has level => (is => 'rw');
has _body => (is => 'rw');
has abort => (is => 'rw');
-has autoflush => (is => 'rw');
+has autoflush => (is => 'rw', default => sub {1});
has _psgi_logger => (is => 'rw', predicate => '_has_psgi_logger', clearer => '_clear_psgi_logger');
has _psgi_errors => (is => 'rw', predicate => '_has_psgi_errors', clearer => '_clear_psgi_errors');
if( $self->autoflush && !$self->abort ) {
$self->_flush;
}
+ return 1;
}
sub _flush {
if ($self->can('_has_psgi_errors') and $self->_has_psgi_errors) {
$self->_psgi_errors->print(@_);
} else {
+ binmode STDERR, ":utf8";
print STDERR @_;
}
}
=head2 autoflush
-When enabled, messages are written to the log immediately instead of queued
-until the end of the request. By default, autoflush is enabled during setup,
-but turned back off thereafter. This is done purely for legacy support,
-specifically for L<Catalyst::Plugin::Static::Simple>, and may be changed in
-the future.
+When enabled (default), messages are written to the log immediately instead
+of queued until the end of the request.
+
+This option, as well as C<abort>, is provided for modules such as
+L<Catalyst::Plugin::Static::Simple> to be able to programmatically
+suppress the output of log messages. By turning off C<autoflush> (application-wide
+setting) and then setting the C<abort> flag within a given request, all log
+messages for the given request will be suppressed. C<abort> can still be set
+independently of turning off C<autoflush>, however. It just means any messages
+sent to the log up until that point in the request will obviously still be emitted,
+since C<autoflush> means they are written in real-time.
+
+If you need to turn off autoflush you should do it like this (in your main app
+class):
+
+ after setup_finalize => sub {
+ my $c = shift;
+ $c->log->autoflush(0) if $c->log->can('autoflush');
+ };
=head2 _send_to_log