X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/f4e2e8e975b8f538375acd9a6d41a20b22f2ac27..b24f236dfa2903805bc233a7fc97101d912ae3ee:/IkiWiki/Render.pm diff --git a/IkiWiki/Render.pm b/IkiWiki/Render.pm index c7849a2dd..8e4c8af71 100644 --- a/IkiWiki/Render.pm +++ b/IkiWiki/Render.pm @@ -7,198 +7,109 @@ use strict; use IkiWiki; use Encode; -sub linkify ($$$) { #{{{ - my $lpage=shift; # the page containing the links - my $page=shift; # the page the link will end up on (different for inline) - my $content=shift; - - $content =~ s{(\\?)$config{wiki_link_regexp}}{ - $2 ? ( $1 ? "[[$2|$3]]" : htmllink($lpage, $page, titlepage($3), 0, 0, pagetitle($2))) - : ( $1 ? "[[$3]]" : htmllink($lpage, $page, titlepage($3))) - }eg; - - return $content; -} #}}} - -sub htmlize ($$) { #{{{ - my $type=shift; - my $content=shift; - - if (exists $hooks{htmlize}{$type}) { - $content=$hooks{htmlize}{$type}{call}->($content); - } - else { - error("htmlization of $type not supported"); - } - - run_hooks(sanitize => sub { - $content=shift->($content); - }); - - return $content; -} #}}} - -sub backlinks ($) { #{{{ - my $page=shift; - - my @links; - foreach my $p (keys %links) { - next if bestlink($page, $p) eq $page; - if (grep { length $_ && bestlink($p, $_) eq $page } @{$links{$p}}) { - my $href=abs2rel(htmlpage($p), dirname($page)); - - # Trim common dir prefixes from both pages. - my $p_trimmed=$p; - my $page_trimmed=$page; - my $dir; - 1 while (($dir)=$page_trimmed=~m!^([^/]+/)!) && - defined $dir && - $p_trimmed=~s/^\Q$dir\E// && - $page_trimmed=~s/^\Q$dir\E//; - - push @links, { url => $href, page => pagetitle($p_trimmed) }; +our %backlinks=(); +our $backlinks_calculated=0; + +sub calculate_backlinks () { + return if $backlinks_calculated; + %backlinks=(); + foreach my $page (keys %links) { + foreach my $link (@{$links{$page}}) { + my $bestlink=bestlink($page, $link); + if (length $bestlink && $bestlink ne $page) { + $backlinks{$bestlink}{$page}=1; + } } } + $backlinks_calculated=1; +} - return sort { $a->{page} cmp $b->{page} } @links; -} #}}} - -sub parentlinks ($) { #{{{ +sub backlinks ($) { my $page=shift; - - my @ret; - my $pagelink=""; - my $path=""; - my $skip=1; - return if $page eq 'index'; # toplevel - foreach my $dir (reverse split("/", $page)) { - if (! $skip) { - $path.="../"; - unshift @ret, { url => $path.htmlpage($dir), page => pagetitle($dir) }; - } - else { - $skip=0; - } - } - unshift @ret, { url => length $path ? $path : ".", page => $config{wikiname} }; - return @ret; -} #}}} - -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 $handle=sub { - my $escape=shift; - my $command=shift; - my $params=shift; - if (length $escape) { - return "[[$command $params]]"; - } - elsif ($preprocessing{$page}++ > 10) { - # Avoid loops of preprocessed pages preprocessing - # other pages that preprocess them, etc. - return "[[$command preprocessing loop detected on $page at depth $preprocessing{$page}]]"; - } - elsif (exists $hooks{preprocess}{$command}) { - # 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, ''; - } - } - my $ret=$hooks{preprocess}{$command}{call}->( - @params, - page => $page, - destpage => $destpage, - ); - $preprocessing{$page}--; - return $ret; - } - else { - return "[[$command not processed]]"; - } - }; - - $content =~ s{(\\?)\[\[(\w+)\s+((?:(?:\w+=)?(?:""".*?"""|"[^"]+"|[^\s\]]+)\s*)*)\]\]}{$handle->($1, $2, $3)}seg; - return $content; -} #}}} + calculate_backlinks(); -sub add_depends ($$) { #{{{ - my $page=shift; - my $pagespec=shift; - - if (! exists $depends{$page}) { - $depends{$page}=$pagespec; - } - else { - $depends{$page}=pagespec_merge($depends{$page}, $pagespec); + my @links; + foreach my $p (keys %{$backlinks{$page}}) { + my $href=urlto($p, $page); + + # Trim common dir prefixes from both pages. + my $p_trimmed=$p; + my $page_trimmed=$page; + my $dir; + 1 while (($dir)=$page_trimmed=~m!^([^/]+/)!) && + defined $dir && + $p_trimmed=~s/^\Q$dir\E// && + $page_trimmed=~s/^\Q$dir\E//; + + push @links, { url => $href, page => pagetitle($p_trimmed) }; } -} # }}} + return @links; +} -sub genpage ($$$) { #{{{ +sub genpage ($$) { my $page=shift; my $content=shift; - my $mtime=shift; - my $template=template("page.tmpl", blind_cache => 1); + my $templatefile; + run_hooks(templatefile => sub { + return if defined $templatefile; + my $file=shift->(page => $page); + if (defined $file && defined template_file($file)) { + $templatefile=$file; + } + }); + my $template=template(defined $templatefile ? $templatefile : 'page.tmpl', blind_cache => 1); my $actions=0; if (length $config{cgiurl}) { - $template->param(editurl => cgiurl(do => "edit", page => $page)); + $template->param(editurl => cgiurl(do => "edit", page => $page)) + if IkiWiki->can("cgi_editpage"); $template->param(prefsurl => cgiurl(do => "prefs")); - if ($config{rcs}) { - $template->param(recentchangesurl => cgiurl(do => "recentchanges")); - } $actions++; } - - if (length $config{historyurl}) { + + if (defined $config{historyurl} && length $config{historyurl}) { my $u=$config{historyurl}; $u=~s/\[\[file\]\]/$pagesources{$page}/g; $template->param(historyurl => $u); $actions++; } if ($config{discussion}) { - $template->param(discussionlink => htmllink($page, $page, "Discussion", 1, 1)); - $actions++; + my $discussionlink=gettext("discussion"); + if ($page !~ /.*\/\Q$discussionlink\E$/ && + (length $config{cgiurl} || + exists $links{$page."/".$discussionlink})) { + $template->param(discussionlink => htmllink($page, $page, gettext("Discussion"), noimageinline => 1, forcesubpage => 1)); + $actions++; + } } if ($actions) { $template->param(have_actions => 1); } + my @backlinks=sort { $a->{page} cmp $b->{page} } backlinks($page); + my ($backlinks, $more_backlinks); + if (@backlinks <= $config{numbacklinks} || ! $config{numbacklinks}) { + $backlinks=\@backlinks; + $more_backlinks=[]; + } + else { + $backlinks=[@backlinks[0..$config{numbacklinks}-1]]; + $more_backlinks=[@backlinks[$config{numbacklinks}..$#backlinks]]; + } + $template->param( title => $page eq 'index' ? $config{wikiname} : pagetitle(basename($page)), wikiname => $config{wikiname}, - parentlinks => [parentlinks($page)], content => $content, - backlinks => [backlinks($page)], - mtime => displaytime($mtime), + backlinks => $backlinks, + more_backlinks => $more_backlinks, + mtime => displaytime($pagemtime{$page}), + ctime => displaytime($pagectime{$page}), baseurl => baseurl($page), ); @@ -207,106 +118,124 @@ sub genpage ($$$) { #{{{ }); $content=$template->output; + + run_hooks(postscan => sub { + shift->(page => $page, content => $content); + }); run_hooks(format => sub { - $content=shift->($content); + $content=shift->( + page => $page, + content => $content, + ); }); return $content; -} #}}} - -sub check_overwrite ($$) { #{{{ - # Important security check. Make sure to call this before saving - # any files to the source directory. - my $dest=shift; - my $src=shift; - - if (! exists $renderedfiles{$src} && -e $dest && ! $config{rebuild}) { - error("$dest already exists and was not rendered from $src before"); - } -} #}}} +} -sub displaytime ($) { #{{{ - my $time=shift; +sub scan ($) { + my $file=shift; - eval q{use POSIX}; - # 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))); -} #}}} + my $type=pagetype($file); + if (defined $type) { + my $srcfile=srcfile($file); + my $content=readfile($srcfile); + my $page=pagename($file); + will_render($page, htmlpage($page), 1); -sub mtime ($) { #{{{ - my $file=shift; - - return (stat($file))[9]; -} #}}} + if ($config{discussion}) { + # Discussion links are a special case since they're + # not in the text of the page, but on its template. + $links{$page}=[ $page."/".gettext("discussion") ]; + } + else { + $links{$page}=[]; + } -sub findlinks ($$) { #{{{ - my $page=shift; - my $content=shift; + run_hooks(scan => sub { + shift->( + page => $page, + content => $content, + ); + }); - my @links; - while ($content =~ /(? sub { - $content=shift->(page => $page, content => $content); - }); +} - return $content; +sub fast_file_copy (@) { + my $srcfile=shift; + my $destfile=shift; + my $srcfd=shift; + my $destfd=shift; + my $cleanup=shift; + + my $blksize = 16384; + my ($len, $buf, $written); + while ($len = sysread $srcfd, $buf, $blksize) { + if (! defined $len) { + next if $! =~ /^Interrupted/; + error("failed to read $srcfile: $!", $cleanup); + } + my $offset = 0; + while ($len) { + defined($written = syswrite $destfd, $buf, $len, $offset) + or error("failed to write $destfile: $!", $cleanup); + $len -= $written; + $offset += $written; + } + } } -sub render ($) { #{{{ +sub render ($) { my $file=shift; my $type=pagetype($file); my $srcfile=srcfile($file); if (defined $type) { - my $content=readfile($srcfile); my $page=pagename($file); delete $depends{$page}; + will_render($page, htmlpage($page), 1); + return if $type=~/^_/; - $content=filter($page, $content); - - $links{$page}=[findlinks($page, $content)]; - - $content=preprocess($page, $page, $content); - $content=linkify($page, $page, $content); - $content=htmlize($type, $content); + my $content=htmlize($page, $page, $type, + linkify($page, $page, + preprocess($page, $page, + filter($page, $page, + readfile($srcfile))))); - check_overwrite("$config{destdir}/".htmlpage($page), $page); - writefile(htmlpage($page), $config{destdir}, - genpage($page, $content, mtime($srcfile))); - $oldpagemtime{$page}=time; - $renderedfiles{$page}=htmlpage($page); + my $output=htmlpage($page); + writefile($output, $config{destdir}, genpage($page, $content)); } else { - my $content=readfile($srcfile, 1); - $links{$file}=[]; delete $depends{$file}; - check_overwrite("$config{destdir}/$file", $file); - writefile($file, $config{destdir}, $content, 1); - $oldpagemtime{$file}=time; - $renderedfiles{$file}=$file; + will_render($file, $file, 1); + + if ($config{hardlink}) { + # only hardlink if owned by same user + my @stat=stat($srcfile); + if ($stat[4] == $>) { + prep_writefile($file, $config{destdir}); + unlink($config{destdir}."/".$file); + if (link($srcfile, $config{destdir}."/".$file)) { + return; + } + } + # if hardlink fails, fall back to copying + } + + my $srcfd=readfile($srcfile, 1, 1); + writefile($file, $config{destdir}, undef, 1, sub { + fast_file_copy($srcfile, $file, $srcfd, @_); + }); } -} #}}} +} -sub prune ($) { #{{{ +sub prune ($) { my $file=shift; unlink($file); @@ -314,140 +243,208 @@ sub prune ($) { #{{{ while (rmdir($dir)) { $dir=dirname($dir); } -} #}}} +} + +sub refresh () { + # security check, avoid following symlinks in the srcdir path by default + my $test=$config{srcdir}; + while (length $test) { + if (-l $test && ! $config{allow_symlinks_before_srcdir}) { + error(sprintf(gettext("symlink found in srcdir path (%s) -- set allow_symlinks_before_srcdir to allow this"), $test)); + } + unless ($test=~s/\/+$//) { + $test=dirname($test); + } + } + + run_hooks(refresh => sub { shift->() }); -sub refresh () { #{{{ # find existing pages my %exists; my @files; eval q{use File::Find}; + error($@) if $@; find({ no_chdir => 1, wanted => sub { $_=decode_utf8($_); - if (/$config{wiki_file_prune_regexp}/) { + if (file_pruned($_, $config{srcdir})) { $File::Find::prune=1; } - elsif (! -d $_ && ! -l $_) { + elsif (! -l $_ && ! -d _) { my ($f)=/$config{wiki_file_regexp}/; # untaint if (! defined $f) { - warn("skipping bad filename $_\n"); + warn(sprintf(gettext("skipping bad filename %s"), $_)."\n"); } else { $f=~s/^\Q$config{srcdir}\E\/?//; push @files, $f; - $exists{pagename($f)}=1; + my $pagename = pagename($f); + if ($exists{$pagename}) { + debug(sprintf(gettext("%s has multiple possible source pages"), $pagename)); + } + $exists{$pagename}=1; } } }, }, $config{srcdir}); - find({ - no_chdir => 1, - wanted => sub { - $_=decode_utf8($_); - if (/$config{wiki_file_prune_regexp}/) { - $File::Find::prune=1; - } - elsif (! -d $_ && ! -l $_) { - my ($f)=/$config{wiki_file_regexp}/; # untaint - if (! defined $f) { - warn("skipping bad filename $_\n"); + foreach my $dir (@{$config{underlaydirs}}, $config{underlaydir}) { + find({ + no_chdir => 1, + wanted => sub { + $_=decode_utf8($_); + if (file_pruned($_, $dir)) { + $File::Find::prune=1; } - else { - # Don't add files that are in the - # srcdir. - $f=~s/^\Q$config{underlaydir}\E\/?//; - if (! -e "$config{srcdir}/$f" && - ! -l "$config{srcdir}/$f") { - push @files, $f; - $exists{pagename($f)}=1; + elsif (! -l $_ && ! -d _) { + my ($f)=/$config{wiki_file_regexp}/; # untaint + if (! defined $f) { + warn(sprintf(gettext("skipping bad filename %s"), $_)."\n"); + } + else { + $f=~s/^\Q$dir\E\/?//; + # avoid underlaydir + # override attacks; see + # security.mdwn + if (! -l "$config{srcdir}/$f" && + ! -e _) { + my $page=pagename($f); + if (! $exists{$page}) { + push @files, $f; + $exists{$page}=1; + } + } } } - } - }, - }, $config{underlaydir}); - - my %rendered; + }, + }, $dir); + }; + my (%rendered, @add, @del, @internal); # check for added or removed pages - my @add; foreach my $file (@files) { my $page=pagename($file); - if (! $oldpagemtime{$page}) { - debug("new page $page") unless exists $pagectime{$page}; - push @add, $file; - $links{$page}=[]; - $pagecase{lc $page}=$page; - $pagesources{$page}=$file; - if ($config{getctime} && -e "$config{srcdir}/$file") { - $pagectime{$page}=rcs_getctime("$config{srcdir}/$file"); + if (exists $pagesources{$page} && $pagesources{$page} ne $file) { + # the page has changed its type + $forcerebuild{$page}=1; + } + $pagesources{$page}=$file; + if (! $pagemtime{$page}) { + if (isinternal($page)) { + push @internal, $file; } - elsif (! exists $pagectime{$page}) { - $pagectime{$page}=mtime(srcfile($file)); + else { + push @add, $file; + if ($config{getctime} && -e "$config{srcdir}/$file") { + eval { + my $time=rcs_getctime("$config{srcdir}/$file"); + $pagectime{$page}=$time; + }; + if ($@) { + print STDERR $@; + } + } + } + $pagecase{lc $page}=$page; + if (! exists $pagectime{$page}) { + $pagectime{$page}=(srcfile_stat($file))[10]; } } } - my @del; - foreach my $page (keys %oldpagemtime) { + foreach my $page (keys %pagemtime) { if (! $exists{$page}) { - debug("removing old page $page"); - push @del, $pagesources{$page}; - prune($config{destdir}."/".$renderedfiles{$page}); - delete $renderedfiles{$page}; - $oldpagemtime{$page}=0; + if (isinternal($page)) { + push @internal, $pagesources{$page}; + } + else { + debug(sprintf(gettext("removing old page %s"), $page)); + push @del, $pagesources{$page}; + } + $links{$page}=[]; + $renderedfiles{$page}=[]; + $pagemtime{$page}=0; + prune($config{destdir}."/".$_) + foreach @{$oldrenderedfiles{$page}}; delete $pagesources{$page}; + foreach (keys %destsources) { + if ($destsources{$_} eq $page) { + delete $destsources{$_}; + } + } } } - - # render any updated files + + # find changed and new files + my @needsbuild; foreach my $file (@files) { my $page=pagename($file); - - if (! exists $oldpagemtime{$page} || - mtime(srcfile($file)) > $oldpagemtime{$page} || + my ($srcfile, @stat)=srcfile_stat($file); + if (! exists $pagemtime{$page} || + $stat[9] > $pagemtime{$page} || $forcerebuild{$page}) { - debug("rendering $file"); - render($file); - $rendered{$file}=1; + $pagemtime{$page}=$stat[9]; + if (isinternal($page)) { + push @internal, $file; + # Preprocess internal page in scan-only mode. + preprocess($page, $page, readfile($srcfile), 1); + } + else { + push @needsbuild, $file; + } } } + run_hooks(needsbuild => sub { shift->(\@needsbuild) }); + + # scan and render files + foreach my $file (@needsbuild) { + debug(sprintf(gettext("scanning %s"), $file)); + scan($file); + } + calculate_backlinks(); + foreach my $file (@needsbuild) { + debug(sprintf(gettext("rendering %s"), $file)); + render($file); + $rendered{$file}=1; + } + foreach my $file (@internal) { + # internal pages are not rendered + my $page=pagename($file); + delete $depends{$page}; + foreach my $old (@{$renderedfiles{$page}}) { + delete $destsources{$old}; + } + $renderedfiles{$page}=[]; + } - # if any files were added or removed, check to see if each page - # needs an update due to linking to them or inlining them. - # TODO: inefficient; pages may get rendered above and again here; - # problem is the bestlink may have changed and we won't know until - # now + # rebuild pages that link to added or removed pages if (@add || @del) { -FILE: foreach my $file (@files) { - my $page=pagename($file); - foreach my $f (@add, @del) { - my $p=pagename($f); - foreach my $link (@{$links{$page}}) { - if (bestlink($page, $link) eq $p) { - debug("rendering $file, which links to $p"); - render($file); - $rendered{$file}=1; - next FILE; - } - } + foreach my $f (@add, @del) { + my $p=pagename($f); + foreach my $page (keys %{$backlinks{$p}}) { + my $file=$pagesources{$page}; + next if $rendered{$file}; + debug(sprintf(gettext("rendering %s, which links to %s"), $file, $p)); + render($file); + $rendered{$file}=1; } } } - # Handle backlinks; if a page has added/removed links, update the - # pages it links to. Also handles rebuilding dependant pages. - # TODO: inefficient; pages may get rendered above and again here; - # problem is the backlinks could be wrong in the first pass render - # above - if (%rendered || @del) { + if (%rendered || @del || @internal) { + my @changed=(keys %rendered, @del); + + # rebuild dependant pages foreach my $f (@files) { + next if $rendered{$f}; my $p=pagename($f); if (exists $depends{$p}) { - foreach my $file (keys %rendered, @del) { + # only consider internal files + # if the page explicitly depends on such files + foreach my $file (@changed, $depends{$p}=~/internal\(/ ? @internal : ()) { next if $f eq $file; my $page=pagename($file); - if (pagespec_match($page, $depends{$p})) { - debug("rendering $f, which depends on $page"); + if (pagespec_match($page, $depends{$p}, location => $p)) { + debug(sprintf(gettext("rendering %s, which depends on %s"), $f, $page)); render($f); $rendered{$f}=1; last; @@ -456,8 +453,10 @@ FILE: foreach my $file (@files) { } } + # handle backlinks; if a page has added/removed links, + # update the pages it links to my %linkchanged; - foreach my $file (keys %rendered, @del) { + foreach my $file (@changed) { my $page=pagename($file); if (exists $links{$page}) { @@ -479,22 +478,60 @@ FILE: foreach my $file (@files) { } } } + foreach my $link (keys %linkchanged) { my $linkfile=$pagesources{$link}; if (defined $linkfile) { - debug("rendering $linkfile, to update its backlinks"); + next if $rendered{$linkfile}; + debug(sprintf(gettext("rendering %s, to update its backlinks"), $linkfile)); render($linkfile); $rendered{$linkfile}=1; } } } + # remove no longer rendered files + foreach my $src (keys %rendered) { + my $page=pagename($src); + foreach my $file (@{$oldrenderedfiles{$page}}) { + if (! grep { $_ eq $file } @{$renderedfiles{$page}}) { + debug(sprintf(gettext("removing %s, no longer rendered by %s"), $file, $page)); + prune($config{destdir}."/".$file); + } + } + } + if (@del) { run_hooks(delete => sub { shift->(@del) }); } if (%rendered) { run_hooks(change => sub { shift->(keys %rendered) }); } -} #}}} +} + +sub commandline_render () { + lockwiki(); + loadindex(); + unlockwiki(); + + my $srcfile=possibly_foolish_untaint($config{render}); + my $file=$srcfile; + $file=~s/\Q$config{srcdir}\E\/?//; + + my $type=pagetype($file); + die sprintf(gettext("ikiwiki: cannot render %s"), $srcfile)."\n" unless defined $type; + my $content=readfile($srcfile); + my $page=pagename($file); + $pagesources{$page}=$file; + $content=filter($page, $page, $content); + $content=preprocess($page, $page, $content); + $content=linkify($page, $page, $content); + $content=htmlize($page, $page, $type, $content); + $pagemtime{$page}=(stat($srcfile))[9]; + $pagectime{$page}=$pagemtime{$page} if ! exists $pagectime{$page}; + + print genpage($page, $content); + exit 0; +} 1