Tux

Just built a binary with curl support inside for windows, for once I used the binaries from the curl site because it's impossible to build curl with my current cross compiler, and it saved me the hassle to reinstall the whole mingw system in windows (plus compiling curl is really complex). The page is here : https://curl.haxx.se/windows/ just copy the include and lib dirs to the same directories, the dlls where they are needed, you'll need also the openssl archive listed below curl, and voila, it works. I was ready to make curl support optional because of the complexity of building it, but thanks to this, it remains easy !

It's exaggerated to call this "os", it's still linux the os here, just take one of thes live distributions about games, change the contents to add raine + some roms, and you're done. I'd suggest using the latest git version too... !

Mainly because they change all the time in mame, and it's tiresome to follow... ! I always thought the main set should be the oldest one, and any newer set should have another name, this way you keep compatibility with the old sets, and those who want to try the newest one still can. But it's too late for that ! Anyway they seem to have all the neogeo roms, some cps1 (except the clones then). Those notably missing : puckman and bubble bobble (but they have puzzle bobble, arcade and neogeo). It's not ideal for someone who wants everything of course, but it can be convenient for the casual gamer...

I added internet archive support directly in raine, that is, if a rom it tries to load is not found, it tries to get it from the internet archive. They don't have everything, they seem to have only the parent sets, and it's mame material which means some roms are incompatible, but still, there is quite a lot of stuff over there, and no copyright problem (they state they eventually remove the file in case of copyright problem). The transfer is a little slow but I added a nice progress bar for that, it's quite fun ! I've been having this idea to add internet support in raine for quite a few years already, it's based on curl to make it work easily and on all platforms (no dos, but dos is not considered as a platform !). When the internet archive started to make these roms available it was in a single huge file which was unusable, but they seem to now provide separate links for each rom, and it's been like this for at least 1 year but I just found out. Quite nice !

Switching to sdl2 would not be easy anyway and it's not planned at all for now, even if it would be a good idea on some aspects. It's the problem with raine : too full of low level stuff, it pushes things to the limit and here changing from sdl1 to sdl2 would be almost as hard as what it was to change from allegro to sdl1 ! Ok, thanks for the report, if I am not mistaken it's something from november 2014 (!) which was supposed to help to load broken cue files where somebody renames all the files to the same base name without changing the cue file. But the surprise here is that the bug is only if the iso given in the cue file is not exactly the one on disk, like an iso.gz file on disk while there is an "iso" file in the cue file. So most of my cue files worked, only the one testing an iso.gz file could not work. To be sure I tested again most of my cue files and everything works. It makes sense, in 5 years I guess I should have a few reports about broken cue files, it had to be specific, but I am not 100% sure I found the right bug for you, you'll have to test again. Also I fixed the error case where the iso really can't be found from a cue file, in this case raine displayed an error message but crashed anyway after it. So I am not 100% certain I found your specific bug, so you get a new binary, very long delay for the reply because of broken 1emulation, the new one looks better... rainew.7z

And you get a link to this compiled windows version, full package with locales updated inside, including the italian one, that's the least I can do : https://mega.nz/#!KRcyxYqY!hQvPgdeea5SRZo6LYk9tY7SSfc0dNhe-Rpb6SqSnp3w since 1emulation seems unreliable for now, I keep mega... ! and thanks again for the time you spent on this !

Added crc info in the error message ok, 7z doesn't display the crc but anyway... For the button 2 to cancel, it's "standard" behavior, it works also with a pad/joystick, the 2nd button is assumed to be used to cancel, it's esc on the keyboard. A little weird at start, but convenient when you are used to it.

wow I finally reach the end of this (except for the video bugs which will be one day... !). For the neocd screenshots : not convinced, it would have to scan the directories before you select them for that to be any useful, so it would slow things down with some input/output. Then when you are in the directory it's too late, you are seconds away of seeing the real thing in action... ! For compressing isos : there are actually 2 ways, you can gzip the iso, but it's quite unefficient. The best way is to extract the files from it and put them in a zip, it will be better compressed than simply ziping the iso and super fast. At this point you can just get rid of the cue file, just point raine to the zip you just made, it should guess about the audio tracks when needed (assuming of course their track number is in their name !).

cpu frameskip : it was a way to slow down the game before having something to tune precisely the fps, now using the fps setting in game options is probably better. benchmark : if using opengl just disable double buffer in the opengl options, the limit framerate of course, and you are good to go.

4-9 : can't reproduce. daioh pause -> cpu at around 6% for raine, 100 being 1 core fully loaded. Tried the windows version in wine too with the same result... 4-10 : not at alll normal, raine is super efficient on the cpu, it's more like cpu in power saving mode at 1 GHz and using 2% cpu power... ! No idea what causes this, never seen or heard anything like that. I just booted in win10 to make sure I get the same results as in linux for that : yes of course. I don't have any xp boot so I can't test xp. You can eventually try to play with the UseRdtsc seeting in the config file but it shouldn't change anything, the rdtsc is only used for the profiler display (f11 key 3 times), there is an sdl timer for the synchronization now. I have run it on quite a few computers now, and never seen anything like that until now... edit : after checking in the code to refresh my memory on this, on its normal behavior it goes explicitely to sleep waiting for the next frame to start which makes cpu usage super low. Start by checking the profiler (f11 key until you see many % numbers in game). The last one, Free should always be the greatest one. If you see it go down while blit goes up, it means your screen has a refresh rate < the game refresh rate, but normally this should not put any load on your cpu, video drivers normally just wait without loading the cpu in this case, unless you have a problem on this side. You can also lower the game refresh rate in this case, 59 fps instead of 60 can be enough for a monitor with 59.25 fps for example. I can't say anything more on this, you'll have to investigate, I can only say it's not normal at all. A last thing though : raine uses only 1 core, it was made for old cpus so there is not much point to use more than 1 core considering its speed.

3-2 : 2 very stupid bugs when rewriting stuff in a more compact way... ! 3-3 : it's about loading a game which has no region after one which has some. Amazing it got unnoticed for so long, you are good at finding these bugs, you probably spent quite a lot of time on this... ! Graphic bugs : too long to fix, sorry, it's known raine has some graphic bugs on some games, lately I fixed syvalion (controls and zoom effect on the title screen), but I don't do that often because it takes a lot of time. Well since I don't kinow the arcade version of darius 2, and I never played it on any other emulator, I never noticed anything, but it's possible I guess... ! Here too, too long to do now, we'll keep that somewhere for some other time... 3-9 : yes I have sound, and from the very first seconds (music during attract mode). 3-11 : More detail ? Is it really obvious ? 4-1 : you need to actually run something to get some info here ! You weren't really intuitive on this one... ! 4-2 : I'd say that sometimes you loose the background for some unknown reason, yeah I know, hard to track and to fix, maybe one day... 4-3 : ah yes it keeps on fixing the aspect ratio when resizing the window, forgot about this. It's more intended for the difficult fullscreen modes where you want to get rid of these black borders at the cost of some quality. 4-4 : Normal blits are for fullscreen, even if they can work in windowed modes they are not meant to be used in windowed modes, it will work but just don't resize your window all the time then... ! 4-5 : the shaders never resize the window ! The hq2x shader produces an hq2x output but then it's resized to fit the window, it's your job to choose a mode where it makes sense, but anyway you usually get a better zoom wihtout even caring about the mode. In this case yeah the window should probably resize to its previous size all the time, the easiest fix is to always have a window big enough so that it won't resize. This kind of thing is a pain to fix, especially with sdl-1.2, switch to sdl2, one day, in distant future... ! 4-6 : I'd say it's more a windows bug related to sdl-1.2, it works fine in linux, don't expect a fix soon, I hate tracking windows bugs like this... ! 4-7 : ok, there is really a problem here, it seems to use the setting from normal blit to adjust the window size when in opengl, it shouldn't touch the window size at all in this case, I'll try to look into it... ! edit : ok, it's done, now the window is resized only when using normal blits and choosing to set the mode automatically, which should always have been the case, sorry ! Actually this code is some old code from allegro adapted to work with sdl, and it's quite messy. Anyway, much nicer to use when in windowed mode now.

3-1 : all fixed, these were all very old typo errors in the dsw definition, for the cps1 one (mtwins) it was because my converter didn't like a comment which was added in the middle of the definition and it got lost because of it. Very old bugs, I suspect that at a time they were not making the emu to stop because it's hard to believe they would have gone unnoticed for so long, probably an effect of the not-so-new dipswitches dialog. Much later for the rest, need to get out now !

1-5 : I don't know how to detect the default language in windows. Raine is multi platform and each platform has a different way to do it, even osx which is an unix doesn't do it the same way as linux (although it's rather similar). But in the case of windows you have to call some windows api function for that apparently, no way to find it from the environment, and since I almost never program the windows api directly I have no idea which one to use. I would need a real windows programmer to take care of the windows version, but again, nobody seems motivated enough for that. So you'll just have to choose the language in the gui, I added the choice for italian now... 1-6 : it's a samurai spirits rpg option, only when translating, if you see it anywhere else, report... and you shouldn't try to use it, it's some really low level stuff. 1-7 : yeah I know, fed up at editing the year all the years so I just stopped, don't complain, I am still here anyway ! 2-1 : not for me, and probably for lots of other people, I only tested on 1 game though, I guess something is broken in your cue, send it, just the cue file, you can use mail for that (or send a link to the whole package eventually). 2-2 : U key usually, check the raine keys, it's associated with "Test"... ! 2-3 : default key is LCTRL s, I even removed my config file to check, reported as LCTRL in the gui and working with LCTRL, and indedd if you put RCTRL instead, it works. 2-4 : yep it's for those who can compile only, reserved to debug builds, you wouldn't know what to do with this anyway, and the cause of the crash is not always there, most of the time you need a debuger for that. 3-1 : if you had a proper console (like the one from mingw, but you are not a dev so I guess it's normal), you would have seen an error when raine quits : "unknown extension 0", it's a collision with some new way to write dipswitches, thanks for finding this one, I'll need more time to look into it... later again then !!!

I didn't really understand what you meant about the logo, your version has more colors which creates a slower color scrolling, it's nice but I also like the old one (faster scrolling = more "lively"). Both are good... The ascii version is a good idea, I never tried to create one... But I wouldn't want to use that in the exe though.... For the translation, yes all the strings are not translated, because I used some script to guess which strings to translate, and they are now encased in _("...") in the source, all the strings which have no _("...") are not translated (and there are some strings which must not be translated). Year : 2006 is just the 1st initialisation of the year field, it's overwritten later by the real thing... this one should not be in the po file... ! That's where the fact that you can't edit the sources is too bad, here you could have added the _(...) where appropriate and then propose a patch for that... ! About categories I am thinking about ditching this entierly, I had never any feedback on this and it can make things awkward when handling a lot of game drivers to update and you don't have the information... Getting rid of it is probably the easiest fix here. For controls / language in the dsw, I don't see why to change the string heree, it's still about controls and language even if it's in the dipswitches... lol for the disappeared "pause when focus lost", I never used this because on my system it thinks it looses focus when switching from gui to game mode in windowed mode, maybe I should add a few frames of lost focus instead of just 1 to make this useful. Anyway I took it back, this dialog is a little crazy taking settings from everywhere and it lost this one when I added the "no hud messages"... ! For the command line in windows I guess it probably works anyway but I can't test it here, you'll have to wait some more for that... ! Later again then ! edit : a last one, for the sprite viewer for vertical games, just switch to fullscreen !!! By the way this feature is just a curiosity for most games, it was added when trying to translate "samurai spirits rpg" neocd game, and that's the only game where it can be useful... but you can use it on any game anyway... After testing quickly in wine : of course the command line works even in windows, maybe you just get no output in your console, in this case use it like that : raine32 -h > output and then check the output file. I might test this later in a real windows, but I am sure it will work. For the .po, you decided to create it without asking anything, good, but normally .po files are created from an initial raine.pot file which is created automatically from the sources using some commands. Then you rename the .po to something related to your locale, edit it, and then pass it to some other commands to generate the .mo file which is actually used by raine, this last pass must be done by the makefile and not by any other tool even if you can use another tool for testing, because these files should be generated on demand. That's why there must be no error in this last pass and why I had to edit your .po. I can keep the ordering of course, if some people want to take the italian .po file as a base to create a new one they'll get something to start from, but the tool which creates the default .po file doesn't care about ordering, it just adds a comment to say where the string comes from in the sources, that's all... ! 1-4 : well I tried to load 1941 first, got the option to appear in the menu, then loaded buble boble, and the option was gone... you'll need to be more specific then !

I have fixed the po as I said, you can grab it from git to see the differences from your file by yourself... Later... For the old dos link, in case you are interested by the fixes inside I am afraid I forgot them completely, this was 4 years ago ! Maybe some of them were merged into the main source, I am not sure about that. Anyway yes the download link is broken for some weird reason, they seem to have installed something to limit the bandwidth I would say, and as usual it seems to fail miserably... ! Sorry I don't have a backup of this specific file (at least none that I know of !), and I can't retrieve it, even using the ftup access I get the same error, the download begins to stall at about 66% and then it takes forever. It seems to progress though, but very slowly, I'll try to keep it to see if I get a usable file in the end.... Yeah I guess I should migrate all this to a saner place one day, but I miss a free host, you could do that on sourceforge once, but they limited so much the shell access that it was crazy to handle it. Oh well... ! I am at 75%, maybe 1 chance left ! Sorry it's totally broken, it never ends and fills the file with crap. Well I guess we can only wait at this point ! edit : here is a link to the very latest dos version (0.64.16) : https://mega.nz/#!3JkxWIoA!DEoZ2SwyylRWYLcpaOVCwZaaNvqBavRC6zSTW-xCikc there is only the exe inside, it needs at least raine.dat from a previous dos installation. And it includes at least the change for vsync discussed in september 2015, that is, there is an explicit call to vsync for dos now (which probably makes things slower, but it also probably works better now that the allegro hacks are mostly lost).