From Newsgroup: alt.bbs.pcboard

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT
2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated
though ...


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,

and how to give up timeslices. To detect Windows, they use INT 2F/1600,

but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all

NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT

2F/1680 is called. The call itself tells you whether it's supported by

the return value in the AL register; 0=supported, 80=not. It's a better

way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,

when using async COM port support. The fossil code is more complicated

though ...

even in dosbox?
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sat, 18 Oct 2014 07:41:42 -0700 (PDT), corey blake
<ainewslv@gmail.com> wrote:

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated
though ...

even in dosbox?

I'm interested in

a) Desqview
b) Windows 3.1
c) Windows XP
d) Windows 7


because they have networking support I can use with a fossil driver I
wrote. I don't even know what networking support dosbox has. Does it
have any?

You don't need dosbox to compile pcboard on Windows XP. The dos prompt provided by XP works. Works on Windows 7 too (32 bit only, not 64 bit).
Not sure about Windows 8.



--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,

and how to give up timeslices. To detect Windows, they use INT 2F/1600,

but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all

NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT

2F/1680 is called. The call itself tells you whether it's supported by

the return value in the AL register; 0=supported, 80=not. It's a better

way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,

when using async COM port support. The fossil code is more complicated

though ...

true.
my old computer died and I ended up with 64bit win 8.1
dosbox 074 is suppost to have networking now.
dosbox.org. I think.
where is giveup.c? I looked in the library and main source trees.
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sat, 18 Oct 2014 10:08:53 -0700 (PDT), corey blake
<ainewslv@gmail.com> wrote:

my old computer died and I ended up with 64bit win 8.1

Only one computer? What happens to your work when a hard drive dies?

where is giveup.c? I looked in the library and main source trees.

lib\source\screen


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sat, 18 Oct 2014 10:08:53 -0700 (PDT), corey blake
<ainewslv@gmail.com> wrote:

I ended up with 64bit win 8.1

I don't know, but I've heard that some people ...

Download VirtualBox 4.3.18, install it, then find a torrent of Windows
2000 and install it on VirtualBox. Windows 2000 includes a DOS prompt
that runs 16-bit apps. The network speed of Windows 2000 running inside VirtualBox, on a 64-bit host with VT-x/AMD-V support, is impressive.

--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,

and how to give up timeslices. To detect Windows, they use INT 2F/1600,

but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all

NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT

2F/1680 is called. The call itself tells you whether it's supported by

the return value in the AL register; 0=supported, 80=not. It's a better

way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,

when using async COM port support. The fossil code is more complicated

though ...

I have a dvdrw disc system.
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sat, 18 Oct 2014 12:22:56 -0700 (PDT), corey blake
<ainewslv@gmail.com> wrote:

I have a dvdrw disc system.

OK but what difference does that make


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,

and how to give up timeslices. To detect Windows, they use INT 2F/1600,

but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all

NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT

2F/1680 is called. The call itself tells you whether it's supported by

the return value in the AL register; 0=supported, 80=not. It's a better

way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,

when using async COM port support. The fossil code is more complicated

though ...

err, to backup my files and stuff.
You asked what I do in a hd crash.
dig out last disc and restore.
err, how do I make a VM disc for vbox?
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sat, 18 Oct 2014 14:08:22 -0700 (PDT), stbenedictmission@gmail.com
wrote:

err, to backup my files and stuff.
You asked what I do in a hd crash.
dig out last disc and restore.

OK, I thought we were onto another topic ... it's hard to be sure when
you don't quote the text you're replying to.

I recommend a dedicated news client like Forte Agent. That's what I
use, with a free news account at Eternal September. Much better than
google groups.

err, how do I make a VM disc for vbox?

I'm not sure what you're asking. But after you install vbox, you start
by creating a virtual machine and disk. It walks you thru the steps.

When you want to install an OS, you download an ISO of the install disk
(search for torrents, you can find almost anything). Vbox can mount an
ISO file on its virtual CD/DVD drive. No need to burn a CD or DVD and
use the physical drive on your computer. Though you can if you want to.


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,

and how to give up timeslices. To detect Windows, they use INT 2F/1600,

but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all

NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT

2F/1680 is called. The call itself tells you whether it's supported by

the return value in the AL register; 0=supported, 80=not. It's a better

way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,

when using async COM port support. The fossil code is more complicated

though ...

I am just using the web site for the google groups.
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sat, 18 Oct 2014 04:13:20 +0000, Trifle Menot
<triflemenot@protocol.invalid> wrote:

it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support.

I fixed giveup.c to release timeslices on XP, and cleaned up the code
segment self modification mess too.

With the fix, PCBoard no longer hogs the CPU while idling at the call
waiting screen. Unfortunately, the COM driver of NT/2000/XP does not
work well with DOS apps; I could not reliably run the COM port faster
than 19200. And at that slow speed, file transfers still hog the CPU.

By now, it's clear that the only good way to run multinode PCBoard is
the old fashioned way: one node per DOS computer, to handle high speed
COM port I/O, and a network server (i.e, Novell) to store common files.

CDC claimed that PCBoard could multitask, but that was just marketing
talk. Their design and coding was too ad hoc for multitasking to be a practical reality with more than 2 nodes.

That's why, back in the day, big users of PCBoard ran one node per DOS computer, and a Novell server to store common files.

Have fun with that.

--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT 2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated though ...

no thanks. I have had netware for a decade.
still have a 50 user netware 6.0 now.
wont run on modern hardware.
but the net client will. but the server has to be on some old dos machine.

--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Fri, 24 Oct 2014 07:48:23 -0700 (PDT), stbenedictmission@gmail.com
wrote:

no thanks. I have had netware for a decade.
still have a 50 user netware 6.0 now.
wont run on modern hardware.

New is not always improved. I like old hardware.

but the net client will. but the server has to be on some old dos machine.

Microsoft's Netware clone FPNW (file and print services for Netware, a
free download) will run on Windows 2000 server, or 2003 server. I have
it running on Windows 2000 server, seems to work fine as a Netware 3.x
server.

But if you're going to throw out your Netware 50 user license, I'll take
it. I have old hardware.


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Fri, 24 Oct 2014 07:48:23 -0700 (PDT), stbenedictmission@gmail.com
wrote:

I still have a 50 user netware 6.0

I downloaded 6.5 (NW65SP8_OVL_DVD.iso) from:

http://download.novell.com/Download?buildid=n6SkQz52x7s~

It installs with unlmited license, so you can test it out. Of course
real corporate customers who actually use it, must pay for licenses to
be legal.

wont run on modern hardware.

6.5 works here. Maybe it will work on that pawn shop box you got.



--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT 2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated though ...

that's unlimited server licenses.
then you have to buy user seat licenses.
everyone that connects to the server needs a user license.
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sun, 9 Nov 2014 06:51:55 -0800 (PST), corey blake
<ainewslv@gmail.com> wrote:

that's unlimited server licenses.
then you have to buy user seat licenses.
everyone that connects to the server needs a user license.

No connection limit is enforced by the server. Novell "audits" your
license paperwork.

But Netware development is dead; they have said there will never be a
Netware 7. Does Novell have the manpower to audit licenses for Netware
6.5? How about the hacked Netware 4.11, with unlimited user connections available on the net? From Russia, with love.


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT 2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated though ...

no I mean it's hard coded.
after it sees so many logins, like if you had 25 people online, and a 26th logs in, it wont let them.
hmm, if netware is done, I better download all the fixes for my version.
I see they bought suse Linux and that is their new system
poor Suse, I knew her well.
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT 2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated though ...

well, I checked and you are right.
they no longer care. the system does install with a limited license,
but they also say a full MLA unlimited license is on the sp8 discs.
they also say it's now on trust with the licenses.
they say they hope you will buy a paper license for each license you use. but a full non expiring license is in the license dir on the discs


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Sun, 9 Nov 2014 16:52:24 -0800 (PST), corey blake
<ainewslv@gmail.com> wrote:

they also say it's now on trust with the licenses.
they say they hope you will buy a paper license for each license you use.
but a full non expiring license is in the license dir on the discs

6.5 looks good, once I learned how to export the self signed Certificate Authority from Netware, and import it into Firefox. That eliminates the annoying "untrusted connection" message at login. The web management is
nice.

Then I tested some clients. The DOS client works. But the Windows 3.x
client fails to login. Maybe there's a security configuration option to
fix that, but I don't know what it might be.

The Windows 3.x client does work with 4.11. I got 4.11 installed on an
80gb hard drive (large for the DOS era) by updating the install CD with
a newer version of ideata.ham.

I hear there are 100 and 250 user licenses for 4.11, if you can ride the
mule to find them.


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT 2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated though ...

I liked 5 better. and I think I had a 1000 user license somewhere once.
--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Tue, 21 Oct 2014 18:50:39 +0000, Trifle Menot
<triflemenot@protocol.invalid> wrote:

By now, it's clear that the only good way to run multinode PCBoard is
the old fashioned way: one node per DOS computer, to handle high speed
COM port I/O, and a network server (i.e, Novell) to store common files.

Novell NetWare 5.1
Server Version 5.00 revision L, January 19, 2005
NetWare Management Portal Version 1.10 revision E, December 7, 2004
Server Up Time: 0:07:25:55

I put pcbsrc on a Netware 5.1 server, and compiled pcbsrc\lib, to see
how well Netware performs with file I/O. It ran almost as fast as when
the files are on my local hard drive.

I ran the same test with linux/Samba serving the files; Samba was much
slower.

So the old stories about Netware file server speed, are true.


--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Wed, 12 Nov 2014 18:27:41 +0000, Trifle Menot
<triflemenot@protocol.invalid> wrote:

I put pcbsrc on a Netware 5.1 server, and compiled pcbsrc\lib, to see
how well Netware performs with file I/O. It ran almost as fast as when
the files are on my local hard drive.

I ran the same test with linux/Samba serving the files; Samba was much >slower.

I also tested FPNW, Microsoft's clone of Netware 3.x:

Netware 5.1 15 seconds
Microsoft FPNW 106 seconds

Both tested from the same client machine, to ensure fairness of the test

Netware, quick; Microsoft, big fail. But the pretty GUI is mesmerizing!

--- Synchronet 3.17a-Linux NewsLink 1.108

From Newsgroup: alt.bbs.pcboard

On Friday, October 17, 2014 9:13:21 PM UTC-7, Trifle Menot wrote:

The source file giveup.c determines what multitasker you are running on,
and how to give up timeslices. To detect Windows, they use INT 2F/1600,
but that only works for Windows 3.1/9X. It fails with 2000/XP/etc, all
NT based versions.

All versions of Windows release the timeslice (of a DOS task) when INT 2F/1680 is called. The call itself tells you whether it's supported by
the return value in the AL register; 0=supported, 80=not. It's a better
way to detect Windows, than the failing code in giveup.c.

So it looks fairly easy to make PCBoard more CPU friendly on Windows XP,
when using async COM port support. The fossil code is more complicated though ...

only weird thing with novell was.
It replaced the computer os.
but still needed dos to boot.
then they bought a dos from someone and used that for the last few versions.

now they bought SUSE Linux. thank god.
May Univel rest in hell.
p.s. that was their first try at a unix system. god awful.
--- Synchronet 3.17a-Linux NewsLink 1.108