Attract Mode 2.6.0 has been released

[rand0m]

Attract-Mode 2.6 has been released and can be downloaded from:

Official Website: http://attractmode.org/download.html
Official Github Repo: https://github.com/mickelson/attract/releases

Release Notes/ Changelog:

Commits from v2.5.1 to v2.6.0
Andrew Mickelson (21 commits)

    Bump version to 2.6
    Issue #582 - Fix layout freezing after a month of inactivity
    Small refactor of code for getting artwork images
    Fixed problematic usage of c_str();
    Issue #558 - Added ability to get sort value with fe.game_info()
    Issue #566 - fix filter info update when display changes
    Improved error status checking for thegamesdb.net scraping
    Fix compile if libcurl not found
    Issue #531 - fix linux packaging typo (libcurl3 -> libcurl4)
    Issue #538 - update compile instructions with curl (windows cross-compile)
    Issue #532 - Initial work to update scraper to new thegamesdb.net API
    minor re-arrange to minimize windows.h header includes
    Issue #534 - Fixed magic tokens returning incorrect result
    The play state of swf files can now be controlled from a script using the video_playing property
    Speed up swf rendering by reusing a single OpenGL context
    fe.get_art() now returns fallback artwork from the layout (if present)
    Fix crash when error encounterd in the 'script.nv' file
    Refactor window creation and fix black screen on launch
    Include OpenGL version information with 'attract --version' command
    Fixed crash for certain invalid attract.cfg settings
    [Windows] fix a hang when returning from emulator (w/ Intel graphics)

Keil Miller (2 commits)

    Scrape "buttons" elements from hyperspin .xml (#563)
    Added Transition.EndNavigation trigger (#567)

Radek Dutkiewicz (16 commits)

    Fixed reggression in video rendering framerates on Nvidia (#552)
    Additional window modes fixes (#551)
    Fixed low framerate on AMD cards (#550)
    Improved decoding quality of videos with non 1:1 pixel aspect ratio (#548)
    Added Nvidia profile management with optimizations (#529)
    Fixed for windowed modes being rendered at desktop resolution (#528)
    Increased video timeout to 5 seconds (#546)
    Updated manifest for multi monitor mode (#526)
    Fixed compiler errors on non Windows builds
    Replaced Windows version check with macros
    Additional fixes
    Fixed black screen in fullscreen mode on game launch
    Fixed high CPU on video playback and video freezing after resuming from sleep (#520)
    Updated nowide to fix crashing on Windows (#523)
    Layout offset on rotation fix (#512)
    Added new fe_text property msg_wrapped (#507)

Happy Gaming 

[raygun]

Ha you beat me to it!!

I’m still finishing off testing and uploading packages, so there may be some changes still to what is posted now.  The official announcement should be ready soon^TM

[zpaolo11x]

Oh I love this moment, when you run through the list of changes and you think "wow this was fixed, that's great! Wait, what was the workaround I put in place to avoid this issue? Oh my god I have to re-check the whole layout!!!"

[jedione]

woot woot,,,,long live AM

[kent79]

Thank you 

[raygun]

Ok the download links on the website are now up to date and point to version 2.6 as well:

http://attractmode.org/download.html

[zpaolo11x]

I'm having issues with 2.6.0 related to the cropping of videos, probably due to the non square pixel aspect ratio. This can be replicated with a simple layout like this:

Code: [Select]

local image = fe.add_artwork("snap",0,0,400,400)

fe.add_transition_callback( this, "on_transition" )

function on_transition( ttype, var, ttime ) {

   print(fe.game_info(Info.Title) + image.subimg_width+" x "+image.subimg_height + "\n")
   print(fe.game_info(Info.Title) + image.texture_width+" x "+image.texture_height + "\n")

}

In 2.5.1 the output is:

Code: [Select]

1943: The Battle of Midway Mark II (US)192 x 256
1943: The Battle of Midway Mark II (US)192 x 256

while in 2.6.0 it is:

Code: [Select]

1943: The Battle of Midway Mark II (US)224 x 256
1943: The Battle of Midway Mark II (US)224 x 256

So it seems to me that 2.6.0 returns the actual pixel resolution of the video _before_ applying the display aspect ratio, which is all good for scanlines filters (thanks Oomek! That's fantastic). Sadly I'll have to find another way to scale and crop my videos, but can you confirm that this is the intended new behavior? And there's no way to get the intended display aspect ratio from the video?

[CSOne]

in my case, AM 2.6.0 is not respond.
I don't know why not.
AM 2.5.1 is not problem.

※ my OS build is Windows10Pro 64bit 1903 18362.418

last_run.log
Attract-Mode v2.6.0 (Windows, SFML 2.5.1 +SWF +7z +Curl)
avcodec 58.54.100 / avformat 58.29.100 / swscale 5.5.100 / avutil 56.31.100 / swresample 3.5.100

[rand0m]

Ha you beat me to it!!

I’m still finishing off testing and uploading packages, so there may be some changes still to what is posted now.  The official announcement should be ready soon^TM

Sorry for jumping the gun Andrew. I have updated the first post with download links.

[dark1973]

New version!! New Version!!  thank you very much everyone. I'm very happy

[xbs]

Thank you!

[connconn1]

Hey I'm glad to see the emulator get updated, but there's still a huge (at least making the program unusable for me) bug.

Someone opened this issue almost 3 years ago, and it seems like a simple enough bug fix: https://github.com/mickelson/attract/issues/294

Basically, when using a controller (at least tested with wireless xbox 360 andd wireless xbox one), the "exit emulator hotkey" doesn't work. For example I can map the hotkey to activate when I press the start and select button at the same time, but it simply has no effect. Using a key on the keyboard works, but since I want to use attract mode from the couch, the program isn't usable. But with a controller, it doesn't matter if I map it as a button combination or single button, exit hotkeys won't function

I'm sure you're busy but if you could implement this fix, I would really enjoy using this program for my living room system. I already love attract mode on my MAME cabinet! If you need any more details then let me know, I'll be happy to provide them

[raygun]

So it seems to me that 2.6.0 returns the actual pixel resolution of the video _before_ applying the display aspect ratio, which is all good for scanlines filters (thanks Oomek! That's fantastic). Sadly I'll have to find another way to scale and crop my videos, but can you confirm that this is the intended new behavior? And there's no way to get the intended display aspect ratio from the video?

I see yes that is the consequence of the fix that was applied here, but we should have a way to get both the pixel dimensions and the intended display dimensions.  I think leaving the texture_width and texture_height dimensions in pixels (as it is with 2.6.0) but also adding new functions to get the actual pixel size should fix this?? that would provide a way to figure out the intended display dimensions and would be easy enough to add I think.

[raygun]

Hey I'm glad to see the emulator get updated, but there's still a huge (at least making the program unusable for me) bug.

Someone opened this issue almost 3 years ago, and it seems like a simple enough bug fix: https://github.com/mickelson/attract/issues/294

Basically, when using a controller (at least tested with wireless xbox 360 andd wireless xbox one), the "exit emulator hotkey" doesn't work. For example I can map the hotkey to activate when I press the start and select button at the same time, but it simply has no effect. Using a key on the keyboard works, but since I want to use attract mode from the couch, the program isn't usable. But with a controller, it doesn't matter if I map it as a button combination or single button, exit hotkeys won't function

I'm sure you're busy but if you could implement this fix, I would really enjoy using this program for my living room system. I already love attract mode on my MAME cabinet! If you need any more details then let me know, I'll be happy to provide them

Yeah I don't have time to get to everything!  I suspect this issue is because SFML can't access joystick state when it isn't the foreground process, but I'm guessing.  The issue you linked is pretty light, could you at least tell me what operating system you are on? thanks

[zpaolo11x]

I see yes that is the consequence of the fix that was applied here, but we should have a way to get both the pixel dimensions and the intended display dimensions.  I think leaving the texture_width and texture_height dimensions in pixels (as it is with 2.6.0) but also adding new functions to get the actual pixel size should fix this?? that would provide a way to figure out the intended display dimensions and would be easy enough to add I think.

I really appreciate that now texture wiidth and height are the actual resolution, that's such a great plus that I already fixed the consequent issues. But it would be great to be able to read the intended display resolution, no idea how it is stored in the media, maybe as pixel aspect ratio? Or as display aspect ratio? Any way will be good an helpful