X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/5591d621b90855beeb5460f0ac11279d19835699..1c5a9c0cc8265a6a552b1500eb3a3d2aa17b9720:/IkiWiki/Render.pm diff --git a/IkiWiki/Render.pm b/IkiWiki/Render.pm index 7d1e8ee53..246c2260d 100644 --- a/IkiWiki/Render.pm +++ b/IkiWiki/Render.pm @@ -1,376 +1,254 @@ +#!/usr/bin/perl + package IkiWiki; use warnings; use strict; -use File::Spec; +use IkiWiki; +use Encode; + +my %backlinks; +our %brokenlinks; +my $links_calculated=0; + +sub calculate_links () { + return if $links_calculated; + %backlinks=%brokenlinks=(); + foreach my $page (keys %links) { + foreach my $link (@{$links{$page}}) { + my $bestlink=bestlink($page, $link); + if (length $bestlink) { + $backlinks{$bestlink}{$page}=1 + if $bestlink ne $page; + } + else { + push @{$brokenlinks{$link}}, $page; + } + } + } + $links_calculated=1; +} -sub linkify ($$) { #{{{ - my $content=shift; +sub backlink_pages ($) { my $page=shift; - $content =~ s{(\\?)$config{wiki_link_regexp}}{ - $2 ? ( $1 ? "[[$2|$3]]" : htmllink($page, titlepage($3), 0, 0, pagetitle($2))) - : ( $1 ? "[[$3]]" : htmllink($page, titlepage($3))) - }eg; - - return $content; -} #}}} + calculate_links(); -sub htmlize ($$) { #{{{ - my $type=shift; - my $content=shift; - - if (! $INC{"/usr/bin/markdown"}) { - no warnings 'once'; - $blosxom::version="is a proper perl module too much to ask?"; - use warnings 'all'; - do "/usr/bin/markdown"; - } - - if ($type eq '.mdwn') { - return Markdown::Markdown($content); - } - else { - error("htmlization of $type not supported"); - } -} #}}} + return keys %{$backlinks{$page}}; +} -sub backlinks ($) { #{{{ +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 $href=File::Spec->abs2rel(htmlpage($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 => $p_trimmed }; - } - } - - return sort { $a->{page} cmp $b->{page} } @links; -} #}}} - -sub parentlinks ($) { #{{{ - my $page=shift; - - my @ret; - my $pagelink=""; - my $path=""; - my $skip=1; - foreach my $dir (reverse split("/", $page)) { - if (! $skip) { - $path.="../"; - unshift @ret, { url => "$path$dir.html", page => $dir }; - } - else { - $skip=0; - } + foreach my $p (backlink_pages($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) }; } - unshift @ret, { url => length $path ? $path : ".", page => $config{wikiname} }; - return @ret; -} #}}} - -sub rsspage ($) { #{{{ - my $page=shift; - - return $page.".rss"; -} #}}} + return @links; +} -sub postprocess { #{{{ - # Takes content to postprocess followed by a list of postprocessor - # commands and subroutine references to run for the commands. +sub genpage ($$) { my $page=shift; my $content=shift; - my %commands=@_; - - my $handle=sub { - my $escape=shift; - my $command=shift; - my $params=shift; - if (length $escape) { - "[[$command $params]]"; - } - elsif (exists $commands{$command}) { - my %params; - while ($params =~ /(\w+)=\"([^"]+)"(\s+|$)/g) { - $params{$1}=$2; - } - $commands{$command}->($page, %params); - } - else { - "[[bad directive $command]]"; - } - }; - - $content =~ s{(\\?)$config{wiki_processor_regexp}}{$handle->($1, $2, $3)}eg; - return $content; -} #}}} -sub blog_list ($$) { #{{{ - my $globlist=shift; - my $maxitems=shift; - - my @list; - foreach my $page (keys %pagesources) { - if (globlist_match($page, $globlist)) { - push @list, $page; + my $templatefile; + run_hooks(templatefile => sub { + return if defined $templatefile; + my $file=shift->(page => $page); + if (defined $file && defined template_file($file)) { + $templatefile=$file; } - } - - @list=sort { $pagectime{$b} <=> $pagectime{$a} } @list; - return @list if ! $maxitems || @list <= $maxitems; - return @list[0..$maxitems - 1]; -} #}}} + }); + my $template=template(defined $templatefile ? $templatefile : 'page.tmpl', blind_cache => 1); + my $actions=0; -sub get_inline_content ($$) { #{{{ - my $parentpage=shift; - my $page=shift; - - my $file=$pagesources{$page}; - my $type=pagetype($file); - if ($type ne 'unknown') { - return htmlize($type, linkify(readfile("$config{srcdir}/$file"), $parentpage)); - } - else { - return ""; - } -} #}}} - -sub postprocess_html_inline { #{{{ - my $parentpage=shift; - my %params=@_; - - if (! exists $params{pages}) { - return ""; - } - if (! exists $params{archive}) { - $params{archive}="no"; - } - if (! exists $params{show} && $params{archive} eq "no") { - $params{show}=10; - } - $inlinepages{$parentpage}=$params{pages}; - - my $ret=""; - - if (exists $params{rootpage}) { - my $formtemplate=HTML::Template->new(blind_cache => 1, - filename => "$config{templatedir}/blogpost.tmpl"); - $formtemplate->param(cgiurl => $config{cgiurl}); - $formtemplate->param(rootpage => $params{rootpage}); - my $form=$formtemplate->output; - $ret.=$form; - } - - my $template=HTML::Template->new(blind_cache => 1, - filename => (($params{archive} eq "no") - ? "$config{templatedir}/inlinepage.tmpl" - : "$config{templatedir}/inlinepagetitle.tmpl")); - - foreach my $page (blog_list($params{pages}, $params{show})) { - next if $page eq $parentpage; - $template->param(pagelink => htmllink($parentpage, $page)); - $template->param(content => get_inline_content($parentpage, $page)) - if $params{archive} eq "no"; - $template->param(ctime => scalar(gmtime($pagectime{$page}))); - $ret.=$template->output; - } - - return "
$ret"; -} #}}} - -sub genpage ($$$) { #{{{ - my $content=shift; - my $page=shift; - my $mtime=shift; - - $content = postprocess($page, $content, inline => \&postprocess_html_inline); - - my $title=pagetitle(basename($page)); - - my $template=HTML::Template->new(blind_cache => 1, - filename => "$config{templatedir}/page.tmpl"); - if (length $config{cgiurl}) { - $template->param(editurl => cgiurl(do => "edit", page => $page)); - $template->param(prefsurl => cgiurl(do => "prefs")); - if ($config{rcs}) { - $template->param(recentchangesurl => cgiurl(do => "recentchanges")); - } + $template->param(editurl => cgiurl(do => "edit", page => $page)) + if IkiWiki->can("cgi_editpage"); + $template->param(prefsurl => cgiurl(do => "prefs")) + if exists $hooks{auth}; + $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}) { + if ($page !~ /.*\/\Q$config{discussionpage}\E$/ && + (length $config{cgiurl} || + exists $links{$page."/".$config{discussionpage}})) { + $template->param(discussionlink => htmllink($page, $page, $config{discussionpage}, noimageinline => 1, forcesubpage => 1)); + $actions++; + } } - if ($config{rss} && $inlinepages{$page}) { - $template->param(rssurl => rsspage(basename($page))); + 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 => $title, + title => $page eq 'index' + ? $config{wikiname} + : pagetitle(basename($page)), wikiname => $config{wikiname}, - parentlinks => [parentlinks($page)], content => $content, - backlinks => [backlinks($page)], - discussionlink => htmllink($page, "Discussion", 1, 1), - mtime => scalar(gmtime($mtime)), - styleurl => styleurl($page), + backlinks => $backlinks, + more_backlinks => $more_backlinks, + mtime => displaytime($pagemtime{$page}), + ctime => displaytime($pagectime{$page}), + baseurl => baseurl($page), ); + + run_hooks(pagetemplate => sub { + shift->(page => $page, destpage => $page, template => $template); + }); + + $content=$template->output; - return $template->output; -} #}}} + run_hooks(postscan => sub { + shift->(page => $page, content => $content); + }); -sub date_822 ($) { #{{{ - my $time=shift; + run_hooks(format => sub { + $content=shift->( + page => $page, + content => $content, + ); + }); - eval q{use POSIX}; - return POSIX::strftime("%a, %d %b %Y %H:%M:%S %z", localtime($time)); -} #}}} + return $content; +} -sub absolute_urls ($$) { #{{{ - # sucky sub because rss sucks - my $content=shift; - my $url=shift; +sub scan ($) { + my $file=shift; - $url=~s/[^\/]+$//; - - $content=~s/new(blind_cache => 1, - filename => "$config{templatedir}/rsspage.tmpl"); - - my @items; - my $isblog=0; - my $gen_blog=sub { - my $parentpage=shift; - my %params=@_; - - if (! exists $params{show}) { - $params{show}=10; + if ($config{discussion}) { + # Discussion links are a special case since they're + # not in the text of the page, but on its template. + $links{$page}=[ $page."/".lc($config{discussionpage}) ]; } - if (! exists $params{pages}) { - return ""; - } - - $isblog=1; - foreach my $page (blog_list($params{pages}, $params{show})) { - next if $page eq $parentpage; - push @items, { - itemtitle => pagetitle(basename($page)), - itemurl => "$config{url}/$renderedfiles{$page}", - itempubdate => date_822($pagectime{$page}), - itemcontent => absolute_urls(get_inline_content($parentpage, $page), $url), - } if exists $renderedfiles{$page}; + else { + $links{$page}=[]; } - - return ""; - }; - - $content = postprocess($page, $content, inline => $gen_blog); - $template->param( - title => $config{wikiname}, - pageurl => $url, - items => \@items, - ); - - return $template->output; -} #}}} - -sub check_overwrite ($$) { #{{{ - # 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 rendered from ". - join(" ",(grep { $renderedfiles{$_} eq $dest } keys - %renderedfiles)). - ", before, so not rendering from $src"); - } -} #}}} + run_hooks(scan => sub { + shift->( + page => $page, + content => $content, + ); + }); -sub mtime ($) { #{{{ - my $file=shift; - - return (stat($file))[9]; -} #}}} - -sub findlinks ($$) { #{{{ - my $content=shift; - my $page=shift; - - my @links; - while ($content =~ /(?) { + 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); @@ -378,125 +256,261 @@ sub prune ($) { #{{{ while (rmdir($dir)) { $dir=dirname($dir); } -} #}}} +} + +sub srcdir_check () { + # 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); + } + } + +} -sub refresh () { #{{{ - # find existing pages - my %exists; - my @files; +sub find_src_files () { + my (@files, %pages); eval q{use File::Find}; + error($@) if $@; find({ no_chdir => 1, wanted => sub { - if (/$config{wiki_file_prune_regexp}/) { - no warnings 'once'; + $_=decode_utf8($_); + if (file_pruned($_, $config{srcdir})) { $File::Find::prune=1; - use warnings "all"; } - 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 ($pages{$pagename}) { + debug(sprintf(gettext("%s has multiple possible source pages"), $pagename)); + } + $pages{$pagename}=1; } } }, }, $config{srcdir}); + foreach my $dir (@{$config{underlaydirs}}, $config{underlaydir}) { + find({ + no_chdir => 1, + wanted => sub { + $_=decode_utf8($_); + if (file_pruned($_, $dir)) { + $File::Find::prune=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 (! $pages{$page}) { + push @files, $f; + $pages{$page}=1; + } + } + } + } + }, + }, $dir); + }; + + # Returns a list of all source files found, and a hash of + # the corresponding page names. + return \@files, \%pages; +} - my %rendered; +sub refresh () { + srcdir_check(); + run_hooks(refresh => sub { shift->() }); + my ($files, $exists)=find_src_files(); + my (%rendered, @add, @del, @internal); # check for added or removed pages - my @add; - foreach my $file (@files) { + foreach my $file (@$files) { my $page=pagename($file); - if (! $oldpagemtime{$page}) { - debug("new page $page") unless exists $pagectime{$page}; - push @add, $file; - $links{$page}=[]; - $pagesources{$page}=$file; - $pagectime{$page}=mtime("$config{srcdir}/$file") - unless exists $pagectime{$page}; + 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; + } + 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) { - if (! $exists{$page}) { - debug("removing old page $page"); - push @del, $pagesources{$page}; - prune($config{destdir}."/".$renderedfiles{$page}); - delete $renderedfiles{$page}; - $oldpagemtime{$page}=0; + foreach my $page (keys %pagemtime) { + if (! $exists->{$page}) { + 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; + foreach my $old (@{$oldrenderedfiles{$page}}) { + prune($config{destdir}."/".$old); + } delete $pagesources{$page}; + foreach my $source (keys %destsources) { + if ($destsources{$source} eq $page) { + delete $destsources{$source}; + } + } } } - - # render any updated files - foreach my $file (@files) { + + # find changed and new files + my @needsbuild; + foreach my $file (@$files) { my $page=pagename($file); - - if (! exists $oldpagemtime{$page} || - mtime("$config{srcdir}/$file") > $oldpagemtime{$page}) { - debug("rendering changed file $file"); - render($file); - $rendered{$file}=1; + my ($srcfile, @stat)=srcfile_stat($file); + if (! exists $pagemtime{$page} || + $stat[9] > $pagemtime{$page} || + $forcerebuild{$page}) { + $pagemtime{$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_links(); + foreach my $file (@needsbuild) { + debug(sprintf(gettext("building %s"), $file)); + render($file); + $rendered{$file}=1; + } + foreach my $file (@internal) { + # internal pages are not rendered + my $page=pagename($file); + delete $depends{$page}; + delete $depends_simple{$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("building %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 handle inlining here. - # 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) { - foreach my $f (@files) { + if (%rendered || @del || @internal) { + my @changed=(keys %rendered, @del); + + my %lcchanged = map { lc(pagename($_)) => 1 } @changed; + + # rebuild dependant pages + foreach my $f (@$files) { + next if $rendered{$f}; my $p=pagename($f); - if (exists $inlinepages{$p}) { - foreach my $file (keys %rendered, @del) { - my $page=pagename($file); - if (globlist_match($page, $inlinepages{$p})) { - debug("rendering $f, which inlines $page"); - render($f); + my $reason = undef; + + if (exists $depends_simple{$p}) { + foreach my $d (keys %{$depends_simple{$p}}) { + if (exists $lcchanged{$d}) { + $reason = $d; last; } } } + + if (exists $depends{$p} && ! defined $reason) { + D: foreach my $d (keys %{$depends{$p}}) { + my $sub=pagespec_translate($d); + next if $@ || ! defined $sub; + + # only consider internal files + # if the page explicitly depends + # on such files + foreach my $file (@changed, $d =~ /internal\(/ ? @internal : ()) { + next if $file eq $f; + my $page=pagename($file); + if ($sub->($page, location => $p)) { + $reason = $page; + last D; + } + } + } + } + + if (defined $reason) { + debug(sprintf(gettext("building %s, which depends on %s"), $f, $reason)); + render($f); + $rendered{$f}=1; + } } + # 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}) { foreach my $link (map { bestlink($page, $_) } @{$links{$page}}) { if (length $link && - ! exists $oldlinks{$page} || - ! grep { $_ eq $link } @{$oldlinks{$page}}) { + (! exists $oldlinks{$page} || + ! grep { bestlink($page, $_) eq $link } @{$oldlinks{$page}})) { $linkchanged{$link}=1; } } @@ -504,21 +518,67 @@ FILE: foreach my $file (@files) { if (exists $oldlinks{$page}) { foreach my $link (map { bestlink($page, $_) } @{$oldlinks{$page}}) { if (length $link && - ! exists $links{$page} || - ! grep { $_ eq $link } @{$links{$page}}) { + (! exists $links{$page} || + ! grep { bestlink($page, $_) eq $link } @{$links{$page}})) { $linkchanged{$link}=1; } } } } + 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("building %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 built 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 build %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