index: add skippedPathsFnmPathname variable to enable disabling the use of FNM_PATHNAME while matching in skippedPaths. Closes issue #67

src/common/rclconfig.cpp

@@ -43,6 +43,7 @@
 #include "smallut.h"
 #include "textsplit.h"
 #include "readfile.h"
+#include "fstreewalk.h"
 
 #ifndef NO_NAMESPACES
 using namespace std;
@@ -200,6 +201,12 @@ bool RclConfig::updateMainConfig()
         m_rmtstate.init(this, 0, "indexedmimetypes");
 	return false;
     }
+    delete m_conf;
+    m_conf = newconf;
+    m_skpnstate.init(this, m_conf, "skippedNames");
+    m_rmtstate.init(this, m_conf, "indexedmimetypes");
+
+
     setKeyDir(cstr_null);
     bool nocjk = false;
     if (getConfParam("nocjk", &nocjk) && nocjk == true) {
@@ -218,10 +225,12 @@ bool RclConfig::updateMainConfig()
 	TextSplit::noNumbers();
     }
 
-    delete m_conf;
+    bool fnmpathname = true;
-    m_conf = newconf;
+    if (getConfParam("skippedPathsFnmPathname", &fnmpathname)
-    m_skpnstate.init(this, m_conf, "skippedNames");
+	&& fnmpathname == false) {
-    m_rmtstate.init(this, m_conf, "indexedmimetypes");
+	FsTreeWalker::setNoFnmPathname();
+    }
+
     return true;
 }
 

src/doc/user/usermanual.sgml

@@ -3441,12 +3441,19 @@ skippedNames = #* bin CVS  Cache cache* caughtspam  tmp .thumbnails .svn \
               <programlisting>
 skippedPaths = ~/somedir/&lowast;.txt
               </programlisting>
-                <para>The values in the <literal>*skippedPaths</literal>
+            </listitem>
-                variables are currently matched with
+          </varlistentry>
-                <literal>fnmatch(3)</literal>, with the FNM_PATHNAME and
+
-                FNM_LEADING_DIR flags. This means that '/' characters must
+          <varlistentry id="rcl.install.config.recollconf.skippedpathsfnmpathname">
-                be matched explicitely, which is probably
+            <term><literal>skippedPathsFnmPathname</literal></term>
-                unfortunate.</para>
+                <listitem><para>The values in the
+                <literal>*skippedPaths</literal> variables are matched by
+                default with <literal>fnmatch(3)</literal>, with the
+                FNM_PATHNAME and FNM_LEADING_DIR flags. This means that '/'
+                characters must be matched explicitely. You can set
+                <literal>skippedPathsFnmPathname</literal> to 0 to disable
+                the use of FNM_PATHNAME (meaning that /*/dir3 will match
+                /dir1/dir2/dir3).</para>
 
             </listitem>
           </varlistentry>

src/sampleconf/recoll.conf.in

@@ -29,6 +29,12 @@ skippedNames = #* bin CVS  Cache cache* caughtspam  tmp .thumbnails .svn \
 # not set, the daemon uses skippedPaths.
 #daemSkippedPaths = 
 
+# Recoll uses FNM_PATHNAME by default when matching skipped paths, which
+# means that /dir1/dir2/dir3 is not matched by */dir3. Can't change the
+# default now, but you can set the following variable to 0 to disable the
+# use of FNM_PATHNAME (see fnmatch(3) man page)
+# skippedPathsFnmPathname = 1
+
 # Option to follow symbolic links. We normally don't, to avoid duplicated
 # indexing (in any case, no effort is made to identify or avoid multiple
 # indexing of linked files)

src/utils/Makefile

@@ -11,7 +11,7 @@ $(BIGLIB): force
 	cd $(depth)/lib;$(MAKE)
 force:
 
-FSTREEWALK_OBJS= trfstreewalk.o fstreewalk.o $(BIGLIB)
+FSTREEWALK_OBJS= trfstreewalk.o $(BIGLIB)
 trfstreewalk : $(FSTREEWALK_OBJS)
 	$(CXX) -o trfstreewalk $(FSTREEWALK_OBJS) $(LIBICONV) \
 	       $(LIBSYS)

src/utils/circache.cpp

@@ -1301,7 +1301,7 @@ int main(int argc, char **argv)
               exit(1);
           }
           string udi;
-          make_udi(fn, cstr_null, udi);
+          make_udi(fn, "", udi);
           sprintf(dic, "#whatever...\nmimetype = text/plain\nudi=%s\n", 
                   udi.c_str());
           string sdic;

src/utils/fstreewalk.cpp

@@ -40,6 +40,8 @@
 using namespace std;
 #endif /* NO_NAMESPACES */
 
+bool FsTreeWalker::o_useFnmPathname = true;
+
 const int FsTreeWalker::FtwTravMask = FtwTravNatural|
     FtwTravBreadth|FtwTravFilesThenDirs|FtwTravBreadthThenDepth;
 
@@ -153,7 +155,7 @@ bool FsTreeWalker::setSkippedPaths(const list<string> &paths)
 }
 bool FsTreeWalker::inSkippedPaths(const string& path, bool ckparents)
 {
-    int fnmflags = FNM_PATHNAME;
+    int fnmflags = o_useFnmPathname ? FNM_PATHNAME : 0;
 #ifdef FNM_LEADING_DIR
     if (ckparents)
         fnmflags |= FNM_LEADING_DIR;
@@ -420,6 +422,8 @@ FsTreeWalker::Status FsTreeWalker::iwalk(const string &top,
 
 #include <iostream>
 
+#include "rclinit.h"
+#include "rclconfig.h"
 #include "fstreewalk.h"
 
 using namespace std;
@@ -542,7 +546,13 @@ int main(int argc, const char **argv)
       opt |= FsTreeWalker::FtwTravFilesThenDirs;
   else if (op_flags & OPT_m)
       opt |= FsTreeWalker::FtwTravBreadthThenDepth;
-  
+
+  string reason;
+  if (!recollinit(0, 0, reason)) {
+      fprintf(stderr, "Init failed: %s\n", reason.c_str());
+      exit(1);
+  }
+
   FsTreeWalker walker(opt);
   walker.setSkippedNames(patterns);
   walker.setSkippedPaths(paths);

src/utils/fstreewalk.h

@@ -38,6 +38,18 @@ struct stat;
  */
 class FsTreeWalker {
  public:
+    // Global option to use FNM_PATHNAME when matching paths (for
+    // skippedPaths).  
+    // We initially used FNM_PATHNAME, and we can't change it now
+    // (because of all the config files around). So add global option
+    // to not use the flag, which can be set from rclconfig by adding
+    // a value to the config file (skippedPathsNoFnmPathname)
+    static bool o_useFnmPathname;
+    static void setNoFnmPathname()
+    {
+	o_useFnmPathname = false;
+    }
+
     // Flags for call to processone(). FtwDirEnter is used when
     // entering a directory. FtwDirReturn is used when returning to it
     // after processing a subdirectory.

website/index.html.en

@@ -1,18 +1,13 @@
 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
-
 <html>
   <head>
-    <title>RECOLL: a personal text search system for
+    <title>RECOLL: a personal text search system for Unix/Linux</title>
-    Unix/Linux</title>
     <meta name="generator" content="HTML Tidy, see www.w3.org">
     <meta name="Author" content="Jean-Francois Dockes">
-    <meta name="Description" content=
+    <meta name="Description" content="Recoll is a personal desktop full text search application for Unix, Linux and Mac OS X, based on the Xapian search engine library.">
-    "recoll is a simple full-text search system for unix and linux based on the powerful and mature xapian engine">
+    <meta name="Keywords" content="full text search,fulltext,full text,search,desktop search,unix,linux,open source,free">
-    <meta name="Keywords" content=
-      "full text search,fulltext,desktop search,unix,linux,solaris,open source,free">
     <meta http-equiv="Content-language" content="en">
-    <meta http-equiv="content-type" content=
+    <meta http-equiv="content-type" content="text/html; charset=iso-8859-1">
-    "text/html; charset=iso-8859-1">
     <meta name="robots" content="All,Index,Follow">
     <link type="text/css" rel="stylesheet" href="styles/style.css">
   </head>
@@ -41,31 +36,30 @@
 	a personal full text search tool for Unix/Linux.</h1>
 
       <p>It is based on the very strong <a href=
-	  "http://www.xapian.org">Xapian</a> back-end, for which it
+	  "http://www.xapian.org">Xapian</a> search engine library, for
-	provides a feature-rich yet easy to use front-end with a Qt
+	  which it provides a powerful text extraction layer and a
-	graphical interface.</p>
+	  complete, yet easy to use, Qt graphical interface.</p>
 
       <p><span class="application">Recoll</span> is free, open source,
 	and licensed under the GPL. The current version is
-	<a class="important" href="download.html">1.16.2</a> 
+	<a href="download.html">1.16.2</a> 
 	(<a href="release-1.16.html">Release notes</a>).</p>
 
       <p class="remark">Recoll will index an <b>MS-Word</b> document stored as
 	  an <b>attachment</b> to an <b>e-mail message</b> inside
 	  a <b>Thunderbird folder</b> archived in a <b>Zip file</b>
-	  (and more...). It will also let you open a copy of the file
+	  (and more...). It will also help you search for it with a
-	  without any fuss. There is little that will remain hidden on 
+	  friendly and powerful interface, and let you open a copy of the file
+	  without much fuss. There is little that will remain hidden on 
 	  your disk. 
         <a class="important" href="features.html">More details &hellip;</a></p>
 
-      <p></p>
-
       <p>If you have problems with
         Recoll, <a href="doc.html">documentation</a>
         and <a href="support.html">support</a> are available.</p> 
 
       <p><b><i>Recoll user ?</i></b> Maybe there are still a few useful
-	  tricks that you don't know about. A quick look at
+      	  search tricks that you don't know about. A quick look at
 	the <a href="usermanual/rcl.search.tips.html">search
 	tips</a> might prove useful ! Also the 
         <a href="http://bitbucket.org/medoc/recoll/wiki/FaqsAndHowTos">