Hi!
I'm trying to install AROS-20251213-pc-i386-boot-iso to either a VM or an old Atom based machine, but GRUB gives me the following errors:
error: no multiboot header found
error: you need to load kernel first
error: you need to load kernel first
error: you need to load kernel first
error: you need to load kernel first
This error has been present for some time now as I have not found a version that boots in the last few weeks.
Anybody got any idea how to fix it so I can install the latest clean version?
In fact, this ISO is also incompatible on my Pentium PC; with VMware, I get a black screen.
Please note that your ISO is AROS ABI-v1, which is incompatible with AROS distributions and almost all software stored on Aros Archive. There is no third-party software for AROS ABI-v1.
Distributions such as my AROS One are based on ABI-v0 and ABI-v11, You could try these Nightly Build here:
@AMIGASYSTEM - In fact, this ISO is also incompatible on my Pentium PC; with VMware, I get a black screen.
Please note that your ISO is AROS ABI-v1, which is incompatible with AROS distributions and almost all software stored on Aros Archive. There is no third-party software for AROS ABI-v1.
Distributions such as my AROS One are based on ABI-v0 and ABI-v11, You could try these Nightly Build here:
Since it's a nightly build, the date will be different.[/quote]
Kalamatee is fixing it. I would recommend hopping in the AROS slack channel and reporting bugs there too as Kalamatee doesn't look at arosworld often and if you want ABIv1 fixed he's who you need to notify.
I don't know, I have plenty of third-party software on my ABIv1 system
There is relatively too little software available on the Internet for ABIv1 compared to the software available for ABIv0 and ABIv11.
If there were more software available, I would have already created a version of AROS One ABIv1.
That is the fault of the distro maintainers and abiv0 in the first place. and with you people continuing to push a fork it will continue to be the case. I know why the fork was created I don't agree with it but I understand it. but as I said I have plenty on my system and all of what I own shall stay firmly in ABIv1 land if it means I lose customers so be it I spent $1000's on AROS probably more than almost anyone in the community and I stand by my decision to keep supporting AROS main branch and only the main branch.
Edited by terminills on 19-12-2025 09:17, 29 days ago
I have nothing against ABIv1, but creating a distribution with only Night makes little sense. As I said, if there were software available, I would immediately create an ABIv1-based distribution. Years ago, I created an ABIv1-based AROS One distribution, but then, since there was no software, I decided to abandon it.
The same thing happened with AROS 68k. There was no software available, and the GCC compiler on the main AROS is not executable. On the fork, however, the executable works but crashes, so for these reasons, you cannot blame the distribution maintainers.
I just downloaded the ABIv0 image and this one works.
Now, I wonder why the ABIv1 team never checked to see if their fork worked in the first place?
With people disliking Windows more and more, hardware prices going up and up, and the excess of perfectly good hardware floating around, I see it as a great opportunity for AROS to become an interesting alternative for the oldest gear.
I have nothing against ABIv1, but creating a distribution with only Night makes little sense. As I said, if there were software available, I would immediately create an ABIv1-based distribution. Years ago, I created an ABIv1-based AROS One distribution, but then, since there was no software, I decided to abandon it.
The same thing happened with AROS 68k. There was no software available, and the GCC compiler on the main AROS is not executable. On the fork, however, the executable works but crashes, so for these reasons, you cannot blame the distribution maintainers.
I know the history of AROS quite well. The problem I have with the fork is this... It's a lot of work that depends on one man back porting work from ABIv1 to it. If something happens to Deadwood it becomes a dead end. Which would mean one of two outcomes. Everything would need to be recompiled/ported over to AROS Again or everything that was done bringing applications to it will be stuck in limbo until someone decides to pick it back up if they decide to.
[quote name=Build-0-Matic post=10345]@Build-0-Matic - I just downloaded the ABIv0 image and this one works.
Now, I wonder why the ABIv1 team never checked to see if their fork worked in the first place?
With people disliking Windows more and more, hardware prices going up and up, and the excess of perfectly good hardware floating around, I see it as a great opportunity for AROS to become an interesting alternative for the oldest gear.[/quote]
ABIv1 isn't a fork it's the mainline branch. Much of the work has been focusing on 64 Bit that's where the main drive was/is. Add to that it's an open source platform so it relies on users to report bugs. Unfortunately that is where the contention comes in from my side. ABIv0 was the Original ABI, ABIv1 was designed to fix mistakes they made when ABIv0 was being developed. Unfortunately most applications at the time were ABIv0 so the Distro maintainers stuck with what was available. ABIv0 stagnated until Deadwood started back porting parts of ABIv1 to it which one could argue was good for the users but the flip side is ABIv1 lost in the end due to lack of users and no one will port software to a platform with no users so here we are in a chicken and the egg situation.
I must admit that it's the first time I hear of people backporting parts of an OS to an older version of said OS...
Usually, when backwards compatibility is required, people create an abstraction layer for the old software. But that slows things down.
This somehow reminds me of the C64 vs C128 thing. Almost nobody produced software for the C128 because it was backwards compatible with the C64 which had a much wider user base.
Unfortunately that's what happened with AROS. We can argue about if it was the right decision all day long. However the outcome is a mainline which became a playground more than the future because without testers bugs go unnoticed. When bugs go unnoticed you have stability and incompatibility issues and when that happens no one will port software to it for basically one user. But if it's to ever fully stabilize it needs testers I'm not saying everyone needs to download a copy every day and test ... but the Distro maintainers actively steering people away from it causes a lot of harm imho.
I've been playing with it (ABIv0) and discovered that they dropped BIOS boot support.
I'm now wondering what would be the best way to put it on my venerable netbook...
Or maybe it would be just better to use ArosONE (which does support BIOS) and remove what I don't want on it (like amistart) ?
You can try arosone ... but you can also try added noacpi to the grub line to see if that helps... but afaik the latest AROS does have Bios and EFI support... but iirc kalamatee fixed the build issue on ABIv1 so tonight or tomorrows build should have the fix.
When I boot with it, it just gives me a black screen with a blinking cursor.
So, the noacpi command is pretty useless in that case.
On Rufus, the target system is locked on UEFI no CSM when I make the USB stick, so this is a no-go.
I could hop on my linux box and try and fix the grub install, but since Aros uses a modified version of grub, I'm not too confident that it would fix it.
You can view all discussion threads in this forum. You cannot start a new discussion thread in this forum. You cannot reply in this discussion thread. You cannot start on a poll in this forum. You cannot upload attachments in this forum. You cannot download attachments in this forum.