2 # Copyright (C) 2014-2021 all contributors <meta@public-inbox.org>
3 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
5 # Stupid script to make HTML from preformatted, utf-8 text versions,
6 # only generating links for http(s). Markdown does too much
7 # and requires indentation to output preformatted text.
10 use PublicInbox::Linkify;
11 use PublicInbox::Hval qw(ascii_html);
24 lei-forget-external(1)
25 lei-forget-mail-sync(1)
39 lei-mail-sync-overview(7)
44 lei-refresh-mail-sync(1)
53 public-inbox-config(5)
54 public-inbox-config(5)
55 public-inbox-convert(1)
56 public-inbox-daemon(8)
59 public-inbox-glossary(7)
67 public-inbox-overview(7)
69 public-inbox-v1-format(5)
70 public-inbox-v2-format(5)
74 my ($n) = (/([\w\-\.]+)/);
75 $xurls{$_} = "$n.html";
76 $xurls{$n} = "$n.html";
79 for (qw[make(1) flock(2) setrlimit(2) vfork(2) tmpfs(5)]) {
80 my ($n, $s) = (/([\w\-]+)\((\d)\)/);
81 $xurls{$_} = "http://www.man7.org/linux/man-pages/man$s/$n.$s.html";
100 gitrepository-layout(5)
103 my ($n) = (/([\w\-\.]+)/);
104 $xurls{$_} = "https://kernel.org/pub/software/scm/git/docs/$n.html"
113 my ($n) = (/([\w\-\.]+)/);
114 $xurls{$_} = "https://www.freedesktop.org/software/systemd/man/$n.html";
117 # favor upstream docs if they exist, use manpages.debian.org if they don't
118 $xurls{'netrc(5)'} = 'https://manpages.debian.org/stable/ftp/netrc.5.en.html';
119 $xurls{'mbsync(1)'} =
120 'https://manpages.debian.org/stable/isync/mbsync.1.en.html';
121 $xurls{'offlineimap(1)'} =
122 'https://manpages.debian.org/stable/offlineimap/offlineimap.1.en.html';
124 'https://spamassassin.apache.org/full/3.4.x/doc/spamc.html';
125 $xurls{'grok-pull'} =
126 'https://git.kernel.org/pub/scm/utils/grokmirror/grokmirror.git' .
127 '/tree/man/grok-pull.1.rst';
128 $xurls{'git-filter-repo(1)'} = 'https://github.com/newren/git-filter-repo'.
129 '/blob/master/Documentation/git-filter-repo.txt';
130 $xurls{'ssoma(1)'} = 'https://ssoma.public-inbox.org/ssoma.txt';
131 $xurls{'cgitrc(5)'} = 'https://git.zx2c4.com/cgit/tree/cgitrc.5.txt';
132 $xurls{'prove(1)'} = 'https://perldoc.perl.org/prove.html';
133 $xurls{'mbox(5)'} = 'https://manpages.debian.org/stable/mutt/mbox.5.en.html';
134 $xurls{'mmdf(5)'} = 'https://manpages.debian.org/stable/mutt/mmdf.5.en.html';
135 $xurls{'torsocks(1)'} =
136 'https://manpages.debian.org/stable/torsocks/torsocks.1.en.html';
137 $xurls{'curl(1)'} = 'https://manpages.debian.org/stable/curl/curl.1.en.html';
138 $xurls{'copydatabase(1)'} =
139 'https://manpages.debian.org/stable/xapian-tools/copydatabase.1.en.html';
140 $xurls{'xapian-compact(1)'} =
141 'https://manpages.debian.org/stable/xapian-tools/xapian-compact.1.en.html';
143 my $str = do { local $/; <STDIN> };
144 my ($title) = ($str =~ /\A([^\n]+)/);
145 if ($str =~ /^NAME\n\s+([^\n]+)/sm) {
146 # don't link to ourselves
148 if ($title =~ /([\w\.\-]+)/) {
152 $title = ascii_html($title);
153 my $l = PublicInbox::Linkify->new;
154 $str = $l->linkify_1($str);
155 $str = ascii_html($str);
157 # longest matches, first
158 my @keys = sort { length($b) <=> length($a) } keys %xurls;
159 my $xkeys = join('|', map { quotemeta } @keys);
160 $str =~ s,(?<![>\w_])($xkeys)(?!(?:[\w<\-]|\.html)),
161 qq(<a\nhref=").$xurls{$1}.qq(">$1).($2//'').'</a>',sge;
163 $str = $l->linkify_2($str);
165 print '<html><head>',
166 qq(<meta\nhttp-equiv="Content-Type"\ncontent="text/html; charset=utf-8"\n/>),
167 "<title>$title</title>",
168 "</head><body><pre>", $str , '</pre></body></html>';
171 # keep mtime on website consistent so clients can cache
172 if (-f STDIN && -f STDOUT) {
173 my @st = stat(STDIN);
174 utime($st[8], $st[9], \*STDOUT);