS - fixing cyclic references - benchmark results update - less verbose recursion test - benchmark suite fixes - Test::Memory::Cycle test implemented - clearer benchmark reasoning - major breakthrough in Net::Curl::* "unreferenced scalar" fixing! - attempt to fight 1-millisecond timeout (again) v0.035 2012-10-22T18:49:04 [Stanislaw Pusep ] - Net::Curl::Share instance destruction code is FUBAR; storing it inside a Mo[ou]se attribute was a really bad idea :~( v0.034 2012-10-22T12:46:49 [Stanislaw Pusep ] - weak_ref unnecessary, after all - removed unnecessary & broken weak_ref declarations - replaced time() from Time::HiRes by AE::tie - a bit less deterministic t/61-nest.t v0.033 2012-10-18T18:31:27 [Stanislaw Pusep ] - better compliance with libcurl docs - a bit cleaner Net::Curl::Share handling - very bad fix for Net::Curl::Share instance "weakness" - getting rid of "Attempt to free unreferenced scalar: SV 0xdeadbeef during global destruction." v0.032 2012-10-16T15:14:13 [Stanislaw Pusep ] - version bump - fix release: dependencies were stripped accidentally v0.031 2012-10-16T14:20:54 [Stanislaw Pusep ] - split Test::HTTP::AnyEvent::Server into a separate module - finally Test::HTTP::AnyEvent::Server can detach & run as separate process v0.030 2012-10-11T22:31:11 [Stanislaw Pusep ] - YADA DWIM: documented & tested - updated test for nasty empty HTTP::Message condition - minor edge case condition adjustments - better Moose encapsulation - better type constraints - added common_opts to AE::N::C::Q and fixed CURLOPT_(SHARE|TIMEOUT) propagation for YADA v0.029 2012-10-03T19:44:25 [Stanislaw Pusep ] - test fixes (no need to rely on cpan.org) [Stanislaw Pusep ] - eg/simple.pl updated to show the new sugar syntax - do *NOT* return HTTP::Response if scheme is not HTTP/HTTPS v0.028 2012-09-26T19:52:31 [Stanislaw Pusep ] - minor compatibility fixes [Stanislaw Pusep ] - attempt to fix FreeBSD test errors - YADA append()/prepend() are now chainable (updated example in SYNOPSIS) - experimental YADA DWIM extensions for append()/prepend() - added AE::N::C::Q::Easy->opts attribute to quickly setopt() stuff - more verbose environment info - benchmark results updated - update process name in eg/benchmark.pl - refactored eg/benchmark.pl v0.027 2012-09-11T04:54:09 [Stanislaw Pusep ] - test cloning - AE::N::C::Q::Easy->clone($param) fixed - fix: no socket_action(CURL_SOCKET_TIMEOUT) if no handles are open - reimplemented eg/benchmark/mojo.pl to work the non-blocking way - documentation update - implemented AnyEvent::HTTP::Tiny client to test Test::HTTP::AnyEvent::Server - use Any::Moose for Test::HTTP::AnyEvent::Server - gave up, Test::HTTP::Server is more reliable for blocking tests - enhance Test::HTTP::AnyEvent::Server coverage - more failing tests - more AE::N::C::Easy cleanup - removed Test::HTTP::Server dependency (now for sure) - test incorrect usage of AE::N::C::Q::Easy setopt()/getinfo() - migrate N::C::Multi & N::C::Easy object capability checks into test - test boundaries fine-tuned - removed unmet conditions - test watchdog & timeout - clean up duplicate URL condition - removed Test::HTTP::Server dependency [Stanislaw Pusep ] - benchmark updated - eg/queue: filtered URLs which lead to redirectors - implemented Test::HTTP::AnyEvent::Server, as fork() breaks Net::Curl :( - use AE::postpone instead of "instant timer" hack v0.026 2012-09-05T21:24:33 [Stanislaw Pusep ] - correctly reenqueue POST requests [Stanislaw Pusep ] - updated t/42-loopback-retry.t to use internal POST retry mechanism - attempt double GET at t/30-queued-single.t (to verify queue deduplication) - stripped dead code from AE::N::C::Q::Stats - fixed AE::N::C::Const cache typo - split queue accessors POD - Mojo::UserAgent benchmark v0.025 2012-07-27T18:08:19 [Stanislaw Pusep ] - benchmark fixed - removed Memoize dependency v0.024 2012-07-23T19:22:24 [Stanislaw Pusep ] - squash setopt calls during init into one - Trait::Array is too slow to handle huge queues v0.023 2012-07-20T19:35:04 [Stanislaw Pusep ] - automatically POST as JSON v0.022 2012-07-04T18:54:54 [Stanislaw Pusep ] - raise CURLMOPT_MAXCONNECTS - example fixes - eg/CrawlApache.pm using Web::Scraper v0.021 2012-05-21T17:18:08 [Stanislaw Pusep ] - Changes [Stanislaw Pusep ] - removed common::sense - using ChangelogFromGit::CPAN::Changes for changelog v0.020 2012-05-12T21:18:51 [Stanislaw Pusep ] - documentation updated with alternatives; prereq common::sense 3.5 [Stanislaw Pusep ] - Changelog v0.019 2012-04-05T20:26:29 [Stanislaw Pusep ] - Changelog [Stanislaw Pusep ] - version bump - fix for perl-5.15.9 v0.018 2012-03-31T19:04:45 [Stanislaw Pusep ] - minor fixes [Stanislaw Pusep ] - Changelog v0.017 2012-03-27T13:54:30 [Stanislaw Pusep ] - Changelog [Stanislaw Pusep ] - ++yada options - yada: stupid typo fix - benchmark is flawed as YADA also can't tell initialization from fetching apart :( v0.016 2012-03-25T21:46:55 [Stanislaw Pusep ] - yada: filename from content-disposition/location tags - better socket watcher recycling - benchmark now separates initialization from the download; using remote server [Stanislaw Pusep ] - yada: get_filename() - yada timeout parameter - --MooseX::NonMoose from dist.ini - verbose yada help v0.015 2012-03-23T15:34:09 [Stanislaw Pusep ] - ++yada CLI tool - all tests under Any::Moose - meta->class_precedence_list => meta->name - ++MooseX::NonMoose to dist.ini - eg/benchmark.pl: unsure about the way to use AnyEvent::Curl::Multi; copied from Template::Parser::RemoteInclude - Changelog v0.014 2012-03-22T19:38:15 [Stanislaw Pusep ] - MouseX::NativeTraits dependency [Stanislaw Pusep ] - README version bump - ++CURL_LOCK_DATA_SSL_SESSION - Changelog v0.013 2012-03-21T20:23:37 [Stanislaw Pusep ] - Changes [Stanislaw Pusep ] - Replaced Moose by Any::Moose - benchmark update - documentation update - fixes - benchmark fix - cleanup test - uninitialization fixes - Mouse prereqs - BUILDARGS hack - benchmark update - Moose => Any::Moose v0.012 2012-02-01T16:33:34 [Stanislaw Pusep ] - "Attempt to free unreferenced scalar: SV 0xdeadbeef." warning was not Test::HTTP::Server's fault in 01-net-curl-compatibility.t, at all. - Changes [Stanislaw Pusep ] - AENCQ object now reusable - fixed kwalitee test due to new Parse::RecDescent release