Permalink
Commits on Mar 26, 2008
  1. At long last, it is time.

    committed Mar 26, 2008
Commits on Mar 23, 2008
  1. Addressed several of the CPAN testers failures. Releasing a new

    test version to flush out any new issues.
    committed Mar 23, 2008
  2. Simulate a driver get() error by reading from a write-only filehandle.

    I suspect that Perl 5.6.2 is reporting the act of reading from a
    closed fielhandle as an EOF rather than an error, so reading from a
    write-only filehandle may trigger different behavior.
    committed Mar 23, 2008
  3. Looks like Perl 5.8.7 mistreats my @x = $o->method() =~ m//g; as an

    lvalue subroutine motification, or something.  Split it into a scalar
    assign, and bind the match to the scalar rather than a method return
    value.
    committed Mar 23, 2008
  4. Contrary to the docs, IO::Pty's clone_winsize_from() seems to die

    noisily if the source filehandle isn't a tty.
    committed Mar 23, 2008
Commits on Mar 20, 2008
Commits on Mar 19, 2008
  1. Solve POD test coverage issues.

    committed Mar 19, 2008
Commits on Mar 14, 2008
Commits on Mar 10, 2008
  1. Apply SREZIC's patch to work around an issue on 64-bit FreeBSD 6.2.

    Resolves rt.cpan.org #32034.
    committed Mar 10, 2008
  2. Use IO::Pty helper methods to condition the slave pty rather than the

    APITUE based low-level calls.  Sorry, Mr. Stevens.
    committed Mar 10, 2008
  3. Add a reference to POE's subversion repository, per Zoffix's

    suggestion.  Resolves rt.cpan.org #33835.
    committed Mar 10, 2008
  4. Minor tweaks.

    committed Mar 10, 2008
  5. Redocument POE::Loop.

    committed Mar 10, 2008
  6. Socket6 and Socket both export AF_INET6, but their prototypes don't m…

    …atch. Ask
    
    Socket6 to keep its symbols to itself.
    
    Skip even more tests that crash perl on Win32.
    committed Mar 10, 2008
Commits on Mar 8, 2008
  1. Make the sense of the network test use positive logic

    interactive use defaults to "do not grant permission"
    non interactive defaults to "grant permission"
    cfedde committed Mar 8, 2008
Commits on Mar 1, 2008
Commits on Feb 29, 2008
  1. Make SIGDIE asynchronous. It will no longer interrupt long chains of

    events (such as _parent/_child chains during session creation and
    destruction).
    
    Exceptions in _stop is currently untrappable, as the asynchronous
    SIGDIE will have nowhere to be delivered.  They will cause POE::Kernel
    to "gracefully" die at the end of _stop's dispatch.
    
    SIGDIE propagation follows the normal rules with one new addition: The
    signal also proragates from the target session upwards through parent,
    grandparent, and so on.  It will not be dispatched to children of
    ancestor sessions.  It will not be dispatched to POE::Kernel itself.
    
    Nearly all regression tests pass.  The notable exception is Jonathan
    Steinert's recursive signal tests.  They have become moot since SIGDIE
    is not recursive anymore.
    committed Feb 29, 2008
Commits on Feb 25, 2008
Commits on Feb 20, 2008
  1. Patch from Martin Ferrari to fix tests not honouring the "don't use t…

    …he network" flag. RT #33276
    bingos committed Feb 20, 2008
  2. Patch from agaran to fix a problem with his original patch to Filter-…

    …HTTPD. Adjusted so the tests pass.
    bingos committed Feb 20, 2008
Commits on Feb 18, 2008
  1. Applied a patch to Filter::HTTPD from Maciej Pijanka (agaran). Create…

    …d applicable regression test.
    bingos committed Feb 18, 2008
Commits on Jan 22, 2008
  1. Comment on sub-second accuracy requiring Time::HiRes::time(). Based o…

    …n a discussion with and test case from Sjors Gielen.
    committed Jan 22, 2008
Commits on Jan 12, 2008
Commits on Jan 11, 2008
  1. Fixed a pod error that was making pod test fail.

    Applied a patch from pravus to fix double InputEvents being
    triggered when tailing an empty file on *BSD.
    
    <+pravus> basically what i am seeing is that if you create a file and stat() it before you
              write to it, rdev will be 0.  then, once the file is written to, the rdev changes.
    <+pravus> that's causing Wheel::FollowTail to read until EOF (getting the line you just wrote),
              reset the file and then re-read (getting the line again).
    bingos committed Jan 11, 2008
Commits on Jan 7, 2008
  1. Redocument.

    committed Jan 7, 2008