New context var 'local_file' that will point to the on-disk-file that represents the main content of a given page request (if such exists, otherwise it is '' or ()).

Move title extracting functions from sitemap.tpl to werclib.rc.
Rewrite .md title extraction code to only match first lines when 'underlines' with =.

Use the changes above to generate better page titles when possible.

bin/werclib.rc

@@ -90,6 +90,36 @@ fn get_md_file_attr {
     sed -n '/^\* '$2': /p; /^\* '$2': /q; /^$/q' < $1
 }
 
+# File title extraction
+
+fn get_md_title {
+    #sed 's/^(................................................................[^ ]*).*$/\1/g; 1q' < $1 
+    sed -n -e '1N; /^.*\n===*$/N; /.*\n===*\n$/!b' -e 's/\n==*\n//p' < $1
+}
+
+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 *.md)
+        get_md_title $1
+    if not if(~ $1 *.html)
+        get_html_title $1
+    if not 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
+    }
+}
+
+
 ##########################################################################
 ##########################################################################
 #app_blog_methods = ( _post index.rss )