Windows Nightly discussion thread


#1

Windows nightlies are now available. The nightlies have no cores included but you should be able to fetch the cores from the core updater in RGUI

http://buildbot.libretro.com/nightly/win-x86/ http://buildbot.libretro.com/nightly/win-x86_64/

Known issues:

  • invalid content directory set in the latest pack

Relevant changes:

  • 19/03/2015 netplay and the core downloader should work now on x86 windows
  • 08/05/2015 overrides support savedir/statedir overriding
  • 08/05/2015 there is an option to sort savestates / saves automatically on a per-core basis in general settings now
  • 19/10/2015 add experimental support for retroarchievements

#2

Hey, I have 3 little questions to RetroArch since I observe the development of RA.

  1. Why aren’t the system files shared in RetroArch Nightly builds? As you know, they are necessary to run for example Super Mario Kart. I always copy them from the build of lordashram on 30th october.
  2. Why there are not all RA cores in latest? For example, before you reset the buildbot page on 27th february, there was also a beetle_psx_libretro.dll.zip core which is missing now since 27th february. Did the development stop?
  3. Wouldn’t be better to have empty folders in RA builds? The “Load Content” option refers to the “content” folder, which is missing… because it would be empty. Also of course the “system” folder is missing. It could be confusing for new users without these folders.

And thats all :slight_smile:


#3
  1. SYSTEM files are often BIOS files and they may (or not) be copyrighted. Also RA is not an emulator so there is no reason to include those
  2. Because RA is a separate entity from the cores. In the nightly page there are instructions to download all the cores. Beetle is experimental, broken and generally speaking not needed atm.
  3. You’re expected to configure your environment. Set your paths at least.

#4

Is this the correct place to report bugs or inconsistencies in RetroArch? If so…

I found a problem with the save folder in RetroArch from the 2015-03-02 build, and which probably persists even now or was before:

I recently manually set each core configuration with a separate saves folder, like “RetroArch->Saves->Super Famicom or PlayStation”, and so on. However, when I load a core the saves folder path is set only once for the first core I load when first initializing RetroArch, and no longer sets the saves path for the next cores I load afterwards. I hope that’s a clear explanation on this problem. Probably, the fix would involve reading the core’s config file and set the current RetroArch saves path to that once the core has loaded, as it seems now it only makes use of the saves path written in the RetroArch config file, not the core’s one.

I also found some nasty bug regarding menu navigation with the glui menu driver. Basically, sometimes on a long list of menu options, the cursor becomes crazy and the column navigation breaks, not showing the next options when going up or down, or having bad wrapping. It is nasty and hard to replicate. I hope someone finds the cause of this bug.


#5

[QUOTE=Ragna;20455]Is this the correct place to report bugs or inconsistencies in RetroArch? If so…

I found a problem with the save folder in RetroArch from the 2015-03-02 build, and which probably persists even now or was before:

I recently manually set each core configuration with a separate saves folder, like “RetroArch->Saves->Super Famicom or PlayStation”, and so on. However, when I load a core the saves folder path is set only once for the first core I load when first initializing RetroArch, and no longer sets the saves path for the next cores I load afterwards. I hope that’s a clear explanation on this problem. Probably, the fix would involve reading the core’s config file and set the current RetroArch saves path to that once the core has loaded, as it seems now it only makes use of the saves path written in the RetroArch config file, not the core’s one.

I also found some nasty bug regarding menu navigation with the glui menu driver. Basically, sometimes on a long list of menu options, the cursor becomes crazy and the column navigation breaks, not showing the next options when going up or down, or having bad wrapping. It is nasty and hard to replicate. I hope someone finds the cause of this bug.[/QUOTE]

I have the same problem with RA PC, it happens with any separate folder (including the browser directory), the next core selected takes the folders from the last core.

With glui menu driver in my case the cursor become crazy, but my RA PC crash when I try to navigate.


#6

Try disabling mouse for the navigation issue. I don’t use per-core configs so I’m not sure about the issue you’re mentioning


#7

As I said, the problem is RetroArch ignores the saves path set on the different cores config files, which is a problem if people want to have everything well organised and not a mess of different systems save files in the same folder.

Is there any way you or someone else could look for this issue? Thank you.

PS: By the way, I myself never use mouse navigation, only keyboard or gamepad, yet sometimes the menu navigation crazyness appears.


#8

Yes I’ll forward it but I don’t think it will be immediately fixed. I don’t use per-core configs because I find it pointless, I load my own configs with --config or appendconfig About the navigation weirdness, is mouse navigation enabled or not?


#9

Might be related to this


#10

[QUOTE=Radius;20489]Yes I’ll forward it but I don’t think it will be immediately fixed. I don’t use per-core configs because I find it pointless, I load my own configs with --config or appendconfig About the navigation weirdness, is mouse navigation enabled or not?[/QUOTE]Obviously not, as I said. It is a very rare bug (and random, one would say), so I don’t really expect this one to be fixed Soon™…

[QUOTE=Radius;20490]Might be related to this https://github.com/libretro/RetroArch/issues/857[/QUOTE]

Well, actually I don’t see the similarities. It looks more like that other issue is an implementation thing related to, for example, the PPSSPP system folder which holds saves too instead of the default saves folder?

If I could, I would solve my issue, because it seems like a very simple check needs to be done to load the core’s config file save path, but I don’t know how to submit fixes, nor have I touched this source ever. sigh

Anyway, I don’t want to look like I’m here just to make complaints (bug-reports, in this case), so excuse me if I “sound” (read) like I am. It is just that I fell so in love with RetroArch that any issue might become a big issue for me. I just cannot go back to the old ways of emulation anymore.

PS: I use RA on Windows (8.1), if that is of any help, anyway.


#11

PPSSPP is meant to work like that, it doesn’t implement the SRAM interface

BTW, netplay and the core downloader should work now on x86 windows


#12

Starting with the March 7 win-x86_x64 nightly build, RetroArch does not pass the libretro path to rgui. This happens from both the command line and the cfg file.


#13

yes, it’s know and it shouldn’t affect anything, it just makes sure that a core is not loaded at startup


#14

It does affect me. I have many scripts that rely on this feature. Is there any chance that this feature will be restored?


#15

No I don’t think so, that produced some unexpected behavior


#16

All the nightlies have been removed from the server. Any idea whats’ going on?


#17

What are you talking about? http://buildbot.libretro.com/nightly/win-x86_64/


#18

Why it didn’t work in my windows 8.1?


#19

what didn’t work?


#20

Hi! It doesn’t seem like the MAME core option ‘hide_nagscreen’ is working anymore in the latest nightlies.

Can anyone confirm?