shithub: werc

ref: 8cec7f7861dde242d8f3f242723edd81c7181c83
dir: /bin/werc.rc/

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

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

# Expected input: ls -F style, $sitedir/path/to/files/
#          <ls -F+x><symlink hack><Useless?><hiden files  >
dirfilter='s/\*$//; s,/+\./+,/,g; s,^\./,,; /\/[._][^\/]/d; /'$forbidden_uri_chars'/d; /^\/(robots|sitemap)\.txt$|\/index\.(md|html|txt|tpl)$/d; /_werc\/?$/d; '
dirclean=' s/\.(md|html|txt)$//; '

# To be used from config files
fn conf_perm_redirect {
    if(~ $#* 1)
        perm_redirect $1
    if not
        perm_redir_patterns=($perm_redir_patterns $1 $2)
}

fn conf_hide_paths {
    for(i in $*)
       dirfilter=$dirfilter^'/^'$i'$/d; '
}

# Standard handlers
fn nav_tree {
    if(! ~ $#sideBarNavTitle 0)
        echo '<p class="sideBarTitle">'$"sideBarNavTitle':</p>'
    # Ignore stderr, last path element might be a file that doesn't exist (eg., foo for foo.md)
    # /./ to deal with p9p's ls failure to follow dir symlinks otherwise
    ls -F $sitedir/./$req_paths_list >[2]/dev/null \
        | sed 's!^'$sitedir'!!; '$dirfilter'/\/[^_.\/][^\/]*(\.(md|txt|html)|\/)$/!d; '$dirclean \
        | sort -u | awk -F/ ' 
    function p(x, y, s) { for(i=0; i < x-y; i+=1) print s }
    { 
        d = ""
        if(match($0, "/$"))
            d = "/"
        sub("/$", "") # Strip trailing / for dirs so NF is consistent 

        p(NF, lNF, "<ul class=\"side-bar\">")
        p(lNF, NF, "</ul>")
        lNF = NF

        bname = $NF d
        path = $0 d
        gsub("_", " ", bname)

        if(index(ENVIRON["req_path"] "/", path) == 1)
            print "<li><a href=\"" path "\" class=\"thisPage\">&raquo;<i> " bname "</i></a>"
        else 
            print "<li><a href=\"" path "\">&rsaquo; " bname "</a></li>"
    }
    END { p(lNF, 0, "</ul>") }'
}


fn md_handler { $formatter < $1 }

fn tpl_handler { template $* }

fn html_handler {
    # body states: 0 = no <body> found, 2 = after <body>, 1 = after <body></body>, -1 = after </body>
    awk 'gsub(".*<[Bb][Oo][Dd][Yy][^>]*>", "") > 0 {body=2}
        gsub("</ *[Bb][Oo][Dd][Yy][^>]*>.*", "") > 0 {print; body=body-1}
        body==2 {print}
        body==0 {buf=buf "\n" $0}
        END {if(body<=0) {print buf}}' < $1
}

fn txt_handler {
    # Note: Words are not broken, even if they are way beyond 82 chars long
    echo '<pre>'
    sed 's/</\&lt;/g; s/>/\&gt;/g' < $1 | fmt -l 82 -j
    echo '</pre>'
}

fn dir_listing_handler {
    d=`{basename -d $1}
    if(~ $#d 0)
        d='/'
    echo $d|sed 's,.*//,,g; s,/$,,; s,/, / ,g; s,.*,<h1 class="dir-list-head">&</h1> <ul class="dir-list">,'
    # Symlinks suck: '/.' forces ls to list the linked dir if $d is a symlink.
    ls -F $dir_listing_ls_opts $sitedir$d/. | sed $dirfilter$dirclean' s,.*/([^/]+/?)$,<li><a href="\1">\1</a></li>,'
    echo '</ul>'
}

fn notices_handler {
    for(type in notify_errors notify_notes notify_success)
        for(n in $$type)
            echo '<div class="'$type'"><b>'$"n'</b></div>'
}

fn setup_handlers {

    if(test -f $local_path.md)
        handler_body_main=(md_handler $local_path.md)
    if not if(test -f $local_path.tpl)
        handler_body_main=(tpl_handler $local_path.tpl)
    if not if(test -f $local_path.html)
        handler_body_main=(html_handler $local_path.html)
    # Global tpl (eg sitemap.tpl), should take precedence over txt handler!
    if not if(test -f lib^$req_path^.tpl)
        handler_body_main=(tpl_handler lib^$req_path^.tpl)
    if not if(test -f $local_path.txt)
        handler_body_main=(txt_handler $local_path.txt)

    # XXX Should check that $enabled_apps exist in $werc_apps?
    # XXX Should split init of apps that provide main handler (eg., blog) and apps that don't (eg., comments)?
    if(! ~ $#enabled_apps 0)
        for(a in $enabled_apps)
            $a^'_init'

    if(! ~ $#handler_body_main 0)
        { } # We are done
    # Dir listing
    if not if(~ $local_path */index)
        handler_body_main=(dir_listing_handler $req_path)
    # Canonize explicit .html urls, the web server might handle this first!
    if not if(~ $local_path *.html && test -f $local_path)
        perm_redirect `{ echo $req_path|sed 's/.html$//' }
    # Fallback static file handler
    if not if(test -f $local_path)
        static_file $local_path
    if not if(~ $req_path /pub/* && test -f .$req_path)
        static_file .$req_path
    # File not found
    if not {
        handler_body_main=(tpl_handler `{get_lib_file 404.tpl})
        echo 'Status: 404 Not Found'
        dprint 'NOT FOUND: '$SERVER_NAME^$"REQUEST_URI^' - '^$"HTTP_REFERER^' - '^$"HTTP_USER_AGENT
    }
}

fn run_handlers { for(h in $*) run_handler $$h }
fn run_handler { $*(1) $*(2-) }

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

headers=lib/headers.tpl
res_tail='</body></html>'
ll_add handlers_bar_left nav_tree
werc_apps=( apps/* )
werc_root=`{pwd}
sitesdir=sites
for(i in siteTitle siteSubTitle pageTitle extraHeaders)
    $i = ''

. ./etc/initrc

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

for(a in $werc_apps)
    . ./$a/app.rc

fn werc_exec_request {
    site=$SERVER_NAME
    base_url=http://$site
    sitedir=$sitesdir/$site
    master_template=`{get_lib_file default_master.tpl}
    current_date_time=`{date}

    # Note: $REQUEST_URI is not officially in CGI 1.1, but seems to be de-facto
    req_path=`{echo -n $REQUEST_URI | sed 's/\?.*//; s!//+!/!g; s/'^$forbidden_uri_chars^'//g; s/\.\.*/./g; 1q'}
    local_path=$sitedir$req_path
    ifs='/' { args=`{echo -n $req_path} }

    # Preload post args for templates where cgi's stdin is not accessible
    if(~ $REQUEST_METHOD POST) {
        load_post_args
        login_user
    }

    if(~ $req_path */index)
        perm_redirect `{echo $req_path | sed 's,/index$,/,'}

    if(~ $local_path */) {
        if(test -d $local_path)
            local_path=$local_path^'index'
        if not # XXX: This redir might step on apps with synthetic dirs.
            perm_redirect `{echo $req_path|sed 's,/+$,,'}
    }
    if not if(test -d $local_path)
        perm_redirect $base_url^$req_path^'/'

    if(! ~ $#args 0)
        pageTitle=`{ echo $args|sed -e 's/ / - /g' -e 's/_/ /g' }

    cd $sitedir
    req_paths_list='/' # Note: req_paths_list doesn't include 'stnythetic' dirs.
    conf_wd='/' # Used in config files to know where we are in the document tree.
    if(test -f _werc/config)
        . _werc/config
    for(i in $args) {
        conf_wd=$conf_wd^$i
        req_paths_list=($req_paths_list $conf_wd)
        if(test -d $i) {
            conf_wd=$conf_wd'/'
            cd $i
            if(test -f _werc/config)
                . _werc/config
        }
    }
    cd $werc_root

    f=();t=()
    for(i in $perm_redir_patterns) {
        if(~ $#f 0)
            f=$i
        if not {
            t=$i
            from=$base_url^$req_path
            to=`{ echo $from | sed 's!'$f'!'$t'!' }
            if(! ~ $to $from)
                perm_redirect $to
            f=() 
        } 
    }

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

    setup_handlers

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

    template $headers $master_template | awk_buffer
    echo $res_tail
}

werc_exec_request