(2016-04-20, 23:15)EMK0 Wrote: [ -> ]nice DS3 is working on this build and i got games to run 
rumble test works also. Sorry didn't provide debug log from last post
just thought it was to early to report bugs as maybe you where still working on it. Just let me know if you need someone to test things out.
Thanks for working on this awesome Kodi feature
It was a bit premature, I was actually surprised how quickly I fixed the crashes with games in VideoPlayer.
At this point, I'm mostly interested in stability and performance results from playing games in VideoPlayer.
Krypton won't hit feature-freeze for a few months, so merging RetroPlayer in 17 is actually a (small) possibility. The only major work left is fixing FF/RW before I consider RetroPlayer complete.
So now all that stands in our way is FF/RW and testing. I welcome any and all help in the testing department. If we can find and fix enough bugs, I'll open a PR against Krypton. That way, even if we don't make the feature-freeze cutoff, the PR will most likely be ready to merge the day that the v18 merge window opens.
game.libretro.genplus tested Sega CD
won't load .cue files for Sega CD, they do load fine in retroarch and is a supported file extension
i loaded the first .bin and i do have the bios setup correct, games do boot up but the screen graphics are very wrong
you can see here
http://imgur.com/PO9bUAc
then trying to load up the game again you get a black screen
http://imgur.com/h3ru5DX
also games that did boot before now will just show a spinner as if its loading but never does anything.
debug log of all tests above
http://pastebin.com/rhpD7r5f
all emulator cores are missing .zip extension support i guess you still need to implement it i am not sure.
SNES
game.libretro.snes9x-next
Works just at very low FPS - debug info on screen shows a constant 60FPS but the game is very choppy running low at a low FPS.
Opening the game again you can hear the sound but no video its as if the game is a audio file as the music OSD shows up
game.libretro.snes9x
Crashes Kodi every time you try to use it.
Debug log of testing SNES cores up to crash
http://pastebin.com/gLAEWaPv
game.libretro.beetle-bsnes
nothing happens
19:16:00 T:8716 INFO: AddOnLog: SNES (Beetle bSNES): Error opening file: Invalid argument
http://pastebin.com/geRAha9H
i hope this helps and its not to early
(2016-04-21, 02:14)garbear Wrote: [ -> ]So now all that stands in our way is FF/RW and testing. I welcome any and all help in the testing department. If we can find and fix enough bugs, I'll open a PR against Krypton. That way, even if we don't make the feature-freeze cutoff, the PR will most likely be ready to merge the day that the v18 merge window opens.
Have the savegames been fixed. They haven't been working for me for quite a while. I wanted to look into it but realized that RetroPlayer wasn't working anymore due to the VideoPlayer refactoring. But maybe I'll find the time to look into it now that it's in better shape again.
SRAM saves are broken for most cores, but they work for 2048, so that might be a place to start digging. Savestates depend on FF/RW and a game database, though for now we can just save to the filename + .sav extension
(2016-04-21, 03:27)EMK0 Wrote: [ -> ]i hope this helps and its not to early
Thanks, I'll start tracking down these bugs
(2016-04-21, 08:16)garbear Wrote: [ -> ]SRAM saves are broken for most cores, but they work for 2048, so that might be a place to start digging. Savestates depend on FF/RW and a game database, though for now we can just save to the filename + .sav extension
(2016-04-21, 03:27)EMK0 Wrote: [ -> ]i hope this helps and its not to early
Thanks, I'll start tracking down these bugs
are saves going into the userdata folder into each core? would it be hard to have a option to save them where the game is, that way no matter the device you can load a state on different kodi machines.
Is it possible to create a rpi2 build so i could test it too?
I've given the latest git a try and I'm still having the problem that when I try to activate a button to be mapped in the controller configuration that the button press is handled twice and the button gets assigned immediately. That makes it impossible to assign any other button to the first button in the list than the button already mapped to the "Select" action.
EDIT: I also always found it confusing that for the D-Pad the order of buttons is Up, Down, Right, Left and for analogue sticks the order is Up, Right, Down, Left. IMO we should make that consistent because I constantly do one or the other in the wrong order because they don't match.
(2016-04-21, 20:01)Montellese Wrote: [ -> ]I've given the latest git a try and I'm still having the problem that when I try to activate a button to be mapped in the controller configuration that the button press is handled twice and the button gets assigned immediately. That makes it impossible to assign any other button to the first button in the list than the button already mapped to the "Select" action.
Not this again
a problem on master too?
(2016-04-21, 20:01)Montellese Wrote: [ -> ]EDIT: I also always found it confusing that for the D-Pad the order of buttons is Up, Down, Right, Left and for analogue sticks the order is Up, Right, Down, Left. IMO we should make that consistent because I constantly do one or the other in the wrong order because they don't match.
Let's go with the analog stick method - Up, Right, Down, Left. For analog sticks, this is nice because you can simply roll the stick clockwise.
How about using the konami code for mapping? :p
I haven't tried master yet but will try to do that soon. I want to update my "shutdown controller on exit" anyway.
(2016-04-22, 03:03)teeedubb Wrote: [ -> ]How about using the konami code for mapping? :p
that code currently exists
I'm thinking of submitting the RetroPlayer PR early, and you gotta input the konami code to enable it
(2016-04-22, 02:35)garbear Wrote: [ -> ] (2016-04-21, 20:01)Montellese Wrote: [ -> ]I've given the latest git a try and I'm still having the problem that when I try to activate a button to be mapped in the controller configuration that the button press is handled twice and the button gets assigned immediately. That makes it impossible to assign any other button to the first button in the list than the button already mapped to the "Select" action.
Not this again a problem on master too?
Seems to work perfeclty fine in current master. I've also run into a lot of other bad behaviour in the retroplayer-17alpha1 branch where the axis inputs are constantly being triggered while the analogue stick is actually perfectly still. It then goes aroudn crazy in the GUI until I press another button (moving the analogue stick doesn't make a difference) at which point it stops going crazy.
Sega Genesis
game.libretro.genplus
Games are playable just running at low fps, strange though when you run Sega CD game with this emulator the graphics are messed up
Well i figured that one out BIOS was set to disabled
<setting id="genesis_plus_gx_bios" value="disabled" /> in Kodi the setting is called Systen bootrom
I think we should have this enabled by default, like in Retroarch and i don't even see this option in the libretro core
http://wiki.libretro.com/index.php?title...is_Plus_GX
if possible we should notify user if they load a game and don't have the bios, so they know why the game did not load.
anyways the graphics look correct now on Sega CD games
did some side by side tests of some cores on retroarch to retroplayer so what i notice is the games are running slow with frame tearing and lots of input lag compare to retroarch running the same core.