shithub: werc

ref: 5cf42c28a0223319b7744ce57bfeeea7dd70a850
dir: /bin/controller.rc/

View raw version
#!/usr/local/plan9/bin/rc
. cgilib.rc
cd ..

forbidden_uri_chars='[^a-zA-Z0-9_+\-\/\.]'
response_format=html

fn get_lib_file {
    wantedfile = $1
    if (test -f $sitedir/_werc/lib/$wantedfile)
    	echo -n $sitedir/_werc/lib/$wantedfile
    if not if(! ~ $masterSite 0 && test -f sites/$masterSite/_werc/lib/$wantedfile)
    	echo -n       sites/$masterSite/_werc/lib/$wantedfile
    if not if (test -f lib/$wantedfile)
    	echo -n        lib/$wantedfile

    # DEPRECATED
    if not if (~ $#disableDeprecated 0 && test -f $sitedir/_inc/$wantedfile) {
    	echo -n $sitedir/_inc/$wantedfile
        dprint DEPRECATION WARNING: Using file in depracted path: $sitedir/_inc/$wantedfile XXX
    }
    if not if(~ $#disableDeprecated 0 && ! ~ $masterSite 0 && test -f sites/$masterSite/_inc/$wantedfile) {
    	echo -n sites/$masterSite/_inc/$wantedfile
        dprint DEPRECATION WARNING: Using file in depracted path: sites/$masterSite/_inc/$wantedfile XXX
    }
    if not if (~ $#disableDeprecated 0 && test -f inc/$wantedfile) {
    	echo -n inc/$wantedfile
        dprint DEPRECATION WARNING: Using file in depracted path: inc/$wantedfile XXX
    }
}

# DEPRECATED
fn get_inc_file {
    get_lib_file $*
    dprint DEPRECATION WARNING: Called get_inc_file, should call get_lib_file
}


# 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; /'^$forbidden_uri_chars^'/d; /\/robots.\txt$/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 $1 }

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 `{get_lib_file 404.tpl } }

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 {

    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 explicit .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

    # Blog handler
    if not if (~ $body */[bB]log/index */[bB]log//index || ! ~ $#blogDirs 0) {
        if (~ $#blogDirs 0)
            blogDirs = `{basename -d $body}

        rssuri=`{basename -d $uri}
        rssuri=$baseuri`{cleanname $"rssuri^/index.rss}
        extraHeaders=$"extraHeaders ^ \
    	'<link rel="alternate" type="application/rss+xml" title="RSS" href="'$rssuri'" />
'

        set_handler blog_dir_handler $blogDirs
    }

    # Rss feeds. TODO: we should check that the request is for a real blog dir
    if not if (~ $REQUEST_URI */index.rss) {
        uri = `{echo $uri | sed 's/index.rss$//'}
        if(~ $blogDirs 0)
            blogDirs = $sitedir^$uri
        uri=$baseuri$"uri
        master_template=feeds/rss20.tpl
        response_format=raw
    }

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

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

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

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

}


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=$SERVER_NAME
headers=lib/headers.tpl
master_template=default_master.tpl
sidebar=sidebar
baseuri=http://$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 $REQUEST_URI | sed -e 's/\?.*//; s/'^$forbidden_uri_chars^'//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 # DEPRECATED
    if (test -f $fpath/_werc/config)
        . $fpath/_werc/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


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

select_handler

# Template/body selection
master_template= `{get_lib_file $master_template}

if(~ $master_template 0) {
    dprint DEPRECATION WARNING: using $sitedir/_default.tpl, should move to _werc/lib/XXX
    master_template=$sitedir/_default.tpl
}


if(~ $response_format html) {
    # Is awk_buffer really useful?
	template $headers $master_template | awk_buffer
	echo '</body></html>'
}
if not if (~ $response_format raw) {
	template < $master_template 
}