home *** CD-ROM | disk | FTP | other *** search
-
- PRMBS ROSERVER v1.53 04/12/91
-
- CONFIG Changes - version/date
-
- FIXES one - only!
- If you set FORWARDKILL ON and forwarded a flood pointer it was
- killing the text file of the base message. That is fixed.
-
- PRMBS ROSERVER v1.52 04/01/91
-
- CONFIG Changes - version/date
-
- FIXES:
- Diskspace re-comp:
-
- Any DOS command ('!') will now return a 'diskfull' flag forcing
- diskspace recomputation.
-
- NULL Pointer Assignment & Good Ole "ulockf:" probs:
-
- Complicated to explain, but in the myriad of changes that went into the
- seemingly subtle mods to the flood bull pointer generation that caused all
- the problems fixed in 1.51 for "Mail For:" was yet another subtle problem
- which never showed up the same way twice but occasionally would show up
- when the long oustanding string corruption problem " ulockf: unable to
- open" would blow down to DOS it would be accompanied by a message "Null
- pointer assignment". The NULL Pointer problem is fixed, but ..... the
- "ulockf:" problem is, alas, still lurking in the mists!
-
- I have included here a copy of the simple program BOOT.COM. I have modified
- my PBBS.BAT to include an ERRORLEVL test for value 70 and GOTO a label
- which executes the BOOT.COM to re-boot the computer via software every
- morning at about 0530 local time. In that manner the occurrence of the
- "ulock" problem has been reduced to nil on my operational computer. I am
- still looking at the problem and its driving me nuts. This is a near sure
- fire workaround.
-
- Sysop High Message Read:
-
- There was a problem that has come and gone for 18 months. The basic
- symptom was that the SYSOP's user record 'high message read' value would
- get set back to zero periodcially - the problem was essentially eradicated
- in ALMOST all circumstances. The one remaining circumstance was when you
- did an "X" in an EVENT file. I looked the problem over and it was far from
- easy to solve without some restructuring whihc I was faint to do. Instead I
- added to the current 'states' of the BBS of LOCAL (super-sysop), SYSOP
- (remote sysop), and REMOTE (normal user) the additional 'state' EVENT which
- is the state of being online waiting for connect and time comes up to
- execute the event file. In doing it in this fashion the problme was solved
- and I was able to set up a mask to prevent commands whihc expect keyboard
- or RF port response from being executed in the EVENT file. The 'cmds'
- command was modified so that now in addition to the 'cmds -s' (lists
- commands available to remote SYSOP) and 'cmds -u' (commands available to
- the USER) you now have 'cmds -e' which will list commands available to the
- EVENT processor.
-
- I will admit I excluded quite a few commands, that I at least could see
- no reason for needing in the EVENT processor, but maybe you can. Soooooo
- once you bring this version up please get to BBS command line and type
- CMDS -E and then look the list over and let me know if anything needs to
- be 'put back' - you can do a simple CMDS to see what amounts to being the
- full list.
-
- Message #65535 and Re-Numbering:
-
- The problem of what to do when the message count rolls over 65535
- was still not fixed. I found several gotchas even after it appeared to
- be fixed in 1.48. This time I believe (I hope!) I have found all of the
- gotcha's (where is Harry Warren when you really need him??!!??). I made
- a dozen or so different MAIL.DATs with numbers in the 65000 range and
- tried various message forwards, imports, console creation, user port
- creations etc and it all seems to work cleanly now - when it rolls from
- 65535 to 0, the zero is detected and the next message is bumped to 1.
-
- This revision is a MUST HAVE for anyone getting near 65535 message number
- unless you are planning on doing a massive message export, initializing
- a new message file at 1, and massive import.
-
- There will be some problems with simple L and E and RN commands
- during the period when messages exist across the span 65535 and 1. The
- system will still run but list commands will either give 'not founds'
- or the world ... MAILWORK has been heavily overhauled and now supports
- renumbering the mail file. I tested it out very carefully in many
- circumstances and different combinations of Bulletins and un-forwarded
- pointers etc. IT works just fine ..... I got so cocky about it I
- renumbered the 'master file' of interesting bulletins and hate mail I
- keep on my Compaq 386/25 - 1834 messages - it took a while even on the 12ms
- 330 Meg HD on the Compaq - but it came through without a hitch.
-
- I warn those of you with 300 plus messages in your files and still
- running on an XT that the renumber will take some time. In order to
- handle flood pointers correctly, as I read each message, I look at at
- all the records following it for a possible flood pointer ... soooo ....
- if you have 300 messages, the first record is read and the next 299. Then
- the second record is read and the next 298, the third and the next 297 ...
- crude, I know but it works, and as you get closer to the end it has
- less and less to read and really starts to fly. On an XT you may wish
- to use Lightening, the disk cache program with about a 300K cache. this
- would speed the process up by a factor of 5 or 10. I should not have to
- say that this absolutely should not be done in a multi-user environment
- with the mail file active in another window.
-
- After I read the above - I decided to add an additional option to
- MAILWORK Renumber. If you use a '-q' option you get 'quick'
- renumbering, that is, '$' pointer will be ignored and not carried to
- the new MAIL.DAT, so you can use this to purge pointers or if you have
- none and do not want to wait forever for the renumbering.
-
- If you are feeling nervous about this, you may want to protect
- yourself by getting into your BBS directory and type:
-
- pkzip -ex -u -P savemail mail.dat mail\msg?????.mai
-
- ** the '-P' in this particular PKZIP command IS CASE SENSITIVE
- This will archive your whole mail file and text files.
-
- HOLDMAIL:
-
- I fixed a problem in the HOLDMAIL function, which was holding
- everything under some circumstances, regardless of the contents of the
- HOLDTO, HOLDFROM, HOLDAT files. (thanks to WR2E and two other for the
- reports on that)
-
- Shelling to DOS:
-
- I modified the '!' command to use the environmental variable COMSPEC
- instead of a hard coded "command.com" when shelling to a DOS prompt. So
- those of you wanting to use a COMMAND.COM replacement like 4DOS will be
- in luck.
-
- CHANGES:
-
- Speaking of pointers:
-
- ...... I added the option '-$' to select flood pointers for edit,
- list, kill etc. ( '-k$' killed pointers, '-B$' for bulletin pointers,
- '-H$' - held pointers. However '-P$' will not work right, nor is it
- likely you will ever get '-T$', and so on)
-
- Read Count Select:
-
- I added an additional option to the LIST/EDIT/KILL/RN/E/R/V commands.
- That is a 'times read' checker. It is mainly aimed at doing automated
- killing of wasted bulletins. The '-r ###' option basically operates on any
- message whose times read counter is less than the number specified. For
- example:
-
- M 0809 K> 4SALE -ZB -X 20 -R 5 all
- M 0809 K> 4SALE -ZB -X 30 all
-
- These sample EVENT lines are setup to execute early in the morning.The
- first will kill, without querying, all 'for sale' bulletins, over 20 days
- old that have been read 4 or less times. The second will kill any 'for
- sale' bulletin over 30 days old
-
- Doorway to DOS - temporary deactivated:
-
- The 'doorway' to DOS function has been deactivated for a a couple of
- months I have found some problems with it and need to work on it. I
- also need the space it occupied for a major restructuring. I plan to
- re-activate the DOORWAY function in code to be released in June. In its
- place, I have added a COPY command whihc iwll let you move files
- around. It is a full DOS COPY sub-run. You will get the console ouput
- back as well so if you do a COPY *.rs SYS, you will see each filename
- copied, etc. I caution against copying to COM1, COM2, etc that will
- destroy the interrupt IO vectors of BPQ or MBBIOS. Likewise do not copy
- to floppy or printers that are not there - lockup-city! I recommend
- some care here. The password functions should definitely be used.
-
-
- PRMBS ROSERVER v1.51 03/15/91
-
- FIXES:
-
- 'Mail For:' beacon was all screwed up, problem goes back to v1.48. I simply
- didn't allow for some other changes I made which totally screwed this
- process up. A further consequence was fouling up the display of 'you
- have new mail' message.
-
- 'readnews' (rn) command would never find anything, yet another casualty
- of the changes made in v1.48 and propagated in v1.50.
-
- 'putfile' yet one more victim of the above changes was the putfile command.
-
- CHANGES:
-
- In fixing the above problems I saw an opportunity with but a half
- dozen lines of code to add a 'subjects' search. If the command
- 'readnews' or 'rn' is issued with no arguments the system will display
- a list of up to the 64 mosr recently entered TO fields for bulletins.
- This 'list' is compiled each time the system computes the 'Mail For'
- banner, i.e. if it isn't Traffic or Private its To field gets examined
- as a 'subject' - the titles aren't duped. This will make the READNEWS
- command easier to use since it can be used to find out what topics to
- ask for. This command is analogous to the LC (list category) command
- that Hank put in RLI code several months ago.
-
- PRMBS Ver 1.50 03/01/91
-
- PROBLEMS fixed:
-
- - in 1.48 with the change in the manner flood pointers were
- generated, a minor change to the findmsg() function set up a condition
- where under specific circumstances an 'L' command would lock the system.
-
- - for several versions 'km' (kill mine') was blowing everything
- addressed to user away, read or unread - and if that person was sysop,
- all messages addressed to SYSOP as well - it is fixed.
-
- CHANGES:
- There is a SETPARM parameter called HOLDMAIL. If set ON when any
- message is being saved, it will check for files HOLDTO.RS, HOLDFROM.RS,
- and/or HOLDAT.RS if the TO, FROM, or AT fields of the incoming message
- match calls listed in any of these files (like all other PRMBS
- permission type files, they do not have to be present.0if you have no
- files as suchthen setthe parameter OFF and save disk churning. You can
- also setup the files for when you are away and turn the parameter
- OFF/ON based on your general presence - like turn it ON when you are
- away for the weekend and back off when you get back) and save an
- messagethat makes a match as status 'H'. it could be used to filter
- @ALLBBS or @USA (editorial comment here, the concept appalls me, but I
- see the need for the mechanism from a pragmatic standpoint)
-
- When a message is created locally by sysop or a user and the TO field
- contains the letters ALL as the first three characters (ALL, ALLUSA,
- ALLCAN, etc) The system will display the following message and prompt
- the user to re-enter the TO field
-
- *** The TO field of a bulletin should be more specific than $S.
- The following is a list of suggested topic areas that can be used
- ALL 4SALE WANTED HELP MODS CLASS EXAMS CLUB HAMFST MAP
- NETWRK LIST SOFT BEACON LLBBS DX SAT PROP QST SPACE
- AMSAT COMPTR C64 IBM C128 MJF AMIGA AEA PC KANTRN
- CPM PACCOM UNIX COMMDR VIC20 ICOM OSBORN YAESU KENWOD ALINCO
- TANDY RADSHK ATARI
- Please re-enter the TO field now (simply hitting ENTER or CR
- will retain your original TO field):
-
-