19-Apr-2024 01:24 GMT.		 UNDER CONSTRUCTION
Anonymous, there are 21 items in your selection
[Files] New version of UAEANN.lu
Posted on 06-Jan-2004 12:43 GMT by Peter Gordon21 comments
View flat
View list
Richard Drummond has posted new UAE binaries for AmigaOS, Linux (PPC & x86), MacOS X, and BeOS.

This is based on WinUAE 0.8.22-r9, and a version based on 0.8.23 is due soon.

Download them here
New version of UAE : Comment 1 of 21ANN.lu
Posted by Georg Steger on 06-Jan-2004 13:27 GMT
Here this versions of UAE with JIT under Linux/x86 have always been
pretty unstable. With some additional strange things. Like on Picasso
Screens in Console/Shell windows there's something wrong with scrolling.
If I have a shell window open and have this lines of text in it:

Neuer Shell-Prozeß 5
5.Ram Disk:> ;1
5.Ram Disk:> ;2
5.Ram Disk:> ;3
5.Ram Disk:> ;4
5.Ram Disk:> ;5[]

[] = cursor. When pressing return it should scroll one line up, as this is
the last visible line in the window. What I then get is:

5.Ram Disk:> ;2
5.Ram Disk:> ;3
5.Ram Disk:> ;4
5.Ram Disk:> ;5
5.Ram Disk:> ;5
5.Ram Disk:> []

Like it repeated the "scroll-one-up" twice. When I then tough move the
mouse over the text lines, the gfx underneath is rerefreshed to what it
should have looked in the first place:

5.Ram Disk:> ;1
5.Ram Disk:> ;2
5.Ram Disk:> ;3
5.Ram Disk:> ;4
5.Ram Disk:> ;5
5.Ram Disk:> []

Looking around in UAE's picasso96.c source file I for now just
spotted a completely broken AmigaListAddTail() function (it puts
NULL into the addtailed node's ln_Pred. In case of an empty list
also in ln_Succ. Urk!).

And the COMPLEMENT drawmode handling is also not correct. It should
not XOR the background pixel with the FG pen, but simply NOT (invert)
the background pixel. And AFAIK/IIRC INVERSVID needs to be handled
also for COMPLEMENT drawmode.
New version of UAE : Comment 2 of 21ANN.lu
Posted by Peter Gordon on 06-Jan-2004 13:37 GMT
In reply to Comment 1 (Georg Steger):
Well don't tell us, email Richard.
New version of UAE : Comment 3 of 21ANN.lu
Posted by Fabio Alemagna on 06-Jan-2004 13:39 GMT
In reply to Comment 2 (Peter Gordon):
> Well don't tell us, email Richard.

What's wrong with telling us too?
New version of UAE : Comment 4 of 21ANN.lu
Posted by Anonymous on 06-Jan-2004 14:04 GMT
I am slightly confused.

Is the TCP/IP wrapper only for Linux or should it work in OS X as well?
I tried putting a bsdsocket.library in libs in the emulated environment which supposedly is necessary on the Linux port, but programs don't seem to be able to use it.
New version of UAE : Comment 5 of 21ANN.lu
Posted by Georg Steger on 06-Jan-2004 14:12 GMT
In reply to Comment 2 (Peter Gordon):
Those seem to be general problems. Not restricted to single versions
(like Richard's) of UAE. The scrolling problem seems to be because
of blitrect() doing the scroll first in UAE's virtual framebuffer
(chunky buffer) and then calling do_blit() which is supposed to repeat
the blit in the real "framebuffer". From do_blit():

if (can_do_blit) {
/*
* Call OS blitting function that can do it in video memory.
* Should return if it was successful
*/
}

This is obviously still missing so it falls down to the code below,
which does a "manual" blit (memcpy) from "virtual" framebuffer to
real "framebuffer". In case of a scroll-blit causing then the wrong
sort-of-double-scroll-like effect, because the "virtual" framebuffer
has been scrolled already.
New version of UAE : Comment 6 of 21ANN.lu
Posted by Peter Gordon on 06-Jan-2004 14:21 GMT
In reply to Comment 3 (Fabio Alemagna):
Well, nothing really. I just meant that it was more important to pass this information along to Richard, who is likely to be interested in fixing them, than telling ANN visitors :)
New version of UAE : Comment 7 of 21ANN.lu
Posted by Matt Parsons on 06-Jan-2004 14:38 GMT
In reply to Comment 6 (Peter Gordon):
Given your interest in Emulators I would have thought you would have been facinated but Georg's (rather exceptional) bug hunting through the UAE sources :-)
New version of UAE : Comment 8 of 21ANN.lu
Posted by Richard Drummond on 06-Jan-2004 15:32 GMT
In reply to Comment 1 (Georg Steger):
Hi Georg

I'm aware of many such problems with the Picasso96 emulation. I just haven't had the time to fix them yet. Thanks for the bug hunting,

Cheers,
Rich
New version of UAE : Comment 9 of 21ANN.lu
Posted by Richard Drummond on 06-Jan-2004 15:40 GMT
In reply to Comment 4 (Anonymous):
> Is the TCP/IP wrapper only for Linux or should it work in OS X as well?

The bsdsocket emulation doesn't work on OS X yet. The current OS X binaries available on my page do have the bsdsocket emulation built in, but they'll crash on network access. Binaries for the new version (when I get around to building them ;-) will have the emulation disabled.

It's on my huge and growing to-do list to get the bsdsocket emulation working properly. The problem, I suspect (but haven't investigated), is with the thread-local storage used by the bsdsocket emulation. It probably doesn't sit well with SDL threads. Or it may be that the OS X pthread implementation is ropey (certainly something strange is happening with POSIX semaphores on OS X . . .)

Cheers,
Rich
New version of UAE : Comment 10 of 21ANN.lu
Posted by Anonymous on 06-Jan-2004 17:17 GMT
@Rich :),

Would it be possible to use lowlevel.device instead
of gameport.device in the m68k AmigaOS version?
Using lowlevel.device would make it compatible
to Amithlon and AmigaOS 4.

It's especially cool since the Poseidon Amiga
USB Stack recently added joypad support via
lowlevel.library.

The latest version runs fine on WinUAE, but I get
an interlaced screen (one visible line, one
black line, one visible line), so half
of the screen "gets moved out" of the UAE
window (which is running on the WB).
I am sure I haven't run UAE with the correct
command line?!

Also, what GUI do recommend to use with
your version? Would that be Samuael
Dewulder's GUI?

Thanks for modernising UAE for AmigaOS:)
New version of UAE : Comment 11 of 21ANN.lu
Posted by Anonymous on 06-Jan-2004 20:32 GMT
In reply to Comment 9 (Richard Drummond):
Yay!

I'll buy you a pint next time I see you (if ever ;)).
New version of UAE : Comment 12 of 21ANN.lu
Posted by Richard Drummond on 06-Jan-2004 21:42 GMT
In reply to Comment 10 (Anonymous):
> Would it be possible to use lowlevel.device instead
> of gameport.device in the m68k AmigaOS version?

Why did I say it used gameport.device? Obviously my brain's overloaded. It does actually use lowlevel.library.

> The latest version runs fine on WinUAE, but I get
> an interlaced screen (one visible line, one
> black line, one visible line), so half
> of the screen "gets moved out" of the UAE
> window (which is running on the WB).
> I am sure I haven't run UAE with the correct
> command line?!

I'm not sure I followed that. Can you send me your config file?

> Also, what GUI do recommend to use with
> your version? Would that be Samuael
> Dewulder's GUI?

That should still work, I think. Some of the config options have changed since the last official Amiga version. One in particular is the gfx_colour_mode option. I've patched my version so that it always asks for a screenmode - until I can resolve the problem. (BTW pressing cancel on the screenmode requester will make UAE run in a window).

Cheers,
Rich
New version of UAE : Comment 13 of 21ANN.lu
Posted by Kolbjørn Barmen on 06-Jan-2004 22:54 GMT
Any chance of Amithlon-partition support for the linux/x86 version?
New version of UAE : Comment 14 of 21ANN.lu
Posted by SixK on 07-Jan-2004 09:10 GMT
In reply to Comment 13 (Kolbjørn Barmen):
I've tryed an UAE native Port for Amithlon, but Gcc seemed too buggy when I tryed 6 month ago.
Someone know if a new Gcc Package have been done last year ?
New version of UAE : Comment 15 of 21ANN.lu
Posted by SixK on 07-Jan-2004 09:13 GMT
In reply to Comment 14 (SixK):
Ooops, I didn't read the previous post correctly, sorry.
New version of UAE : Comment 16 of 21ANN.lu
Posted by Olegil on 07-Jan-2004 13:32 GMT
In reply to Comment 13 (Kolbjørn Barmen):
Whats an amithlon partition?
New version of UAE : Comment 17 of 21ANN.lu
Posted by Kolbjørn Barmen on 08-Jan-2004 07:09 GMT
In reply to Comment 16 (Olegil):
An "Amithon partition" is just a partition on the PC that has filesystem ID 0x76. Amithlon will treat this partition as an entire disk, letting you set it up, partition it etc via HDToolBox.

WinUAE has had Amithlon-partition support for a while, this means I use the very same filesystems from Amithlon and WinUAE. However, I _very_ rarely run Windows at all :)
New version of UAE : Comment 18 of 21ANN.lu
Posted by Richard Drummond on 08-Jan-2004 08:51 GMT
In reply to Comment 17 (Kolbjørn Barmen):
> Any chance of Amithlon-partition support for the linux/x86 version?

In principle it should work, but I haven't tested it. You should be able to mount the partition as an RDB hardfile (specify /dev/hda5 or whatever as the hardfile). The current version is limited to 2GB hardfiles, however, so it's probably not much use for you yet. I'll fix this very soon.
New version of UAE : Comment 19 of 21ANN.lu
Posted by Chris H on 09-Jan-2004 16:23 GMT
In reply to Comment 17 (Kolbjørn Barmen):
@Kolbjørn Barmen
"WinUAE has had Amithlon-partition support for a while, this means I use the very
same filesystems from Amithlon and WinUAE."

REALLY?!? Where is it? How do I activate it? I have the latest WinUAE
version, but there didn't seem to be much new since the last update of
AmigaForever (from ages ago).
New version of UAE : Comment 20 of 21ANN.lu
Posted by Chris H on 09-Jan-2004 16:24 GMT
In reply to Comment 19 (Chris H):
Sorry, please ignore my last comment - my brain apparently went to the fritz and managed to interpret that WinUAE could read real Amiga HDs! :-(
New version of UAE : Comment 21 of 21ANN.lu
Posted by Matt Parsons on 10-Jan-2004 15:54 GMT
In reply to Comment 20 (Chris H):
AFAIK it can... or at least that's what I've been told... I have not yet wanted to risk my Amiga HD in my WinPC yet...
Anonymous, there are 21 items in your selection
Back to Top