check for empty file name in internfile, else gets stuck later because empty fn is interpreted as read stdin in md5

src/desktop/recoll_index_on_ac.sh

@@ -3,7 +3,14 @@
 # This is a shell script that starts and stops the recollindex daemon
 # depending on whether or not the power supply is plugged in.  It should be
 # called from the file ~/.config/autostart/recollindex.desktop.
-
+#
+# That is: make the script executable (chmod +x) and replace in
+# recollindex.desk the line:
+#   Exec=recollindex -w 60 -m 
+# With
+#   Exec=/path/to/recoll_index_on_ac.sh
+#
+#
 # By: The Doctor (drwho at virtadpt dot net)
 # License: GPLv3
 # 

src/doc/user/usermanual.sgml

@@ -2489,15 +2489,40 @@ text/html       [file:///Users/uncrypted-dockes/projets/bateaux/ilur/factEtCie/r
         <listitem><para><literal>dir</literal> for filtering the
             results on file location (Ex:
             <literal>dir:/home/me/somedir</literal>). <literal>-dir</literal>
-            also works to find results out of the specified directory, only
+            also works to find results not in the specified directory
-            after release 1.15.8. A tilde inside the value will be expanded to
+            (release >= 1.15.8). A tilde inside the value will be expanded
-            the home directory. <literal>dir</literal> is not a regular field
+            to the home directory. Wildcards will <emphasis>not</emphasis>
-            and only one value makes sense in a query (you can't use
+            be expanded. You cannot use <literal>OR</literal> with
-            <literal>dir:dir1 OR dir:dir2</literal>). Relative paths make
+            <literal>dir</literal> clauses (this restriction may go away in
-            sense, for example, 
+            the future).</para>
+
+           <para>Relative paths also make sense, for example, 
             <literal>dir:share/doc</literal> would match either
             <filename>/usr/share/doc</filename> or
             <filename>/usr/local/share/doc</filename> </para>
+
+          <para>Several <literal>dir</literal> clauses can be specified,
+          both positive and negative. For example the following makes sense:
+            <programlisting>
+dir:recoll dir:src -dir:utils -dir:common
+            </programlisting> This would select results which have both
+            <filename>recoll</filename> and <filename>src</filename> in the
+            path (in any order), and which have not either
+            <filename>utils</filename> or
+            <filename>common</filename>.</para>
+
+          <para>Another special aspect of <literal>dir</literal> clauses is
+          that the values in the index are not transcoded to UTF-8, and
+          never lower-cased or unaccented, but stored as binary. This means
+          that you need to enter the values in the exact lower or upper
+          case, and that searches for names with diacritics may sometimes
+          be impossible because of character set conversion
+          issues. Non-ASCII UNIX file paths are an unending source of
+          trouble and are best avoided.</para> 
+
+          <para>You need to use double-quotes around the path value if it
+          contains space characters.</para> 
+
            </listitem>
 
         <listitem><para><literal>size</literal> for filtering the

src/internfile/internfile.cpp

@@ -187,18 +187,27 @@ void FileInterner::tmpcleanup()
 // Empty handler on return says that we're in error, this will be
 // processed by the first call to internfile().
 // Split into "constructor calls init()" to allow use from other constructor
-FileInterner::FileInterner(const string &f, const struct stat *stp,
+FileInterner::FileInterner(const string &fn, const struct stat *stp,
 			   RclConfig *cnf, 
 			   TempDir& td, int flags, const string *imime)
     : m_tdir(td), m_ok(false), m_missingdatap(0)
 {
+    LOGDEB0(("FileInterner::FileInterner(fn=%s)\n", fn.c_str()));
+    if (fn.empty()) {
+	LOGERR(("FileInterner::FileInterner: empty file name!\n"));
+	return;
+    }
     initcommon(cnf, flags);
-    init(f, stp, cnf, flags, imime);
+    init(fn, stp, cnf, flags, imime);
 }
 
 void FileInterner::init(const string &f, const struct stat *stp, RclConfig *cnf,
                         int flags, const string *imime)
 {
+    if (f.empty()) {
+	LOGERR(("FileInterner::init: empty file name!\n"));
+	return;
+    }
     m_fn = f;
 
     // Compute udi for the input file. This is used by filters which
@@ -226,7 +235,7 @@ void FileInterner::init(const string &f, const struct stat *stp, RclConfig *cnf,
         }
         l_mime = *imime;
     } else {
-        LOGDEB(("FileInterner:: [%s] mime [%s] preview %d\n", 
+        LOGDEB(("FileInterner::init fn [%s] mime [%s] preview %d\n", 
                 f.c_str(), imime?imime->c_str() : "(null)", m_forPreview));
 
         // Run mime type identification in any case (see comment above).
@@ -320,6 +329,7 @@ FileInterner::FileInterner(const string &data, RclConfig *cnf,
                            TempDir& td, int flags, const string& imime)
     : m_tdir(td), m_ok(false), m_missingdatap(0)
 {
+    LOGDEB0(("FileInterner::FileInterner(data)\n"));
     initcommon(cnf, flags);
     init(data, cnf, flags, imime);
 }
@@ -384,7 +394,7 @@ FileInterner::FileInterner(const Rcl::Doc& idoc, RclConfig *cnf,
                            TempDir& td, int flags)
     : m_tdir(td), m_ok(false), m_missingdatap(0)
 {
-    LOGDEB(("FileInterner::FileInterner(idoc)\n"));
+    LOGDEB0(("FileInterner::FileInterner(idoc)\n"));
     initcommon(cnf, flags);
 
     DocFetcher *fetcher = docFetcherMake(idoc);

src/rcldb/searchdata.cpp

@@ -368,7 +368,7 @@ bool SearchData::toNativeQuery(Rcl::Db &db, void *d, int maxexp, int maxcl)
 	xq = xq.empty() ? tq : Xapian::Query(Xapian::Query::OP_AND_NOT, xq, tq);
     }
 
-    // Add the directory filtering clause. This is a phrase of terms
+    // Add the directory filtering clauses. Each is a phrase of terms
     // prefixed with the pathelt prefix XP
     for (vector<DirSpec>::const_iterator dit = m_dirspecs.begin();
 	 dit != m_dirspecs.end(); dit++) {