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
19 use PublicInbox::Hval qw(src_escape ascii_html);
22 sub _parse_filetypes ($) {
23 my $ft_conf = $_[0]->getFiletypesConfPath('filetypes') or
24 die 'filetypes.conf not found by highlight';
25 open my $fh, '<', $ft_conf or die "failed to open($ft_conf): $!";
29 my @shebang; # order matters
31 # Hrm... why isn't this exposed by the highlight API?
32 # highlight >= 3.2 format (bind-style) (from ikiwiki)
33 while ($cfg =~ /\bLang\s*=\s*\"([^"]+)\"[,\s]+
34 Extensions\s*=\s*{([^}]+)}/sgx) {
36 foreach my $bit (split(/,/, $2)) {
37 $bit =~ s/.*"(.*)".*/$1/s;
38 $ext2lang{$bit} = $lang;
41 # AFAIK, all the regexps used by in filetypes.conf distributed
42 # by highlight work as Perl REs
43 while ($cfg =~ /\bLang\s*=\s*\"([^"]+)\"[,\s]+
44 Shebang\s*=\s*\[\s*\[([^}]+)\s*\]\s*\]\s*}\s*,/sgx) {
45 my ($lang, $re) = ($1, $2);
47 my $perl_re = qr/$re/;
48 push @shebang, [ $lang, $perl_re ];
51 warn "$lang shebang=[[$re]] did not work in Perl: $@";
54 (\%ext2lang, \@shebang);
57 # We only need one instance, so we don't need to do
58 # highlight::CodeGenerator::deleteInstance
62 my $dir = highlight::DataDir->new;
63 $dir->initSearchDirectories('');
64 my ($ext2lang, $shebang) = _parse_filetypes($dir);
67 -ext2lang => $ext2lang,
73 sub _shebang2lang ($$) {
74 my ($self, $str) = @_;
75 my $shebang = $self->{-shebang};
76 foreach my $s (@$shebang) {
77 return $s->[0] if $$str =~ $s->[1];
83 my ($self, $path) = @_;
84 my ($ext) = ($path =~ m!([^\\/\.]+)\z!);
86 $self->{-ext2lang}->{$ext} || $ext;
90 my ($self, $str, $path) = @_;
91 my $lang = _path2lang($self, $path) if defined $path;
92 do_hl_lang($self, $str, $lang);
96 my ($self, $str, $lang) = @_;
98 my $dir = $self->{-dir};
102 $langpath = $dir->getLangPath("$lang.lang") or return;
103 $lang = undef unless -f $langpath
105 unless (defined $lang) {
106 $lang = _shebang2lang($self, $str) or return;
107 $langpath = $dir->getLangPath("$lang.lang") or return;
108 return unless -f $langpath
110 my $gen = $self->{$langpath} ||= do {
111 my $g = highlight::CodeGenerator::getInstance($highlight::HTML);
112 $g->setFragmentCode(1); # generate html fragment
114 # whatever theme works
115 my $themepath = $dir->getThemePath('print.theme');
116 $g->initTheme($themepath);
117 $g->loadLanguage($langpath);
118 $g->setEncoding('utf-8');
122 # we assume $$str is valid UTF-8, but the SWIG binding doesn't
123 # know that, so ensure it's marked as UTF-8 even if it isnt...
124 my $out = $gen->generateString($$str);
130 # Highlight text, but support Markdown "```$LANG" notation
131 # while preserving WYSIWYG of plain-text documentation.
132 # This is NOT to be enabled by default or encouraged for parsing
133 # emails, since it is NOT stable and can lead to standards
134 # proliferation of email.
136 my ($self, $str) = @_;
138 $$str = join('', map {
139 if (/\A(``` ?)(\w+)\s*?\n(.+)(^```\s*)\z/sm) {
140 my ($pfx, $lang, $code, $post) = ($1, $2, $3, $4);
141 my $hl = do_hl_lang($self, \$code, $lang) || \$code;
142 $pfx . $lang . "\n" . $$hl . $post;
146 } split(/(^``` ?\w+\s*?\n.+?^```\s*$)/sm, $$str));