Poster | Thread |
Hypex
 |  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 24-May-2023 13:50:36
| | [ #21 ] |
|
|
 |
Elite Member  |
Joined: 6-May-2007 Posts: 11351
From: Greensborough, Australia | | |
|
| @kolla
Yes I know, but I recall he was working for IBM at the time or has some association. He is said to help draft the ANSI standard for REXX. He wrote an implementation for the Amiga so I'm sure how this would affect any other versions as REXX itself comes under the IBM banner and produced by Mike Cowlishaw. Suppose it similar to BASIC that one party invents it and others implement there idea of it for computers.
Interestingly or strangely, the wiki entry for Arexx speaks of how it's written in 68000 and so won't run at full speed on a PPC. It then talks about ARexx and MorphOS. I don't see how this directly relates since MorphOS isn't AmigaOS. It reads like a random snippet of info. It doesn't even mention AmigaOS4 for some context.
https://en.wikipedia.org/wiki/ARexx |
|
Status: Offline |
|
|
Hypex
 |  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 24-May-2023 13:52:01
| | [ #22 ] |
|
|
 |
Elite Member  |
Joined: 6-May-2007 Posts: 11351
From: Greensborough, Australia | | |
|
| @tlosm
Right now a RPi is about as uncommon as a PPC and A1222 combined.  |
|
Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 24-May-2023 15:27:01
| | [ #23 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 12960
From: Norway | | |
|
| @Hypex
ARM has 32 GPRS and so do PowerPC.
So should be possible to do a 1 to 1 mapping of the registers, more or less. So if you don't care about support PowerPC libraries.
Then you can do an task-based JIT compiler, that simply exit and calls the native API on an illegal opcode, after calling native API, library calls a return and ends up in the emulators API exception call, and then returns to JIT buffer.
That’s if the host OS has same data structures. (32bit big endian) (so the only real difference here is opcodes and operands.)
So actually the PowerPC support can be added later. _________________ http://lifeofliveforit.blogspot.no/ Facebook::LiveForIt Software for AmigaOS |
|
Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 24-May-2023 18:05:25
| | [ #24 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 12960
From: Norway | | |
|
| @Hypex
never mind, don't need illegal opcode, or API exception call. Because we won’t need to translate any addresses, or registers.
We can jump direct from the JIT cache, and call API, and return to the JIT cahce. Because at that point everything is ARM code.
Of course, it be more complicated if the program is 32bit and OS is 64bit, but even then I’m not sure. (It’s just the legacy stuff, and shared stuff need to be in 0x00000000 -> 0x7FFFFFFF range.) Last edited by NutsAboutAmiga on 24-May-2023 at 06:09 PM. Last edited by NutsAboutAmiga on 24-May-2023 at 06:06 PM.
_________________ http://lifeofliveforit.blogspot.no/ Facebook::LiveForIt Software for AmigaOS |
|
Status: Offline |
|
|
mac6
|  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 26-May-2023 23:10:11
| | [ #25 ] |
|
|
 |
Member  |
Joined: 12-Jan-2021 Posts: 23
From: Western Australia | | |
|
| Well done! That's a very good sample size and interesting results. |
|
Status: Offline |
|
|
kolla
|  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 27-May-2023 8:38:17
| | [ #26 ] |
|
|
 |
Elite Member  |
Joined: 20-Aug-2003 Posts: 3357
From: Trondheim, Norway | | |
|
| @Hypex
Quote:
Hypex wrote: @kolla
Yes I know, but I recall he was working for IBM at the time or has some association.
|
I’ve seen nothing that indicates this, other than people mention it like you do here.
Quote:
He is said to help draft the ANSI standard for REXX. |
His name is right there in the published X3.274-1996 which one can buy from ANSI.
https://rexxla.org/rexxlang/standards/j18pub.pdf
Quote:
He wrote an implementation for the Amiga so I'm sure how this would affect any other versions as REXX itself comes under the IBM banner and produced by Mike Cowlishaw. Suppose it similar to BASIC that one party invents it and others implement there idea of it for computers. |
https://home.rexxla.org/presentations/1995/rexx95-001.pdf
There were (and are) many implementations of Rexx, they are not all “associated” with IBM! Like Regina which was written by a friend of mine. Don’t know if you’ve ever been working with standard development processes (and I only have experience with IETF), but typically the whole point of assembling many people (like Bill Hawes) to make a standard is exactly to have inputs from other than just one party (in this case IBM.). Bill Hawes didn’t just write ARexx, he also made CONMan, WShell and various other Amiga stuff that at the time was rather significant (TraceDump??) According to Randell Jesup (on some forum, cannot find it now) he also did quite a bit of contract work for Commodore Amiga, and had been asked to join the team, but didn’t want to move away from Boston. He is mentioned explicitly in the “easter egg” About window of OS 3.1.
Quote:
Interestingly or strangely, the wiki entry for Arexx speaks of how it's written in 68000 and so won't run at full speed on a PPC. It then talks about ARexx and MorphOS. I don't see how this directly relates since MorphOS isn't AmigaOS. |
Er… and neither is Hyperion OSes then. MorphOS certainly evolved from AmigaOS. For quite some time, MorphOS was the primary candiate to become OS4, and the reason it didn’t is all about egos/money/drama/politics/social and nothing technical.
Quote:
It reads like a random snippet of info. It doesn't even mention AmigaOS4 for some context. |
Maybe because Hyperiion OS4 isn’t relevant in any context? :)
The situation for OS4 is exactly the same as for MorphOS in regard to ARexx.Last edited by kolla on 27-May-2023 at 08:38 AM.
_________________ B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC |
|
Status: Offline |
|
|
Hypex
 |  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 27-May-2023 13:54:18
| | [ #27 ] |
|
|
 |
Elite Member  |
Joined: 6-May-2007 Posts: 11351
From: Greensborough, Australia | | |
|
| @NutsAboutAmiga
Quote:
ARM has 32 GPRS and so do PowerPC. |
From what I read it has 31 registers so slightly less oddly enough but close enough.
Quote:
So should be possible to do a 1 to 1 mapping of the registers, more or less. |
It would except for one thing. ABI. PPC uses SysV so since they follow standards I'd expect they would follow the ARM ABI. Which I would expect to differ slightly but register remapping should be possible.
Quote:
So if you don't care about support PowerPC libraries. Then you can do an task-based JIT compiler, that simply exit and calls the native API on an illegal opcode, after calling native API, library calls a return and ends up in the emulators API exception call, and then returns to JIT buffer. |
That's how it is done on 68K. At least in the static emulator. Not sure how it jumps from native JIT block through 68K function calls to native code.
Quote:
We can jump direct from the JIT cache, and call API, and return to the JIT cahce. Because at that point everything is ARM code. |
Except PPC in JIT and even non-JIT would still need PPC interface just like 68K has 68K jump table.
Quote:
Of course, it be more complicated if the program is 32bit and OS is 64bit, but even then I’m not sure. (It’s just the legacy stuff, and shared stuff need to be in 0x00000000 -> 0x7FFFFFFF range.) |
Though it would be good, at that point, it would need a whole new OS redesign. They would need a whole new API, new includes and new documentation. Somewhat needed 20 years ago really but that's how things go. |
|
Status: Offline |
|
|
Hypex
 |  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 27-May-2023 15:11:01
| | [ #28 ] |
|
|
 |
Elite Member  |
Joined: 6-May-2007 Posts: 11351
From: Greensborough, Australia | | |
|
| @kolla
Quote:
I’ve seen nothing that indicates this, other than people mention it like you do here. |
I think I can explain this, or I can from my perspective. So, we have Mike Cowlishaw, who designed the Rexx language. He was working for IBM at the time and IBM embraced the language. Most searches on Rexx will turn up articles on IBM pages.
Mr. Hawes writes an Amiga implementation known as ARexx that in due course is adopted by Commodore as an official OS scripting language. This then brings two scripting languages into use; DOS scripts for simple scripting and ARexx for more advanced scripting with OS level hooks. It also introduces the revolutionary ARexx port. As well as being a replacement for the depreciated AmigaBASIC.
Mr. Hawes is also involved in drafting the Rexx ANSI standard. This brings the association with IBM. Or an indirect association really. So he wouldn't have been directly working for IBM. But with the link between IBM and Rexx, Mr. Hawes would be associated with IBM. That's how I would make the connection. Which would just be by association. 
Quote:
His name is right there in the published X3.274-1996 which one can buy from ANSI. |
Yes it is. Addressed as Bill. That's a rather large document.
Quote:
There were (and are) many implementations of Rexx, they are not all “associated” with IBM! Like Regina which was written by a friend of mine. Don’t know if you’ve ever been working with standard development processes (and I only have experience with IETF), but typically the whole point of assembling many people (like Bill Hawes) to make a standard is exactly to have inputs from other than just one party (in this case IBM.). Bill Hawes didn’t just write ARexx, he also made CONMan, WShell and various other Amiga stuff that at the time was rather significant (TraceDump??) According to Randell Jesup (on some forum, cannot find it now) he also did quite a bit of contract work for Commodore Amiga, and had been asked to join the team, but didn’t want to move away from Boston. He is mentioned explicitly in the “easter egg” About window of OS 3.1. |
Funky cold Regina? I've likely seen the egg but didn't recall the contents. I've not worked with standard development processes in this case. But I can see the benefit of peer review regarding a large design such as a language.
What I would be unclear on, is the commercial aspects. Especially given Mike was working for IBM and any IP claims they would make on Rexx. There can be a fine line when designing hardware or software while working for a company. The most well known example would surely be BASIC. I can't recall who invented the language but Microsoft made it so popular, producing versions of it, that people commonly thought Bill Gates invented it. Of course he didn't, but he did make money off it.
So, the elephant in the room for me then is, for the people that invented these languages; did they receive any kind of financial compensation, be that profit or royalties?
Quote:
Er… and neither is Hyperion OSes then. MorphOS certainly evolved from AmigaOS. For quite some time, MorphOS was the primary candiate to become OS4, and the reason it didn’t is all about egos/money/drama/politics/social and nothing technical. |
OS4 is built off the OS3.1 sources and translated from. That makes it a relation. Without that it wouldn't be a relation at all. MorphOS isn't built off the official sources and can only copy it based off the include files, just like AROS. And from usage it lacks standard AmigaDOS quirks and operation.
Quote:
Maybe because Hyperiion OS4 isn’t relevant in any context? :) |
It's relevant as a descendant of AmigaOS, even a bastard descendant as it may be. It mentions PPC which has no relevance to ARexx. So it makes sense to mention OS4 since it makes no mention of running Arexx within PowerUP or WarpOS.
Quote:
The situation for OS4 is exactly the same as for MorphOS in regard to ARexx. |
I've read that MorphOS doesn't include ARexx which is strange. OS4 includes it as standard as expected. That's not the same.
Supporting an Arexx port is easier. It's just a message port that can receive commands in a standard ARexx structure. ARexx isn't even needed to process it. |
|
Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 27-May-2023 16:24:45
| | [ #29 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 12960
From: Norway | | |
|
| |
Status: Offline |
|
|
Kronos
|  |
Re: Last Chance: AmigaOS 4.x hardware vs emulation survey Posted on 27-May-2023 16:54:36
| | [ #30 ] |
|
|
 |
Elite Member  |
Joined: 8-Mar-2003 Posts: 2713
From: Unknown | | |
|
| @Hypex
Quote:
Hypex wrote:
I've read that MorphOS doesn't include ARexx which is strange. |
ARexx was licensed for AmigaOS2.0 afterwards the author took issue with C= shipping it with 2.1/3.0/3.1. C= went bust and the issue has never been taken to court.
MorphOS does support ARexx, but does not come with the binary which can be copied over from an AmigaOS install (at which point it is as iffy as the ones shipped with H&P or Hyperion versions of AOS).
All ARexx ports can be accessed from LUA which a much more modern language with a clear license._________________ - We don't need good ideas, we haven't run out on bad ones yet - blame Canada |
|
Status: Offline |
|
|