X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/6551c1e5609967c0c7f947fa0e9d85d24e454d37..fad966e681aca8c81422e4a7383684438297dba9:/IkiWiki.pm diff --git a/IkiWiki.pm b/IkiWiki.pm index 39ceec04e..45a3627cc 100644 --- a/IkiWiki.pm +++ b/IkiWiki.pm @@ -3,20 +3,38 @@ package IkiWiki; use warnings; use strict; -use encoding "utf8"; # force use of utf8 for io layer use Encode; +use HTML::Entities; +use open qw{:utf8 :std}; -use vars qw{%config %links %oldlinks %oldpagemtime %pagectime - %renderedfiles %pagesources %depends %hooks}; +use vars qw{%config %links %oldlinks %oldpagemtime %pagectime %pagecase + %renderedfiles %oldrenderedfiles %pagesources %depends %hooks + %forcerebuild}; + +use Exporter q{import}; +our @EXPORT = qw(hook debug error template htmlpage add_depends pagespec_match + bestlink htmllink readfile writefile pagetype srcfile pagename + displaytime will_render + %config %links %renderedfiles %pagesources); +our $VERSION = 1.01; # plugin interface version + +# Optimisation. +use Memoize; +memoize("abs2rel"); +memoize("pagespec_translate"); + +my $installdir=''; # INSTALLDIR_AUTOREPLACE done by Makefile, DNE +our $version='unknown'; # VERSION_AUTOREPLACE done by Makefile, DNE sub defaultconfig () { #{{{ - wiki_file_prune_regexp => qr{((^|/).svn/|\.\.|^\.|\/\.|\.html?$|\.rss$)}, + wiki_file_prune_regexp => qr{((^|/).svn/|\.\.|^\.|\/\.|\.x?html?$|\.rss$|\.atom$|.arch-ids/|{arch}/)}, wiki_link_regexp => qr/\[\[(?:([^\]\|]+)\|)?([^\s\]]+)\]\]/, - wiki_processor_regexp => qr/\[\[(\w+)\s+([^\]]*)\]\]/, wiki_file_regexp => qr/(^[-[:alnum:]_.:\/+]+$)/, + web_commit_regexp => qr/^web commit (by (.*?(?=: |$))|from (\d+\.\d+\.\d+\.\d+)):?(.*)/, verbose => 0, + syslog => 0, wikiname => "wiki", - default_pageext => ".mdwn", + default_pageext => "mdwn", cgi => 0, rcs => 'svn', notify => 0, @@ -26,10 +44,12 @@ sub defaultconfig () { #{{{ diffurl => '', anonok => 0, rss => 0, + atom => 0, discussion => 1, rebuild => 0, refresh => 0, getctime => 0, + w3mmode => 0, wrapper => undef, wrappermode => undef, svnrepo => undef, @@ -37,21 +57,46 @@ sub defaultconfig () { #{{{ srcdir => undef, destdir => undef, pingurl => [], - templatedir => "/usr/share/ikiwiki/templates", - underlaydir => "/usr/share/ikiwiki/basewiki", + templatedir => "$installdir/share/ikiwiki/templates", + underlaydir => "$installdir/share/ikiwiki/basewiki", setup => undef, adminuser => undef, adminemail => undef, - plugin => [qw{mdwn inline htmlscrubber}], + plugin => [qw{mdwn inline htmlscrubber passwordauth}], timeformat => '%c', + locale => undef, + sslcookie => 0, + httpauth => 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 $config{locale}) { + eval q{use POSIX}; + error($@) if $@; + $ENV{LANG} = $config{locale} + if POSIX::setlocale(&POSIX::LC_TIME, $config{locale}); + } + + if ($config{w3mmode}) { + eval q{use Cwd q{abs_path}}; + error($@) if $@; + $config{srcdir}=possibly_foolish_untaint(abs_path($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}; + } + if ($config{cgi} && ! length $config{url}) { error("Must specify url to wiki with --url when using --cgi\n"); } - if ($config{rss} && ! length $config{url}) { - error("Must specify url to wiki with --url when using --rss\n"); + if (($config{rss} || $config{atom}) && ! length $config{url}) { + error("Must specify url to wiki with --url when using --rss or --atom\n"); } $config{wikistatedir}="$config{srcdir}/.ikiwiki" @@ -67,19 +112,28 @@ sub checkconfig () { #{{{ require IkiWiki::Rcs::Stub; } - foreach my $plugin (@{$config{plugin}}) { - my $mod="IkiWiki::Plugin::".possibly_foolish_untaint($plugin); - eval qq{use $mod}; - if ($@) { - error("Failed to load plugin $mod: $@"); - } + run_hooks(checkconfig => sub { shift->() }); +} #}}} + +sub loadplugins () { #{{{ + loadplugin($_) foreach @{$config{plugin}}; + + run_hooks(getopt => sub { shift->() }); + if (grep /^-/, @ARGV) { + print STDERR "Unknown option: $_\n" + foreach grep /^-/, @ARGV; + usage(); } +} #}}} - if (exists $hooks{checkconfig}) { - foreach my $id (keys %{$hooks{checkconfig}}) { - $hooks{checkconfig}{$id}{call}->(); - } - } +sub loadplugin ($) { #{{{ + my $plugin=shift; + + my $mod="IkiWiki::Plugin::".possibly_foolish_untaint($plugin); + eval qq{use $mod}; + if ($@) { + error("Failed to load plugin $mod: $@"); + } } #}}} sub error ($) { #{{{ @@ -87,12 +141,31 @@ sub error ($) { #{{{ print "Content-type: text/html\n\n"; print misctemplate("Error", "

Error: @_

"); } - 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 ($$) { #{{{ + my $type=shift; + + if ($config{syslog}) { + require Sys::Syslog; + unless ($log_open) { + Sys::Syslog::setlogsock('unix'); + Sys::Syslog::openlog('ikiwiki', '', 'user'); + $log_open=1; + } + eval { + Sys::Syslog::syslog($type, join(" ", @_)); + } + } + elsif (! $config{cgi}) { print "@_\n"; } else { @@ -123,10 +196,10 @@ sub dirname ($) { #{{{ sub pagetype ($) { #{{{ my $page=shift; - if ($page =~ /\.(.*)$/) { + if ($page =~ /\.([^.]+)$/) { return $1 if exists $hooks{htmlize}{$1}; } - return "unknown"; + return undef; } #}}} sub pagename ($) { #{{{ @@ -134,7 +207,7 @@ sub pagename ($) { #{{{ my $type=pagetype($file); my $page=$file; - $page=~s/\Q.$type\E*$// unless $type eq 'unknown'; + $page=~s/\Q.$type\E*$// if defined $type; return $page; } #}}} @@ -162,12 +235,7 @@ sub readfile ($;$) { #{{{ local $/=undef; open (IN, $file) || error("failed to read $file: $!"); - if (! $binary) { - binmode(IN, ":utf8"); - } - else { - binmode(IN); - } + binmode(IN) if ($binary); my $ret=; close IN; return $ret; @@ -199,24 +267,35 @@ sub writefile ($$$;$) { #{{{ } open (OUT, ">$destdir/$file") || error("failed to write $destdir/$file: $!"); - if (! $binary) { - binmode(OUT, ":utf8"); + binmode(OUT) if ($binary); + print OUT $content; + close OUT; +} #}}} + +my %cleared; +sub will_render ($$;$) { #{{{ + my $page=shift; + my $dest=shift; + my $clear=shift; + + # Important security check. + if (-e "$config{destdir}/$dest" && ! $config{rebuild} && + ! grep { $_ eq $dest } (@{$renderedfiles{$page}}, @{$oldrenderedfiles{$page}})) { + error("$config{destdir}/$dest independently created, not overwriting with version from $page"); + } + + if (! $clear || $cleared{$page}) { + $renderedfiles{$page}=[$dest, grep { $_ ne $dest } @{$renderedfiles{$page}}]; } else { - binmode(OUT); + $renderedfiles{$page}=[$dest]; + $cleared{$page}=1; } - print OUT $content; - close OUT; } #}}} sub bestlink ($$) { #{{{ - # Given a page and the text of a link on the page, determine which - # existing page that link best points to. Prefers pages under a - # subdirectory with the same name as the source page, failing that - # 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 { @@ -225,9 +304,11 @@ sub bestlink ($$) { #{{{ $l.=$link; if (exists $links{$l}) { - #debug("for $page, \"$link\", use $l"); return $l; } + elsif (exists $pagecase{lc $l}) { + return $pagecase{lc $l}; + } } while $cwd=~s!/?[^/]+$!!; #print STDERR "warning: page $page, broken link: $link\n"; @@ -260,14 +341,38 @@ sub cgiurl (@) { #{{{ return $config{cgiurl}."?".join("&", map "$_=$params{$_}", keys %params); } #}}} -sub styleurl (;$) { #{{{ +sub baseurl (;$) { #{{{ my $page=shift; - return "$config{url}/style.css" if ! defined $page; + return "$config{url}/" if ! defined $page; $page=~s/[^\/]+$//; $page=~s/[^\/]+\//..\//g; - return $page."style.css"; + return $page; +} #}}} + +sub abs2rel ($$) { #{{{ + # Work around very innefficient behavior in File::Spec if abs2rel + # is passed two relative paths. It's much faster if paths are + # absolute! (Debian bug #376658; fixed in debian unstable now) + my $path="/".shift; + my $base="/".shift; + + require File::Spec; + my $ret=File::Spec->abs2rel($path, $base); + $ret=~s/^// if defined $ret; + return $ret; +} #}}} + +sub displaytime ($) { #{{{ + my $time=shift; + + eval q{use POSIX}; + error($@) if $@; + # strftime doesn't know about encodings, so make sure + # its output is properly treated as utf8 + return decode_utf8(POSIX::strftime( + $config{timeformat}, localtime($time))); } #}}} sub htmllink ($$$;$$$) { #{{{ @@ -288,22 +393,19 @@ sub htmllink ($$$;$$$) { #{{{ $linktext=pagetitle(basename($link)) unless defined $linktext; - return $linktext if length $bestlink && $page eq $bestlink; + return "$linktext" + 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 that makes all new files be rendered twice. - if (! grep { $_ eq $bestlink } values %renderedfiles) { + if (! grep { $_ eq $bestlink } map { @{$_} } values %renderedfiles) { $bestlink=htmlpage($bestlink); } - if (! grep { $_ eq $bestlink } values %renderedfiles) { + if (! grep { $_ eq $bestlink } map { @{$_} } values %renderedfiles) { return " "create", page => $link, from => $page). + cgiurl(do => "create", page => lc($link), from => $page). "\">?$linktext" } - require File::Spec; - $bestlink=File::Spec->abs2rel($bestlink, dirname($page)); + $bestlink=abs2rel($bestlink, dirname($page)); if (! $noimageinline && isinlinableimage($bestlink)) { return "\"$linktext\""; @@ -311,6 +413,119 @@ sub htmllink ($$$;$$$) { #{{{ return "$linktext"; } #}}} +sub htmlize ($$$) { #{{{ + my $page=shift; + my $type=shift; + my $content=shift; + + if (exists $hooks{htmlize}{$type}) { + $content=$hooks{htmlize}{$type}{call}->( + page => $page, + content => $content, + ); + } + else { + error("htmlization of $type not supported"); + } + + run_hooks(sanitize => sub { + $content=shift->( + page => $page, + content => $content, + ); + }); + + return $content; +} #}}} + +sub linkify ($$$) { #{{{ + my $lpage=shift; # the page containing the links + my $page=shift; # the page the link will end up on (different for inline) + my $content=shift; + + $content =~ s{(\\?)$config{wiki_link_regexp}}{ + $2 ? ( $1 ? "[[$2|$3]]" : htmllink($lpage, $page, titlepage($3), 0, 0, pagetitle($2))) + : ( $1 ? "[[$3]]" : htmllink($lpage, $page, titlepage($3))) + }eg; + + return $content; +} #}}} + +my %preprocessing; +sub preprocess ($$$;$) { #{{{ + my $page=shift; # the page the data comes from + my $destpage=shift; # the page the data will appear in (different for inline) + my $content=shift; + my $scan=shift; + + my $handle=sub { + my $escape=shift; + my $command=shift; + my $params=shift; + if (length $escape) { + return "[[$command $params]]"; + } + elsif (exists $hooks{preprocess}{$command}) { + return "" if $scan && ! $hooks{preprocess}{$command}{scan}; + # Note: preserve order of params, some plugins may + # consider it significant. + my @params; + while ($params =~ /(?:(\w+)=)?(?:"""(.*?)"""|"([^"]+)"|(\S+))(?:\s+|$)/sg) { + my $key=$1; + my $val; + if (defined $2) { + $val=$2; + $val=~s/\r\n/\n/mg; + $val=~s/^\n+//g; + $val=~s/\n+$//g; + } + elsif (defined $3) { + $val=$3; + } + elsif (defined $4) { + $val=$4; + } + + if (defined $key) { + push @params, $key, $val; + } + else { + push @params, $val, ''; + } + } + if ($preprocessing{$page}++ > 3) { + # Avoid loops of preprocessed pages preprocessing + # other pages that preprocess them, etc. + return "[[$command preprocessing loop detected on $page at depth $preprocessing{$page}]]"; + } + my $ret=$hooks{preprocess}{$command}{call}->( + @params, + page => $page, + destpage => $destpage, + ); + $preprocessing{$page}--; + return $ret; + } + else { + return "[[$command $params]]"; + } + }; + + $content =~ s{(\\?)\[\[(\w+)\s+((?:(?:\w+=)?(?:""".*?"""|"[^"]+"|[^\s\]]+)\s*)*)\]\]}{$handle->($1, $2, $3)}seg; + return $content; +} #}}} + +sub filter ($$) { #{{{ + my $page=shift; + my $content=shift; + + run_hooks(filter => sub { + $content=shift->(page => $page, content => $content); + }); + + return $content; +} #}}} + sub indexlink () { #{{{ return "$config{wikiname}"; } #}}} @@ -340,15 +555,16 @@ sub unlockwiki () { #{{{ } #}}} sub loadindex () { #{{{ - open (IN, "<:utf8", "$config{wikistatedir}/index") || return; + open (IN, "$config{wikistatedir}/index") || return; while () { $_=possibly_foolish_untaint($_); chomp; my %items; $items{link}=[]; + $items{dest}=[]; 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 @@ -359,9 +575,10 @@ sub loadindex () { #{{{ $oldpagemtime{$page}=$items{mtime}[0]; $oldlinks{$page}=[@{$items{link}}]; $links{$page}=[@{$items{link}}]; - $depends{$page}=join(" ", @{$items{depends}}) - if exists $items{depends}; - $renderedfiles{$page}=$items{dest}[0]; + $depends{$page}=$items{depends}[0] if exists $items{depends}; + $renderedfiles{$page}=[@{$items{dest}}]; + $oldrenderedfiles{$page}=[@{$items{dest}}]; + $pagecase{lc $page}=$page; } $pagectime{$page}=$items{ctime}[0]; } @@ -369,20 +586,23 @@ sub loadindex () { #{{{ } #}}} sub saveindex () { #{{{ + run_hooks(savestate => sub { shift->() }); + if (! -d $config{wikistatedir}) { mkdir($config{wikistatedir}); } - open (OUT, ">:utf8", "$config{wikistatedir}/index") || + open (OUT, ">$config{wikistatedir}/index") || error("cannot write to $config{wikistatedir}/index: $!"); foreach my $page (keys %oldpagemtime) { next unless $oldpagemtime{$page}; my $line="mtime=$oldpagemtime{$page} ". "ctime=$pagectime{$page} ". - "src=$pagesources{$page} ". - "dest=$renderedfiles{$page}"; - $line.=" link=$_" foreach @{$links{$page}}; + "src=$pagesources{$page}"; + $line.=" dest=$_" foreach @{$renderedfiles{$page}}; + my %count; + $line.=" link=$_" foreach grep { ++$count{$_} == 1 } @{$links{$page}}; if (exists $depends{$page}) { - $line.=" depends=$_" foreach split " ", $depends{$page}; + $line.=" depends=".encode_entities($depends{$page}, " \t\n"); } print OUT $line."\n"; } @@ -397,14 +617,17 @@ sub template_params (@) { #{{{ my $text_ref = shift; $$text_ref=&Encode::decode_utf8($$text_ref); }, - filename => "$config{templatedir}/$filename", @_; + filename => "$config{templatedir}/$filename", + loop_context_vars => 1, + die_on_bad_params => 0, + @_; } #}}} sub template ($;@) { #{{{ HTML::Template->new(template_params(@_)); } #}}} -sub misctemplate ($$) { #{{{ +sub misctemplate ($$;@) { #{{{ my $title=shift; my $pagebody=shift; @@ -414,63 +637,216 @@ sub misctemplate ($$) { #{{{ indexlink => indexlink(), wikiname => $config{wikiname}, pagebody => $pagebody, - styleurl => styleurl(), - baseurl => "$config{url}/", + baseurl => baseurl(), + @_, ); + run_hooks(pagetemplate => sub { + shift->(page => "", destpage => "", template => $template); + }); return $template->output; }#}}} -sub glob_match ($$) { #{{{ - my $page=shift; - my $glob=shift; +sub hook (@) { # {{{ + my %param=@_; + + if (! exists $param{type} || ! ref $param{call} || ! exists $param{id}) { + error "hook requires type, call, and id parameters"; + } - if ($glob =~ /^link\((.+)\)$/) { - my $rev = $links{$page} or return undef; - foreach my $p (@$rev) { - return 1 if lc $p eq $1; + return if $param{no_override} && exists $hooks{$param{type}}{$param{id}}; + + $hooks{$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}) { + my @deferred; + foreach my $id (keys %{$hooks{$type}}) { + if ($hooks{$type}{$id}{last}) { + push @deferred, $id; + next; + } + $sub->($hooks{$type}{$id}{call}); } - return 0; - } elsif ($glob =~ /^backlink\((.+)\)$/) { - my $rev = $links{$1} or return undef; - foreach my $p (@$rev) { - return 1 if lc $p eq $page; + foreach my $id (@deferred) { + $sub->($hooks{$type}{$id}{call}); } - return 0; - } else { - # turn glob into safe regexp - $glob=quotemeta($glob); - $glob=~s/\\\*/.*/g; - $glob=~s/\\\?/./g; - $glob=~s!\\/!/!g; - - return $page=~/^$glob$/i; } } #}}} -sub globlist_match ($$) { #{{{ - my $page=shift; - my @globlist=split(" ", shift); +sub globlist_to_pagespec ($) { #{{{ + my @globlist=split(' ', shift); - # check any negated globs first + my (@spec, @skip); foreach my $glob (@globlist) { - return 0 if $glob=~/^!(.*)/ && glob_match($page, $1); + if ($glob=~/^!(.*)/) { + push @skip, $glob; + } + else { + push @spec, $glob; + } } - foreach my $glob (@globlist) { - return 1 if glob_match($page, $glob); + my $spec=join(" or ", @spec); + if (@skip) { + my $skip=join(" and ", @skip); + if (length $spec) { + $spec="$skip and ($spec)"; + } + else { + $spec=$skip; + } } - - return 0; + return $spec; } #}}} -sub hook (@) { # {{{ - my %param=@_; - - if (! exists $param{type} || ! ref $param{call} || ! exists $param{id}) { - error "hook requires type, call, and id parameters"; +sub is_globlist ($) { #{{{ + my $s=shift; + $s=~/[^\s]+\s+([^\s]+)/ && $1 ne "and" && $1 ne "or"; +} #}}} + +sub safequote ($) { #{{{ + my $s=shift; + $s=~s/[{}]//g; + return "q{$s}"; +} #}}} + +sub pagespec_merge ($$) { #{{{ + 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 add_depends ($$) { #{{{ + my $page=shift; + my $pagespec=shift; - $hooks{$param{type}}{$param{id}}=\%param; + if (! exists $depends{$page}) { + $depends{$page}=$pagespec; + } + else { + $depends{$page}=pagespec_merge($depends{$page}, $pagespec); + } } # }}} +sub pagespec_match ($$) { #{{{ + my $page=shift; + my $spec=shift; + + return eval pagespec_translate($spec); +} #}}} + +sub match_glob ($$) { #{{{ + my $page=shift; + my $glob=shift; + + # turn glob into safe regexp + $glob=quotemeta($glob); + $glob=~s/\\\*/.*/g; + $glob=~s/\\\?/./g; + + return $page=~/^$glob$/i; +} #}}} + +sub match_link ($$) { #{{{ + my $page=shift; + my $link=lc(shift); + + my $links = $links{$page} or return undef; + foreach my $p (@$links) { + return 1 if lc $p eq $link; + } + return 0; +} #}}} + +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; + } +} #}}} + +sub match_created_after ($$) { #{{{ + my $page=shift; + my $testpage=shift; + + if (exists $pagectime{$testpage}) { + return $pagectime{$page} > $pagectime{$testpage}; + } + else { + return 0; + } +} #}}} + +sub match_creation_day ($$) { #{{{ + 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