Re: Wrong header checking in recent version(s)

[mea@nic.funet.fi]

> Hi,
> 
> ..49p10-s10 seems to break something in header validation.

	Hmm..  Yes, I agree.  Here is cut&paste from xterm for
	a diff to fix that.  My online open source history contains
	2.2.1, which has same buggy header semantics entry.

$ diff -u router/libdb/header.c{~,}
--- router/libdb/header.c~      Tue Feb 10 23:01:52 1998
+++ router/libdb/header.c       Sun Mar 22 14:34:21 1998
@@ -30,12 +30,12 @@
 { "cc",                AddressList,    Recipient,      normal         },
 { "from",              AMailboxList,   Sender,         normal         },
 { "message-id",        MessageID,      nilUserType,    normal         },
-{ "reply-to",          AMailboxList,   Sender,         normal         },
+{ "reply-to",          AddressList,    Sender,         normal         },
 { "resent-bcc",        Addresses,      Recipient,      Resent         },
 { "resent-cc",         AddressList,    Recipient,      Resent         },
 { "resent-from",       AMailboxList,   Sender,         Resent         },
 { "resent-message-id", MessageID,      nilUserType,    Resent         },
-{ "resent-reply-to",   AMailboxList,   Sender,         Resent         },
+{ "resent-reply-to",   AddressList,    Sender,         Resent         },
 { "resent-sender",     Mailbox,        Sender,         Resent         },
 { "resent-to",         AddressList,    Recipient,      Resent         },
 { "sender",            Mailbox,        Sender,         normal         },

...
> I am pretty sure that the "Reply-To:" header syntax is pretty legal here,
> I specially checked this against RFC822.  Also, I think that previous
> versions (up to .49p9?) where quite happy about this format.

	No, 2.2.1 was unhappy too.

> Eugene

/Matti Aarnio