werc/bin/controller.rc

#!/usr/local/plan9/bin/rc
cd ..


# Useful functions
fn dprint {
    echo $* >[1=2]
}

fn perm_redirect {
    echo 'Status: 301 Moved Permanantly
Location: '^$1^'

'
    exit
}

fn get_inc_file {
    wantedfile = $1
    if (test -f $sitedir/_inc/$wantedfile)
    	echo -n $sitedir/_inc/$wantedfile
    if not if(test -f sites/$masterSite/_inc/$wantedfile)
    	echo -n       sites/$masterSite/_inc/$wantedfile
    if not if (test -f inc/$wantedfile)
    	echo -n        inc/$wantedfile
}


# Title
fn gentitle {
    echo '<h1 class="headerTitle"><a href="/">' ^ $"siteTitle ^ ' <span id="headerSubTitle">' ^ $"siteSubTitle ^ '</span></a></h1>'
}

# Don't change var name or trailing ';', some dirs change the filter!
dirfilter = '/\/[._]/d; /\/robots.\txt$/d; /\/sitemap\.txt$/d; s,^\./,,; s,\.md$,,; s,\.html,,; s,\.txt,,; '

# To be used from _config files
fn hide_paths {
    for(i in $*) {
	dirfilter = $dirfilter^'/^'$i'$/d; '
    }
}

# Sidebar 
fn menu {
    lsArgs = ('-F')
    # XXX will not work properly for $#blogDirs > 1 and matching is hackish
    # XXX Work in progress, doesn't work well enough to be usable: too inconsistent, doesn't handle corner cases well at all
    #if ( ! ~ $#blogDirs 0 && ~ $blogDirs^/ *$1 *$1/ ) { 
    #	lsArgs = ($lsArgs -r)
    #}
    ls $lsArgs $1 | sed $dirfilter | awk -F/ '
    BEGIN { print "<ul class=\"side-bar\">" }
    END { print "</ul>" }
    # Should add \. to the list of allowed chars in file names, but need to find a way to filter out .txt and so on
    /^([a-zA-Z0-9+_\-]+[\/*]?)+$/ && $NF != "index" {
        isdir = match($0, "/$")
        sub("[*/]$", "")

        path = bname = $0
        sub("^(.*/)?([0-9]+_)?", "", bname)
        gsub("_", " ", bname)

        if(isdir) {
            bname = bname "/"
            path = $0 "/"
        }

        if(index(ENVIRON["REQUEST_URI"] "/", "/" path) == 1) {
            if(isdir) {
                print "<li><a href=\"/" path "\" class=\"thisPage\">&raquo;<i> " bname "</i></a>"
                system("rc -c ''menu " path "''")
            } else {
                print "<li><a href=\"/" path "\" class=\"thisPage\">&raquo;<i> " bname "</i></a>"
            }
        } else 
            print "<li><a href=\"/" path "\">&rsaquo; " bname "</a>"

        print "</li>"
    }'
}

fn gensidebar {
    @{
        cd $sitedir
        menu .
    }
}

fn sortedBlogPostList {
    # the /./ is added so we can sort -t. and order only the file name
    if (! ~ $#* 0)
        ls $*^'/./' | grep '/[0-9]+.*\.md$'| sort -r -t. +1
}

fn gen_blog_post_title {
    title=`{basename $1 | sed 's/^[0-9\-]*_(.*)\.md$/\1/; s/_/ /g' }
    permlink= `{echo $1 | sed 's,^/[a-z/]*www/,/,; s,^sites/[^/]*/*/,/,; s/\.md$//' }
    du=`{ls -l $1}
    echo '##<a href="'^$"permlink^'">' $"title^'</a> *('By $du(4) Last mod: $du(7 8 9) ')*'
}




# Handlers
fn set_handler {
    handler = $1
    shift
    handler_args = $*
}

fn md_handler {
    cat $* | $formatter
}

fn tpl_handler {
    template.awk $1 | rc $rcargs
}

fn html_handler {
    cat $1 | /bin/sed '0,/<[Bb][Oo][Dd][Yy][^>]*>/d; /<\/[Bb][Oo][Dd][Yy]>/,$d' 
}

fn txt_handler {
    echo '<pre>'
    # XXX Insering a blank line between lines in input is good for fortunes.txt, but maybe not for other .txt files
    # XXX Words are not broken, even if they are way beyond 82 chars long
    cat $1 |sed 's/$/\n/g; s/</\&lt;/g; s/>/\&gt;/g' |fmt -l 82 -j
    echo '</pre>'
}

fn dir_listing_handler {
    body = $1
    echo '<h1 style="text-transform: capitalize;">' `{basename -d $body|sed -e 's,.*//,,g' -e 's,/$,,' -e 's,/, / ,g' } '</h1>'
    echo '<ul style="text-transform: capitalize;">'
    ls -F `{ basename -d $body } | sed -e $dirfilter' s,^'$sitedir'/.*/([^$].*),<li><a href="\1">\1</a></li>,'
    echo '</ul>'
}

fn 404_handler {
    template.awk `{get_inc_file 404.tpl } | rc $rcargs
}

fn blog_dir_handler {
    blogDirs = $*

    if (! ~ $blogTitle '')
        echo '<h1>'$"blogTitle'</h1>' #" stupid vim syntax highlighting ;P

    echo '<div style="text-align:right">(<a href="index.rss">rss feed</a>)</div>'

    for (f in `{ sortedBlogPostList $blogDirs }) {
        #title=`{basename $f | sed 's/^[0-9\-]*_(.*)\.md$/\1/; s/_/ /g' }
        #du=`{ls -l $f}
        #echo '##' $title '*('By $du(4) Last mod: $du(7 8 9) ')*'
        gen_blog_post_title $f
        cat $f 
        echo 
    } | $formatter
}

fn blog_post_handler {

    gen_blog_post_title $1 | $formatter
    $formatter < $1
}

fn quote_html {
    sed 's/</\&lt;/g; s/>/\&gt;/g'
}
fn debug_handler {
    echo '<pre>'
    env |quote_html
    echo '</pre>'
}


fn select_handler {
dprint $body
    if (test -f $body.md) {
        if (! ~ $#inBlog 0)
            set_handler blog_post_handler $body.md
        if not
            set_handler md_handler $body.md
    }
    if not if (~ $body */_debug)
        set_handler debug_handler 
    if not if (test -f $body.tpl)
        set_handler tpl_handler $body.tpl

    if not if (test -f $body.html)
        set_handler html_handler $body.html

    # Handle eplicit .html urls, this should not happen (the web server will usually handle this anyway)
    # XXX We probably should setup a permanent redirect to $body|sed 's/.html$//' here
    if not if (~ $body *.html && test -f $body)
        set_handler html_handler $body

    # This should probably be merged with the blog_dir_handler
    if not if (~ $body */[bB]log/index */[bB]log//index && ~ $#blogDirs 0)
        blogDirs = `{basename -d $body}

    # Global tpl (eg sitemap.tpl)
    if not if (test -f pub/^$uri^.tpl)
        set_handler tpl_handler pub/^$uri^.tpl

    if not if (test -f $body.txt)
        set_handler txt_handler $body.txt

    # Dir listing
    if not if(~ $body */index && ~ $#blogDirs 0)
        set_handler dir_listing_handler $body

    # File not found
    if not if(~ $#blogDirs 0) {
        set_handler 404_handler $body
        dprint 'NOT FOUND: '$SERVER_NAME^$REQUEST_URI^' - '^$"HTTP_REFERER^' - '^$"HTTP_USER_AGENT
        echo 'Status: 404 Not Found'
    }

    if(! ~ $#blogDirs 0)
        set_handler blog_dir_handler $blogDirs
}


fn genbody {
    # Actually execute request
    $handler $handler_args
}


# Careful, the proper p9p path might not be set until initrc.local is sourced
path=(. $PLAN9/bin ./bin/ /bin/ /usr/bin) 

site=$SERVER_NAME
headers=inc/headers.tpl
template=_default
sidebar=sidebar
baseuri=http://$site/
for(i in siteTitle siteSubTitle pageTitle extraHeaders)
    $i = ''

. etc/initrc

if(test -f etc/initrc.local)
    . etc/initrc.local


# Parse request URL
uri = `{echo -n $REQUEST_URI | sed -e 's/\?.*//; s/[^a-zA-Z0-9_+\-\/\.]//g; s/\.\.*/./g;' -e '1q'}
ifs='/' {
	args = `{echo -n $uri}
}


if(! ~ $#debug 0)
    dprint '  ' $SERVER_NAME^$REQUEST_URI^' - '^$"HTTP_USER_AGENT


if (! ~ $args '') {
    if (~ $args($#args) 'index')
        perm_redirect `{ echo $REQUEST_URI | sed 's,/index$,/,' }
        
    pageTitle=`{echo $args | sed -e 's/ / - /g' -e 's/_/ /g'}
    body=$uri
}
if not {
    body='/'
}

fpath=$sitedir
for (i in ('' $args)) {
    fpath = $fpath/$i
    # We don't want blog settings to cascade into posts, note that we are inBlog instead
    if (! ~ $#blogDirs 0 && ! ~ $body */index.rss */[bB]log */[bB]log/) {
        inBlog = $blogDirs
        blogDirs = () 
    }

    if (test -f $fpath/_config)
        . $fpath/_config

    if (~ $#blogDirs 0 && ~ $#inBlog 0 && ~ $i [Bb]log)
        inBlog = 'yes'
}

# Redirections and other preprocessing
if (~ $#redirectPermanent 1)
    perm_redirect $"redirectPermanent

if not if (~ $#redirectPermanent 2 && {echo $SERVER_NAME^$REQUEST_URI|grep -s $redirectPermanent(1) }) {
    # Experimental regexp sub-based redirect, probably should find a nicer interface
    # For now only used at sites/harmful.cat-v.org/software/OO_programming/_config
    to=`{echo $SERVER_NAME^$REQUEST_URI|sed 's|'^$redirectPermanent(1)^'|'^$redirectPermanent(2)^'|'}
    if(! ~ $to $REQUEST_URI)
	perm_redirect $to
}

# Set Page title
if(~ $pageTitle '')
	pageTitle=$siteTitle^' '^$siteSubTitle
if not
	pageTitle=$"pageTitle^' | '^$"siteTitle^' '^$"siteSubTitle

# Template/body selection
template=$sitedir/$template.tpl

body=$sitedir/$body
rssuri=$uri


if (test -d $body) {
    if(! ~ $body */)
    	perm_redirect $REQUEST_URI^'/'
    body=$body/index
    rssuri=$rssuri/
}

if(! ~ $#blogDirs 0 || ! ~ $#inBlog 0) {
    rssuri=`{basename -d $uri}
    rssuri=$baseuri`{cleanname $"rssuri^/index.rss}
    extraHeaders=$"extraHeaders ^ \
    	'<link rel="alternate" type="application/rss+xml" title="RSS" href="'$rssuri'" />
'
}


select_handler

fn template {
    template.awk | rc $rcargs |
    awk '{
        buf = buf $0"\n"
        if(length(buf) > 8192) {
            printf "%s", buf
            buf = ""
        }
    }
    END{ printf "%s", buf }'
}

if(! ~ $REQUEST_URI */index.rss) {
	cat $headers $template | template
	echo '</body></html>'
	exit
}

# RSS
if (~ $body */[bB]log/index.rss */[bB]log//index.rss && ~ $#blogDirs 0)
    blogDirs = `{basename -d $body}


uri = `{echo $uri | sed 's/index.rss$//'}
uri=$baseuri$"uri

fn statpost {
	f = $1
	uri = `{echo $f | sed 's,^'$sitedir',,'}
	title=`{basename $f | sed 's/^[0-9\-]*_(.*)\.md$/\1/; s/_/ /g' }
	date=`{/bin/date -Rd `{basename $f |sed 's/(^[0-9\-]*).*/\1/; s/-[0-9]$//'}}
	# TODO: use mtime(1) and ls(1) instead of lunix's stat(1)
	stat=`{stat -c '%Y %U' $f}
	#mdate=`{/bin/date -Rd $stat(1)} # Not used because it is unreliable
	uri=$baseuri^`{cleanname `{echo -n $uri | sed 's/\.(md|tpl)//g'}}
	by=$stat(2)
	ifs=() {
		summary=`{awk -v max'='1024 '{
			nc += 1 + length;
			if(nc > max) {
				print substr($0, 1, nc - max) "..."
				exit
			}
			print
		}' $f |fmt -j| sed 's/\]\]>/Fucking goddamn XML garbage/g'}
	}
}

cat pub/feeds/rss20.tpl | template