# Copyright 2014 Jeffrey Kegler
# This file is part of Marpa::R2.  Marpa::R2 is free software: you can
# redistribute it and/or modify it under the terms of the GNU Lesser
# General Public License as published by the Free Software Foundation,
# either version 3 of the License, or (at your option) any later version.
#
# Marpa::R2 is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser
# General Public License along with Marpa::R2.  If not, see
# http://www.gnu.org/licenses/.

=head1 NAME

Marpa::R2::NAIF::Semantics::Phases - Details of NAIF parse evaluation

=head1 Overview of the semantic phases

This document deals with Marpa's low-level NAIF interface.
If you are new to Marpa,
or are not sure which interface you are interested in,
or do not know what the Named Argment InterFace (NAIF) is,
you probably want to look instead at
L<the document on semantics for the SLIF
interface|Marpa::R2::Semantics>.

Most applications will find that the order in which
Marpa's NAIF executes its semantics "just works".
This document describes that order in detail.
These details can matter in some applications,
for example, those which exploit side effects.
And some readers may also find this background material
to be helpful.

An advanced document,
this is written on the assumption that the more low-level NAIF interface
is in use.
The SLIF uses the same code as the NAIF for evaluation
and works in the same way,
except that it hides even more of these details from the user.

=head1 Parse trees, parse results and parse series

As a reminder,
when the semantics are applied to a parse tree,
they produce a value called a B<parse result>.
Because Marpa allows ambiguous parsing,
each parse can produce a B<parse series> --
a series of zero or more parse trees,
each with its own parse result.
The first call to the
L<the recognizer's C<value>
method|Marpa::R2::NAIF::Recognizer/"value">
after the recognizer is created is the
start of the first parse series.
The first parse series continues until there is
a call to the
L<the C<reset_evaluation>
method|Marpa::R2::NAIF::Recognizer/"reset_evaluation">
or until the recognizer is destroyed.
Usually, an application is only interested in a single
parse series.

When the
L<C<reset_evaluation>|Marpa::R2::NAIF::Recognizer/"reset_evaluation">
method
is called
for a recognizer, it begins a new parse series.
The new parse series continues until
there is another
call to the
L<the C<reset_evaluation>
method|Marpa::R2::NAIF::Recognizer/"reset_evaluation">,
or until the recognizer is destroyed.

=head1 Summary of the phases

While processing a parse series, we have:

=over

=item * A Series Setup Phase, which occurs during
the first call of the recognizer's C<value> method
for that series.
It is followed by

=item * the processing of zero or more parse trees.

=back

While processing a parse tree, we have:

=over

=item * A Tree Setup Phase, which occurs during
the call of the recognizer's C<value> method
for that parse tree.
It is followed by

=item * a Tree Traveral Phase.

=back

B<Node Evaluation Time>
is the Tree Traversal Phase, as seen from the point of view of
each rule node.  It is not a separate phase.

=head1 Series Setup Phase

During the Series Setup Phase
all value action names are resolved to
value actions --
constants
or rule evaluation closures.
The rule evaluation closures are never called in the Series Setup Phase.
They will be called later,
in the Tree Traversal Phase.
Also, during the Series Setup Phase,
the logic which
ranks parse trees is executed.

=head1 Tree Setup Phase

In the Tree Setup Phase,
the per-parse-tree variable is created.
If a constructor was found for the C<action_object>,
it is run at this point, and the per-parse-tree variable is
its return value.
Exactly one Tree Setup Phase occurs
for each parse tree.

=head1 Tree Traversal Phase

During the Tree Traversal Phase,
the rule evaluation closures are called.
Node Evaluation Time is the Tree Traversal Phase,
as seen from the point of view of the individual nodes of the parse tree.

=head1 Copyright and License

=for Marpa::R2::Display
ignore: 1

  Copyright 2014 Jeffrey Kegler
  This file is part of Marpa::R2.  Marpa::R2 is free software: you can
  redistribute it and/or modify it under the terms of the GNU Lesser
  General Public License as published by the Free Software Foundation,
  either version 3 of the License, or (at your option) any later version.

  Marpa::R2 is distributed in the hope that it will be useful,
  but WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  Lesser General Public License for more details.

  You should have received a copy of the GNU Lesser
  General Public License along with Marpa::R2.  If not, see
  http://www.gnu.org/licenses/.

=for Marpa::R2::Display::End

=cut

# Local Variables:
#   mode: cperl
#   cperl-indent-level: 4
#   fill-column: 100
# End:
# vim: expandtab shiftwidth=4: