Overview Features Coding ApolloOS Performance Forum Downloads Products Order Contact

Welcome to the Apollo Forum

This forum is for people interested in the APOLLO CPU.
Please read the forum usage manual.
Please visit our Apollo-Discord Server for support.



All TopicsNewsPerformanceGamesDemosApolloVampireAROSWorkbenchATARIReleases
The team will post updates and news about our project here

The AMIGA Future Is NOW! NG-AMIGA-OS!page  1 2 3 4 

Samuel Crow

Posts 424
10 May 2017 00:38


AmigaOS 3 is out of our hands.  We've been patching our Kickstart so it will continue to work with the SAGA chipset but there is only so much we can do.  OS 3 doesn't support chunky modes.  P96 and CyberGraphX don't support planar.  So what is a person supposed to do when you need both planar playfields active at the same time as the sprites and a truecolor chunky backdrop behind the lot of them, also active.  And all this from one chipset core!?  Yes!

AROS has an entirely different driver structure and we can update it since the source code is licensed publicly as open-source.  (The AROS Public License is a rebadged Mozilla Public License version 1.1 in case anyone is interested.)  It still needs a lot of work, but it is legal to do it.  For example, it uses two different base classes for chunky and planar displays.  Also the object-oriented HIDD system based on OOP.library needs some different way to implement things to get them to fit in a smaller amount of memory.


Vojin Vidanovic

Posts 770
10 May 2017 01:02


Samuel Crow wrote:

  AmigaOS 3 is out of our hands.  We've been patching our Kickstart so it will continue to work with the SAGA chipset but there is only so much we can do.  OS 3 doesn't support chunky modes.  P96 and CyberGraphX don't support planar.  So what is a person supposed to do when you need both planar playfields active at the same time as the sprites and a truecolor chunky backdrop behind the lot of them, also active.  And all this from one chipset core!?  Yes!
   
    AROS has an entirely different driver structure and we can update it since the source code is licensed publicly as open-source.  (The AROS Public License is a rebadged Mozilla Public License version 1.1 in case anyone is interested.)  It still needs a lot of work, but it is legal to do it.  For example, it uses two different base classes for chunky and planar displays.  Also the object-oriented HIDD system based on OOP.library needs some different way to implement things to get them to fit in a smaller amount of memory.
 

 
 
  Nothing against Vampire pushing 68k AROS further.
 
  a) All patches and tweaks to make some current 68k build to fully use Vampire
  b) If I am not mistaken, AROS 68k needs to be developed ("backported") to level of AROS x86 in many its components. If I am wrong, I apologize.
  c) Also, there are plenty AROS x86 apps on archives and aminet, I suppose best of them could be "backported"
  d) It would be great if all those would instantly be 080 AMMX optimized, or in the very end bring a patched versions so all 68k AROS userbase can use improvements and Vampires could get Vamparized.
 
  Browser remains greatest concern. There are Odyssey sources, maybe they could be backported to 68k (and further 080-AMMX)
  EXTERNAL LINK    As well as of OWB
  EXTERNAL LINK   
  If in wildest dreams there would be a Linux backporting involved, I vote for Midori as balance between lightweight and modern, and has slim XFCE look that I believe could be Amiga adopted
  https://en.wikipedia.org/wiki/Midori_(web_browser)
 
  If all fails, let us all pray for Ibrowse 2.5 ;-)
 
  And AROS seems to love Vampires.
 
  AROS Vision website says EXTERNAL LINK  "Aros on future amiga hardware
 
  Aros Rom Replacement and Aros 68k perfectly fits to new Vampire 2 / Apollo because sources are free and open and can be adapted to new amiga hardware.
 
  More to come..."
AROS RTG test with Vampire 600 V2
EXTERNAL LINK


Wawa T

Posts 695
10 May 2017 03:22


Vojin Vidanovic wrote:

    a) All patches and tweaks to make some current 68k build to fully use Vampire

  thats same or less amount of work that would be necessary to make os3.x support fully all apollo/vampire features, if it was open that is. aros has the advantage of hardware abstraction layer.
 

    b) If I am not mistaken, AROS 68k needs to be developed ("backported") to level of AROS x86 in many its components. If I am wrong, I apologize.
 

  no. except minor issues aros68k is up to the stand of current aros development altogether. what lacks are platform specific impovements and bottlenecks that becoeme apparent on slower systems.
 

    c) Also, there are plenty AROS x86 apps on archives and aminet, I suppose best of them could be "backported"
 

  aros apps and ports are usually open. they need to be compiled against the current sdk and desired target. certainly some tweaks will be necessary now and then as well as adding platform specific code such as hooks semantics or function register parametrizing where it has been left out, since x86 doennt need that.
 

    d) It would be great if all those would instantly be 080 AMMX optimized, or in the very end bring a patched versions so all 68k AROS userbase can use improvements and Vampires could get Vamparized.
 

  thats a task for apollo team.
 

    Browser remains greatest concern. There are Odyssey sources, maybe they could be backported to 68k (and further 080-AMMX)
    As well as of OWB
 

  aros owb has worked on my a4000 years ago. i even posted screenshots. now it doesnt. trying to debug this. what concerns odyssey kalamatee has the sources building far against the aros build system headers and libs. as far as i see he is concerned with the zpdate of subsystems and libs for now, as ssl4. i have some missing header with 68k support, stand by. not sure if it will ever be fast enough even for vampire.

btw, the current odyssey sources are here:
EXTERNAL LINK 
 


Vojin Vidanovic

Posts 770
10 May 2017 10:19


wawa t wrote:

      a) All patches and tweaks to make some current 68k build to fully use Vampire

    thats same or less amount of work that would be necessary to make os3.x support fully all apollo/vampire features, if it was open that is. aros has the advantage of hardware abstraction layer.

 
  OK, Good is less complicated, surely HAL helps now and would
  help if standalone is going to have any extra features (I am thinking of OS as preparation for "operation go standalone" as well as solution to Classic users).
 
   
 
  no. except minor issues aros68k is up to the stand of current aros development altogether. what lacks are platform specific impovements and bottlenecks that becoeme apparent on slower systems.

 
  I am glad about, public AROS repos show old files. Yes, Classic users used to complain and ditch it. Tried under Amiga Forever when it became feat, was impressed by OS compared to plain OS 3.1, but no AROS 68k apps was a turndown.
 
   
        aros apps and ports are usually open. they need to be compiled against the current sdk and desired target. certainly some tweaks will be necessary now and then as well as adding platform specific code such as hooks semantics or function register parametrizing where it has been left out, since x86 doennt need that. 

 
  Yes, I have hoped it is "kind of" "as easy" as Linux GCC
  target backporting. I am glad AROS is built on same expandability and I hope in some universe some care
  could be taken of AROS PPC32/64 for G4/G5/x1000/x5000 users.
 
 
  d) It would be great if all those would instantly be 080 AMMX optimized, or in the very end bring a patched versions so all 68k AROS userbase can use improvements and Vampires could get Vamparized.
    thats a task for apollo team.
 

 
  Agreed and hope they will pledge to it once CORE is finished (up to some feature list like FPU) or supply willing AROS coders with boards / cards.
 
 
  aros owb has worked on my a4000 years ago. i even posted screenshots. now it doesnt. trying to debug this. what concerns odyssey kalamatee has the sources building far against the aros build system headers and libs. as far as i see he is concerned with the zpdate of subsystems and libs for now, as ssl4. i have some missing header with 68k support, stand by. not sure if it will ever be fast enough even for vampire. 

 
  I am glad, however Odyssey 1.23+ if possible, is a better choice. Faster, has better HTML5 score and seems even more stable.
 
 
  btw, the current odyssey sources are here:
  EXTERNAL LINK 

 
  Could should look at githubs. However that was in public AROS archives as AROS port source.  Hope it will benefit anyone that can port it as separate bounty if needed.
 
  Cant code, willing to add small paypals!
 


Michael R

Posts 281
10 May 2017 20:32


Vojin Vidanovic wrote:

Samuel Crow wrote:

    AmigaOS 3 is out of our hands.  We've been patching our Kickstart so it will continue to work with the SAGA chipset but there is only so much we can do.  OS 3 doesn't support chunky modes.  P96 and CyberGraphX don't support planar.  So what is a person supposed to do when you need both planar playfields active at the same time as the sprites and a truecolor chunky backdrop behind the lot of them, also active.  And all this from one chipset core!?  Yes!
   
    AROS has an entirely different driver structure and we can update it since the source code is licensed publicly as open-source.  (The AROS Public License is a rebadged Mozilla Public License version 1.1 in case anyone is interested.)  It still needs a lot of work, but it is legal to do it.  For example, it uses two different base classes for chunky and planar displays.  Also the object-oriented HIDD system based on OOP.library needs some different way to implement things to get them to fit in a smaller amount of memory.
   

   
   
    Nothing against Vampire pushing 68k AROS further.
   
    a) All patches and tweaks to make some current 68k build to fully use Vampire
    b) If I am not mistaken, AROS 68k needs to be developed ("backported") to level of AROS x86 in many its components. If I am wrong, I apologize.
    c) Also, there are plenty AROS x86 apps on archives and aminet, I suppose best of them could be "backported"
    d) It would be great if all those would instantly be 080 AMMX optimized, or in the very end bring a patched versions so all 68k AROS userbase can use improvements and Vampires could get Vamparized.
   
    Browser remains greatest concern. There are Odyssey sources, maybe they could be backported to 68k (and further 080-AMMX)
    EXTERNAL LINK    As well as of OWB
    EXTERNAL LINK   
    If in wildest dreams there would be a Linux backporting involved, I vote for Midori as balance between lightweight and modern, and has slim XFCE look that I believe could be Amiga adopted
    https://en.wikipedia.org/wiki/Midori_(web_browser)
   
    If all fails, let us all pray for Ibrowse 2.5 ;-)
 
  And AROS seems to love Vampires.
 
  AROS Vision website says EXTERNAL LINK    "Aros on future amiga hardware
   
  Aros Rom Replacement and Aros 68k perfectly fits to new Vampire 2 / Apollo because sources are free and open and can be adapted to new amiga hardware.
 
  More to come..."
  AROS RTG test with Vampire 600 V2
  EXTERNAL LINK 

The compiler I'm using is capable of creating programs for AROS i686, AROS 68k, and Amiga 68k from the same IDE. That surely has some possibilities for cross compiling in the future! I'd like to test some of my programs and compile older apps for use on AROS 68k and Amiga 68k. I'm about to make the first batch..."Icon Info" & "Icon Info 68k". This shell program reads all .info data and displays types, offsets, and image data for IFF ILBM Icons such as OLDICONS, NEWICONS, GLOWICONS, and additionally PNG dual state icons. The future looks very good for Amiga and Vampire accelerators!


Wawa T

Posts 695
10 May 2017 23:24


what a compiler is this??


Michael R

Posts 281
11 May 2017 08:03


wawa t wrote:

what a compiler is this??

wxDevCpp IDE, part of AmiDevCpp package.
Just compiled a program for Amiga 68k, and one for AROS i686.
It works great for smaller programs.


Wawa T

Posts 695
11 May 2017 11:48


i hope you do have the current headers.. i have not been using amidevcpp for years. it has a number of annoying bugs. for instance you need to check when compiling c++ stuff that it properly recognizes that it is c++ source, because it repeatedly wants to compile it as c or other way arround. also the makefile format is a bit odd, which makes programs built with it not that easily portable.


Michael R

Posts 281
11 May 2017 14:29


wawa t wrote:

i hope you do have the current headers.. i have not been using amidevcpp for years. it has a number of annoying bugs. for instance you need to check when compiling c++ stuff that it properly recognizes that it is c++ source, because it repeatedly wants to compile it as c or other way arround. also the makefile format is a bit odd, which makes programs built with it not that easily portable.

It now allows to select "C" or "Cpp" using radio buttons. There is an option to update DevPacks. It's a bit quirky sometimes. When I find a better IDE I will use that instead. I have to learn to crawl before I can walk, and before I can run!

Does AROS fully support original MUI programs or do they need to be adapted to use Zune? If they are written for MUI will they just work?



Olaf Schoenweiss

Posts 690
11 May 2017 15:11


Mostly software like simplemail that require 3.8 work, if not it should be fixed on zune. If you use a desktop that not needs zune (wanderer and scalos do), then you can easily change between zune and mui. I do that in aros vision using magellan and MUI 3.8 as default (aros prefs do not work then though)


Wawa T

Posts 695
11 May 2017 16:17


Michael R wrote:

 
  It now allows to select "C" or "Cpp" using radio buttons.
 

  yes, you must not forget about this. ;)
 
 

  There is an option to update DevPacks. It's a bit quirky sometimes.
 

 
  exactly.
 
 

  When I find a better IDE I will use that instead. I have to learn to crawl before I can walk, and before I can run!
 

 
  meanwhile i think it might not be worst choice not to employ ide at all. simply write a makefile against aros build system, edit the files by hand and compile from the shell. same for amiga, all then you need to do is to rewrite the makefile a bit.
 
 

  Does AROS fully support original MUI programs or do they need to be adapted to use Zune? If they are written for MUI will they just work?
 

 
  as olaf says, simply write against mui, zune needs to be compatible. except if you need some zune extension or class missing on mui, if there even are any such.
 


Mr Niding

Posts 459
13 May 2017 12:15


There is some chatter on EAB about the bounty situation with regards to AROS;

EXTERNAL LINK


Szyk Cech

Posts 191
14 May 2017 12:48


I am still wonder in what is AROS is better than (lets say) AOS 3.9?!?


Olaf Schoenweiss

Posts 690
14 May 2017 13:08


it has some advanced features but the biggest advantage is that it is adaptable


Wawa T

Posts 695
14 May 2017 17:18


I am still wonder in what is AROS is better than (lets say) AOS 3.9?!?

  - no patching necessary.
  - doenst need to be painfully installed with all the boing bags.
  - you can simply copy it over your hd partition and it works.
  - on (almost) every amiga setup without modifications.
  - no problems not booting with inappropriate cpu libraries (060/040).
  - you can move windows off screen, even on planar.
  - usb, rtg, lan infrastructure built in.
  - 1.x - 3.x kickstart compatibility out of the box (may not be complete atm)
  - mui compatibility, up to 4.x (almost complete id say)
  - more ergonomic tools, for instace simple but handy editor.
  - a number of contributions (a webkit based css browser among others)
  - posix compatibility library (comparable to isemul)
 
  .. and a lot more..
 


Edgar Fink

Posts 47
16 May 2017 18:40


Does AROS have fork()? I'd like to finally compile bash for 68k. :)


Wawa T

Posts 695
16 May 2017 18:46


no.
no amiga(like) system has fork(). apparently it is impossible to implement. people work around it with vfork, afaik.
aros has abc-shell afair. it must be somewhere in contribs or developer dir i guess. i only have used the genuine shell to run gcc and such.


Edgar Fink

Posts 47
16 May 2017 19:00


Ok, thanks for the info. Well, my first goal when I get AROS up and running is to replace every instance of fork with vfork in the bash source code. Having bash on an Amiga is an ongoing (masochistically pleasing) obsession of mine. :)
 
  Btw, mksh (MirBSH Korn Shell) compiles fine on AmigaOS 3.1 but also needs to fork apparently.


Wawa T

Posts 695
16 May 2017 19:11


@Edgar Fink

im not very reliable source for technical information. for more educated assistance you might register on aros-exec or even or aros dev ml, if you intend to actually develop or port stuff.


Kolbjørn Barmen
(Needs Verification)
Posts 219/ 2
18 May 2017 14:01


Edgar Fink wrote:

Does AROS have fork()? I'd like to finally compile bash for 68k. :)

You cannot do fork on a system with only one running process (exec).

posts 72page  1 2 3 4