GU>      Within the last month I finally took the plunge and rejected the
GU> herd mentality by switching to OS/2 Warp 3.0.  Unfortunately I am having
GU> a compatibility problem with LapLink version NU-2.01, it refuses to
GU> communicate with my serial ports!
MH>That's right.  LapLink won't work under OS/2 (unless you have a native
MH>version).  Why?  Because OS/2 doesn't allow programs to
MH>controller the hardware ports...that's precisely how LapLink
MH>does it's business.  I've even had problems with LapLink in
MH>Win95 in a DOS box...it runs for a while then craps out.  WinNT
MH>would have the same limitation.
Well that is not 100 percent true. Did you forget about
COM_DIRECT_ACCESS?  That is what that option is for in the DOS settings.
I can only speak to one version of Laplink, which is Laplink-III version
3.0a and only using the serial port because parallel requires that I
crawl under the desk.  However I have used LL3 version 3.0a under OS2
2.0GASP, 2.1beta, 2.1GA, 2.11, Warp 3 blue beta, Warp 3.0, Warp connect,
Warp Server.  I use it to move files to several laptops at 115K turbo
using the standard out of the box OS2 com driver.  I have never tried it
with SIO.  I just laplinked 40 meg this morning of training slides for a
road trip that I am leaving for in a few hours.
For all I know you may be correct about parallel, I have never tried it
for reasons stated. I certainly would expect it to work through IBMNULL
with share access checked; but it may not because that method does try to
use other than the data lines.  The system that I normally do the serial
pulls from is a very old 486DX-50 with 20Meg and a ISA bus 16550 running
Warp 3 fixpack 22.  The laptop that I used this morning was an equally
old 386SX-20 with 8meg, since I only need it for slides.  With a newer
laptop I would just tie it to the network but laplink still has uses.
The only draw back is that LL3 slows down my serial mouse on the 486 and
pegs pulse but it works error free.  If I navigate the WPS using the
keyboard then other work can continue as normally as one can expect with
a pegged CPU.  I have also tried it from my pentiums a few times.  One
case does not mean it works for all people, but this is one guaranteed
case of it working using one specific version and any version of OS2.
MH>Why does OS/2 not let programs control hardware ports?  Crash
MH>Protection.  If a program has completely control over a
MH>hardware port and it hangs on a call...it's locked the machine
MH>because it'll continue hammering that IRQ and nothing else will
MH>get service.  OS/2 intervenes in this case and tries to
MH>eliminate this problem.
Also not true, EISA, MCA and PCI buses have watch dog timers to prevent
this and any newer 486 or pentium or pentium pro lets OS2 virtualize
interrupts in a DOS box.  If you have a 386, older 486 or maybe one of
the clones that do no support intel virtualization enhancements and just
has an ISA, or VL bus then; if a program disables interrupts and then
fails, that would hang the system.  It can, does and OS2 does not prevent
it unless you disable the direct hardware access features which are on
by default. Maybe you never heard of the three line program that when
run in a DOS box would hang older 386/486 ISA systems. I have only had
LL3 fail once on this old un-enhanced 486 ISA bus machine in all these
years and I haven't a clue why it did. It has SCSI disks and it just as
soon could have been a hung SCSI bus. For anyone with a system made in
the last two years, the possibility of total system lockup approaches
zero.
--Lynn
 * SLMR 2.1a * Specify 78rpm, 8-track, Edison cylinder or wire recording
--- DB 1.39/004485
---------------