diff --git a/doc/index_internal.html b/doc/index_internal.html
index 4195808b..315933eb 100644
--- a/doc/index_internal.html
+++ b/doc/index_internal.html
@@ -44,7 +44,7 @@ <h1>Main page for <a href="http://dar.linux.free.fr/">dar</a>'s documentation</h
 	</ul>
 	<p>
 	  If you have questions that could not find their answers in this document or have suggestions, you are
-	  welcome to drop your questions on the <a href="https://lists.sourceforge.net/lists/listinfo/dar-discussions">dar-discussion</a>
+	  welcome to drop your questions on the <a href="https://sourceforge.net/projects/dar/lists/dar-support">dar-support</a>
 	  mailing-list.
 	</p>
       </div>
diff --git a/src/build/ChangeLog b/src/build/ChangeLog
index 62a83f32..1bbc4994 100644
--- a/src/build/ChangeLog
+++ b/src/build/ChangeLog
@@ -35,6 +35,8 @@ from 2.7.15 to 2.7.16
 - fixed mask building of path exclusion (dar's -P option) when used with
   regular expression (problem met while testing or merging a backup)
 - adding support for progressive report to reparing operation at API level
+- warning before processing to the backup if gnupg signatories are provided
+  without any gnupg recipient.
 
 from 2.7.14 to 2.7.15
 - updating libdar about CURLINFO_CONTENT_LENGTH_DOWNLOAD symbol which is
diff --git a/src/libdar/archive_options.hpp b/src/libdar/archive_options.hpp
index 86f560f2..0cbca9c7 100644
--- a/src/libdar/archive_options.hpp
+++ b/src/libdar/archive_options.hpp
@@ -154,7 +154,7 @@ namespace libdar
 	    /// defines the protocol to use to retrieve slices
 	void set_entrepot(const std::shared_ptr<entrepot> & entr) { if(!entr) throw Erange("archive_options_read::set_entrepot", "null entrepot pointer given in argument"); x_entrepot = entr; };
 
-	    /// whether to warn (true) or ignore (false) signature failure (default is true)
+	    /// whether to warn (false) or ignore (true) signature failure (default is false, signature failure is reported)
 	void set_ignore_signature_check_failure(bool val) { x_ignore_signature_check_failure = val; };
 
 	    /// whether libdar is allowed to create several thread to work possibly faster on multicore CPU (need libthreadar to be effective)
@@ -380,7 +380,7 @@ namespace libdar
 	    /// compression is performed without block in one single stream per file.
 	    /// This is the historical way used by libdar, it gives the best result
 	    /// and the lowest compute overhead, though it cannot be parallelized.
-	    /// At the opposite using compresion per block reduce the compression ratio
+	    /// At the opposite using compression per block reduce the compression ratio
 	    /// but allows the block to be compressed/decompressed in parallel and thus
 	    /// leverage multi-core systems. When the block size increase you tend to the
 	    /// same compression ratio as compression ration without block
diff --git a/src/libdar/macro_tools.cpp b/src/libdar/macro_tools.cpp
index 4f2c2b3c..a3bd2f6a 100644
--- a/src/libdar/macro_tools.cpp
+++ b/src/libdar/macro_tools.cpp
@@ -1303,6 +1303,9 @@ namespace libdar
 			dialog->message(gettext("WARNING: support for secure memory was not available at compilation time, in case of heavy memory load, this may lead the password/passphrase provided to be wrote to disk (swap space) in clear. You have been warned!"));
 		}
 
+		if(! gnupg_signatories.empty() && gnupg_recipients.empty())
+		    throw Erange("macro_tools_create_layers", gettext("Archive signature is only possible with gnupg encryption (and at least one recipient)"));
+
 		if(!gnupg_recipients.empty())
 		{
 #if GPGME_SUPPORT