Re: XyWrite and upper memory

["J. R. Fox" jr_fox@xxxxxxxxxx]

Patricia M Godfrey wrote:

> Mike said, "Evidently, at least in earlier versions, XyWrite was making
> use of little nooks and crannies of memory in which one would not
> normally expect to find a running application..." and I think he's right
> on target. Jordan asked, "Does Xy make such direct calls [sc., to memory
> apart from the OS]?"

> What follows is a rather detailed attempt to address

Thanks for the recap on this. I'm sure all of us were doing our respective
computer stuff throughout the period in question -- with varying degrees of
interest or attention to these developments -- though maybe none of us could
chart the history of it out as well as you just did.

> memory above 640 K that DOS couldn't. I have circumstantial evidence that
> XyWrite was among the apps that did that by "sneaking up" and using RAM
> supposedly reserved for hardware or by invoking protected mode, or both.

I'm willing to accept this, on the basis of your circumstantial evidence.
However, the serious crashes I've been experiencing in ECS almost certainly have
little or nothing to do with Xy, much less U2. I found that the VDM (DOS
session) alone, with no app. loading, will do it. Choice of video driver proved
to be a red herring. Most of the other theories I had are turning into dead
ends. More recently, it was looking very much like a kernel defect . . . but
I've now plugged in three different kernels, with no success. So I'm running
short on ideas. Still waiting to hear from some learned OS/2 folks, and I'll
take this to Usenet, if I must. I haven't yet tried out Steve Crutchfield's
idea of turning off the Video BIOS Shadowing, but, frankly, neither Warp 4, nor
WSEB (two other, earlier incarnations of OS/2), nor W2K require this, so I can't
see why ECS would.

And this, dear people, is why you Never EVER abandon your old, well-proven,
works perfectly OS, in favor of the "latest & greatest", until it too has
established a sterling track record, *on your system.* Luckily, I have solid
alternatives still in place.

> Third, Microsoft compatibility issues: Has anyone ever tried installing
> XyWrite 4.017 for DOS on Microsoft DOS 6.22? It aborts and crashes with a
> memory error message. This again sounds like differing and incompatible
> attempts to run in protected mode.

I ran it on 6.22 for several years (probably on a '386, then on a P-Pro 200, as
I tended to skip a couple of cpu generations each system change), with no
problem. What's to install ? Just dump a bunch of files in a directory and
PATH it, plus a batch file. I do recall there being some kind of Xy expanded
mem. setting -- even back in 3.5x -- in order to make use of that LIM thingie.

> Fourth, other issues: The display problems with files with lots of
> footnotes that have recently been discussed seem like further evidence
> that XyW is using High RAM, protected mode, or both.

Maybe yes, and maybe no. I'm sure you've seen that old "Xy display jitterbug"
when navigating or trying to refresh a more elaborate Column Tables document ?
I think that's a bug, and a 32-bit Xy could just as easily have such a problem .
. . except that it would probably be a Win-creature, so it would have _other_
problems.

> or RAM to Expanded or Extended memory, and Xy ran better with it all
> Extended.

Did Xy have an Extended mem. setting too ?

>     I suppose only the people who actually coded XyWrite really know the
> answer to this

Yes, we need to have Tim Russert grill Dave Erickson on these issues for an
hour.


Jordan