HAM Radio 1/HamRadio.cdr/misc/rs150run/rs133upd.rme

Text File | 1990-05-06 | 10.4 KB | 207 lines

  *** NOTE *** always assume CONFIG date/version must be changed unless
  otherwise specified.

PRMBS/ROSERVER v1.33 04/06/90

(1) "~C" command got broken, it is fixed

(2) EXPORT allowed you to export with no routes specified, and it tried to
match nothing to empty routes and would randomly select messages to be
exported. The minimum arguments in command table was corrected - problem gone

(3) ROSEBOOT chains to ROSERVER.PGM. I used the wrong variant of "execl()"
and it demanded ROSERVER.PGM be in same directory - this is a problem to
floppy based systems - it is corrected now and will search PATH.

(4) Minor change in display format for LIST commands.

(5) I am writing a BINARY FILE TRANSFER routine for PRMBS. So the "DB" command
has been changed to "DA" ([D]ownload binary in [A]scii) so "DB" can be used
for the new routine at a later time.

(6) While importing 450K of messages to my system from a floppy, I noticed a
lot of churning with no messages added, it dawned on me that many were bulls
I probably had seen. But, its maddening not knowing - I added a message display
indicating tha a bulletin being imported with BID checking not disabled has
been rejected for BID, so each message read will either generate a RCVD
or a REJECTED message.

(7) If you typed  ROSEBOOT -nm0 CONFIG.xxx - you would get screwy time
delay values - because of stacking several options on one "-" thats fixed.

(8) Old bug - just discovered! (by K2ADJ) if you were in message reply entry
in sweeper or remote user in MAIL reply, and timed out, the system would drop
back to sweeper command entry and require a second timeout to expire before
being dropped. This is fixed.

v1.32 04/28/90

(1) date/time  display in list command changed
    from  0424/1150  to  24-Apr 1150
(2) ROSEBOOT now has a built in delay in executing for multi-user
startup. It will pause 30 seconds times the PID (process ID). If you
have two processes going numbered 0 and 1 - the first will come up im-
mediately and the second will come up 30 seconds later. This eases the
thrashing on the disk. If you have PIDs, 0,1,2,3 The copies will come up
in 0, 30,60,90 second time slots.

v1.31 04/26/90

(1) CONFIG changes
    - remove the 'optional forwarding header trailer' in each port struc-
      ture
    - add, right after the host system specification (i.e. KA2BQE.VT.USA
      or whatever), the nearest network node to your BBS. If you are run-
      ning G8BPQ I suggest you not use your BBSNODE name but you conven-
      tional node name as it propagates farther than the BBSNODE name.
      If you are on ROSENET you would use your rose address (802899).


  My survey of sysops and BBS coders on crunching "R:" headers indicates
  that it is an idea for the future but for now it will not be. In an ef-
  fort to cut down the volume I am removing the 'vanity' opportunity offered
  by the trailer and have come down on a hard coded header

     R:9004251256z [e-mail] Underhill Ctr,VT Z:05490 N:UCV

  This satisfies the requirements of programs written by other coders which
  extract information from these headers as well as the general sysop and
  who gains insight into the network by examining headers. It tells when the
  message went through/was created on the system. It gives full H-route
  information for the system, it gives detailes local information (city,
  state, zip) and network location nearest network node.

  My recent experiences exploring the network shows that for the most part
  if you can get into a node in the state (or sub-region if state has that
  much packet activity) shown in the H-route address you will find the node
  shown in the header on its tables.

  Nothing more is really needed. I urge the sysops to keep their city/state
  entry as terse as possible and use acceptable abbreviations like "Ctr"
  above in place of "Center". or "Floral Pk, NY" for "Floral Park, NY"

(2) Script processor for invalid script token just printed an "INVALID>"
message and went on, leaving you at the mercy of the timeout. I changed
it to dump the script and connect if it encounters an invalid Token. This
way the log will show a 'failed script' and make you look for the problem
and it saves time on the links.

(3) Further consideration of the 'service message' matter prompted me to
the type 'S' message altogther. These messages are not even accepted by
W0RLI (that is if you do an  "SS call @ bbs_path" to a W0RLI v11.xx system
you will get DISConnected.) So if you type "SS call @ bbs_path" it will be
handled like a "S call @ bbs_path" meaning that the TO-CALL will be ex-
amined the message will become P or B depending upon if it is in valid
call format.

(4) I created the command  "PUTFILE path\file call@bbs" this allows the
sysop or remote sysop to send a file to another user on another system.
Without having to go through a complex synthesis of the required inbound
REQFIL message to achieve the same thing.

(5) I cleaned up the LOG line for message creation, the 'from' field is
now in upper case.

(6) The "RCVD" message sent back to the creator generated at the end of
message creation now has a '$' prefixing the BID, if any. If no BID it will
just be a lonesome character. I have gottne a lot of positive feedback about
this because it tells the user immediately if the local system does not know
what to do with the message (i.e. route it) and he can then send a note to
the sysop.

(7) The HELP command has been modified to bring it in line with CMDS and ?.
It now accepts wildcards (* and ?). This may be a problem with users who
are use to typing "H D" and getting help for Downloads. It is no longer
"H ALL" for complete list it is "H *". **8 NOTE *** There must be no blank
spaces on the line after the labels in the help file. Each lable MUST BE
FOLLOWED by a CR-LF.

(8) In earlier versions one could type in as a "call@bbs" address simply
the destination BBS with full H-route info, i.e. WB2MNF.NJ.USA, and the
program would make it  [e-mail]. That feature got 'broke' a
while back and its now fixed.


(9) The route edit string array has been expanded to 32 places. The array
is a series of pointers so if you do not have that many strings you will
use less memory. As the strings are read space is allocated for them.

  In northern Vermont (Burlington) we are adjacent to northern new york.
all of the mail we send out of state and all of the mail we send to new york
goes down to WA2UMX in Saratoga Springs or WA2PVV in Valatie, EXCEPT for
three BBSes in NNY and due east of us. they use the designation #NNY.NY.USA.
Now, of course my edit strings are VT.USA and its relatives. You may however
add to the   #NNY.NY.USA et al. to strip the local BBSes in an adjacent state
and then add their explicit calls to the top of your HROUTES file. That way
all NY stuff goes SOUTH for me, but NNY stuff gets stripped for HROUTE parsing
and goes direct or EAST. NJ BBSes mau have a similar situation with many EPA
BBSes and vice-versa.

(10) I did some playing with windows in the ROSEBOOT program and, while its
'pretty work' it does speed up the shameless advertising thats done there yet
giving the display more time in view. I am using the BOSS Windows pckage and
will be using its data entry abilities in a integrated package to handle
startups and system updates in the future.


v.130 04/23/90

  (1) 'Q' added to port descriptors - indicates a BP type port in
addition to being TNC type. The effect of this descriptor is that whne
a user logs off by using the 'BYE' or 'B' command, after being logged
out instead of being DISConnected he is left CONNECTED to the G8BPQ
switch of that system so he can go on from there if he wishes instead
of having to rebuild the circuit. He will get a message the  reads like
"Returned to Node UCV:KA2BQE-1" typing 'BYE' from there will get the
disconnect, or just the good old 'ctrl-C then D'.

  (2) An additional connect script token is added '$' for G8BPQ
connections from the NODE. Most G8BP sysops prefer to have forwarding
connections made to their node and then have the connecting station
issue a BBS2, or BBS3, or BBS4, etc. the problem being that even when
they have 4 or five copies, the connectee has to rerun his forward
cycle for each script trying to get a connection. The '$' eliminate
this. The line  " $2-4 " is taken to mean, at that point issue the
string "BBS2" and look for connect, if not then "BBS3" and then "BBS4"
if still not, then disconnect.  The first character is the start
number, the second is the end number "$4-9" says start with BBS4 and go
til BBS9. remeber, if you connect to an L2 address you will get some
welcoming stuff which will show up after the CONNECT message and after
the BBSx is sent and they will be construed as 'fail strings', You may
want to issue the CONNECT to the NODE then use the '&x' to 'eat' x
number of lines, the do the '$2-4'. Scripts are exxactly that a
specific script to connect to a specific system, as such they must be
tailored to each system.

  (3) the problem with no routing being done to RMAIL with embedded BID
was not fixed til now;  thought I had and had left out one additional
step to make it right. Its works, and has been checked.


  (4) The 'service reply' message for KT has been removed. It had been kept
only for AFMARS, no other major BBS code is still supporting it. Please remove
the SERVM command in the AUTOEVENT section of the Config file. No problem
of you don't, but since the command no longer exists it generates a HUH.


v1.29 04/21/90

 (1) Do an LU and the first user entry would be displayed 'run-on'
after the column header label line.

 (2) Do a "C {port letter} {CALL} [optional filename]" and when done
manually disconnect (i.e. you do a ctrl-C and type D), if you went to
do a SWAP or FORWARD or POLL before going back online again you would
not try to connct and just start dumping message. You hit F or the
system senses no DCD and it come back up to command line prompt and you
do it again and it goes perfectly OK. (this was a verrryyyy OLLLLLDDD
problem)

 (3) In v1.28, if you did a swap or fwd or poll of another system from
command line then came back up to command line and entered a message
you get no text that says "OK send title" but it took the title and the
etx and maed the message OK.

  73 de brian