2021-04-15 13:31:59 -05:00

892 lines
26 KiB
Plaintext
Raw Blame History

Fido/FidoNet version 10K errata sheet T. Jennings 9 July 85
10K has very few changes over 10I. Mostly one bug<75>
fixed, a fairly obscure one: Fido would allow the entry of,<2C>
but would not transmit, a message destined to <region>/0. It<49>
worked correctly for <net>/0.
One addition, a command line switch:
/Y Disable path display
By default, Fido will display the pathnames of the<68>
various message and file areas, plus the DIR.BBS<42>
descriptions. This supresses all path displays. The path<74>
information is pretty much useless anyways.
Fido/FidoNet version 10I errata Sheet T. Jennings 10 May 85
I apologize for the size of this errata sheet; time<6D>
constraints prevent me from creating a new manual. (The<68>
usual excuse.)
Most of the changes were bug fixes, and wont impact<63>
anything. The biggest change has been in the FidoNet network<72>
system. There is an entire section here about the changes.
If you are running a private-only network, the<68>
changes are no big deal. If you are running in the public<69>
domain FidoNet, you should get a copy of:
NODELIST.EXE
Manages the nodelist for you; it will strip/add area<65>
codes, dialing prefixes, adjust message costs, etc.,<2C>
eliminating all manual editing.
ROUTEGEN.EXE
Generates ROUTE files of any arbitrary complexity<74>
from an english language like script.
These programs can be obtained from Fido #31 in Los<6F>
Angeles.
The quote file wont repeat itself over and over. The<68>
position last used in the quote file is saved in SYSTEM.BBS<42>
(its size increases to 212 bytes) and used every time Fido<64>
is run. The first time it will start at the beginning of the<68>
file.
The Fido log file (SYSOP.LOG) has changed. The same<6D>
information is logged, only its appearance has changed. This<69>
should make the log smaller and easier to search with<74>
software. Extraneous words were removed, and the data<74>
displayed in a more compact form. Each line starts with a<>
specific character. (Most do, anyways.) The characters are:
+ user name and time on
! errors of all sorts
= download and upload info
- line between entries
The /I command line switch now accepts a numeric<69>
argument:
<task>/I Set multi-task environment
This does only two simple things: it changes the<68>
names used for the log files, and affects the way messages<65>
are created.
The names for the logs will incorporate the task ID<49>
number; ie. if you specify 9/I the log files will be:
SYSOP9.LOG
MAILER9.LOG
You cannot use 0 as a number. The other effect of<6F>
setting a task ID is to force Fido to recount the messages<65>
before saving a new one. This hopefully prevents duplicate<74>
message numbers making a mess of things, or at least<73>
minimizes it. It has the unfortunate side effect of greatly<6C>
slowing down saving messages.
If you use /I in test mode, you will not be able to<74>
see anything. Do not use /I with /T.
Four new modem types:
9 Ventel MD212 with WECO option
10 GDC modem
11 DC Hayes Smartmodem 2400
12 US Robotics Courier 2400
The Hayes 2400 is a terrible modem. It will be OK<4F>
for an autoanswer application, but very annoying for general<61>
use.
FidoNet messages have an audit trail: whenever a<>
Fido passes a msg on further (ie. a host) it adds to the<68>
bottom the line:
"Via Fido net/node"
Each Fido a given msg passes through will add its<74>
own audit trail line.
Alternate node ID (net/node) has been added; each<63>
node potentially has two independent net and node number<65>
pairs. The current one will apply like it does now; all<6C>
messages will be marked with the regular net/node numbers.<2E>
The alternate is mainly for hosts, who are a node in a<>
specific net, and node #0 as well. The normal case will be<62>
where both node and net numbers are the same. In the host<73>
case, the numbers will be set like:
Node : nn
Net : mm
Alt Node: 0
Alt Net : mm
This says that this system is node nn in net mm, and<6E>
is also node 0, the host. It can also be used as:
Node : nn
Net : mm
Alt Node: oo
Alt Net : pp
So that the node is in two different nets; mm/nn and<6E>
pp/oo.
File attach capability has been given to EXTRA<52>
privilege users, instead of only SYSOPs. Be careful, it<69>
still allows access to the entire system, including *.*.
No longer will FidoNet call millions of times for<6F>
systems that dont answer. The maximum is now 20 tries per<65>
node per event.
10H now allows 300, 1200 and 2400 for all<6C>
operations; whether your modem can or not is your business.<2E>
This applies to baud rate detect for users as well as<61>
Fidonet incoming and outgoing calls.
Due to a peculiarity in the UARTs used on all serial<61>
ports, at 2400 baud whacking CR will not work to determine<6E>
baud rate; you must hit the space bar instead. Fido will<6C>
accept CR or space at all baud rates, but you must use space<63>
at 2400. It is not a software problem but a hardware one.
The file K)ill command now works. It will prompt for<6F>
a filename, accept wildcards, and will delete the names from<6F>
FILES.BBS.
The Novation SmartCat works now. The problem was<61>
that Novation radically changed the command structure, and<6E>
provided a manual errata sheet; I did not have the errata<74>
sheet. The 300 baud only Hayes should work as well; the<68>
problem was that the modem sends all results with odd or<6F>
even parity, and Fido did not strip it when looking for<6F>
result codes.
For those who care, the packet names for Fidonet<65>
messages and file lsts has changed. The new format is:
mmmmnnnn.OUT
mmmmnnnn.FLO
Where:
mmmm is the Net #, in Hex, zero filled, and
nnnn is the Node #, in hex, also zero filled.
For instance, a packet destined for Node #56 in Net<65>
#78, the packet name is:
004E0038.OUT
004E0038.FLO
On the same subject, a new keyword has been added to<74>
the route list, beyond the changes mentioned below. This is:
EXTERNAL-MAIL
This causes FidoNet to not delete it's packets and<6E>
file lists, and mark all mail packeted as (SENT), or Killed<65>
as necessary. This is specifically for the case where<72>
FidoNet messages are transferred manually through some sort<72>
of gateway, such as a mainframe. This entire process can now<6F>
be automated by:
(1) Creating a schedule to handle the mail that is to be<62>
shipped in an abnormal fashion. Make the schedule only one<6E>
minute long.
(2) Make an External schedule terminate with some<6D>
detectable value, immediately following the special Fidonet<65>
schedule.
(3) In the route list for that schedule, add the keyword<72>
EXTERNAL-MAIL.
This complicated process causes Fidonet to create<74>
the packets and file lists for the specified node(s),<2C>
leaving them on disk, and marking the messages as if they<65>
were sent normally. The external event must be used to<74>
"mail" the packets. When this external mailing is complete,<2C>
Fido can be returned to normal operation, either to Fido<64>
mode or another FidoNet event.
Please note that if you "blow it", and lose the<68>
packets, the process cannot be repeated for the same<6D>
messages, as FidoNet will marked them as sent as if they<65>
were mailed in the usual fashion. Packets should be deleted<65>
when complete. MMMMNNNN.FLO is a string of pathnames; it can<61>
be fed to a comm. program to transfer any attached files as<61>
well. The receiving end will need to correlate the lists of<6F>
files with the message packets.
Changed the message status display from (FORWARDED)<29>
to (IN TRANSIT). Much more meaningful. Only the message is<69>
changed.
FidoNet message entry (picking a node number) has<61>
changed, and there is a new help file; NODES.HLP. This is<69>
displayed by entering ? at the prompt for a net/node number<65>
at the E)nter message prompt in the FidoNet section.
PREVIOUSLY UNDOCUMENTED:
"New" questionaire: QNEWUSER.BBS. Answers go into<74>
ANEWUSER.BBS.
This is in the new-user signon loop. This is for new<65>
or existing users who don't get to log on. (Either a private<74>
system, or they forgot their password.)
There are now two ways to handle this case; the<68>
first is the current way: NOPWD.BBS is displayed, the user<65>
gets to enter a message to the sysop, then is logged off.
Now however, if the questionaire QNEWUSER.BBS<42>
exists, the get to fill out the questionaire instead of<6F>
NOPWD.BBS, and then logged off immediately, instead of a<>
message to the sysop. The questionaire can be only displayed<65>
text, for the case where you do not want any input from new<65>
users.
PREVIOUSLY UNDOCUMENTED:
For all .BBS text files (bulletins, etc) two control<6F>
characters allow disabling and enabling users Control-C and<6E>
Control-K:
^B Prevent Control-C aborts
^C Allow Control-C aborts
If ^B is found, the caller cannot abort the display<61>
of a .BBS text file. ^C reenables it. The default is<69>
enabled.
EXAMPLE:
^B
This cannot be canceled with Control-C
^C
This and all following text can be canceled with<74>
Control-C.
PREVIOUSLY UNDOCUMENTED:
At the prompt for entering a node number, a disk<73>
file can be specified instead of a node number. (This was<61>
SYSOP only in previous versions.) The disk file name must<73>
not start with a digit. The disk file is assumed to contain<69>
a list of node numbers, as would be entered manually. The<68>
limit is 300 characters maximum.
The following structures have been changed. The<68>
route table has been deleted, and its functions moved into<74>
nmap. Most other changes consist of using previously zero<72>
fields.
Starred items were filled in as zeros in previous<75>
versions.
/* Message header structure. The message text is just a long<6E>
string. */
struct _msg {
char from[36]; /* who from, */
char to[36]; /* who to, */
char subj[72]; /* message subject, */
char date[20]; /* creation date, */
int times; /* number of times read, */
int dest; /* destination node, */
int orig; /* originating node */
int cost; /* actual cost of msg */
* int orig_net; /* originating net */
* int dest_net; /* destination net */
int caca[4]; /* extra space, */
unsigned reply;
int attr; /* message type, below */
int up;
};
/* MAIL.SYS file structure */
struct _mail {
int node; /* local node number, */
float fudge;
int rate; /* baud rate */
char mailpath[80]; /* path to find mail in */
char filepath[80]; /* mail file path */
int net; /* net number */
int alt_node;
int alt_net;
};
The net field was added, and the node (.number)<29>
field has a special meaning when the net number is -1.
NET NODE
-1 xx NODELIST.SYS revision
xx 00 Defines a HOST
The first entry in NODELIST.SYS is the revision<6F>
record. Fido 10I checks for a correct revision marker, and<6E>
if it does not match, Fido recompiles NODELIST.SYS.
/* Node descriptor. (NODELIST.SYS) */
struct _node {
int number; /* node number, */
int net; /* net number */
int cost; /* cost per minute to call */
int rate; /* baud rate */
char sched; /* schedule tag */
char name[14]; /* node name */
char phone[40]; /* phone number */
char city[40]; /* city and state, */
};
/* Message packet header. */
#define PKTVER 2
struct _pkthdr {
int orig_node; /* originating Node # */
int dest_node; /* destination node */
int year,month,day,hour,minute,second;
int rate; /* baud rate */
int ver; /* packet version */
* int orig_net; /* originating net number */
* int dest_net; /* destination net number */
int extra[17];
};
.pa
FidoNet network organization change
T. Jennings
10 April 85
This covers the new capabilities of Fido version<6F>
10I. Please do not implement net numbers yet! In order to<74>
use net numbers, we have to get the entire net swapped over<65>
to this new software, and issue net numbers.
If you do not do so already, please make sure you<6F>
read FidoNews, the FidoNet newsletter. All network changes<65>
are being discussed there. Get back issues if necessary.<2E>
This is extremely important if you are part of the public<69>
domain FidoNet.
FidoNews can be downloaded from Fido #375, where it<69>
is published, or from most Fido systems, such as #1 or #51.
10H starts a new type of FidoNet organization. It is<69>
backwards compatible, but we will be swapping over to the<68>
new method as soon as possible. This new organization allows<77>
organizing the net and the node list on a region by region<6F>
basis, instead of the monolithic centrally managed list,<2C>
though there will still be a central list.
In previous versions, each different system was<61>
identified by a single number, the node number. This works<6B>
just fine, and the number of nodes that can be accomodated<65>
is 32,767.
The new way allows a second identifier, the net<65>
number. There can be 32,767 nets, each with 32,767 nodes.<2E>
Not likely we'll use up all the numbers!
One annoyance when trying to enter a Fidonet message<67>
on old Fidos is if you type "?" at the E command prompt, you<6F>
get 250 node listings ... not exactly what most people like<6B>
to see. The prompt now asks:
?=Help, #=List NODES /=List NETS
Entering / will list all the local nets; you will<6C>
probably get no more than 20 listings, a very manageable<6C>
number. Once you select a net, you can then enter # to list<73>
the nodes in that net; once again, a very manageable number,<2C>
the biggest net is SoCal, with 30 something nodes.
It also allows you to enter a specific net/node<64>
combination all at once, like:
55/66
Which means Net #55, Node #66. Handy if you send<6E>
mail to a specific node all the time.
Before everyone gets a copy of 10H, the net will<6C>
operate as it does now. Everyone will belong in Net #1. Busy<73>
areas such as SoCal and STL will eventually become their own<77>
nets; they will be assigned one net number for the entire<72>
region, and after that they can pick node numbers as they<65>
please.
STL will pass out net numbers as it now does node<64>
numbers. Each local net will generate its own node list,<2C>
with routing tables, etc just like Fido 51 in STL does now.<2E>
The format will be axactly the same. STL will then merely<6C>
take those local nodelists, stick them all together, and<6E>
pass it out.
Nodes that dont fall into any local net will remain<69>
part of Net #1, just like it is now. If there is only one of<6F>
you in East Overshoe, you wont be burdened with the extra<72>
work involving a local node list.
NODE LIST CHANGES
The node list has to be changed as well. The change<67>
here is also pretty minimal. There has to be some way to<74>
specify which net a node belongs to.
The current scheme assumes we are in the same net.<2E>
(Well, we ARE, and we will remain so.) An example old<6C>
fashioned node list entry looks like:
....
0050 25 1200 Crystal_Caver1-512-263-5805 Austin.........TX
0051 25 1200 DECUS_Central1-314-432-4129 St_Louis.......MO
0052 25 1200 TOPCC 1-805-499-8378 Thousand_Oaks..CA
0053 25 1200 S.E._Prof_Fid1-404-928-1876 Woodstock......GA
....
All of the nodes are in numeric order as well. The<68>
new scheme involves two changes:
(1) Placing all of a local net's nodes in the
same place in the node list,
(2) Specifying the local nets host at the top of
the list.
IMPORTANT: It is an absolute no-no to "sort" the<68>
node list. This will totally wreck everything. The order of<6F>
the node list is now an important part of its structure. DO<44>
NOT CHANGE THE ORDER OF THE NODE LIST! There were two major<6F>
reasons for doing this, both not needed any more: (1) to put<75>
your own node at the top of the list, so Fido wouldnt be so<73>
damn slow listing message. Repaired: Fido wont list the node<64>
name and location if it is YOUR NODE. (2) sorting by city<74>
and state to make the list manageable. The organization<6F>
itself takes care of this.
Two keywords has been introduced into the node list;<3B>
the word "HOST" and "REGION", these change the meaning of<6F>
the node list entry it is on. Both are handled the same, but<75>
HOST automatically does routing that used to be done in the<68>
route lists. More on this later.
A sample entry for a local net is shown below. Note<74>
that all of the St. Louis nodes follow one another; they can<61>
no longer be sprinkled through the node list.
The special line before the normal listing for node<64>
4 defines the local nets host. The "2" is not the node<64>
number, but the net number. It says that all node entries<65>
that follow it belong to net #2, until another HOST line is<69>
found.
Note also that the host is also a regular node, 51.<2E>
The HOST line merely defines who is the host. This entry, as<61>
you might expect after thinking about it, does affect<63>
routing; this is covered below. A host is also treated as<61>
"node 0", so that it is possible to send a message to "host"<22>
or 51, and the message will get there. This is a special<61>
case, and has a special meaning also covered below.
HOST 2 25 1200 DECUS_Central1-314-432-4129 St_Louis.......MO
0004 25 1200 Bulldog 1-314-441-9297 St_Louis.......MO
0010 25 1200 MDC_RCC 1-314-234-1462 St_Louis.......MO
0016 25 1200 Mikes_Board 1-314-726-3448 St_Louis.......MO
0017 25 1200 DCA_BBS 1-314-962-0395 St_Louis.......MO
0022 25 1200 PCLUG 1-314-576-2743 St_Louis.......MO
0051 25 1200 DECUS_Central1-314-432-4129 St_Louis.......MO
HOW THIS AFFECTS ROUTING
The routeing works as previously documented; the<68>
only real change is that Fido will do some of it<69>
automaticaly now instead of having to be put into the route<74>
list.
HOST basically causes automatic routing to be done<6E>
by Fido, instead of requiring route lists as in previous<75>
versions.
Once we swap over to the new method, you will not<6F>
need to pecify the routing for hosts. When you enter a<>
message to say Fido #4, Fido will automatically route it to<74>
Fido #51 for you.
REGION does not do any of this automatic routing; it<69>
only affects how messages are entered and listed.
Mail within your own net is never affected by this<69>
special routing; if you and another node are both in the<68>
same net, then mail you send there will go directly, just as<61>
in previous versions. (You can of course specify routing<6E>
still in the route list, but you have to do it on purpose.)
ROUTE LISTS
Route lists are the ROUTE.BBS, ROUTE.A, ROUTE.B, etc<74>
files.
You can still specify routing as necessary, like<6B>
before. The only difference is that if the nodes are not in<69>
your own net, you will have to specify the net number as<61>
well as the node number. This is described in detail below.
Fido also understands a special "node number", the<68>
word "ALL". This means "all nodes in all nets". This is<69>
useful for incoming hosts, to allow routing mail from any<6E>
node in the world to their own local nodes. For instance:
send-to 4,16,17,22,10,76
accept-from all
... allows Fido #51 to receive mail from the world<6C>
and send it to it's own local nodes.
There is also a new keyword that Fido understands,<2C>
its use will become obvious later:
Net N
Assume Net #N as the default net for all following<6E>
route list statements. This is only to allow easy entry of<6F>
lots of net/node number combinations. This will be explained<65>
further below.
NET/NODE NUMBERS
Previous versions accepted only node numbers on<6F>
route statements, such as "send-to 11,22,33,44" or "schedule<6C>
B 66,77,88,99", etc. These forms are still accepted, but<75>
Fido assumes that the node numbers belong to your own net.
To specify nodes not in your net, you must use the<68>
"long form" of node/net, such as "send-to 4/11,4/22,4/33" or<6F>
"schedule B 100/66,100/77,100/88", etc.
When you specify the same net number a lot, you can<61>
use the new "NET" keyword. What it does is set the default<6C>
net to NN, instead of assuming your own net number. One of<6F>
the above lines could be done as:
Net 100
schedule b 66,77,88,99
.........................................
DECOUPLED MODE
Fido 10h also has a new mode to accomodate seperate<74>
networks; decoupled mode. Decoupled here means that instead<61>
of knowing all about all nodes in a different (from the one<6E>
you are in) net, it will allow mail to nodes within a<>
different net knowing only the host.
Currently, when you enter a node number, Fido checks<6B>
it againts the list; if not found, Fido complains. This<69>
requires that all nodes in all nets be in the nodelist.
In the future, this may be impractical, especially<6C>
for "private" nets. Decoupled mode basically tells Fido to<74>
not care if it can't find a specific node in a specific net.
Note that this only affects different networks; Fido<64>
still requires knowing all about nodes within the current<6E>
network.
In decoupled mode, when a node number entered cannot<6F>
be found in the list for that net, Fido sends it to node #0,<2C>
the host for that net. The host must be in the nodelist. If<49>
when routing a message that does not appear in the node<64>
tables, it will be sent to the host as well.
Decoupled mode is enabled with /C on the command<6E>
line. Please do not use it for normal FidoNet operation.
--------------------------------------------------------
Fido Version 10C Errata Sheet
T. Jennings 25 Jan 85
These are changes made in version 10c from previous<75>
version 10's. Note that this may conflict with what the<68>
manual says; please note the changes.
MESSAGE AREAS
Finally, yes finally, Fido understands the idea of<6F>
'Last Message Read'. You no longer have to memorize where<72>
you left off reading messages since you were last on.
Fido now remembers the last message you read in the<68>
LAST NINE MESSAGE AREAS, not all of them. This is because<73>
that is all the extra space I could find in the user<65>
records. It should be fine for most Fidos.
When you enter a message area, Fido will tell you:
"Highest msg is HH, last one you read was LL"
This will tell you if there are any messages there<72>
that you haven't read. A little known fact of previous Fidos<6F>
was that "*" meant HIGHEST MESSAGE; when you entered "R *"<22>
it read the highest message.
* now means "The one after the last one you read".<2E>
So, to start reading where you left off, just enter:
R *
And if the last message you read was 50, it will<6C>
start displaying messages at 51. You can enter * anytime<6D>
Fido asks you for a message number.
SYSOPS:
The renumber command (8) will correct the message<67>
numbers now stored in the user list. The old program,<2C>
RENUM.EXE, will not handle the message numbers stored in the<68>
user list; everyone will be confused. Do not use it anymore.
.........................................
MODEMS
Fido is a lot smarter about modems now. Since it<69>
already knows what type of modem you are using (the /J<>
switch) it might as well initialize the modem the way it<69>
wants, instead of forcing you to use the complicated<65>
FIDOMDM.BBS.
Fido will initialize the modem for you<6F>
automatically; AFTER THAT, it will do the initialization<6F>
with FIDOMDM.BBS as specified in the manual. This allows you<6F>
to override Fido's default initialization if yuo want; 99%<25>
of all installations will not need to however. You can now<6F>
(and should) DELETE FIDOMDM.BBS.
The IBM PC Junior has it's own modem type, and<6E>
Novation SmartCat will work properly, even with XMODEM.
NEW MODEM TYPES
These are the modem types specified with the /J<>
switch.
1 Hayes
2 DF03
3 Racal Vadic VA212
4 Novation SmartCat
5 US Robotics
7 Prentice POPCOM
8 IBM PC Junior
NOTE: For PopCom and GDC modems, and all modems that<61>
are a "superset" of the Hayes, any additional initialization<6F>
must be done in FIDOMDM.BBS.
.........................................
FIDONET ROUTING:
These changes afect only those Fidos acting as hosts<74>
for other Fidos. If you are not a host, then ignore this, or<6F>
maybe read it out of curiosity.
Some very substantial changes were made here. If you<6F>
are a host, you will have to make changes; however, the<68>
changes should be a pleasant suprise instead of a pain.
MULTIPLE ROUTE LISTS:
Instead of just the single route list, ROUTE.BBS,<2C>
Fido will support one route list per schedule. It is simple:
Schedule A ROUTE.A
Schedule B ROUTE.B
... ...
Schedule W ROUTE.W
The way it works is: Fido first looks for the<68>
specific route list, if it doesn't find it, it uses<65>
ROUTE.BBS as before.
NEW ROUTING COMMANDS
The contents of the route list have been changed,<2C>
too. One keyword was dropped, and one more added.
FORWARD-TO No longer supported
SEND-TO node node node node ...
Now by default, ALL NODES ARE MARKED "FORWARD-TO".<2E>
The ACCEPT-FROM list is the only thing that controls which<63>
messages go out. It turns out that the FORWARD-TO was<61>
totally redundant in every working route list examined; I<>
won't go into the details here why. Any FORWARD-TOs found<6E>
will be totally ignored.
The big problem was that a host needs to run more<72>
than one schedule, and there was only one route list. This<69>
meant that there had to be conflicting or overlapping<6E>
ACCEPT-FROMs and FORWARD-TOs; with the new multiple route<74>
lists, there are no more conflicts.
Also, since when you are editing a route list, you<6F>
only have to think about the one schedule; you will find<6E>
that they are now very small, and easy to read.
The new keyword, SEND-TO is like SCHEDULE: it means,<2C>
"In this schedule (whatever it is) send mail to these nodes"<22>
followed by a list of nodes. For instance, if it is schedule<6C>
B, and you are to send mail to your local nodes, you would<6C>
say "SEND-TO node node node node ...". It is the same as<61>
saying (in this example) "SCHEDULE B node node node node<64>
...". You can think of it as doing "SCHEDULE (the current<6E>
schedule) node node node ...".
SPECIAL NODE NUMBER:
A special "number" has been added, and can be used<65>
anywhere that you would put a list of node numbers:
ALL all nodes in the list
For instance, if you route all your mail to a single<6C>
outgoing host, instead of putting in all possible numbers<72>
(tedious, when there are 150 nodes), you can now do:
ROUTE-TO xx ALL
This works with SCHEDULE, SEND-TO, ACCEPT-FROM and<6E>
ROUTE-TO.
AN EXAMPLE:
You are both incoming and outgoing host for nodes<65>
101, 102, and 103. You are node 200. (This used to be the<68>
worst thing possible; the way you had to to it before was<61>
ACCEPT-FROM all and FORWARD-TO all; this meant it would send<6E>
anyones mail anywhere!).
SCHEDULES:
B 12:30AM - 12:59 AM local to host
A 1:00 - 2:00 AM national
C 2:01 - 2:30 AM host to local
ROUTE.B:
(empty, host is read only)
ROUTE.A:
ACCEPT-FROM 101 102 103
ROUTE-TO (the usual route-to's)
ROUTE.C:
ACCEPT-FROM ALL
SEND-TO 101 102 103
Nice and simple. Make sure there is no ROUTE.BBS.<2E>
ROUTE.B does nothing in the host; since it's only waiting<6E>
for mail from the locals, it doesnt need (or want!) to send<6E>
to anyone. It is just gathering messages that may be sent<6E>
during Schedule A.
ROUTE.A says it's OK to send out mail only from it's<>
local nodes, plus obviously any mail entered at the host.<2E>
The ROUTE-TOs are the ones out of the node list. Since only<6C>
the locals are listed under ACCEPT-FROM, mail from say node<64>
#1 will not be sent, but marked as ORPHAN.
ROUTE.C says forward mail from any node (ACCEPT-FROM<4F>
...) but send only to the locals 101, 102, 103.
I think I hear some groaning out there, but the<68>
changes will be worth it. It means that you can now<6F>
implement any network of any complexity and number of levels<6C>
with no worries of forwarding mail you do not want to.