Improve rcldoc filter and switch back to using it for indexing instead of direct antiword exec. This is slightly slower but it does catch a number of .doc files which would not be indexed otherwise

tests/msword/msword.sh

@@ -5,8 +5,10 @@ topdir=`dirname $0`/..
 
 initvariables $0
 
-recollq '"Evenements et programme 2006"' 2> $mystderr | 
-	egrep -v '^Recoll query: ' > $mystdout
+(
+recollq '"Evenements et programme 2006"' 
+recollq 'pcx11 manuel de programmation iamactuallyanrtf'
+)  2> $mystderr | egrep -v '^Recoll query: ' > $mystdout
 
 diff -w ${myname}.txt $mystdout > $mydiffs 2>&1
 

tests/msword/msword.txt

@@ -1,3 +1,5 @@
 2 results
 application/msword	[file:///home/dockes/projets/fulltext/testrecoll/msword/programme.doc]	[programme.doc]	58880	bytes	
 application/msword	[file:///home/dockes/projets/fulltext/testrecoll/zip/misc.zip]	[misc.zip]	168155	bytes	
+1 results
+application/msword	[file:///home/dockes/projets/fulltext/testrecoll/msword/IAmActuallyAnRTF.DOC]	[DOC PROGRAMMEUR PCX9]	85381	bytes	

tests/runtests.sh

@@ -13,8 +13,9 @@ makeindex() {
   echo "Indexing" 
   recollindex -z
 }
-
+if test x$noindex = x ; then
 makeindex
+fi
 
 # Yes, we could/should use the $toptmp from shared.sh here, but what if
 # this is unset ?