ref: 12a2d9014545a1574b617c8373e2f793598b99dd
parent: 8fc1781d6e308c78063e916a668fdb6c1311cd19
parent: c070f51c5c0417f32e2cf9aadf6a679106f2a501
author: sl <uriel@engel.se.cat-v.org>
date: Fri Feb 6 16:03:58 EST 2009
Merge.
--- /dev/null
+++ b/bin/fltr_cache.rc
@@ -1,0 +1,35 @@
+#!/usr/bin/env rc
+
+fn fltr_cache {+ a=()
+ tmpf=()
+
+ proc=$1
+ shift
+
+ if(~ $#* 0) {+ tmpf=/tmp/fmttmp.$pid
+ f=$tmpf
+ score=`{{tee $tmpf || exit 1} | sha1sum}+ }
+ if not {+ f=$1
+ if(~ $f */) {+ score=`{du -an $f | sha1sum || exit 1} # XXX using -n(bytes) instead of -t(lastmod) because sitemap proc touches files in tree.+ a=$f
+ f=/dev/null
+ }
+ if not
+ score=`{sha1sum $f || exit 1}+ }
+ cachedir=/tmp/fltr_cache/$score
+ mkdir -p $cachedir >[2]/dev/null
+
+ if(test -s $cachedir/$proc)
+ cat $cachedir/$proc
+ if not
+ if($proc $a < $f | tee $cachedir/$pid)
+ mv $cachedir/$pid $cachedir/$proc
+
+ rm $tmpf $cachedir/$pid >[2]/dev/null &
+}
--- a/bin/fproc_cache.rc
+++ /dev/null
@@ -1,19 +1,0 @@
-#!/usr/bin/env rc
-#. 9.rc # Not really needed when calling from werc, only would be needed if you use fproc_cache.rc standalone
-# TODO: Allow to provide two arguments, second arg is a file name, use filename+size+mtime for hashing.
-# Useful when dealing with many big files (eg., thumb gallery generation).
-
-proc=$1
-tmpfile=/tmp/fmttmp.$pid
-score=`{{tee $tmpfile || exit 1} | sha1sum}-
-cachedir=/tmp/fproc_cache/$score
-mkdir -p $cachedir >[2]/dev/null
-
-if(test -s $cachedir/$proc)
- cat $cachedir/$proc
-if not
- if($proc < $tmpfile | tee $cachedir/$pid)
- mv $cachedir/$pid $cachedir/$proc
-
-rm $tmpfile $cachedir/$pid >[2]/dev/null &
--- a/bin/werc.rc
+++ b/bin/werc.rc
@@ -3,6 +3,7 @@
. ./werclib.rc
. ./wercconf.rc
. ./corehandlers.rc
+. ./fltr_cache.rc
cd ..
forbidden_uri_chars='[^a-zA-Z0-9_+\-\/\.]'
@@ -105,7 +106,7 @@
if(! ~ $#debug 0)
dprint $"SERVER_NAME^$"REQUEST_URI - $"HTTP_USER_AGENT - $"REQUEST_METHOD - $"handler_body_main - $"master_template
- template $headers $master_template #| awk_buffer
+ template $headers $master_template | awk_buffer
echo $res_tail
}
--- a/etc/initrc
+++ b/etc/initrc
@@ -9,17 +9,17 @@
# Location of your Plan 9 from User Space installation (usually /usr/local/plan9)
plan9port=$PLAN9
+#plan9port=/usr/local/plan9
# Path, make sure the plan9port /bin directory is included before /bin
# Keep '.' in path! It is needed.
path=($plan9port/bin/ . ./bin/ ./bin/contrib/ /bin/ /usr/bin/)
-# Set this to your favorite markdown formatter, eg., markdown.pl (fproc_cache
-# is a wrapper around a script it takes as an argument, in the default
-# configuration markdown.pl, that caches output)
-# Note that some werc components assume a markdown-like formatter, but all
-# major functionality should should be formatter agnostic.
-formatter=(fproc_cache.rc markdown.pl)
+# Set this to your favorite markdown formatter, eg., markdown.pl (fltr_cache
+# takes as an argument a filter, in the default configuration markdown.pl, that
+# caches output) Note that some werc components assume a markdown-like
+# formatter, but all major functionality should should be formatter agnostic.
+formatter=(fltr_cache markdown.pl)
# Enable debugging, to disable set to ()
debug=true
--- a/lib/sitemap.tpl
+++ b/lib/sitemap.tpl
@@ -5,65 +5,61 @@
echo '' > $tmpfile
saveddf=$dirfilter
-fn getMdDesc {- sed 's/^(.......................................................................................................[^ ]*).*$/\1/g; 1q' < $1
+fn get_md_title {+ sed 's/^(................................................................[^ ]*).*$/\1/g; 1q' < $1
}
-# XXX Instead of recursion should use du(1) or similar.
+fn get_html_title {+ # H1 is not reliable because htmlroff doesn't use it :(
+ #desc=`{cat $1 | sed 32q | grep '<[Hh]1>' | sed 's/<[Hh]1>(.*)(<\/[Hh]1>|$)/\1/;s/<[^>]*>//g;1q'}+ # Pick the first line of body instead
+ desc=`{sed -n '/<[Bb][Oo][Dd][Yy]/,/./s/(<[^>]*>|$)//gp' < $1}+ if(~ $#desc 0)
+ desc=`{sed 's/<[^>]*>//g; 1q' < $1}+}
+
+fn get_file_title {+
+ if(~ $1 */) {+ if(test -f $1/index.md)
+ get_md_title $1/index.md
+ if not if(test -f $1/index.html)
+ get_html_title $1/index.html
+ }
+ if not if(~ $1 *.md)
+ get_md_title $1
+ if not if(~ $1 *.html)
+ get_html_title $1
+ if not
+ echo ''
+}
+
fn listDir {d=$1
- if(~ $#d 0)
- d=''
dirfilter=$saveddf
- blogDirs=()
if(test -f $d/_werc/config)
. ./$d/_werc/config
echo '<ul class="sitemap-list">'
- # Don't hide blog dirs for now
- #if(! ~ $#blogDirs 0 || ~ $1 */blog */Blog)
- # echo ''
- #if not
- if(! ~ $#redirectPermanent 1)
- {-
- for(i in `{ls -dF $d^*/ $d^*.md $d^*.html $d^*.txt >[2]/dev/null | sed $dirfilter$dirclean}) {- desc=''
- if(test -f $i.md)
- desc=`{getMdDesc $i.md}- if not if(~ $i */ && test -f $i/index.md)
- desc=`{getMdDesc $i/index.md}- if not if(test -f $i.html) {- # H1 is not reliable because htmlroff doesn't use it :(
- #desc = `{cat $i.html |sed 32q | grep '<[Hh]1>' |sed 's/<[Hh]1>(.*)(<\/[Hh]1>|$)/\1/;s/<[^>]*>//g;1q'}- # Pick the first line of body instead
- desc=`{sed -n '/<[Bb][Oo][Dd][Yy]/,/./s/(<[^>]*>|$)//gp' < $i.html}- if(~ $#desc 0)
- desc=`{sed 's/<[^>]*>//g; 1q' < $i.html}- #desc=`{/bin/sed -e '0,/<[Bb][Oo][Dd][Yy]/d;s/<[^>]*>//g;/^$/d' < $i.html >[2]/dev/null | sed 1q}- }
-
+ for(i in `{ls -dF $d^*/ $d^*.md $d^*.html $d^*.txt >[2]/dev/null | sed $dirfilter}) {+ desc=`{get_file_title $i}+ u=`{echo $i|sed 's!'$sitedir'!!; '$dirclean's!/index$!/!; '}if(! ~ $#desc 0 && ! ~ $desc '')
desc=' - '$"desc
- tit=`{echo /$i|sed 's/_/ /g; s,.*/([^/]+)/?$,\1,'}- echo '<li><a href="/'$i'">'^$"tit^'</a>' $"desc '</li>'
- echo -n $base_url^$i >> $tmpfile
- if(test -d $i) {- echo / >> $tmpfile
+ n=`{echo /$u|sed 's/_/ /g; s,.*/([^/]+)/?$,\1,'}+ echo '<li><a href="'$base_url$u'">'^$"n^'</a>' $"desc '</li>'
+ echo $base_url^$u >> $tmpfile
+ if(test -d $i)
@{ listDir $i }- }
- if not
- echo >> $tmpfile
}
- }
echo '</ul>'
}
-cd $sitedir
-listDir ''
+fltr_cache listDir $sitedir/
+#listDir $sitedir/
-cp $tmpfile ./sitemap.txt
-rm $tmpfile
+if(test -s $tmpfile)
+ mv $tmpfile $sitedir/sitemap.txt &
%}