FS-UAE <= 2.9.x development series

[FrodeSolheim]

Originally Posted by ArtMac

Hi. About arcade, what version of a game does it load by default??

Could there be an option for it to launch the WHDLoad version (if exists) by default?? In case there are WHDLoad versions for different languages, there could also be an option to launch by default that specific language version... so the priority could be:

1. WHDLoad (in an specific language)
2. If 1. doesn't exist: WHDLoad (english)
Then, the version that is now defaulted which is the first one of the list of the versions available for an specific game.

thank you!!!

Yes, I have though about a feature where you can specify some criteria for the default variant (based on language or media type, etc).

The default variant selected is that highest rated one in the database - or the one you have starred in FS-UAE Launcher. For now, you can use the star function in FS-UAE Launcher to set the favorite variant per game. Of course, you'll soon be able to see (and change) the selected variant in FS-UAE Arcade as well.

Originally Posted by amigafreak68k

On Mac OS X could we find a solution to set FS-UAE to fullscreen mode automatically.

"fullscreen_mode = fullscreen" set FS-UAE to fullscreen at start up - I know, but than, we cannot swipe to other spaces like in the real OS X fullscreen.

If we set "fullscreen_mode = desktop" we get a simple window at startup and must manually select the fullscreen button in upper right of the window.

I cannot reproduce - fullscreen_mode = desktop (with fullscreen enabled) always causes FS-UAE to open in full-screen on my OS X 10.8 laptop (though you can see FS-UAE opens a window before it transitions to fullscreen). Are you launching from FS-UAE Launcher? Please attach fs-uae.log.txt.

[jbl007]

Originally Posted by FrodeSolheim

Hopefully no n00b mistakes with this release

I think, I've got another one.
When launching a config with harddrive from http://... I get this:

LaunchHandler.prepare_hard_drives
 -- Bereite Festplatten vor... --
<Task "Amiga Launch Task"> failed "'module' object has no attribute 'unquote'"
Traceback (most recent call last):
  File "/usr/share/fs-uae-launcher/fsbc/task.py", line 64, in __run
    self.run()
  File "/usr/share/fs-uae-launcher/fs_uae_launcher/FSUAELauncher.py", line 611, in run
    self.launch_handler.run_sequence()
  File "/usr/share/fs-uae-launcher/fsgs/amiga/LaunchHandler.py", line 109, in run_sequence
    self.prepare()
  File "/usr/share/fs-uae-launcher/fsgs/amiga/LaunchHandler.py", line 93, in prepare
    self.prepare_hard_drives()
  File "/usr/share/fs-uae-launcher/fsgs/amiga/LaunchHandler.py", line 298, in prepare_hard_drives
    name = urllib.unquote(name)
AttributeError: 'module' object has no attribute 'unquote'

I changed "urllib.unquote(name)" to "urllib.parse.unquote(name)". Now it works.

[amigafreak68k]

Originally Posted by FrodeSolheim

I cannot reproduce - fullscreen_mode = desktop (with fullscreen enabled) always causes FS-UAE to open in full-screen on my OS X 10.8 laptop (though you can see FS-UAE opens a window before it transitions to fullscreen). Are you launching from FS-UAE Launcher? Please attach fs-uae.log.txt.

Yes! You're right.

Now it works. At this time, I restartet the Launcher after setup "fullscreenmode = desktop"...

Sorry!

[FrodeSolheim]

Originally Posted by jbl007

I changed "urllib.unquote(name)" to "urllib.parse.unquote(name)". Now it works.

Thanks, fixed for 2.5.5dev

[mr_a500]

Originally Posted by FrodeSolheim

Can you give a more exact description so I can try to reproduce?

(mouse problem)

I start Workbench and move the pointer up and down. The pointer randomly stops at a random vertical position, as if it has reached the screen border. There's no problem with horizontal. If you continue to move the pointer up and down, it eventually goes back to normal, but problem comes back later - randomly. (problem seems very familiar, but I can't remember)

The problem seems to have started in FS-UAE 2.5.2dev and is still in FS-UAE 2.5.4dev. There is no problem in FS-UAE 2.5.0dev. (can't remember if I tried 2.5.1dev)


I'm also hoping some long standing problems will eventually be fixed (don't mean to whine about it, but... ):

1. Crackly audio - I've had this problem for years and I always assumed it was my old, slow computers, but now I've got a kick-ass top-of-the-line computer and I still have crackly audio.

2. Scanline flicker - when flipping between screens interlaced/non-interlaced there is a very noticeable flicker as if the scanlines are being drawn very slowly. (UAE has always had a problem with scanlines)

3. Jerky scrolling - scrolling has improved, but it's still far from perfect. Example - try viewing a text file with MuchMore. (jerk city)

4. Multiple config files - yes, this is Apple's problem, but I would like to eventually double click a config file and not have it open default instead

(OK, bitching over... for now )

[Ariath]

Hi, i have found the FS-UAE source in your GitHub, but the Launcher project is empty.
Are all (emulator, launcher and arcade) sources in the FS-UAE repository?

See ya!

[FrodeSolheim]

A new development version has been released. Changes in FS-UAE 2.5.5dev:

• Dual joystick/mouse mode for joystick port 0, both a specific joystick device and mouse is enabled at the same time (only when joystick_port_0_mode = joystick and a joystick host device is selected).
• Show on-screen gui_message notifications from UAE code.
• Added missing @docdir@ substitution variable.
• More build system updates / removed old cruft.
• BSD make should now also work (in addition to GNU make).
• Environment variable FSGS_RETURN_CURSOR_TO can be used to move the cursor to a specific location when FS-UAE quits.
• (Device Helper) Ignore full negative axis motion events on startup.

Changes in FS-UAE Launcher 2.5.5dev:

• Added GUI control for option database_show_adult.
• Automatically update game list when database_show_[games/adult] changes.
• Fixed (python 3) bug in download hard drive via HTTP feature.

Changes in FS-UAE Arcade 2.5.5dev:

• Improvements when running under Gnome 3.
• Kill fs-uae-device-helper when FS-UAE Arcade is quit.
• Fix image loader so it does not stop handling requests on download errors.

Originally Posted by Ariath

Hi, i have found the FS-UAE source in your GitHub, but the Launcher project is empty. Are all (emulator, launcher and arcade) sources in the FS-UAE repository?

Hi, I'm considering putting up the launcher/arcade sources on github. For now, they are available as source tarballs (FS-UAE download page).

[bcripon]

Has anyone been able to run the OS X version of 2.5.5? I have to force quit it and when I look in the Console, it just says "Exited: Terminated: 15". I am running OS X 10.9.2.

If I need to provide a log file, I can do that if I can be told how to create one. Thanks.


Bob C

[thommys]

I'm using the 32bit version on Archlinux64. I have no jit with 2.54 and 2.55. 2.53 is OK. Is it a Problem with the Source or the PKGBUILD in the AUR?

[jbl007]

Configure seems to detect a 64bit OS and refuses to enable JIT.
Try the following line above "./configure --prefix=/usr" in PKGBUILD.

sed '10295s/if/if !/' -i configure

. Bad solution, but it should do it for now.

[FrodeSolheim]

Originally Posted by bcripon

Has anyone been able to run the OS X version of 2.5.5? I have to force quit it and when I look in the Console, it just says "Exited: Terminated: 15". I am running OS X 10.9.2.

If I need to provide a log file, I can do that if I can be told how to create one. Thanks.

Is this the first version you cannot run? Does for example 2.5.2dev run properly for you? Log files are put in Documents/FS-UAE/Cache/Logs.

Originally Posted by thommys

I'm using the 32bit version on Archlinux64. I have no jit with 2.54 and 2.55. 2.53 is OK. Is it a Problem with the Source or the PKGBUILD in the AUR?

I don't really know how the 32-bit build is performed by the Arch port maintainers. The build system has recently been updated, and configure checks if __i386__ is defined by the compiler (and only then enables JIT). The "JIT check" works fine in my own builds, so it's probably an issue to be fixed in the PKGBUILD file. I'll of course fix the source if it turns out the problem lies there.

[jbl007]

I think, I figured it out: CC and CXX need to be set before running configure. Attached fixed PKGBUILD (use without warranty ).
I'll send a pm to grimi (the maintainer).

One question remains: Why do we have "--enable-jit" option for configure now, if it's enabled by default and we can't enable it.

[FrodeSolheim]

Originally Posted by jbl007

I think, I figured it out: CC and CXX need to be set before running configure. Attached fixed PKGBUILD (use without warranty ).

Yes, if it overrides the compiler, configure needs to know it, since it runs the compiler to detect several settings (many other things besides JIT will be wrong if configure is run using a compiler with another arch the make).

Originally Posted by jbl007

One question remains: Why do we have "--enable-jit" option for configure now, if it's enabled by default and we can't enable it.

It's not a grand design decision at all, I just moved several checks to autoconf, including the JIT on/off decision. Of course, it can be used to disable JIT (--disable-jit) when it would be otherwise be enabled, or someone can specifically use --enable-jit and get a configure error if it couldn't be enabled (instead of getting a surprise later on), but it's not a big deal. Note: it's enabled by default only on architectures where JIT is available.

The general point of the migration to autoconf is to check properly for *features* instead of hard-coding decisions based on operating system checks. This will likely make the source work better on e.g. *BSD systems without patches. In addition, the old Makefile was getting a bit ugly, so it was time for a spring cleaning

[grimi]

Hi, just back from work. Fixed PKGBUILD (with working JIT) is already uploaded to AUR.

[bcripon]

Frode,

I have been able to run every dev version before this. In fact, I can run 2.5.4 without any problem. If 2.5.5 is working for you, I can send a log this evening.

Bob C

[FrodeSolheim]

@bcripon Yes, it runs here. But while testing more thoroughly, I noticed that icons/lsuipresentationmode/app title of FS-UAE.app itself wasn't working properly. Turns out an extra < character had gotten itself into Info.plist making it invalid. This may or may not be relevant to your issue.

[FrodeSolheim]

OS X builds with fixed Info.plist: http://fs-uae.net/devel/2.5.5dev2/

[Retro-Nerd]

Any chance that you check the forced CD32 pad script/function issue again?

[FrodeSolheim]

Originally Posted by Retro-Nerd

Any chance that you check the forced CD32 pad script/function issue again?

Ok, it wasn't the plugin. The plugin loader needed a Python 3 compatibility fix. Should work fine in 2.5.6dev

[Retro-Nerd]

Ah, good to hear. Thanks in advance for the fix then.