the problem is that Gnus 5.[234] insists on inserting its own `date' and
`message-id' headers. I think that this is probably the right thing.
In addition, I insist on sending my mail with mh because the mh alias
mechanism is so superior to the .mailrc/mailx stuff. (I say as someone
that uses the group stuff all the time). So this is a patch to 6.8.4
that will probably apply to nmh straight out of the box (I am only
tweaking the struct's and one line of code). It does two things:
1. allows message-id and date headers to appear in mail.
2. if there is a message-id header alreday and -msgid is specified, no
header is added and otherwise it is.
Here is the patch
Soren
*** 1.1 1997/01/14 07:09:21
--- post.c 1997/01/15 17:37:26
***************
*** 186,191 ****
--- 186,194 ----
#define MVIS 0x0008 /* we've seen sighted addrs */
#define MINV 0x0010 /* we've seen blind addrs */
#define MRPY 0x0020 /* we've seen a Reply-to: */
+ #ifdef CS_UCHICAGO_EDU
+ #define MMID 0x0040 /* we've seen a Message-Id: */
+ #endif
};
/* */
***************
*** 196,208 ****
--- 199,217 ----
"Reply-To", HADR | HNGR, MRPY,
"From", HADR | HNGR, MFRM,
"Sender", HADR | HBAD, 0,
+ #ifndef CS_UCHICAGO_EDU
"Date", HBAD, 0,
+ #endif
"Subject", HSUB, 0,
"To", HADR | HTRY, MVIS,
"cc", HADR | HTRY, MVIS,
"Bcc", HADR | HTRY | HBCC | HNIL, MINV,
"Dcc", HADR | HTRY | HDCC | HNIL, MVIS, /* sorta cc & bcc combined */
+ #ifndef CS_UCHICAGO_EDU
"Message-ID", HBAD, 0,
+ #else
+ "Message-ID", HNOP, MMID,
+ #endif
"Fcc", HFCC, 0,
NULL
***************
*** 1001,1007 ****
--- 1010,1020 ----
break;
fprintf (out, "Date: %s\n", dtime (&tclock));
+ #ifndef CS_UCHICAGO_EDU
if (msgid)
+ #else
+ if (msgid && !(msgflags & MMID))
+ #endif
fprintf (out, "Message-ID: <%d.%ld@%s>\n",
getpid (), tclock, LocalName ());
if (msgflags & MFRM)