mkindex

#!/usr/bin/env perl
#
# mkindex was written by Omar Polo <op@openbsd.org> and is placed in
# the public domain.  The author hereby disclaims copyright to this
# source code.

use open ":std", ":encoding(UTF-8)";
use utf8;
use strict;
use warnings;
use v5.32;
use File::Temp qw(tempfile);

use SMArc qw(parse san urlencode initpage endpage index_header
    search thread_header threntry);

my $outdir = $ENV{'OUTDIR'};
die 'Set $OUTDIR' unless defined $outdir;

my $tfh; # thread file handle
my $pfh; # page file handle
my $page = 0;
my @pages;
my @files;
my $from_day;
my $to_day;
my $threads_seen = 0;

my $last_level = 0;
my $last_tid;
my $last_date;
my $last_from;
my $last_subj;

my $threads = 0;
my $threads_per_page = 100;

sub maxs {
	my ($a, $b) = @_;
	return $a unless defined $b;
	return $a gt $b ? $a : $b;
}

sub mins {
	my ($a, $b) = @_;
	return $a unless defined $b;
	return $a lt $b ? $a : $b;
}

sub pagename {
	my $i = shift;
	return $i == 1 && "index.html" || "$i.html";
}

sub endfile {
	say $pfh '</ul></div>';
	close($pfh);
	push @pages, "$from_day - $to_day";
}

sub nextfile {
	endfile if defined $pfh;
	$page += 1;

	my $path;
	($pfh, $path) = tempfile "/tmp/smarc.index.XXXXXXXXXX";
	binmode($pfh, ':utf8');
	push @files, $path;
	say $pfh "<div class='thread'><ul>";
}

sub nav {
	my ($pfh, $n) = @_;
	my ($first, $last) = (pagename(1), pagename($page));
	my ($next, $prev) = (pagename($n+1), pagename($n-1));

	say $pfh "<nav>";
	say $pfh "<a href='$first'>First</a>" if $n > 2;
	say $pfh "<a href='$prev'>Prev</a>" if $n > 1;
	say $pfh "<a href='$next'>Next</a>" if $n < $page;
	say $pfh "<a href='$last'>Last</a>" if $n < $page - 1;
	say $pfh "</nav>";
}

sub copyfrom {
	my ($path, $fh) = @_;

	# there are probably faster ways to do this like File::Copy,
	# but it bypasses the bufio cache...
	open(my $pfh, '<', $path) or die "can't open $path: $!";
	print $fh $_ while <$pfh>;
}

sub renderpages {
	close($pfh);

	for (my $i = 1; $i <= $page; $i++) {
		my $name = pagename($i);
		my $path = shift @files;
		my $dest = "$outdir/$name";

		open(my $pfh, '>', $dest)
		    or die "can't open $dest for writing: $!";

		my $title = "page $i";
		my $subtitle = $pages[$i-1];

		initpage($pfh, $title);
		index_header $pfh, $i, $subtitle;
		say $pfh "<main>";

		nav $pfh, $i if $page > 1;
		search $pfh;
		copyfrom($path, $pfh);
		nav $pfh, $i if $page > 1;

		say $pfh "</main>";
		endpage($pfh);

		close($pfh);
		unlink $path;
	}
}

sub endthread {
	say $tfh "</ul></li>" x $last_level;
	say $tfh "</ul>\n</div>\n";
	endpage($tfh);
	close($tfh);

	$last_level = 0;
}

sub nextthread {
	endthread if defined $tfh;
	my ($mid, $subj) = @_;
	my $dest = "$outdir/thread/$mid.html";
	open($tfh, '>', $dest) or die "can't open $dest: $!";
	initpage($tfh, $subj);
	thread_header $tfh, ["Thread: $subj"];
	print $tfh "<div class='thread'><ul class='mails'>\n";
}

sub index_entry {
	my ($fh, $mid, $date, $from, $subj) = @_;

	# synthetic mail hash
	my $mail = {
		mid => $mid,
		level => 0,
		date => $date,
		from => $from,
		subj => $subj,
	};

	threntry $fh, "thread", 0, 0, $mail;
}

if (`uname` =~ "OpenBSD") {
	use OpenBSD::Pledge;
	use OpenBSD::Unveil;

	unveil($outdir, "rwc") or die "unveil $outdir: $!";

	# can't use tmppath because File::Temp checks whether /tmp exists.
	unveil("/tmp", "rwc") or die "unveil /tmp: $!";

	# fattr for File::Temp
	pledge("stdio rpath wpath cpath fattr") or die "pledge: $!";
}

nextfile;

while (<>) {
	my $mail = parse $_;

	if ($mail->{level} == 0) {
		nextthread $mail->{mid}, $mail->{subj};

		$threads++;
		if ($threads > $threads_per_page) {
			nextfile;
			$threads = 0;
			$to_day = undef;
			$from_day = undef;
		}

		my $day = $mail->{date} =~ s/ .*//r;
		$to_day = mins $day, $to_day;
		$from_day = maxs $day, $from_day;
	}

	$last_level = threntry $tfh, "mail", 0, $last_level, $mail;
	$threads_seen = 1;

	index_entry $pfh, $last_tid, $last_date, $last_from, $last_subj
	    if defined $last_tid && $mail->{level} == 0;

	# `gt' on dates works because the format used allow for
	# lexicographic comparisons.
	if ($mail->{level} == 0 || $mail->{date} gt $last_date) {
		$last_date = $mail->{date};
		$last_from = $mail->{from};
	}

	if ($mail->{level} == 0) {
		$last_tid = $mail->{mid};
		$last_subj = $mail->{subj};
	}
}

index_entry $pfh, $last_tid, $last_date, $last_from, $last_subj
    if defined $last_tid;

endfile;
endthread if $threads_seen;
renderpages;