ScummVM IRC Log - 03 Jul 2016

[Back to Index]

[00:00] <m_kiewitz> well, the whole implementation is a nightmare
[00:01] <m_kiewitz> we simply need what DOS would have returned on those key presses
[00:01] <m_kiewitz> but we can't simply use scancodes because in other situations that won't work again and i think scancodes are platform specific as well, at least on SDL1
[00:01] <m_kiewitz> if i remember correctly
[00:01] <m_kiewitz> i talked with LordHoto quite a bit back then about it
[00:03] <m_kiewitz> and I also think for Ctrl- keypresses, we actually do not want 'x'
[00:04] <m_kiewitz> as i said, it should behave like in DOS. I guess the whole ScummVM event API should get changed to work like this, but that's also a nightmare, because you can't easily change that part
[00:04] <snover> just do what every programmer does and add a new field, like realAscii ;)
[00:04] <m_kiewitz> that would be .ascii2
[00:04] <m_kiewitz> :P
[00:05] <snover> then when that one is used and it turns out there is a flaw, realRealAscii
[00:05] <m_kiewitz> the real nightmare begins when you have to make it work for SDL1 + SDL2 and all the other non-SDL platforms
[00:05] <m_kiewitz> .asciiThisTimeItWorksISwear
[00:06] <snover> hey, at least its not web platform, where its one platform that is pathologically broken when it comes to device inputs
[00:06] <snover> theyve been working on DOM keyboard events specification since like 2004 and its still not out of Working Draft
[00:06] <snover> actually, its been another year, let me go check on that.
[00:07] <snover> nope, still a working draft.
[00:07] <snover> 12 years and you still cant reliably get keyboard input on the web :)
[00:08] <m_kiewitz> ah right I added some extra code for some platform, maybe even Mac to make cursor keys work when Num-Lock is not enabled
[00:08] <m_kiewitz> i think in that case SDL or the backend returns something incorrect as well, but only for some platfors
[00:08] <m_kiewitz> platforms
[00:09] <m_kiewitz> this should really get cleaned up, but well there's tons of platforms involved.
[00:09] <m_kiewitz> maybe that was the reason why LordHoto went on holiday :P
[00:10] <m_kiewitz> i guess we could try to detect Alt + key and then try to calculate the .ascii code from the scancode
[00:10] <snover> yes
[00:12] <m_kiewitz> can you please try the other Ctrl-keys
[00:12] <m_kiewitz> just to make sure that those are working properly
[00:13] <m_kiewitz> we should really create some test debug command for quick keyboard checking
[00:13] <m_kiewitz> last time that was a nightmare as well
[00:13] <-- waltervn left irc: Read error: Connection reset by peer
[00:14] <m_kiewitz> and well the documentation says "the ascii value". so it should be 7/8-bit
[00:14] <m_kiewitz> and i guess i should say DOS-ASCII or something like that, because that would make some sense
[00:14] <m_kiewitz> it
[00:15] <m_kiewitz> and that's actually the problem with those direction keys/keypad
[00:15] <m_kiewitz> because for some platforms we get an .ascii value, where we actually shouldn't. we should get 0 in that field
[00:16] <m_kiewitz> changing stuff like that in the backend is of course problematic because you would have to check every single other engine to make sure that none rely on such behavior
[00:18] <snover> m_kiewitz: which ones did you want me to try? they should work ok, since there is no OS-level magic to control key
[00:18] <snover> but i am happy to verify
[00:18] <m_kiewitz> well you can try the few in qfg1
[00:18] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[00:18] <m_kiewitz> qfg2 also has some
[00:18] <m_kiewitz> is there OS magic on other key combinations like Ctrl-Alt?
[00:18] <m_kiewitz> and is there a Num Lock? If so, which keys are affected
[00:19] <m_kiewitz> i think i will create some keyboard test command, that would make it way easier
[00:19] <snover> (1) yes (2) no
[00:19] <m_kiewitz> Ctrl-Alt has no OS magic? only Alt, correct?
[00:20] <snover> any time alt is held, the alternate keyboard layout is active
[00:20] <m_kiewitz> ah i see
[00:20] <m_kiewitz> hmm, we should really have some documentation on all the different platforms
[00:21] <m_kiewitz> what happens when you have shift-lock enabled and then press Alt-X?
[00:21] <m_kiewitz> do you get the same character?

[00:21] <m_kiewitz> it seems my IRC client isn't fully capable of unicode
[00:21] <snover> yes it is still H with the caps lock on
[00:22] <m_kiewitz> i hope no other platform does some other special stuff in such cases
[00:22] <snover> however, alt characters that have capital representations will use those

[00:22] <m_kiewitz> but Alt-X with Shift-Lock will still return H, correct?
[00:22] <snover> yep

[00:23] <m_kiewitz> we could also think about just allowing certain keys + key combinations and filter away everything else
[00:25] <snover> https://zetafleet.com/i/57785b600e18d.jpg its happening!
[00:25] <snover> however, i am pretty sure i should be able to skip this video&
[00:26] <m_kiewitz> would be nice to finally play phantasmagoria, i never played it because the original interpreter worked so badly
[00:27] <m_kiewitz> i will look into those SDL issues tomorrow
[00:27] <snover> ok. SDL! woo!
[00:27] <snover> thanks!
[00:27] <m_kiewitz> need some sleep
[00:27] <m_kiewitz> bye
[00:27] <snover> bye!
[00:27] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[00:28] <snover> (also, I would really say, ascii field shouldnt ever contain unicode values, there should be a separate place for those)
[00:46] --> Cheeseness joined #scummvm.
[00:47] --> jammm joined #scummvm.
[00:50] <-- jamm left irc: Ping timeout: 250 seconds
[00:55] <-- SylvainTV left irc: Read error: Connection reset by peer
[00:55] <snover> wow, hax. to implement the black lines for VMD output they just dumped in a new global state variable :(
[01:04] <-- Rasi left irc: Quit: WeeChat 1.5
[01:05] --> Rasi joined #scummvm.
[01:30] --> DominusExult joined #scummvm.
[01:33] <-- Dominus left irc: Ping timeout: 276 seconds
[01:34] Nick change: DominusExult -> Dominus

[01:46] --> DJW|Home joined #scummvm.
[01:46] #scummvm: mode change '+o DJW|Home' by ChanServ!ChanServ@services.
[01:48] <snover> my brain is dying from boredom adding new game options to all the SCI2.1/SCI3 games&
[01:48] <snover> surely there is a less tedious way to do this
[01:48] <-- DJWillis left irc: Ping timeout: 240 seconds
[02:01] --> Strangerke_ joined #scummvm.
[02:03] <-- Strangerke left irc: Ping timeout: 272 seconds
[02:03] Nick change: Strangerke_ -> Strangerke
[02:12] <snover> its also weird to me that engine options dont show up unless you delete and re-add games to your library
[02:12] <snover> that seems like a thing that shouldnt be a thing
[02:17] <snover> oh and it just shows up magically once you run the game once, huh.
[02:34] <-- NuSuey left irc: Quit: Connection closed for inactivity
[02:57] <-- jammm left irc: Read error: Connection reset by peer
[02:57] --> jamm joined #scummvm.
[03:29] --> GitHub75 joined #scummvm.
[03:29] <GitHub75> [scummvm] csnover pushed 5 new commits to master: https://git.io/vKfkw
[03:29] <GitHub75> scummvm/master 07b72c9 Colin Snover: SCI32: Document ScreenItem::_insetRect
[03:29] <GitHub75> scummvm/master 77991b7 Colin Snover: SCI32: Add const to getCurrentBuffer
[03:29] <GitHub75> scummvm/master a4c059b Colin Snover: SCI32: Improve accuracy of frameout throttler
[03:29] GitHub75 (GitHub75@192.30.252.41) left #scummvm.
[03:49] --> GitHub188 joined #scummvm.
[03:49] <GitHub188> [scummvm] csnover opened pull request #776: SCI32: kPlayVMD (master...sci32-kPlayVMD) https://git.io/vKfkF
[03:49] GitHub188 (GitHub188@192.30.252.45) left #scummvm.
[04:03] --> Polynomial-C joined #scummvm.
[04:03] <-- snover left irc: Quit: Leaving.
[04:03] <-- Poly-C left irc: Ping timeout: 244 seconds
[04:04] <-- Cruel` left irc: Ping timeout: 252 seconds
[05:36] --> ajax16384 joined #scummvm.
[05:36] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[05:56] --> Begasus joined #scummvm.
[05:57] <Begasus> morning
[06:23] <-- jamm left irc: Read error: Connection reset by peer
[06:24] --> jamm joined #scummvm.
[06:26] <-- jamm left irc: Read error: Connection reset by peer
[06:26] --> jamm joined #scummvm.
[06:51] <-- jamm left irc: Read error: Connection reset by peer
[06:52] --> jamm joined #scummvm.
[06:59] Nick change: Begasus -> Begas_afk
[07:01] --> jammm joined #scummvm.
[07:01] <-- jamm left irc: Read error: Connection reset by peer
[07:09] --> uruk-hai joined #scummvm.
[07:09] #scummvm: mode change '+o uruk-hai' by ChanServ!ChanServ@services.
[07:15] --> DominusExult joined #scummvm.
[07:16] <-- jammm left irc: Ping timeout: 244 seconds
[07:17] <-- Dominus left irc: Ping timeout: 264 seconds
[07:17] Nick change: DominusExult -> Dominus
[07:20] --> jamm joined #scummvm.
[07:42] <-- uruk-hai left irc: Quit: Leaving
[08:07] Nick change: _sev_ -> _sev
[08:15] --> Strangerke_ joined #scummvm.
[08:17] <-- Strangerke left irc: Ping timeout: 244 seconds
[08:17] Nick change: Strangerke_ -> Strangerke
[08:32] <-- DJW|Home left irc: Read error: Connection reset by peer
[08:32] --> DJWillis joined #scummvm.
[08:32] #scummvm: mode change '+o DJWillis' by ChanServ!ChanServ@services.
[08:39] --> Henke37 joined #scummvm.
[08:47] --> criezy joined #scummvm.
[08:47] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[08:54] <-- criezy left irc: Quit: criezy
[09:06] --> m_kiewitz joined #scummvm.
[09:06] <-- m_kiewitz left irc: Changing host
[09:06] --> m_kiewitz joined #scummvm.
[09:06] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[09:07] --> NuSuey joined #scummvm.
[09:17] --> frankyboy_ joined #scummvm.
[09:22] --> rootfather joined #scummvm.
[09:22] #scummvm: mode change '+o rootfather' by ChanServ!ChanServ@services.
[09:22] <rootfather> Hello world!
[09:23] <-- Lightkey left irc: Ping timeout: 272 seconds
[09:25] --> t0by joined #scummvm.
[09:25] #scummvm: mode change '+v t0by' by ChanServ!ChanServ@services.
[09:35] --> Lightkey joined #scummvm.
[09:40] <-- t0by left irc: Ping timeout: 264 seconds
[09:43] --> t0by joined #scummvm.
[09:43] #scummvm: mode change '+v t0by' by ChanServ!ChanServ@services.
[09:46] <-- frankyboy_ left irc: Quit: #E>6C O >B 20A
[09:56] --> GitHub118 joined #scummvm.
[09:56] <GitHub118> [scummvm] sev- closed pull request #772: Bugfixes (master...bugfixes) https://git.io/vob5W
[09:56] GitHub118 (GitHub118@192.30.252.42) left #scummvm.
[09:56] --> GitHub62 joined #scummvm.
[09:56] <GitHub62> [scummvm] sev- pushed 5 new commits to master: https://git.io/vKfsn
[09:56] <GitHub62> scummvm/master 6a2059e Ben Castricum: SCUMM: Use exact strings in stringtable, fixes bug #6886...
[09:56] <GitHub62> scummvm/master 315ad62 Ben Castricum: SKY: change "savegame" into "saved game"
[09:56] <GitHub62> scummvm/master 8957bf4 Ben Castricum: SCUMM: reset ShakePos on loading, fixes one part of bug #7141
[09:56] GitHub62 (GitHub62@192.30.252.41) left #scummvm.
[10:00] <-- t0by left irc: Ping timeout: 260 seconds
[10:02] --> t0by joined #scummvm.
[10:02] #scummvm: mode change '+v t0by' by ChanServ!ChanServ@services.
[10:37] --> VonGuardian joined #scummvm.
[10:37] --> rootfather_ joined #scummvm.
[10:41] <-- rootfather left irc: Ping timeout: 272 seconds
[10:49] Nick change: rootfather_ -> rootfather
[10:49] <-- rootfather left irc: Changing host
[10:49] --> rootfather joined #scummvm.
[10:49] #scummvm: mode change '+o rootfather' by ChanServ!ChanServ@services.
[11:00] --> GitHub157 joined #scummvm.
[11:00] <GitHub157> [scummvm] wjp closed pull request #741: SCI: Clean up some aspects of call handling (master...sci-call) https://git.io/vVlvN
[11:00] GitHub157 (GitHub157@192.30.252.42) left #scummvm.
[11:00] --> GitHub3 joined #scummvm.
[11:00] <GitHub3> [scummvm] wjp pushed 6 new commits to master: https://git.io/vKfZB
[11:00] <GitHub3> scummvm/master 9f2fff4 Willem Jan Palenstijn: SCI: Remove unneeded copy
[11:00] <GitHub3> scummvm/master 08f1727 Willem Jan Palenstijn: SCI: Improve kernel subfunction logging...
[11:00] <GitHub3> scummvm/master 7f12638 Willem Jan Palenstijn: SCI: Remove unclear &rest handling...
[11:00] GitHub3 (GitHub3@192.30.252.42) left #scummvm.
[11:07] --> WooShell joined #scummvm.
[11:29] <WooShell> meow =^.^=
[11:36] --> criezy joined #scummvm.
[11:36] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[11:40] <eriktorbjorn> Anyone around who's familiar with the Prince engine?
[11:47] <eriktorbjorn> Or at the very least owns the game.
[11:51] <WooShell> PoP?
[11:51] <eriktorbjorn> WooShell: No, this one: http://wiki.scummvm.org/index.php/The_Prince_and_the_Coward
[11:52] <WooShell> ah. no, sorry.
[11:52] <eriktorbjorn> Mind you, Prince of Persia is a nice game. At least, the first one. It's the only one I've completed. The second one looked nice too, for as long as I played it.
[11:55] <eriktorbjorn> WooShell: I bought a copy of the game recently. This time I have an excuse for not playing it, though: It's in German. :-)
[11:56] <eriktorbjorn> Judging by the intro, the author was a firm believer in "tell, not show": https://www.youtube.com/watch?v=54U1LC9Sb4E :-)
[11:58] --> SylvainTV joined #scummvm.
[11:58] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[12:01] <-- criezy left irc: Quit: criezy
[12:02] --> iskrich joined #scummvm.
[12:02] #scummvm: mode change '+v iskrich' by ChanServ!ChanServ@services.
[12:11] --> GitHub125 joined #scummvm.
[12:11] <GitHub125> [scummvm-web] sev- pushed 3 new commits to master: https://git.io/vKfcR
[12:11] <GitHub125> scummvm-web/master 6b12b9d Eugene Sandulenko: WEB: Added Myst screenshots. Patch #1647
[12:11] <GitHub125> scummvm-web/master b733c5d Eugene Sandulenko: WEB: Added Bambou screenshots. Patch #1649
[12:11] <GitHub125> scummvm-web/master d048cbd Eugene Sandulenko: WEB: Added GNAP screenshots. Patch #1651
[12:11] GitHub125 (GitHub125@192.30.252.42) left #scummvm.
[12:13] <-- Cheeseness left irc: Ping timeout: 272 seconds
[12:15] <_sev> yay, one of those rare moments when there are no pending screenshots
[12:19] <-- Strangerke left irc: Ping timeout: 272 seconds
[12:21] --> plipfishy joined #scummvm.
[12:23] <m_kiewitz> eriktorbjorn: I own plenty of Prince of Persia games
[12:23] <m_kiewitz> no idea about the engine though :P
[12:23] <m_kiewitz> PoP 2 is a really great game with quite a few interesting puzzles
[12:23] <m_kiewitz> i own that too
[12:26] <m_kiewitz> ah wait, you were not talking about Prince Of Persia :(
[12:26] <m_kiewitz> WooShell: I had the same thought lol
[12:36] --> Cheeseness joined #scummvm.
[12:45] --> criezy joined #scummvm.
[12:45] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[12:57] <-- criezy left irc: Quit: criezy
[13:21] --> snover joined #scummvm.
[13:21] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[13:32] <-- ajax16384 left irc: Read error: Connection reset by peer
[13:34] <m_kiewitz> snover: hi
[13:34] <snover> m_kiewitz: hello!
[13:35] <m_kiewitz> i remembered that we can't really use keycode, well at least for SDL1, because keycode is the raw scancode, which means when layouts / mappings are active, then those are not applied
[13:35] <m_kiewitz> so for example on my keyboard Alt-Y will result in keycode 'z'
[13:37] <snover> darn. i thought about that, but i was hoping it wasnt the case.
[13:37] --> waltervn joined #scummvm.
[13:37] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[13:37] <m_kiewitz> you got mappings / layouts active?
[13:37] <m_kiewitz> because SDL2 should actually apply those
[13:37] <waltervn> hi
[13:38] <m_kiewitz> im currently trying to switch my build to SDL2, but i have some issues
[13:38] <m_kiewitz> hi walter
[13:38] <snover> just standard default US-ASCII keyboard layout here
[13:38] <m_kiewitz> dang
[13:38] <snover> hey waltervn :) put those blacklines in just for you ;)
[13:38] <m_kiewitz> i created my project files using --sdl2, but im getting linker errors. it seems it somehow still tries to link CD-audio SDL1 functions
[13:39] <snover> m_kiewitz: how odd!
[13:39] <waltervn> snover: \o/
[13:41] <snover> however, possibly less odd than 'configure' not having the sdl2 option here for some reason&
[13:44] <m_kiewitz> im getting these errors here
[13:44] <m_kiewitz> 1>surfacesdl-graphics.obj : error LNK2019: unresolved external symbol _SDL_SetVideoMode referenced in function "protected: virtual bool __thiscall SurfaceSdlGraphicsManager::loadGFXMode(void)" (?loadGFXMode@SurfaceSdlGraphicsManager@@MAE_NXZ)
[13:44] <m_kiewitz> 1>surfacesdl-graphics.obj : error LNK2019: unresolved external symbol _SDL_GetVideoInfo referenced in function "protected: virtual bool __thiscall SurfaceSdlGraphicsManager::loadGFXMode(void)" (?loadGFXMode@SurfaceSdlGraphicsManager@@MAE_NXZ)
[13:46] <wjp> snover: SDL_CONFIG=sdl2-config should do it
[13:46] <snover> thank you!
[13:48] --> ajax16384 joined #scummvm.
[13:48] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[13:52] <snover> hm, no linker errors for me
[13:52] <snover> m_kiewitz: what is your build environment? msys2+mingw-w64?
[13:53] <m_kiewitz> visual c++
[13:53] <m_kiewitz> i think i may have solved it now
[13:53] <snover> oh, nice.
[13:59] <m_kiewitz> finally got it now
[13:59] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[14:02] <m_kiewitz> just checked, SDL2 really applies layout/mapping to keycode
[14:03] <m_kiewitz> im getting "y" as keycode, when I press alt-y
[14:06] --> DominusExult joined #scummvm.
[14:08] <m_kiewitz> this mess is also the reason why we use .ascii when possible
[14:08] <-- Dominus left irc: Ping timeout: 260 seconds
[14:08] Nick change: DominusExult -> Dominus
[14:10] <bgK> Isn't .ascii meant to be used only for text input, and .keycode used for the rest of the cases such as shortcuts?
[14:11] <snover> this sort of input problem is pretty universal& with regards to SDL2, theres a text api when you want to get the text that someone types in, and a key api for when you want to know which key of the layout is pressed, but what seems to be desired here is a what would be the text value of this key if modifiers were not applied to it?
[14:13] <wjp> that's SDL2's keycode
[14:14] <wjp> kind of
[14:14] <m_kiewitz> bgK: yes, keycode is used for special keys. But for Alt-key, Ctrl-key, etc. we have to use .ascii too
[14:14] <m_kiewitz> special keys for example are keypad up down left right, function keys and so on
[14:15] <m_kiewitz> that works, but the Alt-key, Ctrl-key etc. is a nightmare because it doesn't work consistently
[14:20] <bgK> ok, having consistent keycodes is one more reason to switch to SDL2, I guess
[14:22] <m_kiewitz> well yes, but is SDL2 supported on all ScummVM-supported platforms?
[14:22] <m_kiewitz> i mean are there cases where only SDL1 is supported?
[14:22] <m_kiewitz> and i mean if we would change the code to use keycodes for those cases, then all engines that use those keys would have to be changed
[14:23] <m_kiewitz> and sadly there is no engine-specific documentation on keyboard usage
[14:26] <-- ajax16384 left irc: Read error: Connection reset by peer
[14:27] <wjp> that's because this is functionality defined by the backends/osystem layer, and used by the engines
[14:29] <wjp> changing it will be an interesting endeavour
[14:46] <m_kiewitz> snover: which SDLKey is passed to SDLToOSystemKeycode in your case, when you press - what was it Alt-X?
[14:47] <snover> m_kiewitz: one moment.
[14:48] <snover> m_kiewitz: 120
[14:48] <snover> (SDLK_x)
[14:48] <m_kiewitz> and I guess "unicode" is set to that special character
[14:49] <snover> yes
[14:49] <m_kiewitz> omg :/
[14:49] <m_kiewitz> i wonder why unicode is even returned in mapKey()
[14:50] <snover> it would seem advisable to not do that and instead to provide an opt-in API for receiving text
[14:50] <m_kiewitz> wjp: any idea on that? does any engine actually expect unicode in .ascii?
[14:53] <m_kiewitz> snover: not do what? not return unicode in .ascii?
[14:53] <snover> m_kiewitz: yeah.
[14:53] <m_kiewitz> it was merged by Max some ages ago.
[14:53] <m_kiewitz> one would have to research when that code was added and why
[14:53] <m_kiewitz> to me it makes no sense
[14:54] <m_kiewitz> especially when there is something valid in keycode
[15:00] <wjp> Because just as often there isn't
[15:01] <wjp> Our current interface is definitely imperfect, but it's largely based on what SDL1 gives us
[15:02] <m_kiewitz> the question is: why does mapKey return unicode in case there is something in unicode
[15:02] <m_kiewitz> was it some special SDL1 case on some platform?
[15:02] <m_kiewitz> if I removed that, I think it would work out for snover
[15:02] <wjp> Keyboard layouts I would guess
[15:02] <snover> yep.

[15:03] <m_kiewitz> ah, yes that makes sense
[15:03] <m_kiewitz> but then we could check if keycode is something valid and return that
[15:03] <wjp> And even for qwertz/azerty I guess
[15:03] <m_kiewitz> and only use unicode in case there was an invalid keycode
[15:03] --> Cruel` joined #scummvm.
[15:04] <wjp> How would you determine that?
[15:04] <m_kiewitz> SDLToOSystemKeycode is called right before
[15:04] <m_kiewitz> and returns KEYCODE_INVALID on unknown cases
[15:05] <m_kiewitz> need to check, what SDL1 returns in those special cases

[15:09] <m_kiewitz> "y" gives me 121, and KEYCODE_y, so that one would work out on SDL2
[15:09] <m_kiewitz> checking SDL1 now
[15:13] --> wanwan joined #scummvm.
[15:20] <m_kiewitz> right, I'm getting KEYCODE_z when I press "y" using SDL1, unicode is correct - 121

[15:26] <m_kiewitz> wjp: are there engines, that actually use .ascii > 255?
[15:26] <m_kiewitz> we could otherwise allow all text characters < 256 and then use .unicode, otherwise use keycode
[15:33] <-- gapan left irc: Quit: leaving
[15:38] <waltervn> scumm seems too, based on a comment in common/keyboard.h
[15:38] <-- t0by left irc: Quit: Bye!
[15:38] <waltervn> if there's scumm-specific stuff in keyboard.h that should probably be refactored
[15:40] <waltervn> actually, I just noticed there's a todo right below that comment ;)
[15:55] --> ajax16384 joined #scummvm.
[15:55] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[16:00] --> Littleboy joined #scummvm.
[16:00] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[17:08] <snover> i guess its movie week, i am going to work on kShowMovie now I think
[17:08] --> GitHub59 joined #scummvm.
[17:08] <GitHub59> [scummvm] m-kiewitz pushed 1 new commit to master: https://git.io/vKfzr
[17:08] <GitHub59> scummvm/master 9aa7174 Martin Kiewitz: SDL: Fix keyboard on macOS, when using SDL2, fix directional keypad...
[17:08] GitHub59 (GitHub59@192.30.252.34) left #scummvm.
[17:08] <m_kiewitz> fingers crossed that this doesn't break thousand of things :P
[17:18] <snover> well the good news is that builds arent pushed live to the world to use every moment :)
[17:20] <wjp> Woah, you just committed that?
[17:23] <m_kiewitz> why not?
[17:23] <wjp> This is the kind of drastic change that I would've prefered go into a PR
[17:23] <m_kiewitz> well i can revert it
[17:23] <m_kiewitz> although i mean real testing won't happen unless it's commited
[17:23] <m_kiewitz> there are so many platforms involved
[17:25] <m_kiewitz> and i guess there should be some keyboard test debug command to make testing easier
[17:28] <m_kiewitz> wjp: shall i revert it?
[17:29] <snover> i like having working sci32 debug hotkeys :P
[17:32] --> Henke37 joined #scummvm.
[17:34] <wjp> Yes please. Let's give people at least a bit of time to consider the consequences
[17:35] <wjp> When I'm behind a pc again I'll see if I can track down unicode ascii use
[17:35] <wjp> I vaguely recall discussions about it a few years back
[17:40] <m_kiewitz> revert commits - I still call it "SDL:" but then revert something something? correct?
[17:41] --> ny00123 joined #scummvm.
[17:44] --> GitHub48 joined #scummvm.
[17:44] <GitHub48> [scummvm] m-kiewitz pushed 1 new commit to master: https://git.io/vKf2w
[17:44] <GitHub48> scummvm/master 29c31f8 Martin Kiewitz: Revert "SDL: Fix keyboard on macOS, fix directional keypad"...
[17:44] GitHub48 (GitHub48@192.30.252.45) left #scummvm.
[18:14] <wjp> Thanks
[18:16] <snover> nooooo (etc.) :P
[18:25] --> Strangerke joined #scummvm.
[18:35] <snover> change of plans& looking at robots& phant1 startup is just too annoying.
[18:37] --> GitHub73 joined #scummvm.
[18:37] <GitHub73> [scummvm] m-kiewitz opened pull request #777: SDL: Fix keyboard on macOS, when using SDL2, fix directional keypad (master...master) https://git.io/vKfVD
[18:37] GitHub73 (GitHub73@192.30.252.40) left #scummvm.
[18:37] <-- WooShell left irc: Read error: Connection reset by peer
[18:37] --> WooShell joined #scummvm.
[18:38] <-- klusark left irc: Ping timeout: 250 seconds
[18:39] --> klusark joined #scummvm.
[18:43] <-- iskrich left irc: Quit: Page closed
[19:22] <-- ajax16384 left irc: Read error: Connection reset by peer
[19:24] <-- NuSuey left irc: Quit: Connection closed for inactivity
[19:26] <-- ny00123 left irc: Quit: Leaving
[19:30] --> frankyboy_ joined #scummvm.
[19:33] --> Raziel^ joined #scummvm.
[19:33] #scummvm: mode change '+v Raziel^' by ChanServ!ChanServ@services.
[19:38] <-- ced117 left irc: Quit: Quitte
[19:40] <-- frankyboy_ left irc: Remote host closed the connection
[19:42] --> frankyboy_ joined #scummvm.
[19:49] <bgK> what about fixing the keycodes using platform specific code so that they match what SDL2 does?
[19:49] <bgK> say for Windows something along that idea: https://gist.github.com/bgK/f9e6722d1d322180c1f2c0cae0583217
[19:51] <m_kiewitz> bgK: we actually fix .ascii, not the keycode
[19:53] <bgK> yes, but that's wrong IMO, .ascii should be the actual ascii value for printable characters, not something tweaked
[19:54] <bgK> (nor the full unicode value, I agree that's wrong)
[19:54] <-- SylvainTV left irc: Read error: Connection reset by peer
[19:54] --> SylvainTV joined #scummvm.
[19:54] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[19:55] <m_kiewitz> bpK: so .ascii should be 0 for alt-x?
[19:55] <m_kiewitz> such a change would mean changing every engine that uses such keypresses
[19:56] <m_kiewitz> and i actually think it's not that weird that alt-x -> .ascii = 'x' and modifier = KMOD_ALT
[19:56] <m_kiewitz> that makes sense to me
[19:58] <bgK> I'd say for alt-x, the appropriate values would be .keycode = 'x', modifier = KMOD_ALT, .ascii = [whatever the keyboard layout defines for that key combination]
[19:58] <m_kiewitz> and well, we can't fix up keycodes that way, because at least on macOS we get higher unicode values, which means we have to use keycodes instead of unicode/ascii in that case
[19:58] <m_kiewitz> keyboard layout?
[19:59] <bgK> the OS keyboard layout
[19:59] <m_kiewitz> well, for example pressing alt-Y returns KEYCODE_z for me and unicode = 'y'
[19:59] <m_kiewitz> so in such case .ascii = 'y' makes sense, keycode is effectively wrong in SDL1, because it doesn't care about keyboard layout
[20:00] <m_kiewitz> in SDL2 keycode is correct
[20:00] <m_kiewitz> and well, the engine should react as if Alt-Y got pressed and not Alt-Z
[20:00] <m_kiewitz> i mean what SDL1 is doing doesn't make sense at all for anyone except maybe kernel keyboard driver developers :P
[20:01] <bgK> indeed what I propose is to alter the keycodes returned by SDL1 so that they match what is returned by SDL2. Possibly as a per engine option.
[20:02] <m_kiewitz> but that won't work at least on macOS, because as i said - on macOS even SDL2 returns some high unicode char and only keycode is correct in that case
[20:02] <m_kiewitz> so if we fixed the keycode, it would probably become keycode_invalid
[20:02] <bgK> yes, the engines would have to be changed to use the keycodes where appropriate
[20:03] <m_kiewitz> no idea what SDL1 would do, snover would have to try that out
[20:03] <m_kiewitz> but that won't solve the underlying issue
[20:03] <bgK> that is to say the .ascii value would be used only for text input, and the keycodes for the other cases
[20:03] <m_kiewitz> but what's alt-x then?
[20:04] <m_kiewitz> is it text input? or another case?
[20:04] <m_kiewitz> and if it's another case, shouldn't keycode + ascii match up then?
[20:04] <bgK> alt-x is .keycode = 'x' and modifier = ALT
[20:04] <m_kiewitz> and what's .ascii then?
[20:04] <m_kiewitz> and so an engine should use keycode in case specific modifiers are enabled?
[20:04] <bgK> .ascii would be used for users to type save game names, for example
[20:05] --> NuSuey joined #scummvm.
[20:05] <m_kiewitz> i mean what's .ascii supposed to contain?
[20:05] <m_kiewitz> because with that new code, .ascii would indeed contain 'x' as well
[20:06] <bgK> for shift-a, .ascii would contain 'A', .keycode 'a', and .modifier SHIFT
[20:06] <m_kiewitz> but what's supposed to happen in case of alt-y, when non-english keyboard layout is present?
[20:06] <m_kiewitz> i think that's actually also already the case
[20:06] <-- DJWillis left irc: Read error: Connection reset by peer
[20:07] <m_kiewitz> and another question is how should engines behave then?
[20:07] --> DJWillis joined #scummvm.
[20:07] #scummvm: mode change '+o DJWillis' by ChanServ!ChanServ@services.
[20:07] <m_kiewitz> and wouldn't it make more sense in case .keycode was = 'x', modifiers = ALT and .ascii = 0 to force everyone to use keycodes?
[20:07] <bgK> I'd say for alt-x, if the keyboard layout does not define a mapping, then .ascii = 0
[20:07] <m_kiewitz> still then i guess removing .keycode for plain text would also make sense, but such changes would require all engines to get adjusted
[20:07] <m_kiewitz> wait, i think we are talking about different things
[20:08] <m_kiewitz> keyboard layout is QWERTY on the actual keyboard. But some countries have QWERTZ keyboards
[20:08] <m_kiewitz> and in those countries keycode is Y, when you press the button Z
[20:08] <wjp> with SDL1
[20:09] <m_kiewitz> yes
[20:09] <m_kiewitz> with SDL2 that BS got fixed
[20:09] <wjp> _please_ add qualifications to your statements
[20:09] <m_kiewitz> but that's one major issue
[20:09] <wjp> it's impossible to follow the conversation currently
[20:09] <m_kiewitz> i assumed this was already known
[20:09] <wjp> ...
[20:10] <m_kiewitz> well it's documented now even in my push request
[20:10] <m_kiewitz> should have been added as a comment ages ago
[20:11] <m_kiewitz> and i guess plenty of people using DOS + QWERTZ keyboards knew this already, at least that the OS internally gets the "wrong" key and has to map it
[20:12] <m_kiewitz> i don't even understand where SDL1 gets it. I would have assumed that Windows maps those keys already before passing them around to applications
[20:12] --> GitHub49 joined #scummvm.
[20:12] <GitHub49> [scummvm] raziel- opened pull request #778: BASE: Fix Typo (master...patch-1) https://git.io/vKfK4
[20:12] GitHub49 (GitHub49@192.30.252.40) left #scummvm.
[20:13] --> GitHub198 joined #scummvm.
[20:13] <GitHub198> [scummvm] wjp closed pull request #778: BASE: Fix Typo (master...patch-1) https://git.io/vKfK4
[20:13] GitHub198 (GitHub198@192.30.252.40) left #scummvm.
[20:13] --> GitHub8 joined #scummvm.
[20:13] <GitHub8> [scummvm] wjp pushed 2 new commits to master: https://git.io/vKfKE
[20:13] <GitHub8> scummvm/master 866f8ed Hubert Maier: BASE: Fix Typo
[20:13] <GitHub8> scummvm/master ea94196 Willem Jan Palenstijn: Merge pull request #778 from raziel-/patch-1...
[20:13] GitHub8 (GitHub8@192.30.252.46) left #scummvm.
[20:16] <wjp> it would be pretty nice if we keycodes could be fixed up somehow in the SDL1 backend
[20:16] <wjp> s/we //
[20:17] <m_kiewitz> i doubt this is properly possible
[20:17] <m_kiewitz> as i said - macOS on alt-X returns unicode U+2248, but keycode_x
[20:17] <m_kiewitz> if we added code that tries to fix keycodes up, then how would that work in this case?
[20:18] <wjp> it clearly would need OS help
[20:18] <m_kiewitz> we could of course assume that U+2248 doesn't make sense, so keep keycode_x
[20:18] <wjp> but don't ask me how
[20:18] <m_kiewitz> sure, that would be possibly the only "proper" solution, but then we would have to add support for every single OS
[20:18] <bgK> m_kiewitz: SDL also sends the os specific scancode. We can probably then use that scancode to so an OS specific call to get the correct keycode, just as I did for Windows in that gist I posted earlier.
[20:19] <bgK> *do
[20:20] <m_kiewitz> that's quite a lot of work, because we would have to go through every single SDL supported (+ ScummVM supported) OS, and add support for each of them
[20:20] <m_kiewitz> and then we would have to go through all ScummVM engines on top and change all the code for alt- and ctrl- keypresses
[20:21] --> GitHub137 joined #scummvm.
[20:21] <GitHub137> [scummvm] eriktorbjorn pushed 2 new commits to master: https://git.io/vKfKN

[20:21] GitHub137 (GitHub137@192.30.252.40) left #scummvm.
[20:21] <m_kiewitz> and i really don't understand what the benefit would be. Atm we can just check for alt- modifier and .ascii = whatever and then take that as alt-x, etc.
[20:21] <m_kiewitz> i mean SDL1 is also outdated on most platforms on top
[20:22] <m_kiewitz> and for SDL2 the keycodes work properly, at least as far as i checked
[20:22] <bgK> that's indeed a lot of work, but it does not have to be done all in one go, and it's kind of a prerequisite to migrate to SDL2
[20:22] <m_kiewitz> SDL2 works quite well, at least as far as i have checked. the keyboard code could use some clean up of course
[20:23] <m_kiewitz> at least for me and snover all sorts of Ctrl- and Alt- keys work when using SDL2
[20:23] <bgK> yes SDL2 works well, but the engines still need to be fixed to use the keycodes
[20:23] <m_kiewitz> but why? why not use .ascii for those cases?
[20:24] <m_kiewitz> .ascii works for SDL2, at least as far as i have looked
[20:25] <wjp> how would you for example treat pressing ctrl + (key with 1 on it) on an azerty keyboard?
[20:25] <wjp> and then ctrl + shift + (key with 1 on it)?
[20:26] <m_kiewitz> .ascii = '1', modifier = KMOD_CTRL
[20:26] <wjp> nope
[20:26] <-- WooShell left irc: Read error: Connection reset by peer
[20:26] <wjp> the ascii for pressing that key would usually be &
[20:26] <wjp> 1 is shift-&
[20:26] --> WooShell joined #scummvm.
[20:27] <m_kiewitz> ???
[20:27] <wjp> azerty
[20:27] <m_kiewitz> key with "1" on it?, that's 1?
[20:28] <wjp> no
[20:28] <wjp> that's &
[20:28] <m_kiewitz> ???
[20:28] <wjp> azerty
[20:28] <Raziel^> & is "6" here
[20:28] <m_kiewitz> when I'm a user, and I press a button with "1" on it, then i expect it to show me a "1"
[20:28] <m_kiewitz> and when i press Ctrl-1 right now, I'm getting literally nothing in this IRC client
[20:29] <m_kiewitz> shift - 1 would be "!" for me
[20:29] <m_kiewitz> and im pretty sure im getting .ascii = "!" in ScummVM
[20:29] <wjp> Raziel^: sorry, did I mix up the specific grouping of keys? Or are there multiple azerty variants with different rows there?
[20:30] <wjp> m_kiewitz: the azerty layout has numbers hidden behind shift
[20:30] <snover> a problem is that in the case of macOS, the behaviour of shift+1 and the behaviour of alt+1 are the same: they choose a variant character
[20:30] <m_kiewitz> hmm, i just googled that. i see

[20:30] <m_kiewitz> still who uses those? and then right i would expect it to return "&" in that case
[20:31] <m_kiewitz> isn't that the case atm in ScummVM?
[20:31] <wjp> m_kiewitz: all of France and Belgium? :-)
[20:31] <m_kiewitz> it should return "&" in .ascii
[20:31] <m_kiewitz> those people use pretty weird keyboards :P
[20:31] <wjp> oh, I fully agree there :-)
[20:31] <m_kiewitz> i don't even understand it - so you have to hold 1 additional key to type a number?
[20:31] <-- WooShell left irc: Read error: Connection reset by peer
[20:31] <Raziel^> m_kiewitz: pretty weird language too
[20:31] <Raziel^> :-D
[20:31] <m_kiewitz> it makes no logical sense
[20:31] <m_kiewitz> lol
[20:31] --> WooShell joined #scummvm.
[20:32] <m_kiewitz> i think anyone doing number work on such keyboards will probably go crazy :P
[20:32] <bgK> that way we -French people- can get our lovely accentuated characters in a single keypress
[20:32] <m_kiewitz> but anyway, yes in that case .ascii should be "&" and I hope that's already the case
[20:32] <Raziel^> maybe they'll have a keypad which has a Shift-Lock permamnently enabled?
[20:33] <Raziel^> nah, that woul dbe madness
[20:33] <m_kiewitz> bpK: i wondered about that, but how does it help that "&" is the base key?

[20:34] <m_kiewitz> only 4 of those number keys are used for accentuated characters, right?
[20:34] <m_kiewitz> see, that's what i don't understand
[20:35] <bgK> well, there's always the keypad for entering numbers
[20:35] <m_kiewitz> if all of them were accentuated, i would understand the logic behind it
[20:36] <m_kiewitz> but what does happen atm in ScummVM for those keys?
[20:37] <bgK> they work mostly ok for text input, .ascii contains the expected value
[20:37] <m_kiewitz> shift and not shift is simple - those should be "&" (no shift) and "1" (shift)
[20:37] <m_kiewitz> what happens for Ctrl-1?
[20:37] <bgK> let me check
[20:37] <m_kiewitz> although i wonder if there is even an engine that actually uses that combination
[20:38] <m_kiewitz> and well, what did happen in DOS for those cases? that's the main question - if DOS returned those as "1" + Ctrl, then that would be "correct", because that's what games would expect
[20:39] <m_kiewitz> and i guess that should have been the case, of course in ScummVM that's probably not the case. but i don't know how to fix such an issue without even adding keyboard layout stuff inside ScummVM
[20:40] <m_kiewitz> i mean we would have to use the scancode then, but we can't rely on the scancode because of completely different keyboard layouts
[20:40] <m_kiewitz> thanks to SDL1. I even wonder what SDL2 does in that case. does it return scancode '1' and modifier ctrl?
[20:41] <wjp> you mean keycode in the case of SDL2
[20:41] <m_kiewitz> if only keyboards were differently designed
[20:41] <m_kiewitz> yes, keycode
[20:42] <wjp> should be SDLK_1, yes
[20:42] <wjp> SCUMM uses both Ctrl-number and Alt-number key combinations
[20:43] <m_kiewitz> and what does SCUMM use atm? the keycodes already?
[20:43] <wjp> (by checking event.kbd.keycode)
[20:43] <m_kiewitz> still that's another one that would get broken by keycode guessing
[20:44] <m_kiewitz> so the only proper way to do this is to add all sorts of OS specific code... to SDL code... which should be platform independent
[20:44] <m_kiewitz> fun
[20:44] <m_kiewitz> thanks to whoever created the SDL keyboard code :/
[20:44] <m_kiewitz> we could also just fix SDL1 then :P
[20:45] <snover> if it such a problem then maybe accelerate SDL2 out of experimental and make it the primary version :)
[20:49] <m_kiewitz> is there a list with all of our platforms that use SDL?
[20:49] <bgK> mhh, actually, with my azerty keyboard on Linux with both SDL 1 and SDL 2, pressing ctrl-1 results in .keycode = '&'
[20:50] <m_kiewitz> urgh
[20:50] <bgK> thankfully a lot of our SDL using platforms don't have a hardware keyboard
[20:50] <m_kiewitz> well there could be USB support?
[20:50] <bgK> in some cases yes, probably
[20:51] <m_kiewitz> can you please check what SDL1/2 return themselves?
[20:51] <m_kiewitz> maybe we ruin something ourselves
[20:51] <m_kiewitz> but well in case SDL1/2 themselves return "&" then we can basically only throw out SDL support
[20:52] <bgK> '&' is what SDL 1 and 2 return before we mess with it
[20:52] <m_kiewitz> which wouldn't be such a bad idea anyway
[20:52] <m_kiewitz> urgh
[20:52] <m_kiewitz> and they don't return "1" anywhere?
[20:53] <m_kiewitz> i mean they return quite a few variables
[20:53] <m_kiewitz> how is this even possible? i mean they had 1 job
[20:53] <m_kiewitz> bgK: and you are on? Linux? Windows? macOS?
[20:54] <m_kiewitz> ah Morocco uses a similar layout
[20:54] <m_kiewitz> "Tifinagh layout"
[20:55] <bgK> m_kiewitz: those tests results are with Linux
[20:55] <m_kiewitz> well we could of course also accept "&" + Ctrl modifier as "1" and map that into "1" + Ctrl
[20:56] <bgK> I'm currently testing with Windows. so far SDL 1 returns '1' as a keycode when I press ctrl+&, which is inconsistent with Linux
[20:56] <m_kiewitz> bgK: btw. what does DOS-Box do? have you tried it?
[20:56] <m_kiewitz> lol
[20:56] <m_kiewitz> yeah, that's SDL1. i love it.
[20:57] <m_kiewitz> in case SDL2 is inconsistent too, then we really should maybe throw it out
[20:57] <m_kiewitz> and what does it do on Ctrl-Shift-"1"?
[20:57] <Raziel^> does anyone know where users can download the fangames (AGI, SCI)?
[20:58] <m_kiewitz> Raziel^: http://sciprogramming.com/fangames.php
[21:00] <Raziel^> m_kiewitz: thank you, i thought they may be on the scummvm site somewhere
[21:00] <m_kiewitz> it would be an idea to offer download links too
[21:00] <m_kiewitz> the SCI demos are available already
[21:01] <m_kiewitz> it would be a shame if those got lost at some point
[21:01] <Raziel^> that was my thinking exactly
[21:01] <m_kiewitz> i downloaded some larger archive with tons of games
[21:01] <m_kiewitz> no idea where i got it from
[21:01] <Raziel^> iirc the whole pack of fangames were available at one point
[21:03] <m_kiewitz> ah we got a "agi-fanmade.zip" on sourceforge
[21:03] <m_kiewitz> https://sourceforge.net/projects/scummvm/files/demos/agi/
[21:03] <m_kiewitz> under demos
[21:03] <bgK> m_kiewitz: on Linux ctrl+shift+& yelds .keycode = '&' on both SDL1 and 2
[21:03] <m_kiewitz> and Windows?
[21:03] <Raziel^> m_kiewitz: maybe we could linkt o them directly from the homepage?
[21:04] <m_kiewitz> maybe we could keep SDL for platforms without keyboard
[21:04] <bgK> m_kiewitz: on Windows ctrl+& and ctrl+shift+& both return '1' with SDL2
[21:04] <m_kiewitz> Raziel^: i guess maybe we should create another directory on sourceforge and put the games in there. Them being in demos makes no sense
[21:04] <bgK> so SDL2 is inconsistent between Linux and Windows in that case :s
[21:04] <Raziel^> true
[21:05] <m_kiewitz> bgK: I wonder what SDL on macOS does :P
[21:05] <Raziel^> we could also create a zip of all the sci fangames and put them there aswell?
[21:05] <bgK> m_kiewitz: you'll have to find another volunteer for that
[21:06] <m_kiewitz> well, maybe we should offer them separately. all SCI fanmade games in 1 archive would get quite big.
[21:06] <Raziel^> m_kiewitz: i was thinking about updating the SCI/AGI fangames wiki page and link to them (if available)
[21:06] <m_kiewitz> bgK: thanks for testing. And it does return keycode_A etc. correctly? or are those mixed up too?
[21:07] <m_kiewitz> i mean no idea how those keyboards are wired, but if they are wired like regular QWERTY keyboards, then all keycodes should basically be mixed up for azerty
[21:07] <m_kiewitz> talking about SDL1 of course
[21:08] <m_kiewitz> Raziel^: maybe _sev can upload those files, i don't have access to sourceforge files
[21:08] <Raziel^> m_kiewitz: k
[21:09] <m_kiewitz> wouldn't it also be possible to keep SDL, but do keyboard processing entirely ourselves?
[21:10] <bgK> m_kiewitz: for SDL1, pressing a returns keycode = 'z' on Windows and keycode = 'a' on Linux
[21:10] <m_kiewitz> ...
[21:10] <bgK> with SDL2, keycode = 'a' in both cases
[21:10] <m_kiewitz> im scared of the result of SDL2
[21:10] <m_kiewitz> oh thank god
[21:11] <wjp> ('z'?)
[21:12] <m_kiewitz> bgK: so you were never able to play parser games on Windows until SDL2 support was added? Or is Linux giving the proper results?
[21:12] <bgK> wjp: 'q', you're right, I failed to convert the numerical values
[21:12] <m_kiewitz> you pressed "a"?
[21:12] <-- jamm left irc: Ping timeout: 240 seconds
[21:12] --> LyTchet joined #scummvm.
[21:13] <m_kiewitz> so ScummVM parser games were basically impossible to play on Windows when using azerty keyboards?
[21:13] <bgK> m_kiewitz: the parser games properly use .ascii for text input, don't they?
[21:13] <m_kiewitz> ah
[21:13] <bgK> that works fine
[21:13] <m_kiewitz> yes, right
[21:14] <bgK> yes, I pressed "a"
[21:14] <m_kiewitz> as i said - can we do keyboard processing ourselves, but still keep SDL at least for now?
[21:15] <m_kiewitz> although yes, SCUMM doesn't work right, correct? you can't use Ctrl-1, yes?
[21:15] <bgK> what is Ctrl-1 supposed to do?
[21:16] <wjp> load game from slot 1
[21:16] <wjp> and alt-1 should save to slot 1
[21:16] <m_kiewitz> we could of course hack support for that into the current SDL code and simply accept "&" + Ctrl as Ctrl-1 and so on
[21:18] <bgK> I confirm alt-1 and ctrl-1 don't work for me on Linux with The Dig and SDL1
[21:19] <m_kiewitz> :(
[21:21] <m_kiewitz> bgK: are you able to run DOS? or at least DOS-Box?
[21:21] <m_kiewitz> i wonder if those games worked properly under actual DOS or if it never worked
[21:21] <bgK> dosbox, yes
[21:22] <m_kiewitz> and well i also wonder what happens in dosbox
[21:22] <m_kiewitz> i think even if we did keyboard support ourselves, how should we detect such a situation in case the OS translates the key press that way
[21:23] <m_kiewitz> i mean we can't use the actual scancode, because that would ruin the other keys
[21:23] <m_kiewitz> and we can't use the mapped key, in case the OS maps it to Ctrl-&
[21:23] <bgK> in dosbox, without setting a keyboard layout, pressing &/1 displays 1, after setting the layout (keyb.com fr), pressing it displays &
[21:23] <m_kiewitz> and Ctrl-1?
[21:24] <m_kiewitz> so it seems DOSBox uses the scancode? but the other keys work properly, correct?
[21:24] <m_kiewitz> i mean without setting the layout
[21:24] <bgK> ctrl-1 displays nothing in both cases
[21:24] <wjp> Dosbox is a bit special as it emulates a (US, by default) keyboard at a low level
[21:24] <m_kiewitz> well, you need a program that shows the key or well try one of the SCUMM games
[21:25] <m_kiewitz> wjp: but why do the regular letter keys work then? shouldn't those be "a" = "q" and so on?
[21:26] <m_kiewitz> maybe they don't use scancodes for regular letters? but then why do umlauts work?
[21:27] <-- LyTchet left irc:
[21:32] <wjp> It has a usescancodes option. It's probably a bit hybrid if that's disabled. Unsure of the details. As I said, it's a bit of a special case
[21:39] <bgK> https://bugzilla.libsdl.org/show_bug.cgi?id=3188
[21:39] <bgK> this is the bug for SDL2 related to the inconsistent keycodes between OSs for azerty keyboards
[21:50] <-- frankyboy_ left irc: Read error: Connection reset by peer
[21:52] <-- WooShell left irc: Quit: Zu gotdy od mpy nrmy stpimf. Zu drvpmf zrsmd aogy jrt iq pt viy jrt yp yjr htpimf.
[21:59] <-- waltervn left irc: Quit: Leaving
[22:03] <-- abruanese left irc: Read error: Connection reset by peer
[22:07] --> abruanese joined #scummvm.
[22:13] <-- Raziel^ left irc: Quit: AmigaOS 4 (Unregistered copy). Evaluation period is over. Program will now quit. Thank you for using AmigaOS.
[22:20] --> go3m joined #scummvm.
[22:24] <-- go2m left irc: Ping timeout: 258 seconds
[23:04] --> GitHub19 joined #scummvm.
[23:04] <GitHub19> [scummvm] eriktorbjorn pushed 1 new commit to master: https://git.io/vKfy0

[23:04] GitHub19 (GitHub19@192.30.252.45) left #scummvm.
[23:05] <eriktorbjorn> While borrowing code from the BBVS engine, I noticed that it uses g_system in several places, rather than it's own _system. I don't have the time to look at that now, though.
[23:11] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:22] <-- SylvainTV left irc: Read error: Connection reset by peer
[23:36] --> Strangerke_ joined #scummvm.
[23:38] <-- Strangerke left irc: Ping timeout: 240 seconds
[23:38] Nick change: Strangerke_ -> Strangerke
[00:00] --- Mon Jul 4 2016