
Thursday 31-Mar-94 20:29:23

--1600--

Any plans on having CNetToss return valid DOS return-codes for IF/ENDIF
script use?

I'm just guessing.  What would we need/like/use?

RETURN_OK     0 Completed without error and tossed msgs.
RETURN_WARN   5 Completed without error and found NO msgs to toss.
RETURN_ERROR 10 Passed wrong CLI values, options, etc.
RETURN_FAIL  20 Failure, user aborted, can't data, can't get memory,
                unable to continue.  Do safe exit immediately.

--1601--
JS> Bill, you are proceeding based on the assumption that adding local RIP
JS> to CNet would be a "months long" project.  I don't see that.

Yes, that was my assumption.
I hope I am VERY wrong.
I'll admit I've never written any kind of "RIP-viewer C code" in my life.

Who knows.  Maybe parsing the RIP-codes and displaying the stuff is
frightfully simple to achieve.  (But it looks like there are a massive
# of RIP codes to support, and a bazillion different combos to beta-
test.)

I'm also under the assumption that Ken even begun this project yet.
I hope I am VERY wrong.
If he's already started, or already 50% done, or 75% done, please let us know.
I would much rather tell the public that the RIP routines are 90%
done, instead of "0%, hasn't started", or worse "who knows".

I know Ken is very good, and very fast.  But I don't know which are
his priority items.  I would *NOT* put RIP at #1, but maybe #2-3.

Maybe he'll consider releasing a version of CNet with just the 10
basic RIP-codes viewing from local.  Then add 10 more as it
progresses.  Then the rest later.  A 3-4 "step" process.  (Shorter
waiting period for the public, than if nothing is released until RIP
is 100% supported.  And easier to find bugs while beta-testing.)

Keeping in mind that we are only talking about LOCAL viewing of RIP
codes.  (Remote viewing can already be done.)

I don't think it'll really kill any SysOps to run a "partial-local-
RIP" while the "full-local-RIP" is being completed.  You can do quite
a bit with the 10 basic cmds.  And it'll certainly be better than an
"all or nothing" approach.

Am I babbling yet?

--1602--

Use this format...

Bbs Name         cS  Phone Number  City         St  Baud ModemType Nets Line
Amiga Asylum     aXE 508-372-2258  Haverhill    MA 14400 HST/v32b  A/F   [1]
    ^            ^ ^       ^            ^       ^    ^     ^ ^     ^ ^    ^
    |            | |       |            |       |    |     | |     | |    |
 bbs name        | |   bbs number     city    state  |    modem    | |  Line
                 | |                                 |   type(s)   | |  #(s)
                / bbs software (1st 2 letters)       |             | |
               /  * see below for abbreviations   highest         Amiga
              /                                    baud          related
     your bbs is run on ?                        supported      Network(s)
           a=Amiga                                                Only!
           i=IBM

H> So it looks like CNet WORLDWIDE is the preferred AMIGA BBS software.  Now
H> if we can only convert a few more of those DLG guys....

Or just get the 100s of missing CNet systems to add their info into the list.

Be sure and add your BBS onto the master list.

Send in your BBS info to "MS SysOp@1:324/134".
  The AMiGa AsYLuM Bulletin Board System
508) 372-2258 - 24 hrs - 14.4k baud HST/v32
           FidoNet 1:324/134.0
           AmigaNet 40:700/14.0
        GEnie address: E.DUKESHIRE
   InterNet: E.DUKESHIRE@genie.geis.com

Tell everyone, everywhere to do the same.

--1603--

RS> No private reply to the author of a message in a public base (in Cnet,
RS> you have to look for the address in the message and type it manually)

Huh?  Are you running v3.05c?
In a non-networked base, just allow private-msgs too, and answer YES when the
systems asks "Private msg?"

Or in networked bases, just type "MS!".

--1604--

What would be the likelihood of getting the IBM world to allow BOTH "
1" and "01".  That sounds like a pretty petty reason to instantly toss
*thousands* of msgs into the trashcan.  No IBM systems have wondered
"Why is this happening?  Let's fix it."  Or at least "Let's tell
somebody."

Are there other systems that toss out "01" and only allow " 1"?
If so, how do we win?

DA> This is a bug that I gave to Ken about 3 weeks ago, it needed fixed by the
DA> 1st of April and it still is not.

DA> Ken has Toss and X/I exporting the message with the date like follows:
DA> For the first of april it will be   1.
DA> The arugmentably proper way is     01.

DA> This means that if you have wondered why activity in the CNet fido base
DA> kinda slows down from the first to the ninth, your guessed it, this is the
DA> reason.

Wow.  Dave, this is the first that I've ever heard of this in my life.
Have you posted this before today?
Has this been occurring since CNetToss was 1st written? (30-Jul-93)
Or since Xfido/Ifido was written?  (BEFORE 12-Feb-92)

100s of people would have helped push for this change if we all just
knew about it.

What are your opinions on 2 other points:

1) Using times like " 8:47a" instead of "08:47a".
   (CNetToss uses the former.)

2) Using dates like "03-May-2" instead of 03-May-02"
   (CNetToss uses the former for the year 2002.)

Let's get all 1-3 of these fixed i-m-m-e-d-i-a-t-e-l-y.

For CNetToss, all Ken needs to do is find *1* occurrence of
"%s %2d-%s-%d %2d:%02d%c" and change it into "%s %02d-%s-%02d %02d:%02d%c".

It could be done in 1 minute, and re-release TODAY, if Ken wants to.

--1605--

Question or bug report format...

I've never seen any specific suggestions from Ken/Jim (or anyone)
regarding what format should be used when reporting a bug, or asking a
question, about a problem you are having.

These are just my own suggestions...

Please mention as much info as is pertinent regarding:
   1) Your hardware, CPU, accelerator, serial-board, modem, etc.
   2) Exactly what you are trying to do.
   3) Exactly what commands, or methods, you are using to do it.
   4) Exactly what happens when you attempt it.
      (Cutting and posting the exact results will help.)
   5) Your CNet version #.
   6) When posting MCI codes, please change all ^Qs into the "{" character.
      And all ^Ys into the "\" character.  (So that we can read them as-is.)
(If #1 and #5 rarely change for you, you might want to put that info
right into your signature-file.)

You will probably *NOT* have too many people offer helpful, clear
solutions (or maybe no replies at all) if you:
   A) Use ALL UPPERCASE LETTERS.
   B) Omit punctuation characters in your sentences.
   C) Leave out one of more of #1-#5 shown above.

(These opinions are my own.)

--1606--


IW> or text saying "Unable to locate handle" maybe it could say..
IW> "Is Fred Bloggs correct y/n?
IW> if they select no...go back and let them try again, if yes then activate

Only if "NEW" still remains in effect also.
That'll make both groups happy...
1) Those the think "NEW" is how you start a new acct, and
2) Those that don't even know about "NEW"
(I couldn't find any mention of "keeping or removing" #1, in your post.)

I'd also like the user to see "B" also:
A) "Is Fred Bloggs correct y/n?
B) Would you like to open a new acct [Yes/No]:

#B is needed in case you accidently entered the wrong handle.
(Some folks use different handles on different BBSs.)

No point in forcing someone into the new-user procedure, if they don't want
to.  #B would prevent that.

Keep in mind your suggestion will remove a very handy feature of CNet...
The User-List At Log-in.
Do we want that removed entirely?
Currently folks with long-names can find their accts just by entering the 1st
few letters and then picking from the user-list.
Do we want that feature removed entirely?
Better be sure.

--1607--


TB> it some, and then got "That command is not available from this prompt" or
TB> some-such error message.

I wish we COULD use FIND from the response/pass prompt, like we could before.
I believe this feature was removed to avoid re-re-re-re-cursive calls to FIND.
But at least 1 (recursive) call would be handy.
(The others could then give the "can't do" prompt.)
This COULD allow FIND usage at response/pass prompt again.
I'd use that A LOT.

--1608--


HM>    Tell me how you think Ken is going to be able to get CNet to change the
HM> stack for an external program if he doesn't execute that program from a
HM> script!?  He can't.  The ONLY way to do it is affect the priority and
HM> stack of the CLI or shell process that the pfile/program is spawned
HM> from.

I thought that's how Cnet already does it.  No?
I thought that's what all those "shell1" files are:
> prompt ~#`#
> date >NIL:
> Exe:SysInfo
> endcli >NIL:

Ken would just need to read the MCI/BBSMENU/VDE field entry and add the
c:stack/c:taskPri cmds into those shell1 files also.
> Stack 50000
> ChangeTaskPri -2
> prompt ~#`#
> date >NIL:
> Exe:SysInfo -p
> endcli >NIL:

I have no idea what the 'prompt ~#`#' and 'date >NIL:' cmds are for.
Anyone know?

--1609--
MM> When I typed in an alias, the system checked it against the userlist..
MM> It gives you the option to 'use the alias anyways' but is there any way
MM> to disable this in the VDE or something?  Alias's certainly wont be in
MM> the userlist..

I believe this feature is in place to warn you about things like:
"Mark, you are about to use the name "Bill Beogelein" as an alias.
But that name is already being used by some one else."

As with 99% of the prompts in CNet, you can:
1) Change the default to YES
2) Change the default to NO
3) Remove the prompt entirely, making it assume YES
4) Remove the prompt entirely, making it assume NO
5) Change the text into anything that you like

Edit the line "use the alias anyways" in BBSTEXT.
Change the MCI code found there into:
1) \?1
2) \?0
3) \=1
4) \=0
respectively.

Let us know how that works out for you.

--1610--
KP> Reply to this post ONLY with CONCISE requests (with EXAMPLES) of features
KP> that CNet TOSS NEEDS.

What do you think of the idea of allowing netmail to go into a sub
with a unique-dir name of "NETMAIL" or "MATRIX"?

For sorting, record-keeping, auto-purging, searching, signatures,
carbon-copies, threading purposes, checking for illegal netmail moving
through your system, MCI-control, the works.

Of course, sysops wouldn't HAVE TO create that NETMAIL sub if they
didn't want to.  (Netmail would then just behave like it does now.)

CNet would also need to know that all msgs posted in this NETMAIL sub
would be treated as outbound-netmail would.

I get a great deal of netmail and it currently pours into my mailbox (only).
I have to kill great #s and it's gone forever.
I respond and would love to keep a copy for my records.  (By easy
means like, "a copy is saved in the NETMAIL sub".)

--1611--

In the Config/Fidonet/Areas window...
Please consider sliding the 4 string-gadgets down so that "Description" can
run the full width of the window.
(And displaying 40-60 chars instead of just the first 8.)

--1612--

You can file-requ "AmigaRIP" from 1:2410/207 and (finally) get an
Amiga RIP term prg...

AmigaRIPc.lha             147211 ----rwed 10-Apr-94 23:59:59
: Alpha test version of the AmigaRIP comm program by Francisco Moraes 21Jan94

The very first such program for the Amiga that I've ever seen.

--1613--

KP> Reply to this post ONLY with CONCISE requests (with EXAMPLES) of features
KP> that CNet TOSS NEEDS.

-v-v-v- CUT HERE -v-v-v-

A AMIGA_SYSOP              Amiga SysOps' Discussion
  AMIGASALE                Amiga stuff for sale
F CNET                     Fido CNet national backbone echo
  COMPUSALE                Computer Trading Post
U NET2410_SYSOP            Fido region 2410 Sysop News
  VIDEO                    Video Toaster Echo

Key: A=Active F=Feed U=Unauthorized

-^-^-^- CUT HERE -^-^-^-

More than just the current "+" marker in areafix reports.
Can anyone think of any additional handy "AFU" markers?

--1614--

I assume CNetToss reconstructs Fidonet msg threads based on SUBJECT text.
I'd love to see a CNetToss option like:
> CNetToss LEN=10

Only the 1st x chars of the subject-lines would need to be matched.

We could toss together common subjects into 1 item instead of many.
>  # Posted Rsp Title
> == ====== === =====
>  1 13-Nov   1 CD-ROMs
>  2 11-Nov   0 CD-ROM
>  3 11-Nov   0 CD-ROM (Re:)
>  4 10-Nov   2 CD-ROMs and CNet/3
>  5 10-Nov   2 CD-ROMs and CNet/3 (Re:)
>  6 10-Nov   3 CD-ROM & Filelists
>  7 10-Nov   1 CD-ROM and CNET
>  8 10-Nov   2 CD-ROM Players
>  9 10-Nov   0 CD-ROMs made easy
                ^^^^^^

(And it might even speed-up the tosser slightly because it would
only have to compare the 1st x chars, at most.)

Terrible idea?

I would NOT suggest setting LEN to 1 or 2.
But if the 1st 6-10 chars are exactly the same, wouldn't it be common
for the topics/discussions to be similar?

Fewer items, but more similar-subjects grouped together.
More items use more ram, response do not.

You could, of course, always leave "LEN=x" off and full-subject matching
would work as it does now.

Or is this something that should be settable on a per-sub basis?
Or a per-item basis?
Or is there a way for me to manually tack-together all the responses
for 2 items into 1 big item?

--1615--

I finally figured out where these stray "_buser3" files are coming from.
(Or at least the ones I'm getting in my "Mail:" dir.)

I edited CNetToss and changed the first occurrence of the string
"Mail:users/%s/_buser3" into "Mail:users/%s/_BuSeR3".
                                             ^ ^ ^

The stray files then became named "_BuSeR3", too.
It's definitely a bug in CNetToss.

--1616--

DA> You are linked to the following message area(s):
DA> NET3619
DA> PASCO4SALE
DA>              (cut down in size for bandwith conservation)

DA> You are linked to the following file area(s):
DA> NODEDIFF
DA> FIDONEWS

DA> The following message areas are available to you:
DA> *NET3619
DA> *PASCO4SALE

DA> The following file areas are available to you:
DA>  BACKBONE
DA> *NODEDIFF
DA> *FIDONEWS

DA> You are NOT linked to the following message area(s):
DA>  VIRUSSUP
DA>  BCST_MODERATOR
DA>  BCST_WANT_ADS

DA> You are NOT linked to the following file area(s):
DA> CLINKLST
DA> CLINK
DA> BACKBONE
DA> Again, size cut down for file size.

David, would we like to see many separate long lists like that, or
just 1 list and more marker-flags (not just "*") that could tells us
linked/not-linked/available/msg/file/etc?  (L, N, A, M, F whatever)

I'd rather have the latter.
Any preferences?  Pros/cons?

--1617--

CNetToss output text:

> Importing to Fido Amiga: CDROM/CD: 14 item(s); 15 response(s)
> Importing to Fido Amiga: SAS/Latt: 0 item(s); 1 response(s)
> Importing to Fido Amiga: Operatin: 11 item(s); 18 response(s)
> Importing to Fido Amiga: Public D: 1 item(s); 8 response(s)
> Importing to Fido Amiga: Programm: 3 item(s); 3 response(s)
> Importing to Fido Amiga: For Sale: 7 item(s); 12 response(s)
> Importing to Fido Amiga: Amiga Vi: 111 item(s); 14 response(s)

Shorten the phrase "Importing to" and "Exporting to" to just "IN " and "OUT".
Lengthen the diplay of sub-name.
Right justify # of items and # of responses to "3" columns wide.
Drop the "(s)", shorten "response(s)" wording.
Show a grand total at the end.
Number each sub.  Counting downward, based on the over # of subs, minus
the current sub number.  (To show if areafix is nearing completion.)

The result would be:
> 234 IN  Fido Amiga: CDROM/CD ROMs        :  14 item;  15 resp
> 201 IN  Fido Amiga: SAS/Lattice C        :   0 item;   1 resp
> 134 IN  Fido Amiga: Operating Systems    :  11 item;  18 resp
>  78 IN  Fido Amiga: Public Domain        :   1 item;   8 resp
>  34 IN  Fido Amiga: Programming          :   3 item;   3 resp
>  30 IN  Fido Amiga: For Sale (MI)        :   7 item;  12 resp
>  22 IN  Fido Amiga: Amiga Video          : 111 item;  14 resp
>   7 IN, 0 OUT overall                      157 items; 44 resp TOTAL

--1618--

Please post your config/archiver/test strings.

My own are:
  c:lha   >%s -m -n t "%s"
  c:unZip >%s -t "%s"
  c:pkax  >%s -t "%s"
  c:lha   >%s -m -n t "%s"
etc.

Someone please post the DEFAULT strings, too.

--1619--

MS> I need to know the format of how the persons name is passed back to a Cli
MS> door as an argument. If you pass your variable, how is it stored in that
MS> particular variable? ie, argv[2] would be used to store the persons name,
MS> yet their name is two or more words separated by a space so argv[2] would

%1  handle
%3  real name
%39 UUCP id
%39 name of the user's personal directory on the hardddrive
%40 user's acct #
%41 user's serial #
(There are literally 100s and 100s more.)

I don't know of any "first_last" values.
(Your door can check for spaces and convert to "_" itself.)

If you want a user's first-name passed into argv[1], and last-name
passed into argv[2], CNet sysops would add something like this to
their menus:
> Device:path/PrgName %3

If you want BOTH first/last name into argv[1]:
> Device:path/PrgName "%3"            (Note the double-quotes)

(The door itself can be in Door: or Doors: or pFiles: or ram: or
anywhere you like.  CNet doesn't force any limitations like that.)

I received a copy of your door and STILL haven't had time to play with it.
Are you sure you want *ME* to test this thing?
I'm not much of a "game person" at all.
And I can be horribly, horribly thorough, unpolite, blunt, picky,
fussy, and nobody goes over things with a finer tooth comb than mine.

With all due respect, I think it is best that I "pass" on this one.

--1620--

MM> Got a question on CNETTOSS. I noticed in the log that I had a few bad get
MM> tossed but I can't find them on the drive. Can someone tell me were they
MM> go?

Create a regular sub with a unique-dir of BAD_MSGS.
The msgs will be in that sub.
(And where that sub is told to put them on the HD.)

See pages 248 and 87 for more info.

--1621--

A> I bought it from Public Prefection 1-800-716-5000 for $977.50 with postage
A> & handling + COD charge (Pioneer DRM-602X for $959.00, S&H $14.00, COD
A> Charge $4.50)

65,000,000,000 bytes ....

Who's going to be the first CNet system to get one of those 100 CD-ROM
devices made by Sony?  65 gigabytes online at once!  At just a few
pennies per megabyte, it's actually far, far cheaper than any CD-ROM
or hard-drive imaginable.

(But try and tell that kind of logic to my wife.)

--1622--
CS> Any time I try to edit the account (EA) of ANY user on my system and go to
CS> the "preferences/term" the port I am on crashes with #80000005.  This

Did it EVER work OK before?

Make sure all your "systext:vde/*" files are the ones that came with
your current release of CNet.

(All your "pfiles:bbs/*" files, too.)

--1623--

KP> Ken
KP> --- CNet/3
KP>  * Origin: Future World! (Home of CNet!) 313-255-2466 (911:7000/0)

A blank-line right before the "--- CNet/3" would be handy.
(And pretty common among Fidonet systems.)

--1624--

DM> You call this BBS for a while, never have caused problems, then all of
DM> the sudden your get deleted twice without a single message telling you
DM> why, you leave many, many feedbacks to the sysop and never get replies or
DM> an answer as to why you were deleted, you only get two calls a day and you
DM> can't reply to your own messages?

Could any of this be #1 instead of #2:
1) Accidentally deletions, BBS set-up wrong, poorly.
2) The SysOp is deliberately deleting your acct.

Different SysOps run their BBSs differently.
(Thank God.)

1st/2nd/3rd warnings.  Zero warnings.  Zero 2nd chances.  A million 2nd
chances.

No local callers.  No LD callers.
Nothing illegal.  Everything illegal.

SysOps with tons of time to answer 100s of emails.
SysOps with little time.

SysOps deleting callers just because the drop-carriers.

Strict, strict time limits, file/bytes ratios.
Very, very liberal limits.

Overall, I'd say a pretty even mixture.
Luckily, you could always call 1 of 50,000 other BBSs, instead of the ones you
don't like, or don't feel are fair.

--1625--

TB> I don't know what all is going on with this "new" mail system thats
TB> apparently _still_ under development, but it'd be nice if killed
TB> messages were simply flagged with a * or something when you Scan,

"*" will confuse things with "*" mean "new responses" in the bases.
"K" would be a good choice though.

--1626--

MC> I am a DLG owner and have been thinking of moving over to CNET.

Many people are.  It's your choice.

MC> I need to know if there is a way to be able to read messages with
MC> CNET where I don't have to read all of those PESKY original messages
MC> before reading the replies.
MC>
MC> DLG has the option of reading in THREADED mode.
MC> Does CNET offer this?

MC> I read many echos on a daily basis and it takes up alot of time
MC> re-reading old messages.

Which BBS software has you doing that?  I need to know.  Thanks.

I get asked a lot of questions about a lot of Amiga BBS packages.
Instead of answering them 1-by-1, I'd rather just post a list of many (but not
all) of CNet's possible msg read methods.  (And hopefully hit many questions
that many people are asking about CNet.)

Read all msgs online.
Read just entirely new msgs threads.  (Not responses to old threads.)
Read all msgs by a certain user.
Read all msgs by you.
Read msgs that have been marked as "favorite".
Read msgs attached to files that are "free" DLs.
Read msgs threaded or unthreaded.
Read just new msgs since your last call.
Read 1 old, previous response (in each thread) in addition to all the new
     ones.  (As a "lead-in" msg to remind you what the thread is about.)
Jump back/forward one (or #n) thread(s).
Jump back/forward #n msgs in the the current thread.
Jump directly to msg #n in the current thread.
Read only private msgs.
Read msgs posted since a certain date, (and even hr/min/sec).
Read msgs posted after a certain date, (and even hr/min/sec).
Read msgs posted to a certain user.
Read msgs posted to you.
Read msgs that contain a certain string.  (with multiply "AND" and filtering)
Read msgs whose subject line contains a certain string.  (with wildcarding)
Browse a list of threads by subject-titles, skip/read the ones you want.
Read now, mark/search, download now/later, pack, or QWK, etc.
(If you see a certain thread that interests you, download the whole thread with
one cmd.)

You can specify many of the other options to work on a 'per area', or 'per
SIG', or 'all area' basis.  (Or mark which areas are/aren't of interest to you.
Or read new-msgs without even entering ANY msg areas at all.)

Do the above in 'Forward' or 'Backward' or 'Alphabetical' 'Files 1st,
chronological' or 'Files 1st, rev. chrono.' or 'Files 1st, alphabetical', etc.
(Those can be set (and remembered) by CNet.  And different ones for each area.)

When you use combinations of many of the above, you can produce literally
millions of different msg reading methods.  (You can 'alias' long cmds into
short 1-word (or 1-letter) cmds, so that you won't have any long/tricky cmds to
re-type, or remember.)

Needless to say, there's not much "missing" from CNet's msg reading powers.

Get the free CNet demo, try it out.  It costs you nothing to try.

--1627--

S> The 289 page
S> CNET manual is very simple, very easy, well indexed, answered 97.2% of my
S> questions

I found the index to be VERY, VERY, VERY short for a manual of that size.

The replacement index file lists 100s and 100s of items not mentioned
at all in the original index.

File-requ' "INDEX" from 1:2410/207...

CNetIndex12.LHA             8785 ----rwed 14-Feb-94 14:30:18
: CNet/3's Extended Manual Index v1.2, 650 topics listed

--1628--

JR> older messages.  However, C-Net toss doesn't get rid of old items when the
JR> maximum number of items for that area is reached.. They just get LOST...

JR> Ken needs to add a feature that will replace the new FIDO incoming
JR> messages by deleting the OLDEST post in that sub, if of course the maximum
JR> number of items is reached...

I don't know how much that would slow down the tosser.
(Get 1 msg, find 1 old msg to delete, delete it, insert the new msg.
100s and 100s of times over.)  I'm just guessing, but if is slows down
the tosser as much as I would think, I am NOT in favor of that idea.

I would much rather see this done at amaint.
(Since amaint is ALREADY there, looking at each sub, killing posts, etc.)

Let new msgs come in, exceed the max limit if need be, later amaint
deletes x old posts until the max limit is (once again) maintained.

I would love to keep some of my subs at EXACTLY 500 items at all times.
Not drop down to 400, 300, 200, 100 during normal daily/weekly/
seasonly hi-activity VS lo-activity times.  (As CNet does now.)

500 would mean *500*.

So which are you in favor of:
1) Do it at amaint?
2) Do it as the tosser runs, msg by msg?
3) Not support it at all?

--1629--
JS> ANOTHER quick follow up to the theory:  I did some experimenting with
JS> SAS C v6.51 last night and this morning, and I found that the largest
JS> number I could pack into a LONG integer was 2,147,483,647, regardless of
JS> whether it was SIGNED or UNSIGNED.  I could still be mistaken about that
JS> since I was using the printf() function and my results could be due to a
JS> formatting limitation.

Are you using print()'s %ld %d %ul (whatever all of them are).

I have NOT personally tried this, but here are the actually quotes
directly from SAS themselves....

-v-v-v- CUT HERE -v-v-v-

#define LONG_MAX 2147483647       /* max value for long int          */
#define LONG_MIN (-2147483647-1)  /* min value for long int          */
#define ULONG_MAX 4294967295      /* max value for unsigned long int */

-^-^-^- CUT HERE -^-^-^-

Those limits DO appear correct to me.
(They'd be causing some pretty big, big problems, in many parts of
everyone's code, if they were wrong even "just a little off".)

Jim, here's what you tell Ken, and tell'em Bill "said so"...
Remove EVERY single occurrance of int/short/long/char/etc from
"cnet.h" and EVERY CNet line of code.  (I know, easier said than done.)

Be explicit.  (See exec/types.h)  Use only:
GLOBAL
IMPORT
STATIC
REGISTER
VOID
APTR_TYPEDEF
*APTR
LONG
UNLONG
LONGBITS
WORD
UWORD
WORDBITS
BYTE
UBYTE
BYTEBITS
RPTR
*STRPTR
FLOAT
DOUBLE
BOOL
TEXT
BYTEMASK

In fact, technically, you shouldn't even be using these in any new code:
SHORT
USHORT
COUNT
UCOUNT
CPTR

You will NEVER have to worry about CNet being right/wrong, for the
rest of your life.

--1630--

Ken, any comments on the 32 responses regarding the CNet-Tosser thus far?

I'd also like to mention:
 5 08-May   0 Resp Info
 6 08-May   0 YES, I'M A CHICKEN
 7 21-May   0 `K`}xEoploQ|tDk=I졻Kn`k/>B`
 8 20-May   0 Scoliosis surgery
 9 20-May   0 Progeria

Item #7 is NOT the fault of CNet.
Sometimes msg subjects get scrambled.
Would it be possible to toss them to BAD?
(Perhaps by detective more than x% of the chars as outside the range
of normal upper/lower case letters?)

This does NOT happen very often, so if it's going to slow the tosser
down greatly, then it is NOT worth doing.  Just an idea.

--1631--

GS> Ok I am running cnet 3.05c and am hoping someone can tell me how to run
GS> rip on my bbs. Is there an patch or something?

Pick 1 of your graphic sets for RIP use.
(Maybe the "SkyPix" one, if you aren't using it.)

> 0) ASCII                                1) C=64 GFX
> 2) IBM GFX                              3) Amiga Int'l
> 4) Amiga RIP
>
> Choose your graphics set:

Construct all your "tt4" menu-files with RIP codes in them:
> SysText:Menus/tt4.Main
> udbase0:main/data/tt4.entry
> SysText:Menus/tt4.Bases
> Pfiles:data/tt4.entry
> Gfiles:data/tt4.entry
> SysText:Menus/tt4.Bases
> SysText:tt4.access0
> etc....

CNet will show the RIP graphics to the users.
(You won't be able to view them on a CNet screen (yet), but that's coming soon.)

--1632--

TP> The subject line must reflect the FULL PATH to the file you wish to
TP> attach. Works fine here for me...

I've been trying for ages to convince Ken that....

If you are attaching a file, that prompt should change to "Enter filename:".

Why can't msgs have BOTH independant SUBJECT and FILENAME lines?
(They are NOT the same things.)

--1633--
PB> I want to be able to have my users enter any message subboard and enter
PB> the command FAQ and read/download a Frequently Asked Questions list for
PB> that specific area.  This is more importqant on networked areas like
PB> Usenet and Fidonet, however, certain local areas would use it also.

I just added this to my BBSMENU file for msg/file subs:

> FAQ~Browse 'FAQ'
> RULes~Browse 'Rule'

(Of course, you could make it SCAN, or READ, or YANK, or whatever.)

--1634--

SB> Maybe a list should be kept handy of all undocumented commands, and
SB> options for future addedums to the manual. I think I start making one,
SB> then everyone can start adding to it

You might want to take a look at my file called CNetTypos.lha.

Lists many typos and things that forgot to get mentioned in the manual.

(File-Requ "CNetTypos" from 1:2410/207)

(It's from a msg thread I started on Future World.  I want to thank the folks
that helped list things.  Please keep them coming.)

--1636--

Earlier I listed a few options that I feel the tosser should have.

Here's a few more:

Give CNetToss independant "TOSS" and "BUNDLE" options.

If I have 3megs of msgs in InBound: and just 2000 BYTES of msgs in
OutBound:, I might only want to process the Outbound. right now.
(Due to time, speed, diskspace, CPU load, etc.)

Or I might only want to process the Inbound, waiting to bundle my
outbound stuff until after I take a peek at it.

Can anyone think of any other cmd-line options that would be handy for
CNetToss to have?

KLUDGES : Do this 1 toss with kludge-lines showing.
          (Overriding the current CNet/config/tosser setting.)

With CNetToss being a CLI-based prg, often called via DOS-scripts,
often run without human intervention, often via timed-events, or
action-tiggered events (AFTERSESSION), cmd-line options come in VERY,
VERY, handy.  (Much more so than manual-tweeking of CNet/config/tosser
settings.)

--1637--

TB> maintained in parallel with the RAM-based one (which pretty much defeats
TB> the purpose of the RAM-based one while they're online..)

As well as the wasted repeated disk-writing each time EVERY items is
added/removed from the RAM-list.

Here's what I would prefer:

1) When you log-in, your old disk select-list is loaded in to RAM.
   (The file is NOT deleted.)
2) You are free to add/remove items from the RAM-list while online.
3) When you log-off the list is saved to disk.
   (Regardless of normal-logoff, dropped-carrier, amaint log-off, idle-out,
    time-limit exceeded, whatever.)
4) If, and only if, the machine GURUs before #3, then your old
   list (#1) remains fully
   intact.  (But of course, if won't/can't contain any changes made to it
   while online.)

Far, far better than the "lose it all" method, or your "repeated disk-writing
each time" method.

--1638--

Can anyone verify this one....
(I only tried it once.)

I kill a bunch (100+) of unwanted msgs in a big sub (2000+).

I attempt a RR of that sub, to RAM:, but RAM: fills-up, and the
rebuild is safely aborted 1/2 way through.

Many of the msgs that I had earlier killed, are now unkilled for some
reason.

(I even checked the logs, and those were definitely the killed msgs,
coming back to life.)

Ken, would there be a reason that might happen?
(If so, could it be fixed?)

As always, thanks for listening.

--1639--

L> Damn! Isnt anyone out there with SmartOne modems?? Or are you people
L> ignoring me... Please..again... I need the prefix for SmartOne 14.4 to use
L> for CNET!

I wish I could help, but I don't have a SmartOne.  Does that have a standard
Hayes compatible cmd set?

Sure wish more SysOps (including FW) would set-up a Gfile dir like mine.
Any caller could see the modem setup without even bothering the SysOp or
asking anyone.

## Description
== =========================================
 1 USR HST-DS 14400 ROM Checksum      (ATI1)
 2 USR HST-DS 14400 RAM Test          (ATI2)
 3 USR HST-DS 14400 Current Call Time (ATI3)
 4 USR HST-DS 14400 Current Settings  (ATI4)
 5 USR HST-DS 14400 Non-Volatile RAM  (ATI5)
 6 USR HST-DS 14400 Link Diagnostics  (ATI6)
 7 USR HST-DS 14400 Product Config    (ATI7)
 8 USR HST-DS 14400 Product Code      (ATI0)

--1640--

DE> I backup the entire SYSDATA directory EVER NIGHT to the HD that CNet runs
DE> from and another actual SEPERATE Hard Drive that I use just for a UDBASE

I used to do that.  But today's bad/corrupt files just overwrite yesterday's
good files.  You lose your backups.  A pointless method.

Now I nightly copy important files to separate dirs called Mon, Tue, Wed, etc.
As long as I find out about a bad/corrupt file within 7 days, I have a backup.
(In fact, many backups.)

--1641--

Set-up the following event:
Cmd     : DOS-CMD Date >>t:Test
Time    : 141         (Start running at 1:41am)
Valid   : 1
Interval: 1           (Re-run once every minute)
Iterate : 1440        (Run 1440 times per day, which is the # of mins in 1 day)
Days    : ALL

"Date" is execute twice each minute BEFORE 1:41am.

But once (as it should) if it is AFTER 1:41am.

Here's what t:Test looks like if I start it at 1:37am:
> Wednesday 22-Jun-94 01:37:01 |
> Wednesday 22-Jun-94 01:37:01 |
> Wednesday 22-Jun-94 01:38:01 |
> Wednesday 22-Jun-94 01:38:01 |
> Wednesday 22-Jun-94 01:39:01 |- twice each
> Wednesday 22-Jun-94 01:39:01 |
> Wednesday 22-Jun-94 01:40:02 |
> Wednesday 22-Jun-94 01:40:02 |_
> Wednesday 22-Jun-94 01:41:02 <- once
> Wednesday 22-Jun-94 01:42:02
> Wednesday 22-Jun-94 01:43:02
> Wednesday 22-Jun-94 01:44:02
> Wednesday 22-Jun-94 01:45:02
> Wednesday 22-Jun-94 01:46:02
> Wednesday 22-Jun-94 01:47:03

I'm sure the cause/solution is staring me right in the face, but I can't see it.

--1642--

JS> What started me thinking in that direction is the idea that a sysop
JS> could have MANY reasons for wanting to limit the length of any call to a
JS> particular port to 5 or 10 or 30 minutes REGARDLESS of the access level of
JS> the caller.

I like that idea, but am torn between your #1, and my #2 ideas:
1) Effect all callers equally, even paying members are limited to
   the same 5 (or whatever) mins.
2) Don't base the port-time-limit on an absolute x mins, but
   rather x PERCENT of the user's actual time.
   If you set it for 50%, 30 min users get 15.  60 min users get 30, etc.
   I just feel that is "fairer" some how.
   (But defeats #1, of course.)

JS> Um... just off the top of my head, this option could have a few optional
JS> arguments, such as a switch to turn RELOGON on and off...

How about a new marker that could be used in BBSMENU to mark each item
similar to the current ' usage:

3; Main prompt
   MEssages, Base, MSGs
   Gfile, Text          ^2
   Joinlink, CB, PArty
   List, BBSlist        ^3
   News, BULletins
   Pfiles, Doors        ^8
   RELogon              ^7
   Off, BYe, GOOdbye, LOGOFF
The corresponding cmds could NOT be used if that port's event had
those matching flags set to ON.  (In the above example, RELOGON could
NOT be used if port-flag #7 as on during that time period.)

JS> the caller.  In my case, I was thinking that it would be nice to limit the
JS> usage of the Fido line for normal BBS callers.

Would anybody like to see CNet track a new variable (on a per port basis)?
> # of Fidonet calls handled

Since it already tracks the # of overall calls, we would end up with
all three tallies:
A) # of overall calls
B) # of human callers
C) # of Fidonet callers

Jim, any wild, wild guesses as to how many "type B vs type C" calls
your Fidonet line gets?  (It would help to determine how necessary
your port-event-limit idea is.)

--1643--

GS> While I was TDY in Saudi my wife accidentally set the date on the computer
GS> to the year 2004 and it deleted a bunch of accounts that had inactivity
GS> days set. Now CNET just keeps pressing on with the next logical number
GS> instead of filling the holes where the other accounts used to be.  Is this
GS> normal?  I thought it would drop back and fill them in but it doesnt.  Is
GS> there a solution?

It should being "filling the holes".
Try running "Pfiles:Maintenance/pointer" to rebuild your accts-file.
I run mine as an auto-event once per month, just in case.
Set it and forget it.

Also, I tried to get Ken to put a safety-net into CNet...
> If an acct (or file, or msg) is about to be purged (that should have been
> purged >365 days ago) then do NOT purge it.
(The "365" would be settable via 1 global variable in config.)

It would be VERY unlikely that you had forgotten to run amaint for 365
consecutive days.  More likely that "my wife accidentally set the date
to the year 2004".

None of your accts would have been "accidentally deleted".

--1644--

JS> Bill: your argument about having passwords hold up the New Scan and
JS> other stuff is VERY VALID, but it presupposes that the sysop is going to
JS> OVER-USE the password feature!

Let's say 100% of the SysOps read the "over-use warning" and obey it. (they
won't).
0% of them OVER-USE it.  (Hardly.)

Let's say just 2 subs are passworded.
But 15 of the 50 different BBSs that you call, do it.
Do you want to remember all 30 passwords?
Do you want to type all 30 of them in on EVERY call to those BBSs?
(Call them 20 times.  That's 600 passwords keyed in.)
You'll have to remember more than just the password.  You'll have to remember
the BBS name or number, the correct sub name, and the password for that sub.

And that's just to get the VERY commonly-used New-Scan cmd to work through to
completion.

JS> Think about what might happen on Future World if one of the BETA TESTERS
JS> happens to use a lame password like "JOHN5", and then use "JOHN1",
JS> "JOHN2", etc. all over the rest of the BBS community.

That's foolish on that user's part.  I wish we could forcibly stop foolishness
in this world, but NOT if the "solution" causes more trouble than the problem
itself.  (And even penalize all the folks that NEVER use lame passwording
methods like that.)

Jim, would you be in favor of passworded-subs only effecting the actual D/Ling
of the files there?  (New-scan would sail through uninterrupted.)
But I'm still not looking forward to my shoebox full of notes like:
   Number          Sub name             Password
313-255-2466     CNet Beta Files       CANOPENER
313-255-2466     CNet Beta Msg         WALNUT
313-555-1234     Foo Files             DUCKSOUP
313-555-1234     More SysOp Files      DIAMOND
202-555-9393     Beta Test Arexx       BUCKBOARD
304-555-3892     Alpa Support Group    SHOELACE

--1645--
S> I think it would be neat to additional arguments to toss, to process
S> different types of things when you wanted to.  Say, you wanted to
S> aftersessions to call toss into processing netmail only and not the

I mentioned a few options to Ken back in Jan 1994...
(I have no idea if he's going to add them.)

-v-v-v- CUT HERE -v-v-v-

Suggestion #1430...

I would really like to see CNetToss have independant "TOSS" and
"BUNDLE" options.

If I have 2megs of msgs in InBound: and just 2000 BYTES of msgs in
OutBound:, I might only want to process the Outbound. right now.
(Due to time, speed, diskspace, CPU load, etc.)

Or I might only want to process the Inbound, waiting to bundle my
outbound stuff until after I take a peek at it.

-^-^-^- CUT HERE -^-^-^-

Does anybody have any more options they'd find useful in CNetToss?
Post them here, address your msg to "Ken Pletzer".

S> I just tossed a 9 meg packet.  On my 28mhz 2000 it took 7 hours!!!!
S> Is this normal or can I do something to tweak a little speed outta toss?

Approximatly how many msgs is that?
Does anyone have any estimated "msgs per minute" averages?
I seem to be getting about 40 per minute.

I've tried to get Ken to add a line like this to the end of CNet-Toss's output (and log):
> Tossed 519 msgs, in 13.1 mins, 39 msgs/min avg

So a 7 hr toss (on my A3000/25mhz) should cover about 16,380 inbound msgs.
(How many did you have?)

Have you run RR_CHECK?  It shows which of your subs are most badly
fragmented.  (And even has an option to rebuild just those.)  I
rebuild 2 (and only 2) of my most fragmented subs each night,
automatically, as an event.

File Requ RR_CHECK from 1:2410/207.

Have you rebuilt/reorg your HD recently?

Have you run DiskSpeed and seen what speeds your HD is reading/writing at?

Did you format it as a FFS (Fast File System) device?

--1646--

EF> Have you assigned 'Spell' in your C-Net startup sequence? I can not spell
EF> worth a darn and would be lost with out it.

I can't find any reference in CNet that states I have to "c:assign Spell:" some
place.  Why is that needed?

If your spell-path is set correctly in config/paths, that's all that is needed.

(No point in having extra c:assign's that do absolutely nothing but further
confuse things.)

--1648--

I sure wish other CNet pfile authors would do this also.

Just create a DOS-script and run it as an event each week.

Version  >CNet:My_List pfiles:rexx/CmdAlias.rexx    FULL FILE
Version >>CNet:My_List pfiles:rexx/NoVote.rexx      FULL FILE
Version >>CNet:My_List pfiles:rexx/EditMenu.rexx    FULL FILE
Version >>CNet:My_List pfiles:rexx/CnetDial.rexx    FULL FILE
Version >>CNet:My_List pfiles:rexx/FormLett.rexx    FULL FILE
Version >>CNet:My_List pfiles:rexx/SelfDel.rexx     FULL FILE
etc. etc...

(Of course, you'll have to be using the standard "$VER:" strings in
each of your prgs.)

Then make the resulting "CNet:My_List" file, viewable online (with the "??"
cmd), and D/L'able, and file-requ'able.

--1649--

PL> Internet and holding them for 3 days.  After 5 days, you get a 200 meg
PL> partition full of stuff... Amaint doesn't seem to be purging some of the
PL> groups' messages.  I have seen messages listed since 2 months back that
PL> haven't purged.  It's really odd.  I say we do have a problem with this
PL> routine.

Keep in mind that #1 and #2 are VERY different things:
1) Purge items (sometimes also called "posts").
2) Purge the individual responses made to the items.

If you set-up #1 to "10 days", and responses to that thread occur every 9 days,
nothing will get purged for weeks, months, years, ever.

If your set-up #2 to "10 days" and "Purge old responses" then responses older
than 10 days will be deleted regardless of the volume of new responses that
come in.

I haven't tried any of the above with Internet/Usenet.  (I don't have a feed.)

But CNet v3.05c DOES work as it should with all other subs that I've tried.

--EOF--

Saturday 02-Jul-94 21:15:16

-Bill "Mr. BBS" Beogelein, 810-473-2020, 2-line HST 14.4k USR-DS, 1:2410/207

