Tux

Yeah welcome to the wonderful world of utf8 insanity, your 3 bytes is just a weird utf8 sequence which is ignored if seen in a browser but raine can't ignore it of course ! It happens sometimes especially when getting a text from the web, you have to find a way to filter out this kind of sequence in this case. Anyway you found a way around the problem so it's fine.

Ah yes you're right, forgot this "detail". Well good news : it doesn't change anything in linux, because there is some compensation built in the driver for this controller for its horrific dead zone. But in windows where there is no compensation, indeed it happens. I'd say maybe there is a way to improve the dead zone compensation for that in raine, but really, the easiest solution is just to use the left stick in the gui instead of the dpad OR just not relying so heavily on the auto repeat, OR even better, just use the keyboard it's much faster to navigate the menus anyway ! Apparently the dead zone compensation works as long as there is only 1 axis moving, if there is more than 1 at the same time, it creates havoc and since this thing is a dead zone nightmare, moving the left sitck 1st as you do makes sure that raine will receive moves from the left stick forever after this, add either the right stick or one of the triggers, and the dead zone compensation becomes crazy ! I'll see if it can be improved, but it's super low priority because you have many ways to work around that. Well the difference with the ds3 is that you can stop the perpetual movement with the ds3, it's not always easy, but always possible. Here it seems to have become impossible (in windows at least !). What's this horrible light for by the way, the big one they added on the controller ? It's a generation 2 too, yes. edit : after looking at the code, I don't want to make it more complex. If you find a case where it's problematic without moving one axis while pressing the dpad (that is all the axis close to their neutral position), I'll have another look, but for now it's just not worth the trouble. I noticed that usually most games don't even bother adding an auto repeat for the joystick when it's usable in the gui, so it's already better than most here.

On 2nd thought I can add something : if your right stick still interrupts movement in 0.94.2 then it means it sends some left stick movement. I don't know if it comes from a bad mapping or from a hardware problem, but that's the only possibility.

Well in this case you are on your own. Tested here, without any fancy configuration file or anything and it works, if it doesn't for you, it's something on your side, I can't help you there.

- In the gui only the left stick and buttons 0 & 1 are now handled - sound associations improvements for kof94/95, and enable this for samsho4 and real bout fatal fury 2. - and the hiscore.dat file is smaller, got rid of a ton of useless stuff in it. That's all !!! http://raine.1emulation.com/download/latest.html

Ok tested this famous ds4 controller, well nothing fancy here, it didn't require any driver in linux or in windows, recognized immediately. In raine too, no need to add a mapping it's one of the most well known game controller. With 0.94.1 yeah you can interrupt the auto-repeat from the gamepad either by the right stick or by L2 or R2, it's because these are axis movements and axis movement are used to move the selection like the gamepad. With latest binary since only the axis coming from the left stick are recognized, all the other axis are ignored in the gui and so the auto-repeat started by the dpad is never interrupted. Tested in win10 & linux, without any fancy mapping in both cases. I guess I should finally release an official binary for 0.94.2 to finish all this for good... ! (+ the new sound associations need some testing from players anyway). The only slight complaint I have against this controller is that if you are using testjoystick on it from the SDL distribution and move one of the analog sticks, it's impossible to stop the move, it's a dead zone nightmare, but if you use something which handles the dead zones correctly it's not really an issue.

You are in luck, my nephew will lend me one tomorrow for testing ! (but I am almost certain I won't have that many problems !)

yeah my cheap controller has analog triggers, although their resolution is not very high (it jumps very quickly from minimum to maximum, very hard to do something precise with that). So you can interrupt the auto repeat with the right analog stick ? Well I have absolutely no idea how then, and no I can't do that here, with the dual shock ps3 controller or the chinese xbox one ! No I don't plan to buy a new controller specially to test your setup, that's too much to ask, eventually one day, but your problems are really not a motivation !

Same kind of signature, will be enabled next time, and will require some more testing too.

You replied yourself, the detection is done on some signatures in the sound rom, some are not tested. Although "Samurai shodown" is supported here, but "real bout fatal fury 2" was not, it has a signature very similar to some other games though so I can enable it for next time, but it will require some more testing to check if everything is ok. The detection of the 1st tracks is right. For samsho be sure that speed hacks are disabled in neogeo/neocd options otherwise it destroys the intro sequence.

There is a misunderstanding here, you are not saying your triggers can move the selection in the gui, just that they are interrupting the auto repeat feature of the dpad in the gui ? Well assuming they are seen as buttons too, and I guess it's the case here, it can't be avoided, the autorepeat works as long as it does not receive another button press. Eventually send the string for your controller either in userdb.txt or gamecontroller.txt to confirm they are seen as buttons, but if it's the case there's nothing which can be done, learn to live with it, that's all. For info I use a very cheap generic Chinese controller found on amazon as my main one, it doesn't have this problem (although it has some other problems, the dpad is unprecise which can be very enerving at some places where you try to input some precise combination, and it can be used only wirelessly, the cable is only to recharge it, it can't be plugged as an usb controller, except that it's recognized as an xbox controller by default which is fine). Well actually buttons other than buttons 0 & 1 could also be totally ignored (that is buttons A & B). Crazy thing...

No I meant only in the gui, of course these can still be mapped ingame to whatever you like (even though it's not easy to find a good usage for these!). Well what you describe doesn't make any sense (L2 & R2 should be totally ignored if it's a game controller in the gui, if not there is really something broken somewhere, is there a joystick which is not a game controller connected at least ?), there is probably a problem somewhere, but I can't reproduce it here for now, maybe later when the weather will be more "reasonable" !

Sorry can't help you more with that, for me the triggers are completely ignored in the gui now and they were not before, can't do anything more about it. Are you sure your ds4 thing is recognized as a game controller ? If so its inputs for the left stick should appear as "leftx" and "lefty" in the controls.

Yeah aof3 is a "type 2" too, so it's affected by these changes. Your gui "fix" is inside too, so this binary can be considered a 0.94.2... except the code for it was very short, so I'll wait some more but you can use this meanwhile !

Yeah replied the post just before this one, happy testing !

Sorry I can't reproduce anything with this save, the graphics are totally garbled when loading it, I don't know how you did that, but the game is not frozen, it still reacts to inputs, I can press a+b+c to get to full power but it has no effect on any music playing or not. After that a makes the final punch, victory, and still nothing happens for the music. Ok, finally did it, there are random sounds playing, one of them stops the music, which gives us 2 more prefix commands, because I tested some more manually, and another test build. This one should be good everywhere, assuming it doesn't break things in the other games of this kind, but it shouldn't. exp.7z

You are supposed to use "Stop" from the sound commands dialog 1st. If you forget that's not a disaster, just leave the associations dialog, use "Stop" here, and go back to the associations. Sorry stopping the song from here is not so easy for me !

kof98 is a completely different game for the associations, not at all comparable with kof94/95 sorry. So changes for these 2 will not impact kof98 at all. These 2 category includes galaxyfg, 3countb, fatfury2, fatfury3... and so it's a problem to know if the change detected here is specific to kof94/95, or general to all these games, so it will take quite some time to test all that. edit : it seems there is a new prefix for the sound commands, this makes a new experimental build to test, 64 bits only since that's what mer-curious uses, it seems to work for kof95, maybe for kof94 (I was not able to reproduce the problem here), but expect no change for kof98. For info I found this while making no real associations for the game, I just used the debug messages I get in a debug build which tell if the song changes... ! exp.7z

Specific to a particular stage or any battle ? There was no change in the current version for this so it's probably an old problem Well I can't seem able to reproduce this in kof94 (but I am not a good player !), so I'll need a lot more details !

Well the easiest solution is simply to put the controller on a table instead of on your lap ! Yeah raine reacts to any "stick" in the gui, the code is from before game controllers so there was no way to identify a movement to know from which jostick or button it came, and so it reacts to any stick movement. I guess I could at least remove this now so that for game controllers only the left stick is handled for that. I'll do that.

It's just the silly bugs reported in the thread of the previous version fixed, nothing more, no comment... ! Get it from there : http://raine.1emulation.com/download/latest.html

Ok, confirmed, that's if using 2 neocd games which both read their files from an iso, it's a 1 error mistake, forgot to clear a file after freeing it. Workaround : extract all the files from the iso and put that in a zip file, then delete the iso, it will take less space and it will work ! It's fixed anyway...

Ok, thanks for testing, I'll make "force" the default for double buffer then, probably more reasonable, and add "(adaptive)" to the yes. For the mouse bug that's where we see I use this mainly with the keyboard and you use the mouse, we already had the same kind of problem before, and I was caught again, forgetting to test the mouse, ok I'll fix that later. The blank space is just the separation between the header and the dialog, use the feature, you'll see it becomes handy when a few games start to populate the list. The save state is just a 32 bits version loaded in a 64 bits version or the other way around, it already happened to you, maybe I should try to make the message clearer, but it's really what happens it doesn't recognize the ids of the cpu emulators of the other version, and making the message clearer is not so obvious, but I could add something generic like "maybe it's a 32 bits save" when loaded in the 64 bits version, and the other way around for the 32 bits version, I'll do that.

Yeah not a big release, mainly small fixes but I had enough of 0.93 already. Main changes : - the new "most played games" feature, which shows the 5 most played games sorted by time in the game selection dialog. You can get details about that in the forum. It wouldn't be too hard to add a dialog to get the stats for all the games played and not only the 5 tops, eventually for later. More details on how it works can be found in the forum. - A small fix for downloads with curl, if downloading a clone and the parent was not in the last rom directory tested it was downloading the parent even if it was already in another rom directory, should be fine now. - A long overdue update for the profile display (the one where you get % times in the upper right corner of the screen). In all the time since I added support for sdl, I never realized that all the sound processing was done in a separate thread, so the profiling must be moved there. This makes the display more precise, also the results are rounded now instead of truncated. - The slow down in kof97 for the 64 bits version was found and fixed. - double buffer gets a "forced" option to ignore completely adaptive vsync, need more feedback on this from testers to see if it's really useful. - a workaround for sdl-image 2.6.2 which doesn't handle png files with a palette anymore, preventing the colors from the raine logo to cycle ! I can't do anything about it for now, so in this case the logo is displayed statically without color cycling, but it's advised not to update the sdl2_image.dll for now (or equivalent in linux). It will help for linux distributions like arch where 2.6.2 is already the official version, I posted this to their bug tracker so we can hope a solution will be found. - quite a few memwatch fixes, most of them were harmless, a few in the console might have been useful, I fixed them without investigating. - kurikinj is removed because it's exactly the same romset as kurkint except for the region byte, making this version the 1st version which actually supports less roms than the previous one ! More details about that in the forum too. - quite a few hiscores updates in the hiscore.dat file ! Now that it's bundled with the raine binary and in git too, it's easier to make our own updates on it. Actually I had added some scores in it for old games from the multi15 rom a long time ago, around 2019, but they were clearly lost in a hiscore.dat update, so I have put them back in, mostly for some pacman clones, adding xenrev which never had a hiscore entry before (the main goal of this version is to make the player lose !). Also battle bakraid finally gets a hiscore entry for its normal and unlimited versions. There was an update for sdl2 to 2.0.24, nothing essential there, mostly things related to game controllers, maybe I should merge the binary update in the dlls packages. For now I left 2.0.22. Ok, this small update is there : https://www.1emulation.com/forums/forum/126-raine/

Ok done. The game selection dialog now displays the 5 most played games at the top (in the header, which is now separated more clearly from the main dialog), sorted by play time. The number of times a game is loaded and how much time it's played are tracked (watching the game intro behind the menu is not considered playing time !), and it's tracked even if you close the window without using the quit command. All this is saved in a new file "stats" in the savedata directory, the contents is ascii but it's not supposed to be edited by the user. Games stats are stored by the short name of the game, which means that if a game is renamed for its short name the stats are lost (it happens almost never in raine anyway). Neocd games are not supported, only arcade games (well neocd games are not even loaded from this dialog so it would be quite a problem, and they don't have a real short name), the picture behind the game selection dialog is not updated when selecting one of this favorite games in the header because it's not worth the effort and it would make some duplicate code, and I don't like that. 5 only for now, because the header doesn't handle any scrolling bar and can't scroll its content, it should be enough for a start anyway. I'll test this a little more, but it seems good now (fixed some bugs related to the header display at the same time).