new version of pdftotext broke us

src/filters/rclpdf

@@ -1,5 +1,5 @@
 #!/bin/sh
-# @(#$Id: rclpdf,v 1.4 2005-10-20 15:42:29 dockes Exp $  (C) 2004 J.F.Dockes
+# @(#$Id: rclpdf,v 1.5 2005-12-02 16:17:55 dockes Exp $  (C) 2004 J.F.Dockes
 # This is copied almost verbatim from Estraier:
 #================================================================
 # Estraier: a personal full-text search system
@@ -65,8 +65,6 @@ iconv -f UTF-8 -t UTF-8 -c -s |
 awk '
 BEGIN {
   esc = 0
-  mul = 1
-  emp = 0
 }
 {
   if(esc < 1 && $0 ~ /^<title>/ && $0 ~ /title>$/){
@@ -77,10 +75,11 @@ BEGIN {
     gsub(/>/, "\\&gt;", $0)
     printf("<title>%s</title>\n", $0)
   } else if($0 == "<pre>"){
+    # Begin of body text. need to escape some chars from now on as 
+    # pdftotext sometimes doesnt do it
     esc++
     printf("<p>")
-    mul = 1
-  } else if($0 == "</pre>"){
+  } else if ($0 ~ /<\/pre>/){
     esc--
     printf("</p>\n")
   } else if($0 ~ /-$/){
@@ -100,9 +99,4 @@ BEGIN {
   }
 }
 ' 
-# Suppressed code 2 lines above (at the last print $0), which seemed to
-# deal with multibyte character being cut by a newline ? It caused problems
-# (sometimes concatenated last word of a line with first of next, and I
-# didn't really understand its use as iconv -c is supposed to fix the
-# encoding anyway