finish implementing envelopes. it's not completely clear to me that
this is really the way to go, but it seems clearly better than what
i had before. this delta includes many other minor changes, s

finish implementing envelopes. it's not completely clear to me that
this is really the way to go, but it seems clearly better than what
i had before. this delta includes many other minor changes, so it
should probably not be blithely removed.

show more ...

try to install envelopes. is this really worth it???

add "envelopes" to contain the basic information needed as control
info for each message. currently there is only one envelope -- this
being the obvious stupid conversion. later there will be separ

add "envelopes" to contain the basic information needed as control
info for each message. currently there is only one envelope -- this
being the obvious stupid conversion. later there will be separate
envelopes for error messages, return receipts, etc.

show more ...

16 bit changes -- should have no effect on VAX binaries to
speak of.

collapse special character processing into macro processing for
simplicity of code.

add host aliasing; add -p flag. this version doesn't yet know about
replacing the text of the host alias into the message however. syntax
is grotty: "/hostmatch:%s@newhost" or whatever.

cleanup so it will go through lint without any fancy grep -v's;
change the way SCCS Id's are handled.

Edit queue control files when running queue (assuming there are still
recipients left). Also, modify the MsgPriority to be the absolute
number (not the Priority: value); this allows us to include ag

Edit queue control files when running queue (assuming there are still
recipients left). Also, modify the MsgPriority to be the absolute
number (not the Priority: value); this allows us to include aging into
the priority algorithm.

show more ...

change the send queue to be only one queue instead of one per mailer.
this is slightly inefficient but simpler. also, pass this queue
around so we can have multiple send queues. this makes VRFY wor

change the send queue to be only one queue instead of one per mailer.
this is slightly inefficient but simpler. also, pass this queue
around so we can have multiple send queues. this makes VRFY work.

show more ...

convert to SMTP draft 3 -- finishing touches. Punt on the VRFY/
EXPN commands for now; they aren't required anyhow. Move the fullname
into the address structure so it can be inherited.

links together aliases into proper trees. This is in part an
attempt to do the SMTP VRFY command properly, but there are a number
of annoying problems that convince me that this is the wrong way to

links together aliases into proper trees. This is in part an
attempt to do the SMTP VRFY command properly, but there are a number
of annoying problems that convince me that this is the wrong way to
do it. Sigh.

show more ...

ignore case in mailer name comparison

Install new experimental queueing facility -- one stage timeout,
etc. This version is still quite incomplete. It needs to reorder
the queue after some interval, do two-stage timeout, take option
in

Install new experimental queueing facility -- one stage timeout,
etc. This version is still quite incomplete. It needs to reorder
the queue after some interval, do two-stage timeout, take option
info from the queue file instead of the command line, read the
sender's .mailcf file, etc. Some of this is useful for SMTP also.

show more ...

Eliminate magic MN_LOCAL and MN_PROG; change q_mailer item in ADDRESS
to be pointer to mailer rather than index.

fix botch in backup code during rewriting

change rewriting rules to use $N on RHS to match LHS, and include
$=X (class match) in the matching; this will allow us to match the
proposed "user.host@domain" syntax as well as the old syntax, by
p

change rewriting rules to use $N on RHS to match LHS, and include
$=X (class match) in the matching; this will allow us to match the
proposed "user.host@domain" syntax as well as the old syntax, by
putting the known domains into a class to disambiguate.

show more ...

fix the <> syntax, convert "at" to "@" in from addresses also,
plus some minorness to get the multi-machine case going nicely.

enable aliases to self-reference themselves correctly.

fix problem with using macros in rules; change configuration to send
different flags when destined for arpanet vs. berknet hosts, etc.

check MAXATOM overflow; increase MAXATOM value

match on full name for local users

cleanup, commenting, linting, etc.

allow "error" as a net name to print error messages

remove special "at" processing (put it in .cf file)

change processing of <LWSP> in addresses