Tux

Just click on the line one day... ! why do you think it's flashing ? To be clicked !!! But there was an error here that I didn't notice from the conversion from sdl1, the bg color of the text has become green instead of blue, it's a common bug in the color conversions, it's fixed in git.

Quite impressive job, congrats !

Well raine is full of tricks to try to preserve it as much as possible, normally it survives a game reset, even a game load (loading a save state), and sometimes the test mode (but not always, depending on the way the ram is initialized or not). It's saved only when you quit raine or load another game, so if you go to test mode, and it's not detected because of a fancy ram initialization and then quit or load another game, you destroy the hiscore file. Sorry I can't do much about that... ! edit : all this pushed me to update the very old code for this turbo key, it was still trying to display the fps as an int when returning to limited speed when it's been a double for quite a few years already, so the display was very wrong. Now it's correct, and it also displays the average fps reached while the turbo key was pressed (I reach more than 4000 fps during the sfad demo and with a debug build, but it's the assembler version, but it's really too fast at this level !). And finally it automatically disables the opengl double buffer while the turbo key is pressed, so now you can finally use this key without changing anything in the video settings !

You are still very bad, I played a while with the turbo key (del), requires to disable double buffer in opengl rendering otherwise it's blocked by the vbl, but once it works it really goes too fast, half a second press and the demo has already done 1 loop ! Anyway finally reached 2 hiscore tables, they look fine to me except the weird C head for the guy on the right for the 1st picture. I think I am going to ignore your posts for a while... ! I attach 2 screenshots of the 2 hiscore tables... edit : after some short thinking, if it's the contents of your hiscore table which looks weird, you might have corrupted unwillingly your hiscore file for this game while experimenting with your scripts. Try to delete the sfad.hi file in the savedata directory (then restart raine of course). But if that's indeed your problem, next time at least post a screenshot it will save me a lot of time !

You are really very bad at reporting problems, was it with sf2 ? probably not since its settings are by dipswitches, so I don't even know which game you are talking about !!! If it's around the save screen, a savegame could be handy so that it can be reproduced quickly... ! sfad because of your other post ? No japanese language there, I checked the scores anyway, no problem to see. So I just don't know !

More details about this crash ? Related to your poke or not ?

You should know that cps2 roms are encrypted, it forces to separate data from code, I can't even say where your poke goes without checking the code, it must use at the same time a part which is encrypted and one which is decrypted, don't remember which is which. You can use the pheonix edition versions which are decrypted as a workaround.

For sf2 it's simply dipswitches, nothing to see in the test mode for that.

it's pretty self explanatory, see at the bottom of the page, up or down to navigate the menu, button 1 to select entries, what you want is system configuration, then make your changes, then exit (it tells to use 1p start to return to previous menu) then exit, it will save the settings to the eeprom automatically. You just need to read. Some games use a test mod like this instead of dipswitches. Of course here it's for a qsound cps1 game and these games don't have dipswitches, just a test mode. For the non qsound games, they have dipswitches !

Test mode (u key by default), which I tend to call "service mode" but apparently "service mode" is for when you start a game without inserting a coin, anyway, u key !

Of course not, it just changes the loaded rom image, you would need to save explicitely the file for that... ! Also notice that if you do the same patch for neocd, then it's not rom anymore, the code is loaded from cd to ram, but since the ram has the same address as the neogeo rom some patches are compatible.

$8d5d is the real rom, not a copy, poke works because it's handy to be able to patch region bytes, but you can easily crash a game if you poke randomly to rom. But clearly you know what you're doing ! Much better with some user comments. If you want, you can take a look at scripts/neocd/ssrpg.txt for an example of a script which uses variables to make things clearer (which avoids to add comments).

Yeah well don't forget most of these games if not all of them are from Japan so they translate texts to english and so sometimes you get inconsistencies, it's probably not a good idea to always copy from them. But it's a detail you are free to choose here. Yeah I think the trigger idea is a good idea. When I did the original mapping it was without the controllers concept so I couldn't identify which control was the trigger and so I had to do all the controls with the standard directions, it's not the case anymore... But I didn't want to add yet another input just now, but I 'll probably do it for next time.

Yep I second the opinion of mer-curious : comments welcome ! Most of the scripts currently in raine are from a conversion from mame, and there are not lots of comments (sometimes there are texts introduced by comm which appear at the bottom of the screen when you select the cheat...). Anyway I added the ability to add comments anywhere on a line starting by ; or #, and here you are even patching the rom from what I saw, and I have no idea what most of these poke do, that's where I see you are an expert at doing these patches, but comments are really most welcome here...

You never know about that, they are used if they are here even if indirectly, it can be sometimes an important bug fix or a new important feature. Anyway I updated only dlls32-0.92x.7z with the previous libwinpthread-1.dll it should be ok now (on my side at least it still launches, but I don't have xp). I don't think it's worth updating the 64 bits version, xp 64 bits is a rarity I doubt anyone will want that.

Ah yes I forgot about this. Ok then it's from libwinpthread-1.dll. It should work if you replace just this dll by the previous one. I'll update the dlls packages again with the old version of libwinpthread-1.dll. It's a lost cause anyway as already said, xp is closed source and Microsoft decided it must die, so it will die, no matter what with time passing the new packages appearing will require functions which are not there. For now it should still be possible to work around it this way though.

1 Raine's romset ? That's how it's called on internet archive... yeah by the way it downloads most of its roms from there now, so the question doesn't really matter... Except that it's mostly compatible with mame, but mame compatibility is not top priority here. 2 done automatically, if you find a rom which is not downloaded automatically, post ! 3 users who tried to select some sprites where they apply transparency effects. It's not always a super choice, but it's worth testing. The option can be easily enabled/disabled so it's not a problem. I don't remember a game where I can't play without this option, but it was easy to add so I did it, mostly out of curiosity. 4 emudx was an emulator from Mike Green, the idea was to apply some more recent graphics and sounds to some very old games, but it was to disappear after a few years, so I proposed to merge some of the emudx changes in raine, he accepted, and that's the result. You can see that in dkong, pacman clones, miss pacman, frogger. In my opinion the best result for the graphics is probably dkong (and I missed some alpha blending effects which could not be reproduced easily), and the most impressive sounds / musics are for frogger !

Yeah well I won't jump on it all the time, I jumped 10 versions last time... ! Except if someone spots a very interesting change in it !

Yeah 0.94.9 was the big update on a lot of things, here it's back to normal, the latest history.dat is in extras, there : http://raine.1emulation.com/download/extras.html and I should remove the 2.40 version one day... ! If you always uncompress in the same directory, you keep the files which were included in 0.94.9, that's the idea.

The big focus of this one is the fix around the cps2 driver for the win64 binary, but since these are some bugs which can affect all the versions even if it's less obvious for the others, everyone should update. See the forum thread for the gory details about this bug. Except that : - added some rnd() function to the console to get some random number, help included, I also added some help for the basic peek/dpeek/lpeek functions. - New GUI options : "background animation" to eventually stop the background animation for those who suffer from motion sickness, and "Autostart drivers", essentially for me, when I don't want the driver to start in the background as soon as it's loaded. - some minor text changes "Action Replay cheats" -> "Game cheats". Oh yeah there is an optional update for the dlls, I removed 4 dlls from the big list by tweaking the build settings for freetype, these are not the biggest ones, but they sure were useless in windows, so it's a little better. You can keep the old version, the update doesn't bring anything except removing these 4 dlls, and I should probably rename the dlls to some other package because I am bound to have some incompatibilities with old versions, but it will be for next time... ! And I finally used a script to update the windows binaries to avoid the kind of mistake of last time ! Find the update there : http://raine.1emulation.com/download/latest.html

Yeah as usual I made the update and then updated the dlls in the old package so people don't have to redownload all the dlls just for that. Yeah why not, I wondered also from time to time why it was chosen like this, probably because action replay was very famous at the time of raine creation, but it's quite a different system indeed, they usually use some kind of device... It's done, and this time the region is not broken. By the way, shouldn't it be "Game commands list" ? It's a list of many commands and not just 1... ! Hehe, sorry but not this time, because there is a difference with the keyboard : if when you move your pad or your stick on the right you are not on an item with multiple selections then the menu scrolls 1 page down, equivalent to page down key. Yeah I guess the scroll down key could be mapped to something else like triggers, we'll think about it, eventually later, but not for now.

Ok, finally fixed everything. The trick with this kind of problem is that even if debug info is not allowed, if it crashes, you can track it with a debuger and with enough patience you finally find the cause. There were actually 2 problems here : 1) a very very old bug, there is an array for layers and layers priorities in cps1/cps2, so I reserved 3 members for these 2 arrays since there are 3 layers. Except that sprites are seen also as a layer in this driver ! So I needed 4 spaces ! And why did it crash just now ? Apparently because gcc decided to change the way variables were allocated in this version for win64, before it was harmless, I don't know where it landed but it was unnoticeable, and with this change from them, it became the end of the world ! 2) a very optimistic speed hack module, that's why save games were crashing sometimes. It was storing the data it needed in the save by assuming variables were allocated in memory as they appeared in the source. It's clearly not the case anymore in the latest gcc, and anyway it's a bad assumption. So I moved everything to a struct and now this bug is gone too ! (I don't guarantee old savegames using this module will be compatible, it's used at least by cps1/cps2, but not only). So finally gcc was not guilty, it just allowed me to find some old bugs... Now I am a little tired so I don't promise a release just now, but it should be good for tomorrow or maybe a little later... !

Well after some tries : the basic random function in C is not exactly random, it expects some random seed to be fed, so you can initialize the seed yourself, or call repeatedly the rand() function for a real random interval this time to get a chance to get some real random number... ... too complex for that. Luckily for you there is an alternative : openssl provides its version of a random number generator and random numbers are important for cryptography so it's serious this time. Since it's already used by curl, it's just a lib which was already here anyway. The result : a function rnd() which returns a random byte, that is, a number between 0 & 255. So you see, simple request, but not so easy solution... ! But anyway it's done...

oh well... ! I guess it can be done, not top priority, I'd like to find the problems with the cps2 driver 1st, but ok... !

Not really convinced here, when I use a cheat I want it to be either enabled or disabled, not 50% chance.