DOS ain't dead - New 15-Oct-2009 XMGR and UIDE Available.

Jack

Fresno, California USA,
16.10.2009, 14:10
(edited by Jack, 16.10.2009, 19:13)

New 15-Oct-2009 XMGR and UIDE Available. (Announce)

With regret I have had to update the XMGR and UIDE drivers again, due to a
new "Gates & Co." problem.

Japheth noted to me in private E-Mail that protected-mode move "calls" (in
place of a normal Int 15h), DO NOT work with Win95/98, and he deleted such
calls from his HIMEMX driver.

I could simply have changed my drivers' README to note this and so require
the XMGR/UIDE /K switch with Win95/98. But I felt /K was not as valuable
as /H, which I retained because it lets the HMA save a lot of upper-memory
for UIDE/UIDEJR users. /K gains only a small bit of protected-mode speed
and is "NOT worth the risk!" of Win95/98 CRASHES! /K is thus DELETED and
my drivers again do only a normal Int 15h command in protected-mode moves.

As always, the 15-Oct-2009 DRIVERS.ZIP file is posted on Johnson's site:

http://johnson.tmfc.net/dos/driver.html

I am again DISGUSTED at having to delete /K due to uncorrected Gates & Co.
ERRORS, but this was necessary so XMGR and UIDE remain SAFE with Win95/98!

---
(Account disabled on user's request.)

Rugxulo

Usono,
19.10.2009, 10:39

@ Jack

New 15-Oct-2009 XMGR and UIDE Available.

Post reply

> Japheth noted to me in private E-Mail that protected-mode move "calls"
> (in
> place of a normal Int 15h), DO NOT work with Win95/98, and he deleted
> such
> calls from his HIMEMX driver.

Which version had them? Is this for pre-existing versions or future? (Still never officially got 3.33 out the door. Oh well.)

Anyways, if the speedup isn't much, then it's probably not worth it, but you could always do a runtime check for FreeDOS (OEM id: FD) or DR-DOS (int 21h, 4452h aka "DR") and only enable it then. I mean, better than disabling it completely, or am I wrong? Either way is fine, though, I doubt I'd even notice the difference. ;-)

Jack

Fresno, California USA,
19.10.2009, 19:05

@ Rugxulo

New 15-Oct-2009 XMGR and UIDE Available.

Post reply

> > Japheth noted to me in private E-Mail that protected-mode move "calls"
> > (in place of a normal Int 15h), DO NOT work with Win95/98, and he
> > deleted such calls from his HIMEMX driver.
>
> Which version had them? Is this for pre-existing versions or future?
> (Still never officially got 3.33 out the door. Oh well.)

No idea which HIMEMX versions had this (if any!); you must ask Japheth.
UIDE from 1-Jun-09 to 6-Oct-09, and XMGR from 3-Sep-09 to 6-Oct-09, had
protected-mode XMS "moves". Japheth did not note any time limitations
for Win95/98, he just said "Do NOT work", so I assume this means ALL of
Gates & Co.'s versions, and I simply chose to get RID of the /K switch!

> Anyways, if the speedup isn't much, then it's probably not worth it
> but you could always do a runtime check for FreeDOS (OEM id: FD) or
> DR-DOS (int 21h, 4452h aka "DR") and only enable it then. I mean,
> better than disabling it completely, or am I wrong?

I want as little "system-dependent" logic in my drivers as is possible.
Users "play games" with the DOS version numbers, maybe with the variant
type as well, so I have NEVER "trusted" such logic and will NOT use it!
In any case, this problem is caused by Win95/98, apparently not by what
DOS variant is in use, so a variant check might not solve this problem.
We are all "stuck with" leftover/uncorrected Gates & Co. ERRORS, so the
safest idea for XMGR/UIDE is using protected-mode move "calls" NO MORE!

---
(Account disabled on user's request.)

Jack

Fresno, California USA,
20.10.2009, 14:52

@ Jack

Drivers Updated to 19-Oct-2009 -- UIDEJR Corrected.

Post reply

Johnson Lam has posted a new DRIVERS.ZIP file with a corrected UIDEJR, also
new assembly options for the UIDE.ASM file, on his website at:

http://johnson.tmfc.net/dos/driver.html

The UIDEJR driver tested for "free HMA" and decided where to load, after it
tested for CD/DVD drives and either dismissed or saved its CD/DVD routines.
Long before its CD/DVD test, UIDEJR issues a "VDS lock", so its memory will
never get "re-mapped" by JEMM386/EMM386, for use by some other "task", etc.
But, the EMM driver will get an improper UIDEJR size, when HMA space is not
adequate and UIDEJR loads in upper or DOS memory later! This can allow the
EMM driver to "map" other things in UIDEJR's memory space and may cause I-O
errors or a CRASH!

"Silly ME!", but no big problem. Most systems have enough HMA for UIDEJR,
and its /H "lock" sizes were being set properly. To fix this, but not add
a lot more code, UIDEJR will now "dismiss" its CD/DVD logic only if the /N2
switch was given. Without /N2, if no CD/DVD drives are found, its CD/DVD
logic will now remain in memory. So, when its driver "VDS lock" is issued
UIDEJR now "knows" from its switches how much memory to "lock", same as the
logic in UIDE.

Note also that a 6K UIDE2, an XIDE and a 6K XIDE2 can now be created, along
with the original UIDE, as listed at the start of the UIDE.ASM source file.
XIDE and XIDE2 have 25- and 50-MB medium size caches with 1600 cache blocks
and are more efficient than the 15- or 40-MB UIDE caches (960/1280 blocks).
This may help, with a small-memory system. Those who want 80-MB caches or
larger can go on using the normal UIDE or 6K UIDE2. The added drivers are
"special-assembly only", and NOT described in the driver README, to prevent
confusing many users, and to save Johnson Lam a lot of work translating the
README file to Chinese!

My apologies for the error in UIDEJR.

---
(Account disabled on user's request.)

Rugxulo Usono, 20.10.2009, 15:02 @ Jack	Drivers Updated to 19-Oct-2009 -- UIDEJR Corrected. Post reply
	> Johnson Lam has posted a new DRIVERS.ZIP file with a corrected UIDEJR, > also > new assembly options for the UIDE.ASM file, on his website at: > > http://johnson.tmfc.net/dos/driver.html The website is updated, but the file itself seems to still be "Oct. 15". (BTW, personally I always named it jack_drv.zip, less generic, don't you think?) Anyways, thanks again for your diligence.

Jack

Fresno, California USA,
20.10.2009, 15:25

@ Rugxulo

New 19-Oct-2009 Drivers ARE On Johnson's Site.

Post reply

> The website is updated, but the file itself seems to still be "Oct. 15".

Shouldn't be, as I sent it to Johnson dated 19-Oct-2009. I also downloaded
his posted DRIVERS.ZIP file again -- all files in it ARE the new 19-Oct-2009
drivers and "match" the masters on my system. Could your web browser maybe
need a "refresh"??

> (BTW, personally I always named it jack_drv.zip, less generic, don't you
> think?)

Perhaps, but just as Julius Caesar always referred to himself as "He" or as
"Caesar" in his published writings, I also do not like using my own name in
my work, except when necessary! Too much "vanity" in this world, already!

> Anyways, thanks again for your diligence. :-D

You are welcome. Unlike Gates & Co. who "save up bugs" for a big "service
pack", my bugs are "BATTLE Stations!" and get fixed immediately!

---
(Account disabled on user's request.)

Rugxulo

Usono,
20.10.2009, 17:11

@ Jack

19-Oct-2009 Drivers on Johnson's site: Confirmed

Post reply

> > The website is updated, but the file itself seems to still be "Oct. 15".
>
> Shouldn't be, as I sent it to Johnson dated 19-Oct-2009. I also
> downloaded
> his posted DRIVERS.ZIP file again -- all files in it ARE the new
> 19-Oct-2009
> drivers and "match" the masters on my system. Could your web browser
> maybe
> need a "refresh"??

You're right, it was the browser (Chrome) although I have no idea why!! Very strange. Oh well, at least Firefox works. (Even closing + reopening Chrome doesn't help.)

> > (BTW, personally I always named it jack_drv.zip, less generic, don't
> you
> > think?)
>
> Perhaps, but just as Julius Caesar always referred to himself as "He" or
> as
> "Caesar" in his published writings, I also do not like using my own name
> in
> my work, except when necessary! Too much "vanity" in this world,
> already!

Okay, but "drivers.zip" is VERY generic! Then again, I can't think of any better names, and it's obviously not crucial, so ....

> > Anyways, thanks again for your diligence. :-D
>
> You are welcome. Unlike Gates & Co. who "save up bugs" for a big
> "service
> pack", my bugs are "BATTLE Stations!" and get fixed immediately!

Win2k3 won't even get any more service packs, and it (allegedly) still has the blasted DPMI limit! Argh. Luckily I don't use it, but that kinda pisses me off.

Jack

Fresno, California USA,
20.10.2009, 20:30

@ Rugxulo

"Ain't Computers GRAND!!"

Post reply

>> ... I downloaded [Johnson Lam's] posted DRIVERS.ZIP file again --
>> all files in it ARE the new 19-Oct-2009 drivers and "match" the
>> masters on my system. Could your web browser maybe need a
>> "refresh"??
>
> You're right, it was the browser (Chrome) although I have NO IDEA
> why!! Very strange. Oh well, at least Firefox works. (Even
> closing + reopening Chrome doesn't help.)

Try the power switch, every now and then. I NEVER leave my system
on, when it is not being used, so my Opera browser always starts up
"fresh" with its cache "dead blank". "The ONLY way to go!", IMHO!

>> ... Unlike Gates & Co. who "save up bugs" for a big "service
>> pack", my bugs are "BATTLE Stations!" and get fixed immediately!
>
> Win2k3 won't even get any more service packs, and it (allegedly)
> still has the blasted DPMI limit! Argh. Luckily I don't use it
> but that kinda pisses me off.

My favorite computer cartoon, from about 1967, shows a HUGE computer
with tapes/lights/wheels/etc. and two rather-insignificant techs who
are reading a listing from it. One tech says to the other "It says
it wants a GOAT sacrificed to it"!!

Sad, how Gates & Co. seem to have that kind of mentality. They and
Intel have ALWAYS left me feeling "Ain't computers GRAND!!"

---
(Account disabled on user's request.)