[tin 1.7.x] allow "<>" as From:?

below is a patch from Ralf Doebelitz which allows "<>" as
from-address (I have added the code in post.c to disallow empty
Message-IDs). Im uncertain if we should enable the feature as the
gnksa-checker might be used elsewhere in the code where it should
reject an empty address (like in the Message-ID case where I've
fixed this (I did not add such a check to post.c:build_messageid()
as it's very unlikely that a system has no hostname/fqdn and even if
we hit that case the other test will catch this right before
posting)). comments?

--- misc.c	2005-05-04 11:46:33.000000000 +0200
+++ misc.c	2005-05-12 14:55:09.000000000 +0200
@@ -3397,7 +3397,15 @@
 		}
 	}
 
+	/*
+	 * if we allow <> as From: we must disallow <> as Mesage-ID,
+	 * see code in post.c:check_article_to_be_posted()
+	 */
+#if 1
+	if (!strchr(address, '@') && *address) /* check for From: without an @ but allow <> */
+#else
 	if (!strchr(address, '@')) /* check for From: without an @ */
+#endif /* 1 */
 		return GNKSA_ATSIGN_MISSING;
 
 	/* split successful */
--- post.c	2005-05-04 11:46:33.000000000 +0200
+++ post.c	2005-05-12 14:53:31.000000000 +0200
@@ -907,8 +907,17 @@
 		}
 
 		if (cp - line == 10 && !strncasecmp(line, "Message-ID", 10)) {
+#if 1 /* see comment about "<>" in misc.c:gnksa_split_from() */
+			char addr[HEADER_LEN], name[HEADER_LEN];
+			int type;
+
+			i = gnksa_check_from(cp + 1);
+			gnksa_split_from(cp + 1, addr, name, &type);
+			if (((GNKSA_OK != i) && (GNKSA_LOCALPART_MISSING > i)) || !*addr) {
+#else
 			i = gnksa_check_from(cp + 1);
 			if ((GNKSA_OK != i) && (GNKSA_LOCALPART_MISSING > i)) {
+#endif /* 1 */
 				setup_check_article_screen(&init);
 				StartInverse();
 				my_fprintf(stderr, _(txt_error_bad_msgidfqdn), i);