Amiga OS 3.1 Source code leaked yesterday

[xArtx]

Why is there mention of Apple and IBM in it?

EDIT,, all to do with Apple printers & other hardware.

Originally Posted by Leffmann

Someone has already pointed out that the source code for SetPatch and other small things are missing from the archive, so it's definitely not complete.

Isn’t the nature of patches to fix bugs discovered after release?
Anything that should be in ROM missing?

[mark_k]

I've been looking through the files. It's not just AmigaOS 3.1, there is some older/historical stuff which is interesting to me. For example:

Source for the Amiga (1000) keyboard microcontroller: other/keyboard/keyboard.tar.z

ABasiC source code (ABasiC came with the first Amigas before AmigaBASIC): other/abasic/abasic.tar.z

The Tripos kernel of AmigaDOS (written in assembly language),
BCPL source for CLI commands and (maybe) the rest of AmigaDOS, changelogs: other/old_fs/old_fs.tar.z

What is probably some source for Metacomco Lisp (did Commodore market that?): other/lisp/lisp.tar.z

CD32 MPEG device source: other/cd32mpeg/

CD32 early ROMs (v40.5, 40.6)...


Apart from history, some parts could be useful, for example someone might want to debug/bug-fix scsi.device etc. Getting individual modules like that into a buildable state would be much less work than the entire OS source tree.

[ptyerman]

The problem is, anyone that touches it will be opening themselves up to a world of legal hurt! There is no way anything could be derived from these sources legally.

[xArtx]

It’s a hellavalotmore than OS!
Just about anything I can think of is there, references to A590, Amigaterm.. what is that? was there ever a terminal from Commodore?.
the CD authoring tools for CDTV and CD32 (tho I know they were out there). IDE & SCSI controllers it looks like too.
It looks more like a copy of an entire media that someone had.

[jotd]

That leak is good news for the Amiga community.

You can look at it, figure out what the bug is, then try to patch the binary file in ASM just like you reverse engineered it, except that you won't waste time understanding ASM code without symbols.
There will be no proof that you used the source, hence it will be as legal as other ROM fixes done without the source! (we do it all the time in whdload slaves)

What you cannot release is a compiled version from modified sources! (although you can create one for test purposes)

Hey didn't Olaf write the Action Replay MKIII code? back in 1992 it just rocked man, I still have mine somewhere.

[ptyerman]

There was a terminal program from Commodore, I have a adf of it here, (AmigaTerm v1.03 (1986)(Commodore).

[Olaf Barthel]

Originally Posted by jotd

That leak is good news for the Amiga community.

You can look at it, figure out what the bug is, then try to patch the binary file in ASM just like you reverse engineered it, except that you won't waste time understanding ASM code without symbols.
There will be no proof that you used the source, hence it will be as legal as other ROM fixes done without the source! (we do it all the time in whdload slaves)

What you cannot release is a compiled version from modified sources! (although you can create one for test purposes)

I would not recommend this approach. First thing, now that the archive is in the open, everybody would naturally assume how you came by the information on updating the code (why? it's the easiest way), putting you on the defensive to prove that you achieved this by different means. Hyperion is bound by contract to defend the AmigaOS 3.1 source code, as opposed to the game companies which issued what is now considered abandonware for the Amiga: they don't care, Hyperion does.

The other reason being that you'll have the devil of a time testing the changes, verifying that they indeed work on all variations of the hardware which the code shipped for. This is difficult even for guys such as myself who did just this sort of thing for Amiga Technologies GmbH and then for the AmigaOS 3.5 and 3.9 updates.

You break somebody else's setup, you are going to regret it

Hey didn't Olaf write the Action Replay MKIII code? back in 1992 it just rocked man, I still have mine somewhere.

No, wasn't me. There were only a few developers active in the Amiga field by the name of "Olaf", but I never was involved with the Action Replay cartridge.

[idrougge]

Originally Posted by ptyerman

Fixes are always nice I must admit. But most of the things I can think of are already been worked on by Cloanto, large drive support for example. It's obviously been available a long time as a DIY fix but it's nice to see a kickstart ROM released with it already fixed.
Perhaps drop them a line with a list of specific fixes you would like to see?

Perhaps Workbench support for displaying the correct sizes for large partitions?
Oh, but Cloanto instead removed workbench.library entirely.

Originally Posted by Olaf Barthel

Andy Finkel worked as a consultant for Amiga Technologies GmbH in 1995, and as far as I know this would include getting the Amiga operating system code into a usable form again, suitable for product development.

Interesting, I thought Finkel disappeared out of the picture no later than 1994.

Originally Posted by Olaf Barthel

Andy Finkel did something for me which literally changed my life. He pointed me to two books which were new at the time, these being "Debugging the development process" by Steve Maguire and "Dynamics of software development" by Jim McCarthy (Steve Maguire's other book "Writing solid code" blew my mind). Reading these books I realized that there was a professional practice for developing quality software. Andy set me on the path of becoming a professional programmer.

That's also interesting. I think most people in this community would have regarded you as a professional programmer already, but what you're saying is that Commodore actually employed professionals and not just semi-professionals.

Originally Posted by Akira

I for one don't give a shit and hope Hyperion fold and stop trying to cash in on the Amiga scene. I love the work of people trying to fix and improve the ROM and its functions and I hope it keeps on happening unconstrained by bullshit legal matters.

Hyperion? Surely you mean Cloanto? Hyperion (or rather, their contractors) have their own codebase developed by a large effort of their own.

[gulliver]

Originally Posted by ptyerman

There was a terminal program from Commodore, I have a adf of it here, (AmigaTerm v1.03 (1986)(Commodore).

This one is AmigaTerm v1.2k (1990) (Commodore), binary and source

[ptyerman]

No good telling me, I have nothing to do with it.
I suppose they had to take something out the ROM to make some room, at least it's not a problem to stick it in the libs drawer on hard disk. It will make things more difficult when making floppy disk compilations though because of disk space.

[gulliver]

There is also in the leaked archive:

Bridgeboard utilities with source code (PC utilities require MASM to compile, which is freeware)
Envoy earlier releases up to 1.7 with source code
Early TCP-IP Commodore stack code
AmigaVision 2.5 with source code

There is a very wonderful document called futprod.doc (Word document from november 5th 1993). That describes a lot of the projects
and the insights of Commodore before its demise. It provides a great description of the roadmap they were heading.
A lot of it revolved around several CD32 variants, the new PA-RISC architecture and what AAA could and couldn´t do.
Amongst other things, it clearly portrays why Haynie´s AAA project had little or no future. Also, there is mention
of hardware that we never knew existed (eg.: PA-RISC/3D early development, Motivator RTG graphics board)
It also mentions AmigaOS 4.0 was all about RTG for standard Amigas and for AAA and that part of 4.0 work in progress
is the little v42 sourcecode there is in the leaked archive.

[Olaf Barthel]

Originally Posted by idrougge

Interesting, I thought Finkel disappeared out of the picture no later than 1994.

I met several ex-Commodore employees at Amiga Technologies GmbH back then. This was the time when everybody was sort of between jobs, and consulting gigs helped to pay the bills. The Amiga technology was a highly specialized field and those companies who needed your help needed it badly. Not a bad time to be a consultant

That's also interesting. I think most people in this community would have regarded you as a professional programmer already, but what you're saying is that Commodore actually employed professionals and not just semi-professionals.

In product development and engineering, yes, in management, well... not so much after Jack Tramiel quit.

As for "professional programmer", I learned that there can be a world of a difference between programming as a vocation/profession and professional conduct and practices while working as programmer/software developer.

I still believe that developing software is just about the greatest thing ever, but to last in the business, you cannot coast on self-taught practices. You need to learn how to organize your work, how to get better at avoiding mistakes, educate yourself (it helps to have a CS degree, but as contemporary research shows, self-taught software developers tend to be much more productive than those who started with a CS degree and no prior programming experience).

Otherwise you will invariable hit your limits and either burn out or come think that banging your head against the limits is "normal". I met too many programmers in those past 20 years who consider this type of self-inflicted damage the "cost of doing business".

[modrobert]

Originally Posted by gulliver

The leaked source code contains both the Green Hills Compiler version and Peter Cherna´s SASC v6.x intuition.library version.

Originally Posted by gulliver

There is also in the leaked archive:

Bridgeboard utilities with source code (PC utilities require MASM to compile, which is freeware)
Envoy earlier releases up to 1.7 with source code
Early TCP-IP Commodore stack code
AmigaVision 2.5 with source code

There is a very wonderful document called futprod.doc (Word document from november 5th 1993). That describes a lot of the projects
and the insights of Commodore before its demise. It provides a great description of the roadmap they were heading.
A lot of it revolved around several CD32 variants, the new PA-RISC architecture and what AAA could and couldn´t do.
Amongst other things, it clearly portrays why Haynie´s AAA project had little or no future. Also, there is mention
of hardware that we never knew existed (eg.: PA-RISC/3D early development, Motivator RTG graphics board)
It also mentions AmigaOS 4.0 was all about RTG for standard Amigas and for AAA and that part of 4.0 work in progress
is the little v42 sourcecode there is in the leaked archive.

Yes, lots of interesting stuff, not to mention if you do a 'grep -r -i copyright *' to realize there is a lots of source code from other companies, with comments like "remove IBM garbage" as the code development progresses.

[xArtx]

Determination attracts the required help sometimes, so you might not bang your head so much in the hobby arena.
I already know that adopting someone else’s goals for money would ruin the whole thing though, so better to be a labourer.
It’s probably because self taught programmers were labourers

[Leffmann]

Originally Posted by xArtx

Isn’t the nature of patches to fix bugs discovered after release?
Anything that should be in ROM missing?

Yeah, but SetPatch had to be installed with the Workbench so you could easily update it later, and there was just one SetPatch which knew how to fix bugs in all prior releases. The one that came with Workbench 3.1 mainly fixed bugs in Kickstart 3.0.

I don't have the archive myself, and I don't know what else is missing.

[xArtx]

The C commands might only be what’s present in the ROM. I wouldn’t know which ones they are.

That document is interesting, they go ahead and say “We’ll release old games on our new console” in more formal language.

[8bitbubsy]

There's a lot of foul language in the source code. Not surprising, but fun!

This is my favorite:

#if WHAT_THE_FUCK // ????

[nogginthenog]

Originally Posted by 8bitbubsy

There's a lot of foul language in the source code. Not surprising, but fun!

This is my favorite:

#if WHAT_THE_FUCK // ????

I see I'm not the only one who looked for this :-)
There's even a fuckup.c!

[kamelito]

Originally Posted by Leffmann

Yeah, but SetPatch had to be installed with the Workbench so you could easily update it later, and there was just one SetPatch which knew how to fix bugs in all prior releases. The one that came with Workbench 3.1 mainly fixed bugs in Kickstart 3.0.

I don't have the archive myself, and I don't know what else is missing.

Believe it or not someone put all the code to github.
Kamelito

[8bitbubsy]

Originally Posted by kamelito

Believe it or not someone put all the code to github.
Kamelito

The leaked, incomplete one?