dosemu-devel — DOSEMU developers mailing list

Re: [Dosemu-devel] merge devel and release something

[Stas S. <st...@li...>]

21.09.2014 05:55, Bart Oldeman пишет:
> Stas Sergeevwrote:
>
>     Eric Auer wrote:
>     >> change is to allow contributing under "GPLv2 or later".
>     > Did any other DOSEMU experts have a strong opinion
>     > about this? Or did people not care which of the two
>     > variants should be used, except for Bart who prefers
>     > the old variant?
>
>
> Eric, you can see other Eric's (Biederman) replies in the archives here:
> https://www.mail-archive.com/dos...@li.../index.html#00333
>
> I mostly agree with him, and he was involved in DOSEMU before me so 
> knows slightly more about that history before.
With the attached patch the Eric's opinion about a
clarification, should be addressed. There is no problem
if the clarification should stay. I am going to put that to devel.

> As to removing clause 5 and 6, their intent was just to clarify how 
> the main copyright holders of DOSEMU were interpreting the GPL. I 
> agree they are/were not written perfectly but the idea is just the 
> same as Linus' kernel COPYING header of
>
> " NOTE! This copyright does *not* cover user programs that use kernel
> services by normal system calls - this is merely considered normal use
> of the kernel, and does *not* fall under the heading of "derived work".
It is important to note that Linus here is talking about the
program itself, not about the program+kernel combination.
He says the program is not a "derived work", period.
Dosemu, instead, is saying something about a combination
of DOS prog+dosemu, and says that "We grant the right to
use a proprietary DOS together with DOSEMU". Not to repeat
that this is a nonsense, this has nothing to do with what
Linus said.

Also, do you really find sane the statements like this in something
that pretends to have the legal meaning:
---
this view comes from interpreting more
into the current version (2) of the GPL than is actually defined
---
Is this a joke? Since when we started to interpret the legal
documents over than that was written?

Or this:
---
However, recent discussions about the scope of 'library linking' with
GPL code and the possibility that future versions of the GPL may
define this issue in a more restrictive manner, make it necessary to
restrict
---
Who have seen these discussions? Dosemu was being developed
under closed doors these times, with private MLs, private IRC channels.
Referring to some private discussions that no one can read and
evaluate, is not only invalid but is also impolite. We now only have
the Eric's word that these discussions actually took place... but
this is not enough for sure.

So I am not _asking_ you to remove that, its pretty much a matter
of understanding on either side. And if my understanding is valid,
then it should be removed pretty much automatically rather than
because I ask to do so.

Re: [Dosemu-devel] merge devel and release something

[Stas S. <st...@li...>]

21.09.2014 05:55, Bart Oldeman пишет:
> As to removing clause 5 and 6, their intent was just to clarify how 
> the main copyright holders of DOSEMU were interpreting the GPL. I 
> agree they are/were not written perfectly but the idea is just the 
> same as Linus' kernel COPYING header of
OK, to ease an understanding of what I mean, let me
as an example apply Linus's statement to dosemu.

> " NOTE! This copyright does *not* cover user programs that use kernel
> services by normal system calls - this is merely considered normal use
> of the kernel, and does *not* fall under the heading of "derived work".
This will read as:
---
Note: this copyright does not cover the DOS programs you run
under dosemu. They are not a derived work and do not need to
be licensed under GPL for running under dosemu.
---
But this is obvious and is not needed to be clarified.
With linux it may not be so obvious as the linux progs may be
explicitly written for linux and use its headers.
For dosemu this part is simply irrelevant.

> Also note that the only valid version of the GPL as far as the kernel
> is concerned is _this_ particular version of the license (ie v2, not
> v2.2 or v3.x or whatever), unless explicitly otherwise stated.
This will read as:
---
The source files without an explicit copyrights are under
GPLv2 only.
---
It is arguably much better to have that just in readme, instead
of explicitly modifying all such a files. But in dosemu also the
files that explicitly stated otherwise were modified.

So either the above interpretation is completely wrong, or
what was done in dosemu have no relation to the above
quote from Linus.

Please provide your view, please clarify why mine is wrong,
and why it is always stated that this has some relation to
what Linus did.

Re: [Dosemu-devel] merge devel and release something

[Stas S. <st...@li...>]

21.09.2014 05:55, Bart Oldeman пишет:
> Stas Sergeevwrote:
>
>     Eric Auer wrote:
>     >> change is to allow contributing under "GPLv2 or later".
>     > Did any other DOSEMU experts have a strong opinion
>     > about this? Or did people not care which of the two
>     > variants should be used, except for Bart who prefers
>     > the old variant?
>
>
> Eric, you can see other Eric's (Biederman) replies in the archives here:
> https://www.mail-archive.com/dos...@li.../index.html#00333
>
> I mostly agree with him, and he was involved in DOSEMU before me so 
> knows slightly more about that history before.
(adding Eric to CC)
The particular Eric's e-mail on that subject is this one:
https://www.mail-archive.com/dos...@li.../msg00328.html
He basically says that the right to use dosemu with proprietary progs needs
to be _clarified_ (not granted, just clarified).

I agree with that and propose the simple clarification:
---
Q: Can I run a proprietary software, including proprietary DOS, under 
dosemu?
A: Yes!
Q: Can I distribute dosemu with proprietary DOS or other proprietary 
software?
A: No, but we suggest to distribute dosemu with FreeDOS and install the
proprietary software separately.
---
Is this good for clarification?
No weird stuff about dynamic linking, GPLv2 etc.

I asked Eric explicitly about a few things:
https://www.mail-archive.com/dos...@li.../msg00329.html
namely, is that clause validly states the reasons why dosemu should be
licensed as "GPLv2 only" (3) and why does it _grants_ the
right to use rather than to clarify that right (4). But, as you
can see from here:
https://www.mail-archive.com/dos...@li.../msg00334.html
3 and 4 were explicitly skipped (3 is not even quoted, 4 is not answered).
So I can take only Eric's opinion that the _clarification_
should stay. The above example of clarification can be
applied to devel if no one objects.

Later I also asked why the files with explicit "GPLv2 or later"
headings were changed, and this was not addressed as well.

So what other points from Eric's mails we can consider?
He also explains why the files without an explicit copyrights
should be licensed as "GPLv2 only" and I agree:
http://sourceforge.net/p/dosemu/code/ci/b0c0830292330e08df498063a78cb67c99a91017/
(this is just a proposal in a separate branch).
So I think Eric's considerations were taken into account.
Or have I missed some?

>     I am not sure Bart prefers the old variant.
>     It could have been a simple misunderstanding, I don't
>     know how can that be classified as re-licensing. He was
>     very terse.
>
>
> Changing from "GPLv2 only" to "GPLv2 or later" is a relicensing 
> (alternatively all GPLv2-only code needs to be rewritten like you wrote).
Maybe, but this is not what I did.
However, AFAIK this is a "light-weight" re-licensing that
does not require rewriting the code or asking the copyright
holders. Only re-licensing to "GPLv3 or later" will require this.

> As to removing clause 5 and 6, their intent was just to clarify how 
> the main copyright holders of DOSEMU were interpreting the GPL. I 
> agree they are/were not written perfectly but the idea is just the 
> same as Linus' kernel COPYING header of
Well, this is the most controversial statement from you
and other dosemu developers involved, that is always re-occuring
here, but my questions about it are always ignored.
So lets try again.

--- Linus says ---
" NOTE! This copyright does *not* cover user programs that use kernel
services by normal system calls - this is merely considered normal use
of the kernel, and does *not* fall under the heading of "derived work".
---
To me, Linus simply made it clear that the software that
_uses_ linux kernel is not a derived work, and as such,
does not fall under GPL. He basically protects the proprietary
software _authors_ from someone else saying "hey, you use
linux system calls, you include linux headers! Release under GPL!"
I do not think his clarification had something to do with the
linux kernel itself, either its use or its distribution.
Now in dosemu, this somehow infers that:
1. dosemu should be licensed under "GPLv2 only", or otherwise
it may be illegal to run proprietary DOS under it
2. we need to grant an explicit right to _use_ the DOS software
under dosemu.
IMHO this is total and complete misunderstanding of everything.
But I may be wrong. As I said, the direct questions are always
ignored. So now could you _please_ clarify how the below is similar
to what Linus did:
---

5.   The nature of DOSEMU requires the use of (ie "booting") a DOS, which
      may be proprietary. This could be interpreted as 'library linking'
      the DOS functions to DOSEMU (this view comes from interpreting more
      into the current version (2) of the GPL than is actually defined).

      However, past discussions about the scope of 'library linking' with
      GPL code and the possibility that future versions of the GPL may
      define this issue in a more restrictive manner, made it necessary to
      restrict the DOSEMU copyright explicitly to version 2 of the GPL.
                                               ============        ===

      We grant the right to use a proprietary DOS together with DOSEMU.

---
I see no similarities whatsoever.

--- Linus continues ---
Also note that the only valid version of the GPL as far as the kernel
is concerned is _this_ particular version of the license (ie v2, not
v2.2 or v3.x or whatever), unless explicitly otherwise stated.
---
This is a different, and is also a very interesing part.
And it was also taken to dosemu with a serious misunderstandings.
Namely, "unless explicitly otherwise stated" was missed and the
files with the explicit "GPLv2 or later" headings were changed.
My question was and still is: what resemlance this mess could
have with what Linus validly and properly did?

> IANAL but I suspect that somebody can fork the Linux kernel and remove 
> that text just as you can fork dosemu (or you'd become the new 
> maintainer) and remove that text (or I can remove that text as still 
> maintainer,
Before removing something, there is a need of an understanding.
I do not claim the misunderstanding is on you side, it can be on
my side just as well. But the point is that we (major contributors)
need to get to a similar view before changing or removing anything.
That's why I put all my proposals into a separate branch and am
waiting for years before the simple questions are answered. :)
Answering your statement above, my current thinking is that
if I fork linux, I won't remove anything, but if I fork dosemu, I
definitely will. I see ZERO similarities in what was done by Linus
and what was done here.

> Hans wrote me to that was ok, it's at my own legal risk of course).
> In any case the intent of the old copyright holders can be traced back 
> from old versions in the highly unlikely case of any lawsuit just as 
> Linus' intent can be clearly traced back.
OK, I tracked the intent of dosemu copyright holders to an
explicit "GPLv2 or later" headings, but they were then obscured
by the reference to the COPYING.DOSEMU that says "GPLv2 only".
So tracking back didn't help much.

> However nobody can relicense the kernel or DOSEMU from GPLv2 to "GPLv2 
> or later" or "GPLv3" or "GPLv3 or later" without approval of all 
> copyright holders.
But GPL faq disagree with you:
http://www.gnu.org/licenses/gpl-faq.en.html#AllCompatibility
It is only not possible to re-license to "GPLv3 or later", but possible
to "GPLv2 or later":
---
2: While you may release your project (either your original work and/or 
work that you received and modified) under GPLv2-or-later in this case, 
note that the other code you're using must remain under GPLv2 only.
---

> *I* as maintainer think it does not hurt to clarify that we see DOSEMU 
> and DOS as separate independent works and that if somebody sells 
> DOSEMU and MSDOS on a CD I don't consider that copyright infringement 
> (as long as they make DOSEMU's source available).
This is a completely different story.
Lets not dive into that topic for now, or the discussion will
became rusty again. It is perfectly fine to discuss also this,
but right now we can't even merge devel because of the
different views on the very simple questions. I am asking
this new code contributors to ignore the COPYING.DOSEMU
and add the normal GPL headings but this is not a normal
development process. I want to be able to trust dosemu's
licensing style, but its current state is IMHO far from what
the one can trust.

Re: [Dosemu-devel] merge devel and release something

[Bart O. <bar...@gm...>]

Stas Sergeev wrote:

> Eric Auer wrote:
> >> change is to allow contributing under "GPLv2 or later".
> > Did any other DOSEMU experts have a strong opinion
> > about this? Or did people not care which of the two
> > variants should be used, except for Bart who prefers
> > the old variant?
>

Eric, you can see other Eric's (Biederman) replies in the archives here:
https://www.mail-archive.com/dos...@li.../index.html#00333

I mostly agree with him, and he was involved in DOSEMU before me so knows
slightly more about that history before.


> I am not sure Bart prefers the old variant.
> It could have been a simple misunderstanding, I don't
> know how can that be classified as re-licensing. He was
> very terse.
>

Changing from "GPLv2 only" to "GPLv2 or later" is a relicensing
(alternatively all GPLv2-only code needs to be rewritten like you wrote).

As to removing clause 5 and 6, their intent was just to clarify how the
main copyright holders of DOSEMU were interpreting the GPL. I agree they
are/were not written perfectly but the idea is just the same as Linus'
kernel COPYING header of

"   NOTE! This copyright does *not* cover user programs that use kernel
 services by normal system calls - this is merely considered normal use
 of the kernel, and does *not* fall under the heading of "derived work".
 Also note that the GPL below is copyrighted by the Free Software
 Foundation, but the instance of code that it refers to (the Linux
 kernel) is copyrighted by me and others who actually wrote it.

 Also note that the only valid version of the GPL as far as the kernel
 is concerned is _this_ particular version of the license (ie v2, not
 v2.2 or v3.x or whatever), unless explicitly otherwise stated.

                        Linus Torvalds"

IANAL but I suspect that somebody can fork the Linux kernel and remove that
text just as you can fork dosemu (or you'd become the new maintainer) and
remove that text (or I can remove that text as still maintainer, Hans wrote
me to that was ok, it's at my own legal risk of course).
In any case the intent of the old copyright holders can be traced back from
old versions in the highly unlikely case of any lawsuit just as Linus'
intent can be clearly traced back.

However nobody can relicense the kernel or DOSEMU from GPLv2 to "GPLv2 or
later" or "GPLv3" or "GPLv3 or later" without approval of all copyright
holders.

*I* as maintainer think it does not hurt to clarify that we see DOSEMU and
DOS as separate independent works and that if somebody sells DOSEMU and
MSDOS on a CD I don't consider that copyright infringement (as long as they
make DOSEMU's source available).

For others this clarification is not needed. E.g. gog.com sells old DOS
games + DOSBOX and DOSBOX does not have any text clarifying that a DOS game
does not dynamically link to DOSBOX. DOSBOX developers do not complain to
gog.com (on the contrary!).

So I am in favour of keeping our language, maybe clarify it a bit.

However, yes the development should not be thrown away of course I fully
agree, 1.4.0 is already 7 years old, how time flies! I just have two small
children + a day job in HPC where there is little time left for free time
free software development, that's the only reason for my quietness.

Bart

Re: [Dosemu-devel] merge devel and release something

[solarflow99 <sol...@gm...>]

+1 assuming there's isn't any reason against.  The dosemu site shows old
releases as stable, maybe we can update the version to 1.6 at least, 1.4 is
very old now.


On Sat, Sep 20, 2014 at 3:24 PM, Stas Sergeev <st...@li...> wrote:

> 20.09.2014 21:55, Eric Auer пишет:
> >> The only reason I didn't merge devel before, is because
> >> it has this commit:
> >>
> https://sourceforge.net/p/dosemu/code/ci/a1dff2a15698efc4f400812993ec819ea7091439/
> >> which Bart called a re-licensing. I of course do not
> >> want to do any re-licensing, but the intention of that
> >> change is to allow contributing under "GPLv2 or later".
> > Did any other DOSEMU experts have a strong opinion
> > about this? Or did people not care which of the two
> > variants should be used, except for Bart who prefers
> > the old variant?
> I am not sure Bart prefers the old variant.
> It could have been a simple misunderstanding, I don't
> know how can that be classified as re-licensing. He was
> very terse.
>
> >   You could ask license expert Shane.
> I don't know that guy.
> And also there is too much of a context involved.
> For instance there was a change many years ago that
> added a dosemu-specific copyrights to all source files,
> even to those that had an explicit "GPLv2 or later" headings.
> This was explained as "something Linus did for linux",
> but I checked what Linus did - he only changed a readme,
> without touching any source files.
> So there is a lot of confusion on dosemu side, the
> poor COPYING file was changed and tweaked many times,
> including rather recently: I can see a differences even
> between 1.2.2 and 1.4.0! So I don't think someone from
> outside can dive into that mess without a good payment.
>
> >> We already have enough of such a code (all my code,
> >> and not only that), so that clause I removed, was
> >> invalidated long ago. There is simply no other way than
> >> to remove it.
> > Not sure about the context here...
> Me too.
> The original intention of that restriction was _probably_
> to restrict all the source code to GPLv2. At least, as I
> said, even the explicit "GPLv2 or later" headings were
> changed with the reference to COPYING.DOSEMU which
> states "GPLv2 only". I made sure we have enough of "GPLv2 or
> later" code _without_ a reference to COPYING.DOSEMU
> (because I don't trust it and am not sure it is valid for
> "GPLv2 or later" code to refer to it), and, as a side effect,
> removed all the code of the author of that exception, but
> of course this alone does not change anything.
> If this exception is removed however, GPL faq says
> it would be possible to re-license the whole package
> to "GPLv2 or later" but not to "GPLv3 or later" till some
> "GPLv2 only" code remains.
>
>
> ------------------------------------------------------------------------------
> Slashdot TV.  Video for Nerds.  Stuff that Matters.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk
> _______________________________________________
> Dosemu-devel mailing list
> Dos...@li...
> https://lists.sourceforge.net/lists/listinfo/dosemu-devel
>

Re: [Dosemu-devel] merge devel and release something

[Stas S. <st...@li...>]

20.09.2014 21:55, Eric Auer пишет:
>> The only reason I didn't merge devel before, is because
>> it has this commit:
>> https://sourceforge.net/p/dosemu/code/ci/a1dff2a15698efc4f400812993ec819ea7091439/
>> which Bart called a re-licensing. I of course do not
>> want to do any re-licensing, but the intention of that
>> change is to allow contributing under "GPLv2 or later".
> Did any other DOSEMU experts have a strong opinion
> about this? Or did people not care which of the two
> variants should be used, except for Bart who prefers
> the old variant?
I am not sure Bart prefers the old variant.
It could have been a simple misunderstanding, I don't
know how can that be classified as re-licensing. He was
very terse.

>   You could ask license expert Shane.
I don't know that guy.
And also there is too much of a context involved.
For instance there was a change many years ago that
added a dosemu-specific copyrights to all source files,
even to those that had an explicit "GPLv2 or later" headings.
This was explained as "something Linus did for linux",
but I checked what Linus did - he only changed a readme,
without touching any source files.
So there is a lot of confusion on dosemu side, the
poor COPYING file was changed and tweaked many times,
including rather recently: I can see a differences even
between 1.2.2 and 1.4.0! So I don't think someone from
outside can dive into that mess without a good payment.

>> We already have enough of such a code (all my code,
>> and not only that), so that clause I removed, was
>> invalidated long ago. There is simply no other way than
>> to remove it.
> Not sure about the context here...
Me too.
The original intention of that restriction was _probably_
to restrict all the source code to GPLv2. At least, as I
said, even the explicit "GPLv2 or later" headings were
changed with the reference to COPYING.DOSEMU which
states "GPLv2 only". I made sure we have enough of "GPLv2 or
later" code _without_ a reference to COPYING.DOSEMU
(because I don't trust it and am not sure it is valid for
"GPLv2 or later" code to refer to it), and, as a side effect,
removed all the code of the author of that exception, but
of course this alone does not change anything.
If this exception is removed however, GPL faq says
it would be possible to re-license the whole package
to "GPLv2 or later" but not to "GPLv3 or later" till some
"GPLv2 only" code remains.

Re: [Dosemu-devel] merge devel and release something

[Eric A. <e....@jp...>]

Hi Stas, Bart and other DOSEMU experts,

while I am not experienced with the fine details,

http://sourceforge.net/p/dosemu/code/ci/a1dff2a15698efc4f400812993ec819ea7091439/tree/COPYING.DOSEMU?diff=7a762fd24a3a53268b856874ffe15d4094ef8208

sounds as if you dropped the limitation to only
version 2 of GPL, along with the explanation of
why that restriction existed? Sounds ok for me.

> I think enough is enough, and I am going to merge
> devel branch into master and then perhaps upload
> the tarballs.

> People do not know about the devel branch, but it
> works much better than master.

I agree that people would miss quite some changes
by being stuck to the non-devel branch.

> The only reason I didn't merge devel before, is because
> it has this commit:
> https://sourceforge.net/p/dosemu/code/ci/a1dff2a15698efc4f400812993ec819ea7091439/
> which Bart called a re-licensing. I of course do not
> want to do any re-licensing, but the intention of that
> change is to allow contributing under "GPLv2 or later".

Did any other DOSEMU experts have a strong opinion
about this? Or did people not care which of the two
variants should be used, except for Bart who prefers
the old variant? You could ask license expert Shane.

> We already have enough of such a code (all my code,
> and not only that), so that clause I removed, was
> invalidated long ago. There is simply no other way than
> to remove it.

Not sure about the context here...

> So if there is no evidence this is something that should
> not be done, I am going to merge things. devel right
> now includes over a year of development! It can't be
> allowed to rot.

Sounds like a plan :-)

Regards, Eric

[Dosemu-devel] merge devel and release something

[Stas S. <st...@li...>]

Hi.

I think enough is enough, and I am going to merge
devel branch into master and then perhaps upload
the tarballs.
People do not know about the devel branch, but it
works much better than master.

The only reason I didn't merge devel before, is because
it has this commit:
https://sourceforge.net/p/dosemu/code/ci/a1dff2a15698efc4f400812993ec819ea7091439/
which Bart called a re-licensing. I of course do not
want to do any re-licensing, but the intention of that
change is to allow contributing under "GPLv2 or later".
We already have enough of such a code (all my code,
and not only that), so that clause I removed, was
invalidated long ago. There is simply no other way than
to remove it.

So if there is no evidence this is something that should
not be done, I am going to merge things. devel right
now includes over a year of development! It can't be
allowed to rot.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> 08.06.2014 03:25, Julius Schwartzenberg пишет:
>> On the Wikipedia page about Merge (Win4Lin was the Linux port of Merge)
>> it is mentioned that the company had access to the Windows source code
>> under the WISE program.
>> https://web.archive.org/web/20120615111330/http://hyper.sunjapan.com.cn/~hz/win32/wise.htm
> Sounds much like wine to me.
> I don't think this had a sources of vxds.
> And, as wikipedia says,
> ---
> which allowed later versions of Merge to run Windows /Shrink wrapped/ 
> applications without a copy of Windows.
> ---
> So I think they speak about wine-alike approach here.
> The win4lin page mentions there were 3 flavours of win4lin,
> one being a wine-alike emulator.

Yes, I noticed that too, but Win4Lin Home was just a home-use version of
their full Win4Lin product. I think those citations are incorrect. I
have never seen a Wine-like version of Merge.
Indeed it's unknown how much documentation they had through WISE.


>> With dosemu dual-booting is easy though :) 
> How?
> Manipulating a symlink for drive C?
> But this will still require an installation of a second DOS.

Yep! I guess using multiple configuration files would work as well.


>> Maybe when SCO is completely dead, more interesting bits will come out 
>> from there :) 
> Even if it is dead, the copyrights are not automatically
> re-assigned to someone else. Maybe the original author
> will release it as abandonware, but it is still impossible to
> port such a code to freedos I suspect.

You're right, it's not likely it could immediately benefit any free
software.


>> Yep, I still need to try that, but considering the HDPMI results it
>> doesn't look promising. But in the regular case, that's how should be
>> able to run it in DOSEMU if it would work?
> At least for win31 that works right, but win.com can be used
> as well: dosemu cheats it and forces to load krnl386.exe instead
> of anything else.
> Essentially, krnl386.exe is an unpriviledged part of the kernel,
> while win386.exe is a priviledged part. dosemu skips the
> priviledged part since it doesn't do much but to provide DPMI
> and a couple of vxds.

Thanks I just tried this with DOSEMU and the earliest Chicago builds.
They both fail with:
KERNEL: Unable to initialize heap

I had the same error with hdpmi16 in QEMU when I used a Chicago version
which did not have a mismatch with its DOS version.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

08.06.2014 03:25, Julius Schwartzenberg пишет:
> On the Wikipedia page about Merge (Win4Lin was the Linux port of Merge)
> it is mentioned that the company had access to the Windows source code
> under the WISE program.
> https://web.archive.org/web/20120615111330/http://hyper.sunjapan.com.cn/~hz/win32/wise.htm
Sounds much like wine to me.
I don't think this had a sources of vxds.
And, as wikipedia says,
---
which allowed later versions of Merge to run Windows /Shrink wrapped/ 
applications without a copy of Windows.
---
So I think they speak about wine-alike approach here.
The win4lin page mentions there were 3 flavours of win4lin,
one being a wine-alike emulator.

> With dosemu dual-booting is easy though :) 
How?
Manipulating a symlink for drive C?
But this will still require an installation of a second DOS.

> Maybe when SCO is completely dead, more interesting bits will come out 
> from there :) 
Even if it is dead, the copyrights are not automatically
re-assigned to someone else. Maybe the original author
will release it as abandonware, but it is still impossible to
port such a code to freedos I suspect.

>>> When I do not let it start win.com and try to start system\krnl386 I
>>> shows this message:
>>> KERNEL: Unable to enter Protected Mode
>> This simply means DPMI is unavailable.
>> Under dosemu it should pass much further than this.
> Yep, I still need to try that, but considering the HDPMI results it
> doesn't look promising. But in the regular case, that's how should be
> able to run it in DOSEMU if it would work?
At least for win31 that works right, but win.com can be used
as well: dosemu cheats it and forces to load krnl386.exe instead
of anything else.
Essentially, krnl386.exe is an unpriviledged part of the kernel,
while win386.exe is a priviledged part. dosemu skips the
priviledged part since it doesn't do much but to provide DPMI
and a couple of vxds.

>>> Then I run hdpmi16
>> Is win95 still 16 bit? I think maybe hdpmi32 would be better.
> Then I get the 'KERNEL: Unable to enter Protected Mode' though. It seems
> only hdpmi16 makes a difference.
This likely means that the 32bit services are provided with
vwin32.386 itself and not with DPMI. And we don't know their protocol.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> 07.06.2014 03:17, Julius Schwartzenberg пишет:
>>> Some IDA work I guess.
>>> Or maybe the documentation on these vxds is available
>>> in some SDK/DDK, or they got it through some partner
>>> contract with MS.
>> Reading their history a bit, it indeed seems there was some more
>> documentation.
> Where do you read this?

On the Wikipedia page about Merge (Win4Lin was the Linux port of Merge)
it is mentioned that the company had access to the Windows source code
under the WISE program.
https://web.archive.org/web/20120615111330/http://hyper.sunjapan.com.cn/~hz/win32/wise.htm

It seems that at least they used the VtoolsD development kit from Vireo
as well when looking at the included vxds. Probably in addition to
Microsoft's DDK. So I guess those had some documentation with it as well.


>>   There are multiple patches involved to get it all working.
>  From your findings it seems they patch some vxds to
> use LDT where they otherwise would probably use GDT.

There are also patches for some DLLs. They mostly change the addressing
I understand.


>>> I think if freedos could run win95, there would be slightly
>>> more reasons to support it under dosemu. I don't think
>>> it does, though.
>> Is there a special reason to use it with freedos?
> Not to take a troubles of dual-boot setup?

With dosemu dual-booting is easy though :)


>>   I seems the only other
>> DOS that could run Windows 95 was an unreleased version of DR-DOS:
>> http://www.msfn.org/board/topic/109018-windows-98-in-dr-dos/
> Wow, great reading! :)
> So now we know there is at least one person in this world
> who regularly runs Win98 in non-MS DOS, and even writes
> a forum messages from that installation, so it works rather
> well. :)
> Of course the freedos devs will unlikely do so much of the
> RE work for free, as he did for Caldera.

Maybe when SCO is completely dead, more interesting bits will come out
from there :)


>> When I do not let it start win.com and try to start system\krnl386 I
>> shows this message:
>> KERNEL: Unable to enter Protected Mode
> This simply means DPMI is unavailable.
> Under dosemu it should pass much further than this.

Yep, I still need to try that, but considering the HDPMI results it
doesn't look promising. But in the regular case, that's how should be
able to run it in DOSEMU if it would work?


>> Then I run hdpmi16
> Is win95 still 16 bit? I think maybe hdpmi32 would be better.

Then I get the 'KERNEL: Unable to enter Protected Mode' though. It seems
only hdpmi16 makes a difference.


>> Incorrect MS-DOS version. MS-DOS 7.0 or greater required.
>> I'm using the included DOS version.
>>
>> Is there a way it could try to detect its DOS version through the DPMI
>> server?
> This is quite strange, and on dosemu you don't see this, right?

That's on QEMU as well. I'll try it in dosemu & see how it goes.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

07.06.2014 03:17, Julius Schwartzenberg пишет:
>> Some IDA work I guess.
>> Or maybe the documentation on these vxds is available
>> in some SDK/DDK, or they got it through some partner
>> contract with MS.
> Reading their history a bit, it indeed seems there was some more
> documentation.
Where do you read this?

>   There are multiple patches involved to get it all working.
 From your findings it seems they patch some vxds to
use LDT where they otherwise would probably use GDT.

>> I think if freedos could run win95, there would be slightly
>> more reasons to support it under dosemu. I don't think
>> it does, though.
> Is there a special reason to use it with freedos?
Not to take a troubles of dual-boot setup?
Also, if freedos guys would reverse-engineer so much of
win95, they'll certainly share some tricks with us. :)
It is very unlikely that someone of who contributes to
dosemu, uses windows, so it is much better to get the
useful tricks from the right people who actually at least
used it and know how it works.
In fact, dosemu was not very compatible with win95+ DOS.
I've spent some time fixing bugs for that DOS some 10
years ago, but I haven't tested it since. But freedos should
always be properly supported.

>   I seems the only other
> DOS that could run Windows 95 was an unreleased version of DR-DOS:
> http://www.msfn.org/board/topic/109018-windows-98-in-dr-dos/
Wow, great reading! :)
So now we know there is at least one person in this world
who regularly runs Win98 in non-MS DOS, and even writes
a forum messages from that installation, so it works rather
well. :)
Of course the freedos devs will unlikely do so much of the
RE work for free, as he did for Caldera.

> When I do not let it start win.com and try to start system\krnl386 I
> shows this message:
> KERNEL: Unable to enter Protected Mode
This simply means DPMI is unavailable.
Under dosemu it should pass much further than this.

> Then I run hdpmi16
Is win95 still 16 bit? I think maybe hdpmi32 would be better.

> Incorrect MS-DOS version. MS-DOS 7.0 or greater required.
> I'm using the included DOS version.
>
> Is there a way it could try to detect its DOS version through the DPMI
> server?
This is quite strange, and on dosemu you don't see this, right?

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> So you can try running Chicago under hdpmi (by starting
> krnl386.exe directly I think, at least this is how win31 was
> running under it). If it works then dosemu can do that too.
> But I don't think it does.

I set it all up in QEMU (Chicago version 56). Chicago doesn't have a
problem to come up when not doing anything special.

When I do not let it start win.com and try to start system\krnl386 I
shows this message:
KERNEL: Unable to enter Protected Mode

Then I run hdpmi16 and then system\krnl386 which gives:
Incorrect MS-DOS version. MS-DOS 7.0 or greater required.
I'm using the included DOS version.

Is there a way it could try to detect its DOS version through the DPMI
server?

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> 05.06.2014 23:27, Julius Schwartzenberg пишет:
>> Stas Sergeev wrote:
>>> 05.06.2014 01:51, Julius Schwartzenberg пишет:
>>>> Stas Sergeev wrote:
>>>>> Ah, it already does, it says vwin32.vxd must be re-implemented.
>>>> Searching online about this, it seems someone mentioned that Win4Lin has
>>>> its own implementation of vwin32.vxd as well to run Windows within
>>>> Linux. I checked the files included with Win4Lin, but I couldn't quickly
>>>> find that they have a special version of that file or patches for it.
>>> It is very unlikely to be in that form.
>>> Most likely they implement it inside their own sources, which
>>> are unavailable. The same way dosemu implements a few vxd's
>>> too.
>> Yes, it seems they patch this internally.
> Maybe they patch some vxds but unlikely vwin32.
> This should be re-implemented completely I am afraid.
> 
>> _VWIN32_BlockForTermination
>> _VWIN32_CopyMem
>> _VWIN32_Set_Thread_Context
>> _VWIN32_Get_Thread_Context
>> _VWIN32_QueueUserApc
> Just so few?

That's all I found with strings in emdbg.rom regarding VWIN32. I've
uploaded the file in case you're interested:
https://yadi.sk/d/JvihPmPbSYGqX/roms/
I only searched through it with strings.


>> It still intrigues me they managed to get it all to work.
> Some IDA work I guess.
> Or maybe the documentation on these vxds is available
> in some SDK/DDK, or they got it through some partner
> contract with MS.

Reading their history a bit, it indeed seems there was some more
documentation. There are multiple patches involved to get it all working.


> I think if freedos could run win95, there would be slightly
> more reasons to support it under dosemu. I don't think
> it does, though.

Is there a special reason to use it with freedos? I seems the only other
DOS that could run Windows 95 was an unreleased version of DR-DOS:
http://www.msfn.org/board/topic/109018-windows-98-in-dr-dos/

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

05.06.2014 23:27, Julius Schwartzenberg пишет:
> Stas Sergeev wrote:
>> 05.06.2014 01:51, Julius Schwartzenberg пишет:
>>> Stas Sergeev wrote:
>>>> 04.06.2014 00:11, Stuart Axon пишет:
>>>>> Funnily enough, googling Eric Auer Dosemu, led me straight to
>>>>> http://sourceforge.net/p/dosemu/mailman/message/2149777/
>>>>>
>>>>> Which seems to be a your discussion of the Win31 implementation..
>>>>>
>>>> Indeed. :)
>>>>
>>>> I guess some googling on win95+dosemu will enlighten
>>>> us how to run win95 under dosemu, too.
>>>> Ah, it already does, it says vwin32.vxd must be re-implemented.
>>> Searching online about this, it seems someone mentioned that Win4Lin has
>>> its own implementation of vwin32.vxd as well to run Windows within
>>> Linux. I checked the files included with Win4Lin, but I couldn't quickly
>>> find that they have a special version of that file or patches for it.
>> It is very unlikely to be in that form.
>> Most likely they implement it inside their own sources, which
>> are unavailable. The same way dosemu implements a few vxd's
>> too.
> Yes, it seems they patch this internally.
Maybe they patch some vxds but unlikely vwin32.
This should be re-implemented completely I am afraid.

> _VWIN32_BlockForTermination
> _VWIN32_CopyMem
> _VWIN32_Set_Thread_Context
> _VWIN32_Get_Thread_Context
> _VWIN32_QueueUserApc
Just so few?

> It still intrigues me they managed to get it all to work.
Some IDA work I guess.
Or maybe the documentation on these vxds is available
in some SDK/DDK, or they got it through some partner
contract with MS.
I think if freedos could run win95, there would be slightly
more reasons to support it under dosemu. I don't think
it does, though.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> 05.06.2014 01:51, Julius Schwartzenberg пишет:
>> Stas Sergeev wrote:
>>> 04.06.2014 00:11, Stuart Axon пишет:
>>>> Funnily enough, googling Eric Auer Dosemu, led me straight to
>>>> http://sourceforge.net/p/dosemu/mailman/message/2149777/
>>>>
>>>> Which seems to be a your discussion of the Win31 implementation..
>>>>
>>> Indeed. :)
>>>
>>> I guess some googling on win95+dosemu will enlighten
>>> us how to run win95 under dosemu, too.
>>> Ah, it already does, it says vwin32.vxd must be re-implemented.
>> Searching online about this, it seems someone mentioned that Win4Lin has
>> its own implementation of vwin32.vxd as well to run Windows within
>> Linux. I checked the files included with Win4Lin, but I couldn't quickly
>> find that they have a special version of that file or patches for it.
> It is very unlikely to be in that form.
> Most likely they implement it inside their own sources, which
> are unavailable. The same way dosemu implements a few vxd's
> too.

Yes, it seems they patch this internally. I found some files em.rom and
emdbg.rom, the latter containing strings like:
emLLDT: Setting new LDTR to 0
emLLDT: Win31: Skipping Zero load of LDT
emLLDT: Win95: Skipping Zero load of LDT
emLLDT: doing SetUnixLDTDescriptor for 0
emLLDT: Inslen = %x,
PatchVxDs: version Win_31
PatchVxDs: version WFW_311
PatchVxDs: version WIN_40a (Win98)
PatchVxDs: version WIN_40 (Win95)
PatchVxDs: version WINME
PatchDynamicVxd: DeviceName:[%s] ModName:[%s] LogSeg:[%d]
PatchDynamicVxd: CHK DeviceName:[%s] ModName:[%s] LogSeg:[%d]
PatchDynamicVxD: No patches for Device %s, Module %s, lseg %x
_VWIN32_BlockForTermination
_VWIN32_CopyMem
_VWIN32_Set_Thread_Context
_VWIN32_Get_Thread_Context
_VWIN32_QueueUserApc

It still intrigues me they managed to get it all to work.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

05.06.2014 01:51, Julius Schwartzenberg пишет:
> Stas Sergeev wrote:
>> 04.06.2014 00:11, Stuart Axon пишет:
>>> Funnily enough, googling Eric Auer Dosemu, led me straight to
>>> http://sourceforge.net/p/dosemu/mailman/message/2149777/
>>>
>>> Which seems to be a your discussion of the Win31 implementation..
>>>
>> Indeed. :)
>>
>> I guess some googling on win95+dosemu will enlighten
>> us how to run win95 under dosemu, too.
>> Ah, it already does, it says vwin32.vxd must be re-implemented.
> Searching online about this, it seems someone mentioned that Win4Lin has
> its own implementation of vwin32.vxd as well to run Windows within
> Linux. I checked the files included with Win4Lin, but I couldn't quickly
> find that they have a special version of that file or patches for it.
It is very unlikely to be in that form.
Most likely they implement it inside their own sources, which
are unavailable. The same way dosemu implements a few vxd's
too.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> 04.06.2014 00:11, Stuart Axon пишет:
>> Funnily enough, googling Eric Auer Dosemu, led me straight to
>> http://sourceforge.net/p/dosemu/mailman/message/2149777/
>>
>> Which seems to be a your discussion of the Win31 implementation..
>>
> Indeed. :)
> 
> I guess some googling on win95+dosemu will enlighten
> us how to run win95 under dosemu, too.
> Ah, it already does, it says vwin32.vxd must be re-implemented.

Searching online about this, it seems someone mentioned that Win4Lin has
its own implementation of vwin32.vxd as well to run Windows within
Linux. I checked the files included with Win4Lin, but I couldn't quickly
find that they have a special version of that file or patches for it.
There are multiple other VXD files included however, but I guess they
are for different things.

Chicago seems to include a vwin32.386. I hope to try it one of the
coming days with hdpmi.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

04.06.2014 03:46, Eric Auer пишет:
> Hi Stuart, Stas,
>
>> Funnily enough, googling Eric Auer Dosemu, led me straight to
>> http://sourceforge.net/p/dosemu/mailman/message/2149777/
> Wow, that was 2004... But alledgedly before HX & HDPMI came out?
No.
All the dpmi development was a matter of 90s.

> Oddly, I got Stuart's reply but not your initial reply? 
Yes.
I've got the delivery failure from your mailbox, because of a spam.

>>> For example he suggested to alter the exception routing
>>> rules. While dpmi spec says unhandled exceptions 1..5 and 7
>>> should generate real-mode interrupts, he suggested to
>>> call the protected-mode handlers instead, and if there
>>> are none - terminate the client, but never call the real-mode
>>> interrupt the way spec does.
> Sounds quite non-spec indeed...
As this article says:
http://virtuallyfun.superglobalmegacorp.com/?p=1188
---

It is a grave mistake to treat "DPMI Version
1.0" as an improvement upon "DPMI Version 0.9", it is merely a further move
away from True DPMI.

---
By "True DPMI" they mean a windows version.
The dpmi 1.0 specs are not covering it properly.

>> making them able to run windows. I even wrote the
>> code and it mostly worked. But now its dropped
>> because it is difficult to compile DOS code under linux.
> As far as I remember, a few FreeDOS parts can be cross-
> compiled from Linux with OpenWatcom for daily builds...
> Another way would of course be compiling in DOSEMU :-)
I've spent a few hours trying to google my aforementioned code.
And, quite surprisingly, here it is:
http://sourceforge.net/p/dosemu/feature-requests/_discuss/thread/0a75f20e/de18/attachment/pmdapi-0.2.tar.gz
I'd be happy if you try to compile it under linux.
Maybe I'll find some time to complete it if it can ever be compiled
these days. I am sure it was almost complete when I dropped it.
Of course the question still stays can it be of any use at all. :)

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Eric A. <e....@jp...>]

Hi Stuart, Stas,

> Funnily enough, googling Eric Auer Dosemu, led me straight to
> http://sourceforge.net/p/dosemu/mailman/message/2149777/

Wow, that was 2004... But alledgedly before HX & HDPMI came out?

>>>> The proper win31 support happened because an author of hx
>>>> extender shared a few tricks with us how to do that. That even
>>>> involved things contradicting with the dpmi specs. He wrote
>>>> hdpmi server that was able to run win31 before dosemu did.
>>> Interesting! When was that and which anti-specs features
>>> were required? Maybe you can email details to the list?

>> Eric, there is no secret here, but this was so long ago
>> that it is mostly lost.
>> What do you want to do with that info? Write your own
>> dpmi server?

Oh just, make sure that the information stays alive :-)
Oddly, I got Stuart's reply but not your initial reply?

>> For example he suggested to alter the exception routing
>> rules. While dpmi spec says unhandled exceptions 1..5 and 7
>> should generate real-mode interrupts, he suggested to
>> call the protected-mode handlers instead, and if there
>> are none - terminate the client, but never call the real-mode
>> interrupt the way spec does.

Sounds quite non-spec indeed...

>> Also there are many undocumented things, for instance,
>> the limit of LDT selector must dynamically grow when
>> you allocate the descriptors in LDT.

Ah. Thanks for implementing all that, DOSEMU experts :-)

>> There is a summary on undocumented DPMI:
>> ftp://ifctfvax.harhan.org/pub/micro/msdos/above640k/TrueDPMI/dpmiext.txt
>> but it documents actually nothing of what we didn't know
>> ourselves. The most important and difficult to discover
>> things are not in that doc.

Oh well, undocumented compatibility requirements which
were not even in a document about undocumented DPMI? :-p

>> After that, I did an attempt to port dosemu's PM api
>> translator to DOS. The idea was to make it runnable
>> under different "weak" dpmi servers (like cwsdpmi),

Interesting approach!

>> making them able to run windows. I even wrote the
>> code and it mostly worked. But now its dropped
>> because it is difficult to compile DOS code under linux.

As far as I remember, a few FreeDOS parts can be cross-
compiled from Linux with OpenWatcom for daily builds...
Another way would of course be compiling in DOSEMU :-)

Regards, Eric

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

04.06.2014 00:11, Stuart Axon пишет:
> Funnily enough, googling Eric Auer Dosemu, led me straight to
> http://sourceforge.net/p/dosemu/mailman/message/2149777/
>
> Which seems to be a your discussion of the Win31 implementation..
>
Indeed. :)

I guess some googling on win95+dosemu will enlighten
us how to run win95 under dosemu, too.
Ah, it already does, it says vwin32.vxd must be re-implemented.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stuart A. <stu...@gm...>]

Funnily enough, googling Eric Auer Dosemu, led me straight to
http://sourceforge.net/p/dosemu/mailman/message/2149777/

Which seems to be a your discussion of the Win31 implementation..


On 3 June 2014 20:32, Stas Sergeev <st...@li...> wrote:

> 03.06.2014 22:31, Eric Auer пишет:
> > Hi!
> >
> >> The proper win31 support happened because an author of hx
> >> extender shared a few tricks with us how to do that. That even
> >> involved things contradicting with the dpmi specs. He wrote
> >> hdpmi server that was able to run win31 before dosemu did.
> > Interesting! When was that and which anti-specs features
> > were required? Maybe you can email details to the list?
> Eric, there is no secret here, but this was so long ago
> that it is mostly lost.
> What do you want to do with that info? Write your own
> dpmi server?
> For example he suggested to alter the exception routing
> rules. While dpmi spec says unhandled exceptions 1..5 and 7
> should generate real-mode interrupts, he suggested to
> call the protected-mode handlers instead, and if there
> are none - terminate the client, but never call the real-mode
> interrupt the way spec does.
> Also there are many undocumented things, for instance,
> the limit of LDT selector must dynamically grow when
> you allocate the descriptors in LDT.
> There is a summary on undocumented DPMI:
> ftp://ifctfvax.harhan.org/pub/micro/msdos/above640k/TrueDPMI/dpmiext.txt
> but it documents actually nothing of what we didn't know
> ourselves. The most important and difficult to discover
> things are not in that doc.
>
> After that, I did an attempt to port dosemu's PM api
> translator to DOS. The idea was to make it runnable
> under different "weak" dpmi servers (like cwsdpmi),
> making them able to run windows. I even wrote the
> code and it mostly worked. But now its dropped
> because it is difficult to compile DOS code under linux.
>
>
> ------------------------------------------------------------------------------
> Learn Graph Databases - Download FREE O'Reilly Book
> "Graph Databases" is the definitive new guide to graph databases and their
> applications. Written by three acclaimed leaders in the field,
> this first edition is now available. Download your free book today!
> http://p.sf.net/sfu/NeoTech
> _______________________________________________
> Dosemu-devel mailing list
> Dos...@li...
> https://lists.sourceforge.net/lists/listinfo/dosemu-devel
>

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

03.06.2014 22:31, Eric Auer пишет:
> Hi!
>
>> The proper win31 support happened because an author of hx
>> extender shared a few tricks with us how to do that. That even
>> involved things contradicting with the dpmi specs. He wrote
>> hdpmi server that was able to run win31 before dosemu did.
> Interesting! When was that and which anti-specs features
> were required? Maybe you can email details to the list?
Eric, there is no secret here, but this was so long ago
that it is mostly lost.
What do you want to do with that info? Write your own
dpmi server?
For example he suggested to alter the exception routing
rules. While dpmi spec says unhandled exceptions 1..5 and 7
should generate real-mode interrupts, he suggested to
call the protected-mode handlers instead, and if there
are none - terminate the client, but never call the real-mode
interrupt the way spec does.
Also there are many undocumented things, for instance,
the limit of LDT selector must dynamically grow when
you allocate the descriptors in LDT.
There is a summary on undocumented DPMI:
ftp://ifctfvax.harhan.org/pub/micro/msdos/above640k/TrueDPMI/dpmiext.txt
but it documents actually nothing of what we didn't know
ourselves. The most important and difficult to discover
things are not in that doc.

After that, I did an attempt to port dosemu's PM api
translator to DOS. The idea was to make it runnable
under different "weak" dpmi servers (like cwsdpmi),
making them able to run windows. I even wrote the
code and it mostly worked. But now its dropped
because it is difficult to compile DOS code under linux.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Stas S. <st...@li...>]

03.06.2014 01:53, Julius Schwartzenberg пишет:
>> Of course I admit that running a kind of win95 would
>> be a good demo of dosemu's abilities. :)
> For sure!! I remember when you got Windows 3.1 running after all those
> years DOSEMU didn't support it. For some reason I wouldn't be surprised
> to see newer Windows versions working one day :)
The proper win31 support happened because an author of hx
extender shared a few tricks with us how to do that. That even
involved things contradicting with the dpmi specs. He wrote
hdpmi server that was able to run win31 before dosemu did.
So you can try running Chicago under hdpmi (by starting
krnl386.exe directly I think, at least this is how win31 was
running under it). If it works then dosemu can do that too.
But I don't think it does.

Re: [Dosemu-devel] Running win.com from Chicago (alpha version of Win95)

[Julius S. <jul...@gm...>]

Stas Sergeev wrote:
> 30.05.2014 01:39, Julius Schwartzenberg пишет:
>> Yes, of course. But it'd still be really cool if it could work in
>> DOSEMU :) 
> Were you running it after booting in its own DOS?
> It will not work with freedos or even MS-DOS other
> than its own.

I was running it in its own DOS.


> Also, are you aware about HX dos extender?
> It has a win32 support built-in and works in dosemu.

Yes, I've read about it, but never tried it. It's also quite impressive.
I imagine it could significantly enhance old systems with some newer tools.


> Of course I admit that running a kind of win95 would
> be a good demo of dosemu's abilities. :)

For sure!! I remember when you got Windows 3.1 running after all those
years DOSEMU didn't support it. For some reason I wouldn't be surprised
to see newer Windows versions working one day :)