That's not fully true. There was my HDOS extender (initially based on PMODE extender, the same that was origin for PMODE/W), which was running first fasm versions before (about one year later) I changed it to run in flat real mode. It was written in syntax of my previous assembler ASM32 (you may find some info about it here), and because it was so similar to fasm's syntax, it was compilable with fasm almost immediately then.
I still have the sources of several HDOS versions, I recovered them some time ago - if anyone's interested I may share them.

Only 99.99% true



Thanks for pointing out ... but since it seems to never have been public
and you did not mention it in my DPMI thread some months ago, I had
no reason to assume that it could exist



Licensing problems ???



Could be of some interest ... what are the features ?

- DPMI only or "extender" ?
- Residentable ?
- Stubable ?
- Memory (bad: cca 32 MiB | good: 4 or at least 2 GiB)
- DPMI version (minimal 0.9 , 0.9+ , 0.9++ , 1.0) ?
- CPL&IOPL ?
- Bits (16 , 32 , both) ?
- Zero / Non-Zero based
- EXE formats (COFFee , LE , PE , flat , ...) ?
- "truth" / "API" translation ?
- Reliability ?

If it is far behind HDPMI32 or DOS/32A then I am probably not smart
enough to turn it into a serious competitor by now

I'll update the FAQ ASAP

WARN: I've just split this from DOS FAQ

I just realized I haven't answered some important questions.
So here we go.


Might be. Some parts of PMODE code are still left there (just converted to ASM32/early fasm syntax).


Extender, with some additional own interrupt services (with int 48h, if I remember correctly)

No, because it was more a loader-extender variant, than just a DPMI service

Yes.

There were some small issues, but generally should go for a whole memory without problem.

0.9 with some of the 1.0 functionalities (perhaps it's one of the + or ++ variants you mentioned, I just don't know those conventions). It was also aspiring to be the "True DPMI" host - that is, with 32-bit MS-DOS services (only emulated, though).

Program was executed at ring 0 - just to keep the "DOS spirit".

32 bits only

Non-Zero based and this is the important flaw of it, and one of the reasons why I no longer use it.

My own HF format. It had HF DLL support, and HDOS was loading, relocating, and processing DLL imports - all from the stub loader. You could also load DLLs on fly with the HDOS interrupt services.

As I said above - generally yes; some of the less important DOS functions may be missed.

Well, never tested it on more than three-four machines. But on those it was very reliable.


No, I don't plan to revive it and even try to make it any serious comptetitor of today's nice and shiny DPMI hosts and extenders. It's just a piece of my coding history, also related to the history of fasm.

Thanks.

Tomasz wrote:

> No, I don't plan to revive it and even try to make it any serious comptetitor

Yeah ... you just previously neglected my statement that "no DOS extender or DPMI host compiles with FASM" and offered to "share" it. I don't really need it badly but if you release it I will download it and have a look. OTOH possible licensing problems ("Might be. Some parts of PMODE code are still left there") are an argument against releasing it.

> Program was executed at ring 0 - just to keep the "DOS spirit"

DPMI with 3/3 and GDT and IDT "open" also keeps it sufficiently

I doubt using parts of PMODE are an issue since it's no longer sold commercially. Lots of people have used PMODE for various things (e.g. XPL0, CC386). IIRC, even PMODE/W is free for commercial use in your own projects these days (as long as you don't try selling it outright as your own).

BTW, wasn't FASM at one time sorta using WDOSX? (I seem to remember some mention in the docs a long time ago.) It's stub-able, 32-bits, extended int 21h API, supports DLLs, ring 0, etc. And there are others similarly available too for assembly (e.g. D3X). At least the latter uses DJGPP + NASM (freely available tools), so maybe you could just tweak that for your old projects if you got really bored.

> doubt using parts of PMODE are an issue since

Well, rechecked:

Original PMODE: liberal BSD-like license

PMODE/W: closed source freeware

PMODE/DJ: GPL

> too for assembly (e.g. D3X). At least the latter uses DJGPP + NASM (freely available tools)

D3X extender itself is 100% ASM and uses NASM only ... DGJPP/C is for the tools

Yes, but you need something (e.g. WLINK ?? or its own STUBX) to bind the stub to the .EXE. And STUBX is compiled by DJGPP (but the actual extender core(s) themselves are fully NASM). It work with most DJGPP, Watcom (32-bit), and via pure assembly.

But, you, DOS386, claim that FreeBASIC tools don't like it. Oh well, nothing's perfect.

Appendix: I was reluctant to make HDOS public at the time of this discussion, but I finally released it few years later in this thread: http://board.flatassembler.net/topic.php?p=138815