1 # Copyright (C) 2019 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # I have no idea how stable or safe this is for handling untrusted
5 # input, but it seems to have been around for a while, and the
6 # highlight(1) executable is supported by gitweb and cgit.
8 # I'm also unsure about API stability, but highlight 3.x seems to
9 # have been around a few years and ikiwiki (apparently the only
10 # user of the SWIG/Perl bindings, at least in Debian) hasn't needed
11 # major changes to support it in recent years.
13 # Some code stolen from ikiwiki (GPL-2.0+)
14 # wrapper for SWIG-generated highlight.pm bindings
15 package PublicInbox::HlMod;
18 use highlight; # SWIG-generated stuff
21 sub _parse_filetypes ($) {
22 my $ft_conf = $_[0]->searchFile('filetypes.conf') or
23 die 'filetypes.conf not found by highlight';
24 open my $fh, '<', $ft_conf or die "failed to open($ft_conf): $!";
28 my @shebang; # order matters
30 # Hrm... why isn't this exposed by the highlight API?
31 # highlight >= 3.2 format (bind-style) (from ikiwiki)
32 while ($cfg =~ /\bLang\s*=\s*\"([^"]+)\"[,\s]+
33 Extensions\s*=\s*{([^}]+)}/sgx) {
35 foreach my $bit (split(/,/, $2)) {
36 $bit =~ s/.*"(.*)".*/$1/s;
37 $ext2lang{$bit} = $lang;
40 # AFAIK, all the regexps used by in filetypes.conf distributed
41 # by highlight work as Perl REs
42 while ($cfg =~ /\bLang\s*=\s*\"([^"]+)\"[,\s]+
43 Shebang\s*=\s*\[\s*\[([^}]+)\s*\]\s*\]\s*}\s*,/sgx) {
44 my ($lang, $re) = ($1, $2);
46 my $perl_re = qr/$re/;
47 push @shebang, [ $lang, $perl_re ];
50 warn "$lang shebang=[[$re]] did not work in Perl: $@";
53 (\%ext2lang, \@shebang);
56 # We only need one instance, so we don't need to do
57 # highlight::CodeGenerator::deleteInstance
61 my $dir = highlight::DataDir->new;
62 $dir->initSearchDirectories('');
63 my ($ext2lang, $shebang) = _parse_filetypes($dir);
66 -ext2lang => $ext2lang,
72 sub _shebang2lang ($$) {
73 my ($self, $str) = @_;
74 my $shebang = $self->{-shebang};
75 foreach my $s (@$shebang) {
76 return $s->[0] if $$str =~ $s->[1];
82 my ($self, $path) = @_;
83 my ($ext) = ($path =~ m!([^\\/\.]+)\z!);
85 $self->{-ext2lang}->{$ext} || $ext;
89 my ($self, $str, $path) = @_;
90 my $lang = _path2lang($self, $path) if defined $path;
91 do_hl_lang($self, $str, $lang);
95 my ($self, $str, $lang) = @_;
97 my $dir = $self->{-dir};
101 $langpath = $dir->getLangPath("$lang.lang") or return;
102 $lang = undef unless -f $langpath
104 unless (defined $lang) {
105 $lang = _shebang2lang($self, $str) or return;
106 $langpath = $dir->getLangPath("$lang.lang") or return;
107 return unless -f $langpath
109 my $gen = $self->{$langpath} ||= do {
110 my $g = highlight::CodeGenerator::getInstance($highlight::HTML);
111 $g->setFragmentCode(1); # generate html fragment
113 # whatever theme works
114 my $themepath = $dir->getThemePath('print.theme');
115 $g->initTheme($themepath);
116 $g->loadLanguage($langpath);
117 $g->setEncoding('utf-8');
121 # we assume $$str is valid UTF-8, but the SWIG binding doesn't
122 # know that, so ensure it's marked as UTF-8 even if it isnt...
123 my $out = $gen->generateString($$str);