Great interview with Michal Schulz Michal joined AROS in late 90's and was instrumental in making it run on real hardware: x86, PPC, ARM. He did a lot of low-level work with AROS before moving to Emu68. [url]https://www.youtube.com/watch?v=mXaKDhtTTqQ[/url]

I tried to check it on my puppy miga distro which is Fossapup64 9.6 and Ubuntu 20.04 based but dependencies missing I think libSDL3 releated , all other amiga emulation on puppy works nice

Now, that is an unexpected surprise. Thank you for reporting AMIGASYSTEM. I always use Frodos' emulator and was in the process to get (an auto-creating) amiberry going on my Linux rig (not much luck on the latter there though as configuration seems to be a biatch).

First version of WinUAE for macOS and Linux https://github.com/reinauer/WinUAE/releases/tag/unix_r1

Hi again. Quite a hard time … moving from the old code to the new library code was not as easy as reuse the code, it is doing all the way again. The new code is somehow better than the old (running more things) but still i cant get rid from the old as it sometime still have better code than the new one. About the rom/adf pair of aros, is there some way that i can check the compatibility programatic so i can return the propper error? Erverything is far from usable, but all kickstarts allready gives signals of life (KS31 needs 68020 emulation) AROS, that was my main goal, no success yet. I have the bootscreen without media and some disk activity with media (ADF), but no workbook. I got much more things to somehow work (still lots of glitches) and the bate metal is not running as good as the hosted version. So, i have 4 options: 1) focus on AROS 2) focus on fix the glitches 3) focus on the bare metal 4) focus on the EMU68 integration My feeling is that number 2 would be the better approach atm so i could get rid of the old code. (Anyway, somehow i got ks1.3 booting into wb13, budbrain megademo, sota, etc …. With bad mouse, keyboard and bad sound. There is tons of glitches, i can spent quite some time on it. AROS becomes a frustrating effort, i spent time and cant progress. Have no idea what is happening. Probabbly tomorrow i Will try again with aros, and if possible at least get out a propper error of version mismatches and monday start to focus on fixing the visíble glitches of the chipset) I had add usb msc (fat32) support on the bare metal, so the flow would be load iso or adf from pen drive (no need for extra space on the boot partition of the sdcard), still need to test it. Regards,

Keep in mind that supported host version is 24.04 - other version might or might not work.

Interestingly, the version of Ubuntu newly installed where compiling stops is 20.04 focal fossa. Both the original Ubuntu and the new one (20.04) stop at flexcat. However, the version installed for ABIv0 which is compiling correctly is 18.04 bionic beaver which is an older version. I could try installing 18.04 again in a new virtual machine. Or I could also try installing 24.04. The difference that I see mostly is the version of GCC for 18.04 is 7.5.0 but for 20.04 it's GCC 9.4.0 so maybe I can try downgrading GCC.

You are correct there retrofaza in that things are build on a shared folder. Again, this is otherwise not a problem and virtualbox has provisions in place to circumvent some of the more pressing issues (such as the inability to create links on a shared folder). This particular VM however seem to suffer from issues regarding clock skew (seem to have something todo with with cpu/thread-scheduling) which I discovered just now and also not have encountered when using virtualbox before (and yes, I know it is not the best of virtualization solutions for linux but I got used to VBox when I was still using windows many moons ago so, sue me :) ). Thanks for the input though because the issues for this particular VM are still haunting me just as they seem to do for miker and his VMWare install (and imho oracle documentation sucks big time as you seem to need to read all nooks and crannies to be able to have a chance to discover such an issues is able to exist in the first place)

I suspect (judging by the path) that you're using a folder shared by the host. The problem might lie in the permissions for that folder (or the file system?). But that's just a guess on my part :) I never build anything in a virtual machine myself.

[quote name=retrofaza post=12387]@retrofaza - /media/sf_builder ? Shared folder, permission issue. This has nothing to do with the Aros build system[/quote] Thank you for the suggestion retrofafza. I Would agree if it was not for the fact that I use this vm for all my 3th party builds (and otherwise works like a charm), though in all honesty I do that with a debian 13 VM while this is an (recently installed) debian 12 VM. I do not deny the fact that I (also) suspect it is a VM related issue (this particular debian 12 setup caused other/earlier headaches for some reason or another). Or, is the suggestion that aros should be build as someone with root permission ?

/media/sf_builder ? Shared folder, permission issue. This has nothing to do with the Aros build system

@miker: Debian Buster used to work without any issues but can imagine it being a bit out of date in order to be able to build AROS these days though in such case I would not expect the error you are currently reporting. A small piggyback on/in this thread if allowed (if it requires more attention then please say so and I'll open up another thread on this topic). @deadwood (or someone else that known): My build stops here (debian 13, virtualbox): [code] Fetching zstd-1.5.7 ... Trying /media/sf_builder/arosbuilds/portssources/zstd-1.5.7.tar.gz... unpack_cached: `/media/sf_builder/arosbuilds/toolchain-core-x86_64-build/bin/linux-x86_64/Ports/zstd' does not exist. Making it. Unpacking `zstd-1.5.7.tar.gz'... tar: zstd-1.5.7/tests/cli-tests/bin/unzstd: Cannot utime: Operation not permitted tar: zstd-1.5.7/tests/cli-tests/bin/zstdcat: Cannot utime: Operation not permitted tar: Exiting with failure status due to previous errors fetch: Error while unpacking `zstd-1.5.7.tar.gz'. make[1]: *** [mmakefile:39: workbench-libs-zstd-fetch] Error 1 [MMAKE] make --no-print-directory TOP=/media/sf_builder/arosbuilds/toolchain-core-x86_64-build SRCDIR=/media/sf_builder/arosbuilds/AROS CURDIR=workbench/libs/zstd TARGET=workbench-libs-zstd-fetch -s --file=mmakefile workbench-libs-zstd-fetch failed: 512 [MMAKE] Error: Error while running make in workbench/libs/zstd: No such file or directory make: *** [Makefile:178: /media/sf_builder/arosbuilds/toolchain-core-x86_64/.installflag-crosstools] Error 10 [/code] I have the suspicion it is more distro/vbox related than it is an issue regarding aros build-system (but I might very well be wrong there).

The version installed is ubuntu 20.04 focal fossa. Maybe it's too old? I'll update it to 22.04 or 24.04 and try again. I'll check the version for ABIv0 which seems to compile correctly.

I am waiting for the release of new builds with the new GL driver and to insert the new game packs into my site, thanking Retrofaza and also Deadwood, everyone who appreciates it

There doesn't seem to be any error in the build. Flexcat built correctly and then the process stopped - not really sure why. ./rebuild.sh has been used by many people in the past to get crosscompiler + AROS to build. One last thing that comes to my mind is the Ubuntu version you are using. Which version is it? The process was not tested on anything newer then 24.04.

Ok. Something is wrong with building the toolchain. I made a new virtual machine in VMWare then installed a fresh Ubuntu 64 from an ISO image. Then I followed INSTAL.md instructions. I installed all the required packages. Then I ran the ./rebuild sh as instructed chosing option 1 to build the toolchain. The result was exactly the same as the previous two virtual machines that I tried. It stopped at flexcat. See screenshot. Can anyone confirm that this isn't working? Or am I the only one with these strange issues? If I can't get ABIv11 to build correctly I'll go back to using the Build System for ABIv0 for development and testing. It seems to be just fine. I'm losing momentum with these issues.

Be sure to install dependencies (sudo apt-get install line in tutorial). If you are missing autoconf, it suggests dependencies were not installed.

For some reason I believe my virtual machine for building ABIv11 is haunted or cursed! :-) It's strange because my ABIv0 virtual machine is working fine. But it's a copy of the original VMDK that I also used for the ABIv11. So I have a few other options. I am currently making a copy of the original again to delete the sources and try again. If that doesn't work I will setup a completely new virtual machine in VMWare then install Ubuntu and follow Install.md That should definitely work correctly!!

@deadwood Thanks for the information. I didn't know that was part of Ubuntu. I thought it was an AROS file. I'll try again this evening. :-) I've had some family matters to take care of so I haven't had much time for AROS programming. But I'll get back to it later today or tomorrow.

Today I decided to test an old multi-card reader of mine (see screenshot) on my Acer Aspire 5920 and found that the Micro SD card had been mounted correctly.