Amiga.org
Amiga computer related discussion => General chat about Amiga topics => Topic started by: persia on August 13, 2009, 04:13:36 PM
-
Just a curiosity question about Amiga OS 4.1. I thought all the source code to Amiga OS disappeared in the bankruptcy/insolvency movements of it's owners. What is the kernel that Amiga OS 4.1 is based on? Is it using the same kernel as Morphos (Quark)? Or the AROS kernel?
Is it using AROS code?
-
It's based on AmigaOS4.0! :-)
What it's based on is is the OS3.1 sourcecode and whatever OS3.9 sourecode Amiga managed to scrap together for Hyperion.
It isn't the MorphOS kernel, otherwise there would be no red team and big team competing to be The Sorest Looser. :-D
And AROS? Are you serious? Hehe. LOL! :-D In this case it would be an Amiga inspired kernel, but not The Amiga kernel. And you would see no OS4 on x86 threads. I regret saying that "phrase" but it was a logical progression.
-
I thought it was based on sunshine and lollipops
-
OS 4.1 (http://amigakit.leamancomputing.com/catalog/advanced_search_result.php?keywords=4.1) is based on 3.1 sourcecode and 3.9 sourcecode but with a new replacement ExecNG which was developed by the Frieden brothers. This replaces the older Exec with a more modern/scalable implemetation.
-
What it's based on is is the OS3.1 sourcecode and whatever OS3.9 sourecode Amiga managed to scrap together for Hyperion..
I wonder how much 3.5/3.9 code they did manage to re-use given all the legal issues with H&P; I know that Amidock is a completely new application, for example, but I'm guessing they somehow managed to get the code for things like ReAction...
-
A lot of the team that contributed to 3.9, also work on 4.1
-
I wonder how much 3.5/3.9 code they did manage to re-use given all the legal issues with H&P; I know that Amidock is a completely new application, for example, but I'm guessing they somehow managed to get the code for things like ReAction...
According to the Friedens, very little of the original 3.1 source code was usable (search ann.lu for the quotes). Which would make sense given that it was all 68k asm.
-
It's based on AmigaOS4.0! :-)
What it's based on is is the OS3.1 sourcecode and whatever OS3.9 sourecode Amiga managed to scrap together for Hyperion.
It isn't the MorphOS kernel, otherwise there would be no red team and big team competing to be The Sorest Looser. :-D
And AROS? Are you serious? Hehe. LOL! :-D In this case it would be an Amiga inspired kernel, but not The Amiga kernel. And you would see no OS4 on x86 threads. I regret saying that "phrase" but it was a logical progression.
Don't comment on things you don't understand. AROS is not Amiga inspired, it is an AmigaOS clone. You imply that AROS is a modern kernel, made to look like AmigaOS it isn't. Bear in mind that MorphOS used AROS code, and there has been quite a bit of code sharing between the projects.
-
I thought it was based on sunshine and lollipops
Funny, JJ, but I thought that was what's coming out of your arse! =D
-
According to the Friedens, very little of the original 3.1 source code was usable (search ann.lu for the quotes). Which would make sense given that it was all 68k asm.
Huh?
-
Don't comment on things you don't understand.
Which would make sense given that it was all 68k asm.
Uhh.. except for all the C code. Which required 3 different C compilers in its original state :-)
They ditched Exec, and ditched large chunks of some other modules, but the code was not as useless as you imply.
-
Uhh.. except for all the C code. Which required 3 different C compilers in its original state :-)
And almost all the C stuff was from 3.5 and 3.9.. I was talking about 3.1 (as I stated in my original post).
They ditched Exec, and ditched large chunks of some other modules, but the code was not as useless as you imply.
Not useless, but only a little more useful than the RKMS... And contrary to what some people think, OS4.x is as new as MorphOS and AROS, these people should think about that before they make comments.
-
Huh?
Highlight the parts of my post you are having difficulty reading, and I will elaborate for you.
-
It's based on AmigaOS4.0! :-)
What it's based on is is the OS3.1 sourcecode and whatever OS3.9 sourecode Amiga managed to scrap together for Hyperion.
Could you please refrain from making assumptions? I do recall hearing that Amiga Inc has no source code. And thats from unofficial contacts. I could be wrong as this was 4 years ago.
OS4/4.1 is based indeed on source code from os3.1/3.9. But it came from another source, not A Inc.
-
AROS is not Amiga inspired
Do you mean to say it wasn't "Amiga Inc." inspired? Certainly its inspired by the AmigaOS
Plaz
-
Do you mean to say it wasn't "Amiga Inc." inspired? Certainly its inspired by the AmigaOS
Plaz
No, I mean it is a clone. A bug for bug clone. If it were just "inspired", then you wouldn't clone the braindead single address space design for example.
-
Ah, I follow your meaning now.
Plaz
-
Based on years of vapourware too, sadly.
-
Highlight the parts of my post you are having difficulty reading, and I will elaborate for you.
Ok.
Which would make sense given that it was all 68k asm.
No difficulty reading it, I just think someone may have fed you some BS, maybe you could tell us where you got this info from? A link perhaps, to some reliable source?
-
Could you please refrain from making assumptions? I do recall hearing that Amiga Inc has no source code. And thats from unofficial contacts. I could be wrong as this was 4 years ago.
OS4/4.1 is based indeed on source code from os3.1/3.9. But it came from another source, not A Inc.
I head much the same thing, so what did happen to the 3.1 source code...it's it still on a couple of floppy disks in a dusty filing cabinets at Gateway? Or did it (at least officially) get 'lost in the post' along the way after Commodore exploded? Who actually owns it these days?
Our of interest. How much of the 3.1 code was available to the 3.5 developers, I wonder if anything was missing that had to be re-done from scratch.
-
OS4/4.1 is based indeed on source code from os3.1/3.9. But it came from another source, not A Inc.
It came from Olaf Barthel.
http://www.merlancia.us/amiga-hyperion/26exhibit8show_case_doc.pdf
-
I wonder how much 3.5/3.9 code they did manage to re-use given all the legal issues with H&P; I know that Amidock is a completely new application, for example, but I'm guessing they somehow managed to get the code for things like ReAction...
A few things are are missing. Arexx is 68K, and likely to remain that way since they only had a license to the 68K source code, and can only go as far as debugging that same source code. Hyperion brought Python in, but it just doesn't look as easy as AmigaDOS and Arexx scripts. Not to matter, as even if Arexx gets bumped off, programs can still support Arexx ports and commands. Since it is just a standard message port and messages in Arexx format; just a "string in a structure" really, so easily kept going.
The Installer was lost, although they have made their own for certain things; thus the OS4 installer can't be used to install OS3.9 programs that make use of it's multimedia or other features for instance.
The Printer modules were lost. They had to spend time reimplementing features there and not all were put in. So the OS4 printer.device is not perfectly backwards compatible to the OS3.9 one .
I thought Palette used to be 68k, it's now native. IIRC it had some screen colour bug. It's still not right, as it looks dated and has no colour wheel? :-? Perhaps that was the sacrifice for becoming native.
There there is Unarc, some decompressors are 68k.
The Shell. Well I never liked the ones used to replace it in the past. And always thought they should just bite the bullet and get the sources to KingCON. Well now I guess it's too late and the Shell we have is slightly better than what we had in OS3 and 100 times better than what you get in WindowsXP!
From my above descriptions it would look like AmigaOS4 is a big mess! I would have liked a more cleaner rebirth to the Amiga, but it was pulled apart after Commodore already broken it. so this is what we ended up with. It still here at least. :-)
-
Don't comment on things you don't understand. AROS is not Amiga inspired, it is an AmigaOS clone. You imply that AROS is a modern kernel, made to look like AmigaOS it isn't. Bear in mind that MorphOS used AROS code, and there has been quite a bit of code sharing between the projects.
I understand all right. There may have been no reverse engineering, but it's obvious AROS uses the AmigaOS include files to recontruct it in portable C. Now if that isn't some serious inspiration I don't know what is!
-
I understand all right. There may have been no reverse engineering, but it's obvious AROS uses the AmigaOS include files to recontruct it in portable C. Now if that isn't some serious inspiration I don't know what is!
AROS doesn't use the AmigaOS includes, they are copyrighted... Come on keep up, you are showing your ignorance here. And, AROS is reversed engineered, and using the publically available docs.
-
Funny, JJ, but I thought that was what's coming out of your arse! =D
?????????????????
-
AROS doesn't use the AmigaOS includes, they are copyrighted... Come on keep up, you are showing your ignorance here.
Then where does it get all its structures from? You say AROS is a clone, but if it is a clone then it needs AmigaOS includes to be one. If it doesn't use AmigaOS includes then we are stuck. Where do the AROS system includes come from?
And, AROS is reversed engineered, and using the publically available docs.
So you mean to say that the writers examined 68k disassemblies to see what the code does and convert that to C? If so that's not much better than using copyrighted files as your basis.
I thought it would have been easier for them if they used the AmigaOS include files and autodocs as their basis. Reimplementing all the core functions themselves by needessly reinventing the wheel and writing code to mirror the Amiga API behavour.
-
So you mean to say that the writers examined 68k disassemblies to see what the code does and convert that to C?
Yes, exactly.
If so that's not much better than using copyrighted files as your basis.
Legally there is a difference. If you don't believe me, consult a copyright lawyer.
-
Yes, exactly.
:-O
Legally there is a difference. If you don't believe me, consult a copyright lawyer.
I thought reverse engineering would be worse. What's more, I thought the basic system structures would be copied, since if you want to compile Amiga C programs you need those same systtem structures anyway.
I thought AROS would be done just reading docs and public structures and recreating it. Well that seemed the clean approach to me, if you don't want legacy issues, and a better framework for forward moving features.
-
I thought..
.. I thought ..
I thought ..
That whole "imagination rules the world" thing isn't always true.
-
:-O
I thought reverse engineering would be worse. What's more, I thought the basic system structures would be copied, since if you want to compile Amiga C programs you need those same systtem structures anyway.
All system structures are copied, otherwise AmigaOS software wouldn't compile. But they must be written from scratch by someone not involved with the original OS.
I thought AROS would be done just reading docs and public structures and recreating it. Well that seemed the clean approach to me, if you don't want legacy issues, and a better framework for forward moving features.
AROS is a bug for bug clone to ensure as much software compatibility as possible.
AmigaOS design does not make any distinction between the Operating system and the User space, so you have to copy the system design right down to a fundamental level to get any sort of compatibility with existing software.
AROS copies the AmigaOS totally and suffers the same legacy issues that plague the original... though a lot of work has been done to try and mitigate them.
I feel the MorphOS team had the right idea to use a microkernel and then run an Amiga flavour (Like AROS hosted) on top.
That said I am an AROS guy, I like to run my OSes on cheap x86/x86-64 hardware.
-
And almost all the C stuff was from 3.5 and 3.9.. I was talking about 3.1 (as I stated in my original post).
Uhh... yeah. Haage & Partner decided to use C compilers for Sun workstations. Oh no, that would be commodore...
-
It came from Olaf Barthel.
http://www.merlancia.us/amiga-hyperion/26exhibit8show_case_doc.pdf
That is also what I heard, just didn't feel confident enough to mention this. ;-)
-
Uhh... yeah. Haage & Partner decided to use C compilers for Sun workstations. Oh no, that would be commodore...
Yes of course, that's why I didn't say ALL source code... but now you are just pointlessly pointing out irrelevancies!
-
Don't comment on things you don't understand. AROS is not Amiga inspired, it is an AmigaOS clone. You imply that AROS is a modern kernel, made to look like AmigaOS it isn't. Bear in mind that MorphOS used AROS code, and there has been quite a bit of code sharing between the projects.
One might expect a Windows clone to run software written for Windows -- or an AmigaOS clone to run software for ? (just a harmless stab)
-
Yes of course, that's why I didn't say ALL source code... but now you are just pointlessly pointing out irrelevancies!
Oh ok.
According to the Friedens, very little of the original 3.1 source code was usable (search ann.lu for the quotes). Which would make sense given that it was all 68k asm.
all
-
One might expect a Windows clone to run software written for Windows -- or an AmigaOS clone to run software for ? (just a harmless stab)
It tries to be a clone of the operating system, not the underlying hardware, which is also a requirement for binary level compatability with Amiga applications.
-
But aros isn't binary compatiable so what your point ?
-
One might expect a Windows clone to run software written for Windows -- or an AmigaOS clone to run software for ? (just a harmless stab)
Talking about OS clones, please try running apps for older commercial Unices on Linux, and they won't work: AROS aims to be source-level compatible with AmigaOS on different platforms, and binary-compatible on the same (68K) platforms, but since there is no AROS for 68K in a standalone shape, nobody can say exactly how the two are effectively compatible. Some parts of AROS are available for AmigaOS 3.x, though, to enhance it (look at bernd's AFA-OS), and they are ironed out to mantain compatibility with existing 68K software.
And, just to hit the nail on the coffin of this discussion, just try to port ReactOS on a PPC motherboard, and look if x86 binaries will run on it...
-
@bloodline
"that it was all 68k asm."
you are wrong. It used quite a lot of C code. The low level stuff used asm but you aren't going to use a600 ide driver with an A1...
-
@bloodline
"that it was all 68k asm."
you are wrong. It used quite a lot of C code. The low level stuff used asm but you aren't going to use a600 ide driver with an A1...
Not just C, but also BCPL (http://en.wikipedia.org/wiki/BCPL).
So, they yanked out all the ASM, redid the C stuff to deal with differences in cpu arch, pulled the BCPL...
There really isn't a whole lot left...
-
Not just C, but also BCPL (http://en.wikipedia.org/wiki/BCPL).
So, they yanked out all the ASM, redid the C stuff to deal with differences in cpu arch, pulled the BCPL...
There really isn't a whole lot left...
The BCPL stuff was mostly down to dos.library and it's TripOS heritage, if I recall. I don't think there was a great deal of TripOS code in it but remnants such as BPTR stem from it. The dos.library required a total rewrite for OS4 anyway.
-
Do you mean to say it wasn't "Amiga Inc." inspired? Certainly its inspired by the AmigaOS
Plaz
I liked JJ's answer best. As an analogy I'm guessing it's pretty spot on. But that half licked lolly pop is stuck underneath someone's shoe... somewhere...
Gertsy