parse altar alignment from the floor message

[doy]

See TAEB/TAEB-AI-Behavioral#77.

[seankelly]

There is an announcement for it:

    [
        qr/^There is an altar to [\w\- ]+ \((law|neu|cha|unaligned)\w*\) here\./ =>
            ['dungeon_feature' => sub { ucfirst($1) .' altar' } ],
    ],

This appears to affect the tile properly:

<T9318> 2013-12-24 12:13:01 [DEBUG:publisher] Announcing dungeon_feature with arguments Law altar.
<T9318> 2013-12-24 12:13:01 [DEBUG:cartographer] msg_dungeon_feature('Law altar') caused the current tile to be updated from ('@', 'obscured') to ('_', 'altar')
<T9318> 2013-12-24 12:13:01 [DEBUG:publisher] Announcing tile_type_change with arguments TAEB::Announcement::Dungeon::Tile::TypeChange=HASH(0x80cfc5018).
<T9318> 2013-12-24 12:13:01 [DEBUG:moose] Reblessing a TAEB::World::Tile into TAEB::World::Tile::Altar.
<T9318> 2013-12-24 12:13:01 [DEBUG:publisher] Announcing check with arguments tile [TAEB::World::Tile::Altar: (47,5) Altar t=altar g<@> i< > f<_> i=9].

[seankelly]

From the msg_dungeon_feature method:

    elsif ($feature =~ /\baltar$/) {
        $floor = '_';
        $type = 'altar';
        $subtype = shift;
    }
    ...
    $tile->change_type($type => $floor, $subtype);

But change_type doesn't do anything with the subtype:

sub change_type {
    my $self     = shift;
    my $newtype  = shift;
    my $newglyph = shift;

(No other shifts in the method.)

So, the message is parsed but the alignment is lost.

[seankelly]

Wait. The change_type method ends with:

    $self->rebless("TAEB::World::Tile::\L\u$newtype", @_);

Going to Altar->rebless:

sub reblessed {
    my $self = shift;
    my ($old_class, $align) = @_;

    if ($align) {
        $self->align($align);
        return;
    }

    TAEB->send_message(check => tile => $self);
}

So, maybe the alignment isn't lost?