This version was released on 2003-02-07.

  2003-02-07 17:48:47 by rcaputo; POE.pm 1.155; README 1.103

    Finalize things for the imminent 0.25 release. 

  2003-02-07 05:03:43 by hachi
  POE/Component/Client/TCP.pm 1.30; POE/Component/Server/TCP.pm 1.35

    Documentation cleanup in PoCo::Client::TCP and PoCo::Server::TCP

    First time I committed CVS (outside of my own hard drive) Yay me! 

  2003-02-06 20:07:15 by rcaputo; t/04_selects.t 1.17

    This fix is against revision 1.14. Forcing that new test to
    use "inet" pipes breaks on systems where localhost is firewalled.
    Tweaked the test to use whichever pipes it can under non-Windows
    operating systems, and forced it to "inet" for MSWin32. A real
    solution would be to skip the test entirely if the pipe can't be
    built. I suspect we'll need to revisit this later. 

  2003-02-06 03:30:38 by rcaputo; t/06_tk.t 1.40

    It is unfortunate, but a bug in ActivePerl 5.8.0's version of Tk
    prevents us from supporting Tk at all in this build. We regret the
    inconvenience and hope that support can resume in a future ActivePerl
    build. Meanwhile, you may follow the status of the bug at
    http://bugs.activestate.com/show_bug.cgi?id=22619 

  2003-02-06 03:25:52 by rcaputo; POE/Loop/Tk.pm 1.39

    Collateral refactoring while trying to work around ActivePerl's
    problem with Tk, and Tk's problem with my preferred workaround. 

  2003-02-05 16:22:11 by rcaputo; POE/Kernel.pm 1.233

    Nancy Grady pointed out that Randal Schwartz's listing at
    http://www.stonehenge.com/merlyn/LinuxMag/col41.listing.txt was not
    exiting when it was done. This turned out to be a missing garbage
    collection test in POE::Kernel, introduced during the Great
    Refactoring of 2002. This fix adds back the GC test that was removed. 

  2003-02-04 03:32:01 by rcaputo; POE/Pipe.pm 1.9

    Bad Rocco! Always verify copied/pasted/tweaked code has actually been
    tweaked, or you'll keep doing things like what was fixed here.
    Luckily it hadn't caused any trouble to speak of. 

  2003-02-01 19:13:27 by cwest
  POE/Component/Client/TCP.pm 1.29; POE/Component/Server/TCP.pm 1.34

    The code to validate SessionParams was broken. Even when
    SessionParams was defined correctly, it was overwritten with an empty
    array reference. This has been fixed. 

  2003-02-01 04:52:07 by cwest; t/001_queue.t 1.2; t/00_coverage.t 1.17;
  t/01_sessions.t 1.22; t/02_alarms.t 1.12; t/03_aliases.t 1.12;
  t/04_selects.t 1.16; t/05_macros.t 1.6; t/06_tk.t 1.39;
  t/07_event.t 1.16; t/08_errors.t 1.42; t/09_wheels_unix.t 1.14;
  t/10_wheels_tcp.t 1.13; t/11_signals_poe.t 1.24;
  t/12_signals_ev.t 1.17; t/13_wheels_udp.t 1.4; t/14_wheels_ft.t 1.11;
  t/15_filter_block.t 1.4; t/16_filter_stream.t 1.3;
  t/17_filter_ref.t 1.6; t/18_filter_line.t 1.7;
  t/19_filterchange.t 1.11; t/20_accept.t 1.5; t/21_gtk.t 1.16;
  t/22_wheel_run.t 1.30; t/23_nfa.t 1.5; t/24_filter_stack.t 1.3;
  t/25_detach.t 1.4; t/26_comp_tcp.t 1.3; t/27_poll.t 1.6;
  t/28_windows.t 1.4; t/29_sockfact6.t 1.7

    Resolving Ticket #1597 on rt.cpan.org

    All tests now look in ./lib, ../lib, .., and . for libraries so as to
    not go testing already installed libraries. 

  2003-02-01 04:17:41 by cwest; POE/Component/Client/TCP.pm 1.28

    Added new() parameters SessionType and SessionParams. SessionType
    allows you to override the default POE session handler, POE::Session,
    with your choice of drop in replacement such as
    POE::Session::MultiDispatch. SessionParams allows you to pass extra
    arguments to the SessionType's constructor. Most useful are things in
    the options hash. 

  2003-02-01 04:04:39 by cwest; POE/Component/Server/TCP.pm 1.33

    Added create() parameters SessionType and SessionParams. SessionType
    allows you to override the default POE session handler, POE::Session,
    with your choice of drop in replacement such as
    POE::Session::MultiDispatch. SessionParams allows you to pass extra
    arguments to the SessionType's constructor. Most useful are things in
    the options hash. 

  2003-01-26 03:30:39 by rcaputo
  POE/Kernel.pm 1.232; POE/Wheel/Run.pm 1.44

    The old HP-UX SIGRTMIN bug in Perl bit us again on Gentoo Linux. It
    appears the hints for that platform allow all these non-signal
    /SIG.*/ symbols into the %SIG hash, and setting them causes perl to
    segfault. Normally we wouldn't know about it, but recent signal
    reforms have reintroduced the errors. This time we build a single
    list of safe signals and use that instead of keys(%SIG). If we're
    lucky, this will be the last of that. 

  2003-01-21 22:22:35 by rcaputo; POE/Wheel/FollowTail.pm 1.42

    One cannot assume that filehandles have methods such as clearerr().
    If one does, someone (Hi, Rob Bloodgood) using Perl 5.005_03 will
    inform one that one's code doesn't work. One will therefore be
    required to commit patches such as this one that replaces
    $handle->clearerr() with IO::Handle::clearerr($handle). One might be
    somewhat bitter about all these tethers holding one from using more
    modern coding techniques. Gulliver probably felt similar upon
    awakening Lilliput. 

  2003-01-21 07:27:59 by rcaputo; t/08_errors.t 1.41

    Wheel::Run can change input/output filters now, so remove the tests
    which no longer generate errors. 

  2003-01-21 07:26:43 by rcaputo
  POE/Kernel.pm 1.231; POE/Pipe.pm 1.8; POE/Wheel/SocketFactory.pm 1.68

    The introduction of $handle->blocking() for ActivePerl 5.8.0 breaks
    perl 5.005_03 on two counts. First, filehandles aren't blessed into
    IO::Handle by default. Second, blocking() doesn't exist in the older
    Perl. Reintroduce the homegrown non/blocking code for legacy Perl,
    but use the portable IO::Handle version for modern Perl. 

  2003-01-17 18:26:57 by cwest; POE/Component/Server/TCP.pm 1.32

    POE::Component::Server::TCP->new() now accepts an 'Args' parameter.
    It does the same thing as 'Args' does for
    POE::Component::Client::TCP. A list of arguments are passed to the
    ClientConnected callback filling @{ARG0..$#_}. This is helpful for
    setting up sessions with important data. 

  2003-01-11 04:33:35 by rcaputo; POE/Component/Server/TCP.pm 1.31

    Bob Maccione discovered that ECONNABORTED (a connection was closed
    before it could be accepted) was killing off PoCo::Server::HTTP. This
    change ignores ECONNABORTED, so clients can't use it to kill servers. 

  2003-01-09 17:56:09 by rcaputo; POE/Kernel.pm 1.230

    Mark each "resource" section with the name of the module it will
    become. 

  2003-01-04 04:23:55 by rcaputo; t/22_wheel_run.t 1.29

    Halfjack discovered six subtests that failed. We tracked them down to
    a bogus $Configure{perlpath}. I changed the tests to discover the
    perl path via $^X instead, which exists at least as far back as
    5.005_03. 

  2003-01-02 06:19:03 by rcaputo; POE/Component/Client/TCP.pm 1.27

    Document Started in the SYNOPSIS. 

  2003-01-01 04:24:40 by rcaputo; POE/Session.pm 1.83

    Doc patch for Averell. 

  2002-12-31 22:59:57 by rcaputo; POE/Wheel/Run.pm 1.43

    Applied Jason Stillwell's patch to finish several features in this
    wheel. We can now pause and resume Stdout and Stderr events, and the
    wheel now supports changing filters at runtime. 

  2002-12-31 22:09:46 by rcaputo; POE/Kernel.pm 1.229

    Jason Stillwell discovered that event handlers can't return undef.
    Nothing seems to rely on this behavior, so now it's removed. 

  2002-12-24 03:53:02 by rcaputo; POE/Kernel.pm 1.228

    Scott Beck pointed out that statement-modifier foreach isn't
    supported in older versions of Perl. This patch replaces them with
    C<foreach BLOCK> versions, which should work much farther back in
    Perl history. 

  2002-12-21 04:39:13 by rcaputo; POE/Kernel.pm 1.227

    Damien Neil discovered a temporary Session reference leak in the new
    signal handling code. The leak prevents sessions from stopping when
    they should (they do stop on subsequent signals, though). We were
    able to fix the problem thanks to his information. 

  2002-12-19 05:32:06 by rcaputo
  POE/Kernel.pm 1.226; POE/Pipe.pm 1.7; POE/Wheel/SocketFactory.pm 1.67

    Replace the homegrown blocking and nonblocking handle code with
    IO::Handle's blocking() method. This should do the right thing in
    more places with less work. For example, it fixes the ioctl() errors
    introduced in ActiveState 5.8.0 without breaking things on FreeBSD
    5.6.1. 

  2002-12-18 17:38:25 by rcaputo; POE/Kernel.pm 1.225

    Applied Garrett Goebel's patch that allows call() to honor wantarray
    context. 

  2002-12-18 17:26:46 by rcaputo; POE/Kernel.pm 1.224

    Applied Garrett Goebel's patch to clean up calls to
    _data_ev_refcount_dec(). This does three things: 1) Disambiguates a
    $_ in a foreach loop. 2) Removes a few temporary lexical variables.
    3) Reduces the number of lines of code. 

  2002-12-16 17:32:55 by rcaputo; POE/Wheel/FollowTail.pm 1.41

    Applied Matt Sergeant's globref patch. 

  2002-12-13 03:44:26 by rcaputo; POE/Wheel/FollowTail.pm 1.40

    Removed an unneeded Fcntl module use that was also causing trouble
    with Rob Bloodgood's 5.005_03 system. 

  2002-12-13 03:28:18 by rcaputo; POE/Kernel.pm 1.223

    Apply Rob Bloodgood's patch to support 5.005_03 in POE::Kernel. 

  2002-12-10 03:09:17 by rcaputo; POE/Session.pm 1.82

    Damien Neil pointed out that Session.pm destroys its package_states
    parameter. This is rude and has been fixed. 

  2002-12-09 20:22:41 by rcaputo; POE.pm 1.154

    Bump POE's version to 0.2301 for development. 

Back to POE CHANGES.