X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/af9566ff1ae55d273ff96c7ca39e34265623a55d..a79f5ce802693f029de0cd705c36e4e7482160fb:/IkiWiki.pm?ds=sidebyside

diff --git a/IkiWiki.pm b/IkiWiki.pm
index 960d26de9..990836f8e 100644
--- a/IkiWiki.pm
+++ b/IkiWiki.pm
@@ -3,13 +3,81 @@
 package IkiWiki;
 use warnings;
 use strict;
-use File::Spec;
-use HTML::Template;
+use Encode;
+use HTML::Entities;
+use open qw{:utf8 :std};
+
+# Optimisation.
+use Memoize;
+memoize("abs2rel");
+memoize("pagespec_translate");
+
+use vars qw{%config %links %oldlinks %oldpagemtime %pagectime %pagecase
+            %renderedfiles %pagesources %depends %hooks %forcerebuild};
+
+my $installdir=''; # INSTALLDIR_AUTOREPLACE done by Makefile, DNE
+
+sub defaultconfig () { #{{{
+	wiki_file_prune_regexp => qr{((^|/).svn/|\.\.|^\.|\/\.|\.x?html?$|\.rss$)},
+	wiki_link_regexp => qr/\[\[(?:([^\]\|]+)\|)?([^\s\]]+)\]\]/,
+	wiki_file_regexp => qr/(^[-[:alnum:]_.:\/+]+$)/,
+	verbose => 0,
+	syslog => 0,
+	wikiname => "wiki",
+	default_pageext => "mdwn",
+	cgi => 0,
+	rcs => 'svn',
+	notify => 0,
+	url => '',
+	cgiurl => '',
+	historyurl => '',
+	diffurl => '',
+	anonok => 0,
+	rss => 0,
+	discussion => 1,
+	rebuild => 0,
+	refresh => 0,
+	getctime => 0,
+	w3mmode => 0,
+	wrapper => undef,
+	wrappermode => undef,
+	svnrepo => undef,
+	svnpath => "trunk",
+	srcdir => undef,
+	destdir => undef,
+	pingurl => [],
+	templatedir => "$installdir/share/ikiwiki/templates",
+	underlaydir => "$installdir/share/ikiwiki/basewiki",
+	setup => undef,
+	adminuser => undef,
+	adminemail => undef,
+	plugin => [qw{mdwn inline htmlscrubber}],
+	timeformat => '%c',
+	locale => undef,
+	sslcookie => 0,
+} #}}}
+   
+sub checkconfig () { #{{{
+	# locale stuff; avoid LC_ALL since it overrides everything
+	if (defined $ENV{LC_ALL}) {
+		$ENV{LANG} = $ENV{LC_ALL};
+		delete $ENV{LC_ALL};
+	}
+	if (defined $config{locale}) {
+		eval q{use POSIX};
+		$ENV{LANG} = $config{locale}
+			if POSIX::setlocale(&POSIX::LC_TIME, $config{locale});
+	}
 
-use vars qw{%config %links %oldlinks %oldpagemtime %pagectime
-            %renderedfiles %pagesources %depends %hooks};
+	if ($config{w3mmode}) {
+		eval q{use Cwd q{abs_path}};
+		$config{srcdir}=possibly_foolish_untaint(abs_path($config{srcdir}));
+		$config{destdir}=possibly_foolish_untaint(abs_path($config{destdir}));
+		$config{cgiurl}="file:///\$LIB/ikiwiki-w3m.cgi/".$config{cgiurl}
+			unless $config{cgiurl} =~ m!file:///!;
+		$config{url}="file://".$config{destdir};
+	}
 
-sub checkconfig () { #{{{
 	if ($config{cgi} && ! length $config{url}) {
 		error("Must specify url to wiki with --url when using --cgi\n");
 	}
@@ -30,6 +98,10 @@ sub checkconfig () { #{{{
 		require IkiWiki::Rcs::Stub;
 	}
 
+	run_hooks(checkconfig => sub { shift->() });
+} #}}}
+
+sub loadplugins () { #{{{
 	foreach my $plugin (@{$config{plugin}}) {
 		my $mod="IkiWiki::Plugin::".possibly_foolish_untaint($plugin);
 		eval qq{use $mod};
@@ -37,12 +109,12 @@ sub checkconfig () { #{{{
 			error("Failed to load plugin $mod: $@");
 		}
 	}
-
-	if (exists $hooks{checkconfig}) {
-                foreach my $id (keys %{$hooks{checkconfig}}) {
-                        $hooks{checkconfig}{$id}{call}->();
-                }
-        }
+	run_hooks(getopt => sub { shift->() });
+	if (grep /^-/, @ARGV) {
+		print STDERR "Unknown option: $_\n"
+			foreach grep /^-/, @ARGV;
+		usage();
+	}
 } #}}}
 
 sub error ($) { #{{{
@@ -50,12 +122,31 @@ sub error ($) { #{{{
 		print "Content-type: text/html\n\n";
 		print misctemplate("Error", "<p>Error: @_</p>");
 	}
-	die @_;
+	log_message(error => @_);
+	exit(1);
 } #}}}
 
 sub debug ($) { #{{{
 	return unless $config{verbose};
-	if (! $config{cgi}) {
+	log_message(debug => @_);
+} #}}}
+
+my $log_open=0;
+sub log_message ($$) { #{{{
+	my $type=shift;
+
+	if ($config{syslog}) {
+		require Sys::Syslog;
+		unless ($log_open) {
+			Sys::Syslog::setlogsock('unix');
+			Sys::Syslog::openlog('ikiwiki', '', 'user');
+			$log_open=1;
+		}
+		eval {
+			Sys::Syslog::syslog($type, join(" ", @_));
+		}
+	}
+	elsif (! $config{cgi}) {
 		print "@_\n";
 	}
 	else {
@@ -86,12 +177,10 @@ sub dirname ($) { #{{{
 sub pagetype ($) { #{{{
 	my $page=shift;
 	
-	if ($page =~ /\.mdwn$/) {
-		return ".mdwn";
-	}
-	else {
-		return "unknown";
+	if ($page =~ /\.([^.]+)$/) {
+		return $1 if exists $hooks{htmlize}{$1};
 	}
+	return undef;
 } #}}}
 
 sub pagename ($) { #{{{
@@ -99,7 +188,7 @@ sub pagename ($) { #{{{
 
 	my $type=pagetype($file);
 	my $page=$file;
-	$page=~s/\Q$type\E*$// unless $type eq 'unknown';
+	$page=~s/\Q.$type\E*$// if defined $type;
 	return $page;
 } #}}}
 
@@ -127,7 +216,7 @@ sub readfile ($;$) { #{{{
 	
 	local $/=undef;
 	open (IN, $file) || error("failed to read $file: $!");
-	binmode(IN) if $binary;
+	binmode(IN) if ($binary);
 	my $ret=<IN>;
 	close IN;
 	return $ret;
@@ -159,7 +248,7 @@ sub writefile ($$$;$) { #{{{
 	}
 	
 	open (OUT, ">$destdir/$file") || error("failed to write $destdir/$file: $!");
-	binmode(OUT) if $binary;
+	binmode(OUT) if ($binary);
 	print OUT $content;
 	close OUT;
 } #}}}
@@ -171,7 +260,7 @@ sub bestlink ($$) { #{{{
 	# goes down the directory tree to the base looking for matching
 	# pages.
 	my $page=shift;
-	my $link=lc(shift);
+	my $link=shift;
 	
 	my $cwd=$page;
 	do {
@@ -180,9 +269,11 @@ sub bestlink ($$) { #{{{
 		$l.=$link;
 
 		if (exists $links{$l}) {
-			#debug("for $page, \"$link\", use $l");
 			return $l;
 		}
+		elsif (exists $pagecase{lc $l}) {
+			return $pagecase{lc $l};
+		}
 	} while $cwd=~s!/?[^/]+$!!;
 
 	#print STDERR "warning: page $page, broken link: $link\n";
@@ -215,18 +306,32 @@ sub cgiurl (@) { #{{{
 	return $config{cgiurl}."?".join("&amp;", map "$_=$params{$_}", keys %params);
 } #}}}
 
-sub styleurl (;$) { #{{{
+sub baseurl (;$) { #{{{
 	my $page=shift;
 
-	return "$config{url}/style.css" if ! defined $page;
+	return "$config{url}/" if ! defined $page;
 	
 	$page=~s/[^\/]+$//;
 	$page=~s/[^\/]+\//..\//g;
-	return $page."style.css";
+	return $page;
 } #}}}
 
-sub htmllink ($$;$$$) { #{{{
-	my $page=shift;
+sub abs2rel ($$) { #{{{
+	# Work around very innefficient behavior in File::Spec if abs2rel
+	# is passed two relative paths. It's much faster if paths are
+	# absolute!
+	my $path="/".shift;
+	my $base="/".shift;
+
+	require File::Spec;
+	my $ret=File::Spec->abs2rel($path, $base);
+	$ret=~s/^// if defined $ret;
+	return $ret;
+} #}}}
+
+sub htmllink ($$$;$$$) { #{{{
+	my $lpage=shift; # the page doing the linking
+	my $page=shift; # the page that will contain the link (different for inline)
 	my $link=shift;
 	my $noimageinline=shift; # don't turn links into inline html images
 	my $forcesubpage=shift; # force a link to a subpage
@@ -234,30 +339,30 @@ sub htmllink ($$;$$$) { #{{{
 
 	my $bestlink;
 	if (! $forcesubpage) {
-		$bestlink=bestlink($page, $link);
+		$bestlink=bestlink($lpage, $link);
 	}
 	else {
-		$bestlink="$page/".lc($link);
+		$bestlink="$lpage/".lc($link);
 	}
 
 	$linktext=pagetitle(basename($link)) unless defined $linktext;
 	
-	return $linktext if length $bestlink && $page eq $bestlink;
+	return "<span class=\"selflink\">$linktext</span>"
+		if length $bestlink && $page eq $bestlink;
 	
 	# TODO BUG: %renderedfiles may not have it, if the linked to page
 	# was also added and isn't yet rendered! Note that this bug is
-	# masked by the bug mentioned below that makes all new files
-	# be rendered twice.
+	# masked by the bug that makes all new files be rendered twice.
 	if (! grep { $_ eq $bestlink } values %renderedfiles) {
 		$bestlink=htmlpage($bestlink);
 	}
 	if (! grep { $_ eq $bestlink } values %renderedfiles) {
 		return "<span><a href=\"".
-			cgiurl(do => "create", page => $link, from =>$page).
+			cgiurl(do => "create", page => lc($link), from => $page).
 			"\">?</a>$linktext</span>"
 	}
 	
-	$bestlink=File::Spec->abs2rel($bestlink, dirname($page));
+	$bestlink=abs2rel($bestlink, dirname($page));
 	
 	if (! $noimageinline && isinlinableimage($bestlink)) {
 		return "<img src=\"$bestlink\" alt=\"$linktext\" />";
@@ -302,7 +407,7 @@ sub loadindex () { #{{{
 		$items{link}=[];
 		foreach my $i (split(/ /, $_)) {
 			my ($item, $val)=split(/=/, $i, 2);
-			push @{$items{$item}}, $val;
+			push @{$items{$item}}, decode_entities($val);
 		}
 
 		next unless exists $items{src}; # skip bad lines for now
@@ -313,9 +418,9 @@ sub loadindex () { #{{{
 			$oldpagemtime{$page}=$items{mtime}[0];
 			$oldlinks{$page}=[@{$items{link}}];
 			$links{$page}=[@{$items{link}}];
-			$depends{$page}=join(" ", @{$items{depends}})
-				if exists $items{depends};
+			$depends{$page}=$items{depends}[0] if exists $items{depends};
 			$renderedfiles{$page}=$items{dest}[0];
+			$pagecase{lc $page}=$page;
 		}
 		$pagectime{$page}=$items{ctime}[0];
 	}
@@ -323,6 +428,8 @@ sub loadindex () { #{{{
 } #}}}
 
 sub saveindex () { #{{{
+	run_hooks(savestate => sub { shift->() });
+
 	if (! -d $config{wikistatedir}) {
 		mkdir($config{wikistatedir});
 	}
@@ -336,32 +443,165 @@ sub saveindex () { #{{{
 			"dest=$renderedfiles{$page}";
 		$line.=" link=$_" foreach @{$links{$page}};
 		if (exists $depends{$page}) {
-			$line.=" depends=$_" foreach split " ", $depends{$page};
+			$line.=" depends=".encode_entities($depends{$page}, " \t\n");
 		}
 		print OUT $line."\n";
 	}
 	close OUT;
 } #}}}
 
+sub template_params (@) { #{{{
+	my $filename=shift;
+	
+	require HTML::Template;
+	return filter => sub {
+			my $text_ref = shift;
+			$$text_ref=&Encode::decode_utf8($$text_ref);
+		},
+		filename => "$config{templatedir}/$filename",
+		loop_context_vars => 1,
+		die_on_bad_params => 0,
+		@_;
+} #}}}
+
+sub template ($;@) { #{{{
+	HTML::Template->new(template_params(@_));
+} #}}}
+
 sub misctemplate ($$) { #{{{
 	my $title=shift;
 	my $pagebody=shift;
 	
-	my $template=HTML::Template->new(
-		filename => "$config{templatedir}/misc.tmpl"
-	);
+	my $template=template("misc.tmpl");
 	$template->param(
 		title => $title,
 		indexlink => indexlink(),
 		wikiname => $config{wikiname},
 		pagebody => $pagebody,
-		styleurl => styleurl(),
-		baseurl => "$config{url}/",
+		baseurl => baseurl(),
 	);
 	return $template->output;
 }#}}}
 
-sub glob_match ($$) { #{{{
+sub hook (@) { # {{{
+	my %param=@_;
+	
+	if (! exists $param{type} || ! ref $param{call} || ! exists $param{id}) {
+		error "hook requires type, call, and id parameters";
+	}
+	
+	$hooks{$param{type}}{$param{id}}=\%param;
+} # }}}
+
+sub run_hooks ($$) { # {{{
+	# Calls the given sub for each hook of the given type,
+	# passing it the hook function to call.
+	my $type=shift;
+	my $sub=shift;
+
+	if (exists $hooks{$type}) {
+		foreach my $id (keys %{$hooks{$type}}) {
+			$sub->($hooks{$type}{$id}{call});
+		}
+	}
+} #}}}
+
+sub globlist_to_pagespec ($) { #{{{
+	my @globlist=split(' ', shift);
+
+	my (@spec, @skip);
+	foreach my $glob (@globlist) {
+		if ($glob=~/^!(.*)/) {
+			push @skip, $glob;
+		}
+		else {
+			push @spec, $glob;
+		}
+	}
+
+	my $spec=join(" or ", @spec);
+	if (@skip) {
+		my $skip=join(" and ", @skip);
+		if (length $spec) {
+			$spec="$skip and ($spec)";
+		}
+		else {
+			$spec=$skip;
+		}
+	}
+	return $spec;
+} #}}}
+
+sub is_globlist ($) { #{{{
+	my $s=shift;
+	$s=~/[^\s]+\s+([^\s]+)/ && $1 ne "and" && $1 ne "or";
+} #}}}
+
+sub safequote ($) { #{{{
+	my $s=shift;
+	$s=~s/[{}]//g;
+	return "q{$s}";
+} #}}}
+
+sub pagespec_merge ($$) { #{{{
+	my $a=shift;
+	my $b=shift;
+
+	return $a if $a eq $b;
+
+        # Support for old-style GlobLists.
+        if (is_globlist($a)) {
+                $a=globlist_to_pagespec($a);
+        }
+        if (is_globlist($b)) {
+                $b=globlist_to_pagespec($b);
+        }
+
+	return "($a) or ($b)";
+} #}}}
+
+sub pagespec_translate ($) { #{{{
+	# This assumes that $page is in scope in the function
+	# that evalulates the translated pagespec code.
+	my $spec=shift;
+
+	# Support for old-style GlobLists.
+	if (is_globlist($spec)) {
+		$spec=globlist_to_pagespec($spec);
+	}
+
+	# Convert spec to perl code.
+	my $code="";
+	while ($spec=~m/\s*(\!|\(|\)|\w+\([^\)]+\)|[^\s()]+)\s*/ig) {
+		my $word=$1;
+		if (lc $word eq "and") {
+			$code.=" &&";
+		}
+		elsif (lc $word eq "or") {
+			$code.=" ||";
+		}
+		elsif ($word eq "(" || $word eq ")" || $word eq "!") {
+			$code.=" ".$word;
+		}
+		elsif ($word =~ /^(link|backlink|created_before|created_after|creation_month|creation_year|creation_day)\((.+)\)$/) {
+			$code.=" match_$1(\$page, ".safequote($2).")";
+		}
+		else {
+			$code.=" match_glob(\$page, ".safequote($word).")";
+		}
+	}
+
+	return $code;
+} #}}}
+
+sub pagespec_match ($$) { #{{{
+	my $page=shift;
+	my $spec=shift;
+
+	return eval pagespec_translate($spec);
+} #}}}
+
+sub match_glob ($$) { #{{{
 	my $page=shift;
 	my $glob=shift;
 
@@ -369,35 +609,59 @@ sub glob_match ($$) { #{{{
 	$glob=quotemeta($glob);
 	$glob=~s/\\\*/.*/g;
 	$glob=~s/\\\?/./g;
-	$glob=~s!\\/!/!g;
-	
-	$page=~/^$glob$/i;
+
+	return $page=~/^$glob$/i;
 } #}}}
 
-sub globlist_match ($$) { #{{{
+sub match_link ($$) { #{{{
 	my $page=shift;
-	my @globlist=split(" ", shift);
+	my $link=lc(shift);
 
-	# check any negated globs first
-	foreach my $glob (@globlist) {
-		return 0 if $glob=~/^!(.*)/ && glob_match($page, $1);
+	my $links = $links{$page} or return undef;
+	foreach my $p (@$links) {
+		return 1 if lc $p eq $link;
 	}
+	return 0;
+} #}}}
 
-	foreach my $glob (@globlist) {
-		return 1 if glob_match($page, $glob);
+sub match_backlink ($$) { #{{{
+	match_link(pop, pop);
+} #}}}
+
+sub match_created_before ($$) { #{{{
+	my $page=shift;
+	my $testpage=shift;
+
+	if (exists $pagectime{$testpage}) {
+		return $pagectime{$page} < $pagectime{$testpage};
+	}
+	else {
+		return 0;
 	}
-	
-	return 0;
 } #}}}
 
-sub hook (@) { # {{{
-	my %param=@_;
-	
-	if (! exists $param{type} || ! ref $param{call} || ! exists $param{id}) {
-		error "hook requires type, call, and id parameters";
+sub match_created_after ($$) { #{{{
+	my $page=shift;
+	my $testpage=shift;
+
+	if (exists $pagectime{$testpage}) {
+		return $pagectime{$page} > $pagectime{$testpage};
 	}
-	
-	$hooks{$param{type}}{$param{id}}=\%param;
-} # }}}
+	else {
+		return 0;
+	}
+} #}}}
+
+sub match_creation_day ($$) { #{{{
+	return ((gmtime($pagectime{shift()}))[3] == shift);
+} #}}}
+
+sub match_creation_month ($$) { #{{{
+	return ((gmtime($pagectime{shift()}))[4] + 1 == shift);
+} #}}}
+
+sub match_creation_year ($$) { #{{{
+	return ((gmtime($pagectime{shift()}))[5] + 1900 == shift);
+} #}}}
 
 1