patch 9.1.1223: wrong translation used for encoding failures

Problem:  wrong translation for encoding failures because of using
          literal "from" and "to" in the resulting error message
          (RestorerZ)
Solution: use separate error messages for errors "from" and "to"
          encoding errors.

fixes: #16898
closes: #16918

Signed-off-by: Christian Brabandt <cb@256bit.org>
diff --git a/src/strings.c b/src/strings.c
index 640ad39..d6f1b38 100644
--- a/src/strings.c
+++ b/src/strings.c
@@ -1339,7 +1339,7 @@
 	    vim_free(str);
 	    if (converted_str == NULL)
 	    {
-		semsg(_(e_str_encoding_failed), "from", from_encoding);
+		semsg(_(e_str_encoding_from_failed), from_encoding);
 		goto done;
 	    }
 	}
@@ -1348,7 +1348,7 @@
 	{
 	    if (!utf_valid_string(converted_str, NULL))
 	    {
-		semsg(_(e_str_encoding_failed), "from", p_enc);
+		semsg(_(e_str_encoding_from_failed), p_enc);
 		vim_free(converted_str);
 		goto done;
 	    }
@@ -1414,7 +1414,7 @@
 	    str = convert_string(str, p_enc, to_encoding);
 	    if (str == NULL)
 	    {
-		semsg(_(e_str_encoding_failed), "to", to_encoding);
+		semsg(_(e_str_encoding_to_failed), to_encoding);
 		goto done;
 	    }
 	}