[Back to Index]

[00:05] --> dreammaster joined #scummvm.
[00:05] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[00:13] --> user10 joined #scummvm.
[00:14] <-- user9 left irc: Remote host closed the connection
[00:14] Nick change: user10 -> user9
[00:21] --> Deledrius_ joined #scummvm.
[00:24] <-- WooShell left irc: Quit: If you understand or if you don't, if you believe or if you doubt - There's a universal justice, and the eyes of truth are always watching you.
[00:24] <-- Deledrius left irc: Ping timeout: 268 seconds
[00:34] <TMM> snover, just finished lsl7, only broken thing appears to be the dice game. The rest seems just fine
[00:35] <snover> feat-ture. its pronounced feat-ture.
[00:36] <TMM> snover, ok, the only scummvm-specific feature appears to be that the dice game can't be lost ;)
[00:36] <snover> (i have no idea whats wrong with it)
[00:36] <TMM> well, lsl7 could at least be listed as 'completable'
[00:37] <TMM> I also tested all of the easter eggs I could remember, that also all works
[00:37] <TMM> only thing in the console log is a lot of these: WARNING: kRandom: caller requested to set the RNG seed!
[00:38] <Simei> _sev: t0by: Still having memory errors out of nowhere, I read a little the Lab history for inspiration. I think maybe I should redo all the things I've done by now ? Else, I may spend too much time on debugging.
[00:38] <-- Deledrius_ left irc: Quit: App.Exit
[00:39] <TMM> snover, anyway, great job on this :) I was very happy to play lsl7 again. If you need more testing on something else let me know
[00:39] --> Deledrius joined #scummvm.
[00:39] <snover> well, those warnings are probably why the game doesnt work
[00:40] <TMM> WARNING: mem 0033:0032 is truncated!
[00:40] <TMM> also those
[00:40] <snover> those warnings are going to crash the game if you arent careful where you stick your mouse
[00:40] <TMM> oh, I guess I was lucky then
[00:40] <TMM> I didn't have any crashes
[00:41] <Simei> _sev: t0by: I don't know if this is right. But I imagine to get back to the initial sludge detection code (the very first commit) and then calling the main function directly in the SludgeEngine::run()
[00:41] <snover> i assume that those showed up on the about screen?
[00:41] <snover> (the truncation warnings)
[00:41] <TMM> I don't think so
[00:42] <TMM> I'll replay and find out when those were emitted
[00:43] <Simei> _sev, t0by: I think it may be the better way to do for now.
[00:43] <Simei> _sev, t0by: What's your opinion about it ?
[00:46] <snover> i had a thing to use that seed in one of these stashes that i probably just dropped ho ho timing
[00:50] <-- Axy left irc: Read error: Connection reset by peer
[00:52] <snover> the good news is that i have a savegame at that point in the game already, so lets see what happens.
[00:52] --> omer_mor_ joined #scummvm.
[00:54] <-- omer_mor left irc: Ping timeout: 268 seconds
[01:48] <-- Dominus left irc: Ping timeout: 260 seconds
[01:48] --> Dominus joined #scummvm.
[02:12] --> kaenovsky joined #scummvm.
[03:15] <snover> (hours later) this dice game is actually broken quite a bit.
[03:16] <snover> not only does the opponent always challenge, you always lose when you challenge.
[03:17] <snover> so thats a little terrifying.
[03:17] <snover> also, i never see any warnings about the kRandom seed.
[03:26] <-- dreammaster left irc:
[03:59] kaenovsky (kaen@host144.186-125-230.telecom.net.ar) left #scummvm.
[05:01] <Strangerke> hi guys
[05:17] <-- holdsworth left irc: Ping timeout: 255 seconds
[05:19] --> holdsworth joined #scummvm.
[05:29] <-- Lightkey left irc: Ping timeout: 255 seconds
[05:42] --> Lightkey joined #scummvm.
[05:57] --> Strangerke_ joined #scummvm.
[05:58] --> Begasus joined #scummvm.
[05:59] <-- Strangerke left irc: Ping timeout: 246 seconds
[05:59] Nick change: Strangerke_ -> Strangerke
[06:13] --> Mia joined #scummvm.
[06:17] --> Henke37 joined #scummvm.
[06:30] <-- LittleToonCat left irc: Remote host closed the connection
[06:32] <-- eriktorbjorn left irc: Quit: Leaving.
[06:40] --> omer_mor joined #scummvm.
[06:42] <-- omer_mor_ left irc: Ping timeout: 260 seconds
[06:49] <-- Begasus left irc: Quit: Ex-Chat
[07:12] <Simei> Hi, Strangerke. Hi, everyone :)
[07:14] --> ajax16384 joined #scummvm.
[07:14] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[07:21] --> t0by joined #scummvm.
[07:21] <-- t0by left irc: Changing host
[07:21] --> t0by joined #scummvm.
[07:21] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[07:21] <t0by> Simei: that's precisely what I meant by "try redoing it from scratch from c19b8509"
[07:22] <-- Mia left irc: Read error: Connection reset by peer
[07:23] <t0by> The beauty of working under version control is that you can turn back time.
[07:28] --> eriktorbjorn joined #scummvm.
[07:28] #scummvm: mode change '+o eriktorbjorn' by ChanServ!ChanServ@services.
[07:39] <Simei> _sev, t0by: But at c19b8509, I have already to make some least OOP efforts. It doesn't have any problem then, but I wonder if this temptation of OOP may cause something I can't figure out myself later ? So I think I may leave all things as global functions and variables as they were and see again ?
[07:40] <Simei> _sev, t0by: It's what I want to do next. What's your ideas about it ?
[07:40] <t0by> Simei: please refer to _sev on this. I don't find it likely that memory leaks and danging pointers can be due to that, but wait for and follow _sev's directions.
[07:43] <Simei> yes, sir
[07:43] <t0by> Hey, don't "yes sir" me :)
[07:44] <Simei> ok, just the word "direction" makes me want to say it, haha
[07:45] <t0by> If you are essentially turning all globals into members of Engine and all functions into methods of Engine, as you were trying to do, what you end up with should be logically very close anyway. In fact, I'm surprised there are any issues here.
[07:49] <-- Henke37 left irc: Read error: Connection reset by peer
[07:49] --> Henke37 joined #scummvm.
[08:07] <-- t0by left irc: Quit: t0by
[08:08] --> t0by joined #scummvm.
[08:08] <-- t0by left irc: Changing host
[08:08] --> t0by joined #scummvm.
[08:08] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[08:17] --> m_kiewitz joined #scummvm.
[08:18] <-- m_kiewitz left irc: Changing host
[08:18] --> m_kiewitz joined #scummvm.
[08:18] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[08:22] <-- Harekiet left irc: Remote host closed the connection
[08:24] --> Harekiet joined #scummvm.
[08:54] <m_kiewitz> snover/wjp: it seems I'm currently getting the same issue as TMM some days ago
[08:54] <m_kiewitz> my already detected gk2 is shown as "unknown version" when I try to start the game
[08:54] <m_kiewitz> and md5s plus file sizes seem to 100% match
[08:56] <m_kiewitz> snover: I'm also getting this: WARNING: Resource audio.55730 from 55730.WAV points beyond end of 55730.WAV (146
[08:56] <m_kiewitz> 07808 + 175959 > 175959)!
[08:56] <m_kiewitz> looks like some code going wrong?
[09:27] <-- borosky left irc: Ping timeout: 260 seconds
[09:28] --> borosky joined #scummvm.
[09:33] <m_kiewitz> snover: im able to reproduce the issue. It happens automatically, you just need to wait inside the grotto for a bit without doing anything
[09:33] <m_kiewitz> now checking what original SCI does there
[09:38] <wjp> m_kiewitz: can you trace how the detector is deciding it's an unknown version?
[09:39] <m_kiewitz> havent looked into that for now. i think i had such an issue previously and deleting/adding the game once more fixed it
[09:39] <m_kiewitz> trying to get gk2 to work in dos-box right now :/
[09:40] --> WooShell joined #scummvm.
[09:40] <wjp> since TMM couldn't reproduce it afterwards, it's probably good to make a note of the current situation
[09:41] <WooShell> good meowning =^.^=
[09:42] <Simei> Strangerke, criezy: Hi, I tried to get myself into the dev emailing list
[09:43] <m_kiewitz> wjp: im backing up my scummvm.ini every day, i definitely won't readd the game until it got resolved
[09:43] <Simei> But I keep getting Invalid confirmation strings.
[09:44] <Simei> I guess it's because I've asked for it once around 10th but forgot to send the confirmation
[09:46] <Simei> Strangerke, criezy: So, what should I do now ? :/
[10:13] <TMM> m_kiewitz, I'm *not* crazy!
[10:21] <m_kiewitz> hmm, for gk2 it works fine in dos-box (well "fine", the game runs like a mess and was a pain to even get to work at all :/)
[10:21] <m_kiewitz> it seems a guard is supposed to appear, but in scummvm only the wait mouse cursor is shown and nothing else
[10:23] <TMM> m_kiewitz, fyi: lsl7 is completable. Apart from the dice game not functioning.
[10:24] <m_kiewitz> it's a feature :P
[10:24] <TMM> everyone keeps saying that...
[10:25] <m_kiewitz> whatever it is, i kinda want to keep it in some way, because i hate those kind of mini games
[10:25] <logix> not that my opinion matters much, but I think it should be either as in the original release or selectable
[10:26] <logix> if it's selectable then defaulting to "always win" is ok imho
[10:26] <TMM> we could just add another dialog or button like we originally did with FT for the insane stuff
[10:26] <TMM> before that was working
[10:27] <logix> it's obviously different for actual puzzles, like the bouncer ("if this is 2 fingers, how many fingers is this?") in monkey island 2...
[10:29] <m_kiewitz> logix: it will get fixed at one point, it should work like in the original game. but maybe including an option to change it back to the "broken" state
[10:35] <m_kiewitz> wjp/snover: in gk2 scripts are calling kDoAudio(DoAudioWaitForPlay) in a effective endless loop
[10:35] <m_kiewitz> waiting for it to return 0
[10:36] <m_kiewitz> and it's called with no real parameters, it's just kDoAudio(1)
[10:36] <m_kiewitz> can one of you please look inside the IDB? there has to be some inaccuracy in that case in ScummVM
[10:37] <m_kiewitz> ah, it's returning the number of active channels in ScummVM. but background music is playing, so it's always 1
[10:37] <m_kiewitz> hmmm
[10:40] <m_kiewitz> music is definitely playing when using original interpreter, maybe they changed something for the gk2 interpreter, idk
[10:55] <Henke37> there is the insane subengine. or as I like to call it, BenBikeMiniGame
[11:00] <Henke37> so there is precedense for replicating special minigames
[11:19] <-- t0by left irc: Quit: t0by
[11:33] <TMM> well, there is a way to cheat in the dice game
[11:34] <TMM> you can drag her cup away
[11:34] <TMM> I'm not sure if that really counts as a puzzle though
[11:39] --> t0by joined #scummvm.
[11:39] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[11:53] --> exg joined #scummvm.
[11:55] <exg> Hi, is it expected that the os x intel daily dev build does not work? when I launch it finder complains that the app is incomplete or broken
[12:04] --> jamm joined #scummvm.
[12:04] <-- jamm left irc: Changing host
[12:04] --> jamm joined #scummvm.
[12:15] --> criezy|Work joined #scummvm.
[12:15] #scummvm: mode change '+o criezy|Work' by ChanServ!ChanServ@services.
[12:16] <Mataniko> assume it's not intended
[12:16] <criezy|Work> Hi guys
[12:18] <criezy|Work> Simei: is your issue to register to the scummvm-devel list fixed?
[12:19] <criezy|Work> exg: That is not intended. I will try to remember to have a look in the next few days.
[12:19] <criezy|Work> Can you also tell me which version of OS X you are using?
[12:20] <exg> criezy|Work: great, thank you. I am using El Capitan (10.11.6)
[12:26] <exg> criezy|Work: in case it may help, I have just noticed that the dev build scummvm binary is 32-bit "ScummVM.app/Contents/MacOS/scummvm: Mach-O executable i386"
[12:28] <criezy|Work> Thanks. And I believe that's normal. Some intel macs are 32 bits (I have one).
[12:28] <criezy|Work> But indeed it may not work with the later OS X version. I will need to check that.
[12:29] <criezy|Work> There is actually both a 32 bits and 64 bits release of ScummVM 1.9.0 on our download page.
[12:29] <criezy|Work> Could you check if the 32 bits version works for you?
[12:29] <criezy|Work> (its on http://www.scummvm.org/downloads/)
[12:29] <criezy|Work> Otherwise I am going to do the checks when I am back home...
[12:29] <-- TMM left irc: Ping timeout: 240 seconds
[12:35] <exg> criezy|Work: the 1.9.0 32-bit version seems to work fine.
[12:38] <criezy|Work> exg: Thank you for confirming. So being 32 bits is probably not the issue with the daily build.
[12:43] <Simei> criezy|Work: Yes, it's fixed now. :)
[12:43] <exg> criezy|Work: oh actually, if I execute it from a shell it works.
[12:43] <criezy|Work> Simei: great!
[12:44] <criezy|Work> exg: you mean if you execute the daily build from a shell it works?
[12:44] <exg> criezy|Work: yes
[12:44] <criezy|Work> That's going to be a fun one to debug!
[12:45] <exg> criezy|Work: :)
[12:52] <exg> criezy|Work: one weird thing in the dev build is that the 'save' button in the menu opened by ctrl-f5 is greyed out.
[12:55] <exg> criezy|Work: oh well, seems to be specific to gabriel knight. save is available in other games.
[12:55] <criezy|Work> exg: The save button is only available for some engines and greyed out for others (such as SCI).
[12:56] <exg> criezy|Work: oh didn't know that thanks.
[12:56] <m_kiewitz> exg: saving via GMM is disabled for all of SCI. For SCI16 it's basically impossible to save using that menu simply because how it works internally. Saving is fully controlled by scripts and if scripts do not like a certain situation and you still force a save, all sorts of fun things can happen
[12:57] <m_kiewitz> for SCI32 it seems to be the same case, but maybe it's possible to somehow reliably detect such situations to avoid them and then enable saving via GMM for those
[12:59] <m_kiewitz> so it's really that way for all Sierra games using the SCI engine.
[13:00] <exg> m_kiewitz: i see, thanks for the explanation.
[13:27] <wjp> can we make this button entirely invisible if it will never be available in a game?
[13:28] <wjp> (so normal if you can save right now via the GMM, "greyed out" when there are situations where you can save, but not right now, and entirely invisible if it's just not available in the current game)
[13:32] <criezy|Work> I think it would be good to do something considering the number of users that are confused by this (see also the recent post in the forum on Simon the Sorcerer).
[13:35] <wjp> // TODO: setEnabled -> setVisible
[13:35] <wjp> is it really just that?
[13:36] <wjp> (engine/dialogs.cpp)
[13:38] <criezy|Work> It might be.
[13:38] <criezy|Work> At first glance it looks like this would be just that, but maybe I am missing something...
[13:40] <criezy|Work> I am wondering why whoever wrote that todo did it rather than change the code.
[13:40] <criezy|Work> I have to assume there was a drawback or issue with that change...
[13:41] --> LittleToonCat joined #scummvm.
[13:51] <wjp> at the time apparently there were GUI layout issues
[13:51] <wjp> https://bugs.scummvm.org/ticket/8925#comment:10
[13:54] --> Strangerke_ joined #scummvm.
[13:55] <wjp> I think I'll make this a PR, and we can discuss there to see if we're missing anything
[13:57] <-- Strangerke left irc: Ping timeout: 260 seconds
[13:57] Nick change: Strangerke_ -> Strangerke
[14:01] --> GitHub22 joined #scummvm.
[14:01] <GitHub22> [scummvm] wjp opened pull request #952: GUI: Fully hide GMM save/load buttons if not supported by engine (master...hide_gmm_saveload) https://git.io/vH3zX
[14:01] GitHub22 (GitHub22@192.30.252.45) left #scummvm.
[14:04] <t0by> Simei: hey, any news?
[14:06] <m_kiewitz> wjp: saw my comments on the gk2 chapter 6 game breaking issue?
[14:21] <wjp> no
[14:21] <Simei> _sev, t0by: I'm back at the original commit of detecting .slg files. I'm looking into the origin sludge code to make their formats align with formatting convention and separating the sludge engine related scripts from those of the develop kit.
[14:22] <t0by> Simei: you *are* using astyle, right?
[14:23] <t0by> also it doesn't do too much harm if you pull in one extra useless include from CommonCode
[14:26] <Simei> I installed a plugin for sublime.
[14:27] <Simei> The .vert and .frag files, we simply ignore them or keep them at present?
[14:28] <Simei> Yes, I guess all .h in CommonCode are useful
[14:28] <t0by> I take that as a no?
[14:29] <t0by> http://wiki.scummvm.org/index.php/Code_Formatting_Conventions#Automatically_converting_code_to_our_conventions
[14:29] <t0by> Here, run astyle *cpp with there settings and you're good to go.
[14:30] <Simei> Ah, thanks, I didn't notice this page
[14:30] <Simei> I mean, the Astyle link
[14:31] <t0by> Not the astyle link as much as the ready to go settings file.
[14:31] <t0by> you can astyle with your usual package manager, typically with "apt-get install astyle"
[14:32] <Simei> Ok, this is nice
[14:33] <t0by> Simei: the .frag files are probably GL stuff we won't use, but pull them in if you need them for the thing to compile.
[14:34] <Simei> Ok, I see
[14:51] <t0by> _sev: hah, my copy of Lingo Workshop arrived. Sorry for no activity in that area but as I mentioned I'm pretty swamped until early June.
[14:53] --> omer_mor_ joined #scummvm.
[14:54] <-- jamm left irc: Quit: Leaving
[14:55] <-- omer_mor left irc: Ping timeout: 246 seconds
[14:55] <-- Strangerke left irc: Ping timeout: 260 seconds
[14:55] <m_kiewitz> wjp: please take a look, if your idb has kDoAudio (SCI32) reversed, then please check what kDoAudio(WaitForPlay) is really doing
[14:56] <m_kiewitz> we return the currently active channels, which is 1 in gk2 because of digital background music. and the scripts wait for it to return 0
[15:24] <snover> hi, reading backlog&
[15:27] <t0by> Simei: compiling?
[15:27] <-- omer_mor_ left irc: Read error: Connection reset by peer
[15:27] <snover> m_kiewitz: that warning about 55730.WAV is pretty weird, can you breakpoint on that and share the backtrace? otherwise I will try to reproduce it here in a bit
[15:28] <snover> (break on the verifySignature call in VS)
[15:28] --> omer_mor joined #scummvm.
[15:28] <m_kiewitz> it happens right at the startup
[15:29] <m_kiewitz> and maybe it's localized gk2 only
[15:29] <m_kiewitz> there is a 55730.WAV in the movies directory
[15:29] <m_kiewitz> which isn't there in the UK gk2 release
[15:30] <snover> yeah, but I want to know who is calling, it should be ResourceManager::processPatch for such a file, but maybe it isnt somehow
[15:31] <snover> ugh, in the movies directory, really? does the resource.cfg for the game have the movies directory listed in patchdir?
[15:32] <m_kiewitz> nope
[15:32] <m_kiewitz> movieDir=CD:\movies
[15:32] <m_kiewitz> patchDir=C:\sierra\gk2dos;CD:\
[15:32] --> omer_mor_ joined #scummvm.
[15:33] <-- omer_mor left irc: Ping timeout: 268 seconds
[15:34] <snover> sigh.
[15:34] <snover> sierra.
[15:35] <snover> m_kiewitz: fwiw the dice minigame in LSL7 has a cheat mode that lets you see the opponents dice
[15:36] <snover> so it should really be fixed, even if it not working is the best feature :P
[15:37] <snover> i reversed sci32 audio so let me go see if GK2 changed something
[15:39] <m_kiewitz> well there has to be a change or an inaccuracy on that one kDoAudio command
[15:41] <snover> yeah, bindiff is showing that they changed some stuff&
[15:42] <m_kiewitz> :/
[15:42] --> ny00123 joined #scummvm.
[15:42] <m_kiewitz> btw. is there a way to detect digital music somehow?
[15:42] <m_kiewitz> in larry 7 it's really irritating that you can turn down volume of the music, but it still plays at full volume (obviously)
[15:42] <m_kiewitz> can we somehow detect it?
[15:43] <snover> when you say turn down volume you mean in scummvm launcher, or the in-game panel?
[15:43] <m_kiewitz> both
[15:44] <m_kiewitz> really weird that sierra didn't implement a volume control for the music
[15:44] <snover> they did
[15:44] <m_kiewitz> oh? then i missed it
[15:44] <snover> game -> audio mixer
[15:44] <m_kiewitz> where is it?
[15:45] <m_kiewitz> argh, yeah i now see it :/
[15:46] <m_kiewitz> can we somehow support that? have you looked into it?
[15:46] <snover> i havent programmed guest additions for any of the SCI3 games yet so the settings dont sync up from inside scummvm at the moment.
[15:46] <m_kiewitz> ah i see
[15:47] <m_kiewitz> gk2 plays really well inside ScummVM
[15:47] <m_kiewitz> i had to play it for a bit today in dos box too and wow what a difference
[15:47] <snover> excellent!
[15:47] <m_kiewitz> i hope that gog will add scummvm to gabriel knight 1+2 additionally
[15:48] <m_kiewitz> so people can play those games properly
[15:48] <m_kiewitz> mouse cursor works perfectly fine
[15:48] <m_kiewitz> really weird that larry 7 has those issues
[15:49] <m_kiewitz> at some point it would be really great to fully reverse Pandora Directive etc.
[15:49] <m_kiewitz> The Tex Murphy games
[15:50] <m_kiewitz> i own them since forever, but they are freezing on me and have all sorts of other issues. really a shame
[15:50] Action: t0by BWHAHAHAHAHAHAHAAHAHAHAHAAH gk2
[15:50] <m_kiewitz> probably one of the best FMV games ever
[15:51] <t0by> Simei: hey, is it compiling yet? Got stuck?
[15:54] <t0by> m_kiewitz: it's no plumbers though.
[15:54] <m_kiewitz> plumbers is imo only so bad that it's good again
[15:54] <m_kiewitz> Tex Murphy is just great :P
[15:55] <m_kiewitz> it's even self aware :P
[16:04] --> omer_mor joined #scummvm.
[16:07] <-- omer_mor_ left irc: Ping timeout: 260 seconds
[16:07] <-- criezy|Work left irc: Ping timeout: 260 seconds
[16:13] --> _sev_ joined #scummvm.
[16:13] <-- _sev_ left irc: Changing host
[16:13] --> _sev_ joined #scummvm.
[16:13] #scummvm: mode change '+o _sev_' by ChanServ!ChanServ@services.
[16:14] <-- _sev left irc: Ping timeout: 240 seconds
[16:16] <-- ajax16384 left irc: Read error: Connection reset by peer
[16:24] <snover> ARGH
[16:24] <snover> ok
[16:24] <snover> so, GK2 and maybe some earlier games (havent checked yet), instead of returning the number of active samples, returns the number of active samples that were loaded into memory (vs. ones that are streamed using a file handle)
[16:25] <Simei> _sev_, t0by: I'm compiling it, not succeeded yet. I got "exception handling disabled, use -fexceptions to enable", what does this mean
[16:25] <Simei> comment it ?
[16:25] <_sev_> Simei: yes, we do not use exceptions at all
[16:25] <_sev_> you will have to rewrite that code eventually
[16:25] <_sev_> I recommend not commenting it out, but putting #if 0 #endif
[16:26] <_sev_> then there are much less diffs to your changes
[16:26] <m_kiewitz> snover: what
[16:26] Nick change: _sev_ -> _sev
[16:26] <m_kiewitz> so maybe that's what they did to remove music from the count
[16:27] <m_kiewitz> is streaming / not streaming just a flag somewhere during init/play?
[16:27] <m_kiewitz> afaik we stream basically everything from disc
[16:28] <snover> actually, we stream basically nothing from disc
[16:28] <m_kiewitz> oh we don't?
[16:28] <m_kiewitz> i think wave player etc. are streaming when possible
[16:28] <m_kiewitz> (talking about ScummVM as a whole here)
[16:28] <snover> i complained about this a while ago :) https://bugs.scummvm.org/ticket/9576
[16:29] <m_kiewitz> ah, well I remember making the Miles audio support to be able to stream
[16:29] <Simei> _sev_: Thanks, I get it
[16:30] <snover> yeah. its not a limitation of scummvm but a limitation of the sci resource manager code
[16:31] <snover> so of course, we dont actually *need* to stream audio in order to tell the game that were streaming some samples
[16:31] <snover> but, i need to go remember how the games did this. there is kDoAudioPreload to load samples into memory, but i dont remember if there was also a flag in the play calls to do the same thing
[16:32] --> waltervn joined #scummvm.
[16:32] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[16:33] <waltervn> evening
[16:33] <snover> hey waltervn
[16:36] <m_kiewitz> snover: when streaming gets implemented, I guess we shouldn't use it for everything. small sound effects can be directly loaded into memory imo
[16:37] --> Strangerke joined #scummvm.
[16:37] <Simei> LINK scummvm
[16:37] <Simei> engines/sludge/libsludge.a(detection.o): In function `SludgeMetaEngine::createInstance(OSystem*, Engine**, ADGameDescription const*) const':
[16:37] <Simei> /home/simei/Projects/scummvm/engines/sludge/detection.cpp:88: undefined reference to `Sludge::SludgeEngine::SludgeEngine(OSystem*, Sludge::SludgeGameDescription const*)'
[16:37] <Simei> collect2: error: ld returned 1 exit status
[16:37] <Simei> Makefile.common:81: recipe for target 'scummvm' failed
[16:37] <Simei> make: *** [scummvm] Error 1
[16:38] <Simei> _sev_, t0by: I'm getting a such compiling error. What should I do?
[16:38] <m_kiewitz> i guess resource manager not support it isn't that weird. Originally we had no support for sci1.1 and before that there were almost only small sound effects and no voices
[16:38] <m_kiewitz> *supporting
[16:39] <t0by> Simei: let's see your code.
[16:40] <t0by> Simei: I suspect it's exactly what it says on the tin, i.e. Sludge::SludgeEngine::SludgeEngine(OSystem*, Sludge::SludgeGameDescription const*) is declared but not implemented anywhere.
[16:40] <t0by> Simei: have you updated modules.mk?
[16:40] <t0by> That could be a reason why the linker can't find it.
[16:42] <somaen> I would bet on the relevant file that contains that constructor, not even being compiled
[16:42] <t0by> That would be modules.mk, ainnit?
[16:42] <somaen> Yup
[16:43] <somaen> And that is defined in terms of object-files, which should drive both the compilation AND linking
[16:43] <somaen> So, presumably, the file isn't compiled at all
[16:44] <Simei> https://www.irccloud.com/pastebin/KBV16rpd/module.mk
[16:45] <m_kiewitz> snover: any more game breaking issues open right now?
[16:45] <somaen> Simei: Which file implements your constructor
[16:45] <somaen> ?
[16:45] <Simei> _sev_, t0by: This is my module.mk. I've updated it. But I don't know if it's in a correct way. :/
[16:46] <somaen> I.e. which file contains SludgeEngine::SludgeEngine ?
[16:47] <t0by> hm
[16:47] --> omer_mor_ joined #scummvm.
[16:47] <Simei> True, I forget about that one!
[16:47] <somaen> Pro-tip, if you're unsure if a build system picks up something, add deliberate compile errors to the file
[16:48] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[16:49] <Simei> somaen: note it. thanks :)
[16:49] <-- omer_mor left irc: Ping timeout: 272 seconds
[16:50] <snover> m_kiewitz: yes, https://bugs.scummvm.org/ticket/9761, https://bugs.scummvm.org/ticket/9748 (which seems to not be reproducible), QFG4 is almost certainly full of crashes, missing MIDI in MGDX makes that game basically useless
[16:53] <m_kiewitz> oh oh oh a gabriel knight 1 issue, now I will gladly work on that one for now :P
[16:56] --> LyTchet joined #scummvm.
[17:09] <-- Strangerke left irc: Ping timeout: 268 seconds
[17:13] <-- Strangerke|work left irc: Ping timeout: 260 seconds
[17:22] <t0by> Simei: solved?
[17:23] <-- ny00123 left irc: Ping timeout: 240 seconds
[17:25] <Simei> _sev, t0by: Yes, still a number of glfunctions to be placed between #if 0 #endif. Nothing confusing for now.
[17:26] --> Strangerke joined #scummvm.
[17:28] <Simei> _sev, t0by: I imagine functions begin with "al" like "alureInitDevice" are from some sound library? Deal them the same way as gl functions?
[17:29] <t0by> Stuff from sound_openal.cpp is indeed OpenAL-specific stuff. Deal them the same way as GL functions.
[17:30] <t0by> Needless to say we won't be using OpenAL nor OpenGL-specific stuff.
[17:32] <Simei> Ok, I see
[17:53] --> exmensa_ joined #scummvm.
[17:54] <-- exmensa left irc: Ping timeout: 240 seconds
[18:19] --> omer_mor joined #scummvm.
[18:20] <Lightkey> m_kiewitz: how is Gabriel Knight 2 improved in ScummVM?
[18:20] <m_kiewitz> it works way more smoothly without hickups, palette flashes, etc.
[18:21] <m_kiewitz> and it "just works"[tm]
[18:21] <-- omer_mor_ left irc: Ping timeout: 240 seconds
[18:28] <wjp> m_kiewitz: kIsOnMe should never do an error() due to invalid coordinates, so this won't be a script issue
[18:28] <m_kiewitz> wjp: yes
[18:29] <m_kiewitz> i just told snover that there has be basic checking before calling the view pixel checking code
[18:29] <m_kiewitz> and that's either incorrect (off-by-one) in SSCI or not 100% accurate in ScummVM
[18:29] <wjp> all the rounding going on in there is very error prone
[18:30] <m_kiewitz> oh right, the rounding :/
[18:30] <m_kiewitz> in any case, point was to document from where kOnMe is getting called from and with which parameters
[18:31] <m_kiewitz> still there has to be basic rect verification and that should be double checked for any differences
[18:33] <wjp> basic checks are there, but with all the scaling, those are easy to get wrong
[18:34] <t0by> m_kiewitz: speaking of improvements, feature request: it would be fantastic if there was a way to adjust the gamma of the FMVs. It seems to me those were compressed with CRTs in mind -- specifically the kind where all greys would turn to black.
[18:34] <m_kiewitz> wjp: you can debug that part w/ dos-box and IDA?
[18:35] <wjp> possibly
[18:35] <wjp> should be relatively doable
[18:35] <m_kiewitz> t0by: speaking of improvements, i need to work on that phantasmagoria 2 subtitle patch (create our own)
[18:36] <t0by> What's with phantasmagoria?
[18:36] <t0by> Sorry, I never really ever paid any attention to phantasmagoria.
[18:36] <t0by> Call me a heretic.
[18:36] <m_kiewitz> it has no subtitles, unless you turn audio off afaik. or none at all
[18:36] <m_kiewitz> there is a fan "patch", but it's not really a patch, but all scripts of the game in 1 zip file
[18:37] <m_kiewitz> and it also enforces subtitles
[18:37] <m_kiewitz> and saved games are also not compatible
[18:38] <wjp> m_kiewitz: can you attach the full envelope object? (Including the scaleX, scaleY, maxScale, etc selectors)
[18:38] <m_kiewitz> sure
[18:39] <logix> I hope whatever loads the subtitles from those zips has no problems with ".."s in directory names
[18:39] <m_kiewitz> wjp: it's really weird in any case, i don't click right at the lower right edge of the newspaper, but higher than that.
[18:42] <Simei> _sev, t0by: For the file movie.cpp, there are some vpx headers included (vpx_decoder.h, vp8dx.h) and vpx functions. Treat them the same way as well?
[18:45] <Simei> _sev, t0by: If so, there will be AL, GL and VPX functions stubbed out . Should I stub the SDL functions as well?
[18:46] LyTchet (alexandre@host-85-26-83-225.dynamic.voo.be) left #scummvm.
[18:46] <m_kiewitz> wjp: what's really weird is that if I misclick a tiny bit, the game thinks I clicked on the door
[18:46] <m_kiewitz> although the door is maybe 10 pixels above where I clicked
[18:49] <m_kiewitz> maybe it's intentional and its the doormat, need to verify in ssci
[18:50] --> ny00123 joined #scummvm.
[18:51] <m_kiewitz> hmm it seems a tiny part of the doormat is really seen as the door o_O
[18:54] <wjp> oh, no scaling
[18:54] <wjp> so that should make things at least relatively easy to follow
[18:55] --> Begasus joined #scummvm.
[18:56] --> Mia joined #scummvm.
[18:57] <m_kiewitz> i have only tried in hires mode
[18:57] <m_kiewitz> you tried in low res mode and it also assert()s?
[18:57] --> Begas_VBox joined #scummvm.
[19:06] <wjp> I haven't tried at all
[19:07] <wjp> (other than noticing I can't load the save)
[19:07] <wjp> oh, by the way, I still have this uninitialized read in castlebrain open in the debugger from last week
[19:07] <m_kiewitz> worked for my gk1 Limited Edition copy
[19:08] <m_kiewitz> which is 01.100.000
[19:08] <wjp> GOG here
[19:09] <m_kiewitz> which version is that?
[19:10] <m_kiewitz> you can change the day by setting global 7Bh
[19:10] <wjp> no VERSION file; SIERRA.EXE contains 2.000.000
[19:11] <m_kiewitz> urgh
[19:11] <m_kiewitz> resource.000 date/time?
[19:11] <wjp> in 2010
[19:11] <m_kiewitz> urgh
[19:11] <m_kiewitz> any file from < 2000?
[19:12] <wjp> all files are 2010-01-29, except RESOURCE.CFG, which is 2010-07-22
[19:12] <m_kiewitz> :/
[19:13] <wjp> md5's match the gk1 CD entry in our detection tables
[19:13] <m_kiewitz> just start a new game, set global 7Bh to 6 during intro
[19:13] <m_kiewitz> hopefully the envelope will be there
[19:13] <wjp> (with .map size 10996)
[19:14] <t0by> Simei: sorry, I'm back.
[19:15] <t0by> Simei: question: what do you think?
[19:15] <m_kiewitz> mine is the same size
[19:15] <m_kiewitz> resource.000 is 12581736 in size
[19:15] <wjp> yeah
[19:17] <wjp> any idea where the version in "WARNING: Save game was created for game version 01.100.000, but the current game version is 1.000.051!" comes from?
[19:18] <wjp> (_seems_ to be a kCheckSaveGame32 thing, but I don't know where that's from exactly)
[19:18] <t0by> Comes from version 1.000.051 probably.
[19:18] Action: t0by hides
[19:20] <wjp> odd
[19:20] <wjp> scripts read that from the version file
[19:22] <wjp> except that it doesn't somehow
[19:23] <wjp> oh, urgh
[19:24] <wjp> I seem to have a version file in my savegame directory...
[19:25] <wjp> from 2009. I wonder if that's from something before we started tagging files with their game id
[19:26] <m_kiewitz> lol
[19:26] <m_kiewitz> maybe we should ignore files like those all the time in that directory
[19:30] <wjp> seems my GK1 install is a bit of a mess
[19:31] <wjp> I wonder if it's an older gog one, or if I messed it up myself
[19:31] <wjp> extracting the latest gog installer, things look much better
[19:37] <wjp> ok, fresh copy. Can load the savegame now
[19:38] Action: wjp browses through his savegame directory
[19:40] <wjp> I seem to have a files called "a" and "b" in there too (looking like they're in dreamweb format), blue.egg, HISCORE.DAT, SAVEGAME.{007,009,018}, saveFile.0, iq-points, "kyra1. 0" (with spaces, yes)
[19:46] <Simei> _sev, t0by: sev has mentioned that SDL should be stubbed. But SDL calls don't generate much compiling errors.
[19:47] <t0by> What do you mean by "much"?
[19:48] <_sev> Simei: they must be replaced with OSystem
[19:48] <t0by> Simei: note that "stubbed" and "commented out" don't mean the same.
[19:48] <_sev> Simei: so, currently I recommend to #ifdef them, and then we make it run
[19:48] <m_kiewitz> wjp: btw. wait, you had a version file in your saved game directory? Why was this used? Shouldn't it only use files with the game prefix?
[19:48] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[19:49] <_sev> Simei: one important thing: There must be NO FORBIDDEN_SYMBOLS macros in the engine
[19:49] <t0by> Usually, you stub a function by making it uneffectful, and that can very well be achieved by commenting any preexisting code.
[19:50] <_sev> Simei: like here: https://github.com/yinsimei/scummvm/blob/WIP1-1/engines/sludge/sludge.cpp#L23
[19:50] <_sev> Simei: this is a big no-no
[19:50] <t0by> _sev: about that -- you don't think Simei should first get the script engine with stubbed builting running without being horribly borked and only after that replace fopens with OSystem calls and maybe then squash?
[19:50] <t0by> *stubbed builtins
[19:50] <_sev> t0by: no, I prefer those to be #ifdef'ed rather
[19:51] <_sev> t0by: and then start with rewriting things with ReadStream
[19:51] <t0by> #ifdef'ed out, you mean?
[19:51] <_sev> yes
[19:52] <t0by> But still, first run, then rewrite with OSystem, FileManager and whatnot, right?
[19:52] <_sev> first, make it compile
[19:52] <_sev> then start looking into the game data reading
[19:52] <t0by> Well, that's necessary in order to run - but yes, that's commit #0, I guess.
[19:53] <_sev> then start display
[19:53] <-- Begasus left irc: Quit: Ex-Chat
[19:53] <_sev> commit #0 is the original engine
[19:53] <_sev> with no files added to module.mk
[19:53] <Simei> _sev, t0by: So I get the code compilie now. If I understand well, I test the simple game with it and then add the NO FORBIDDEN_SYMBOLS to make it recompile again ?
[19:53] <_sev> then add file to module.mk, make it compilable, repeat
[19:54] <_sev> Simei: do not put any FORBIDDEN_SYMBOLS-related macros
[19:54] <_sev> but if you have it all compilable with those
[19:54] <_sev> then ok, let's make it commit #0
[19:54] <Simei> what is the order? Reset module.mk to empty first ?
[19:54] <_sev> Simei: do you have it all in the module.mk and compilable?
[19:54] <Simei> Yes
[19:55] <_sev> then congrats!
[19:55] <_sev> commit
[19:55] <_sev> then start looking into file reading routines
[19:55] <_sev> replace them with ReadStream class calls
[19:56] <_sev> once done, remove FORBIDDEN_SYMBOLS
[19:56] <_sev> but when you commit, that will let me start reviewing
[19:57] <_sev> commit and push :)
[19:57] <Simei> https://github.com/yinsimei/scummvm/tree/newWIP1-1
[19:57] <Simei> I make a new branch for it
[19:58] <t0by> _sev: note the thing with that engine is that literally *everything* is handled in scripts with builtin primitives and constructors, there is *no* data files except raw bitmaps and waveforms; so in order to display *something* I guess you'd need the script engine up and running. That's why I had suggested that Simei first check that is working and then start replacing stdio.h with ReadStream. You think she should postpone t
[19:58] <t0by> hat and focus on replacing the stream reading calls?
[19:58] <_sev> Simei: one immediate note. Please fix the commit log message of your last commit in that branch
[19:58] <Simei> Ah, yes
[19:58] <Simei> SLUDGE:
[19:59] <_sev> t0by: good point
[19:59] <_sev> t0by: if it reads the files correctly now, which I highly doubt, then you might consider that way
[19:59] <_sev> t0by: but there could easily be all kinds of troubles if you use stdio
[20:00] <t0by> _sev, Simei reported success in this regard earlier
[20:00] <t0by> (except that then the script engine exploded)
[20:00] <_sev> t0by: e.g. subdirectories, file name cases
[20:00] <_sev> t0by: endianness on reading
[20:00] <t0by> but she was reportedly able to load scripts correctly
[20:00] <t0by> Ew, endinanness.
[20:01] <_sev> let me see if it compiles and runs on Mac
[20:02] <t0by> Anyway, if you read the backlog she *was* able to load a simple script on her other branch, but then with scripts with more complex primitives the interpreter exploded - my guess is that she was overeager in stubbing stuff and left a dangling pointer or broke the stack at some point.
[20:02] <_sev> I wouldn't hold my breath. The engine is originally Win-only and Simei is working on Win too
[20:02] <t0by> Simei: can you confirm?
[20:02] <t0by> _sev: Simei is working on Linux I think
[20:02] <Simei> I'm working on linux
[20:03] <Simei> The engine is good for linux and windows I think
[20:03] <t0by> (I can try to build it on Win, if you need)
[20:03] <_sev> yep, it blows up
[20:04] <_sev> C++ engines/sludge/backdrop.o
[20:04] <_sev> engines/sludge/backdrop.cpp:4:10: fatal error: 'libpng/png.h' file not found
[20:04] <_sev> #include <libpng/png.h>
[20:04] <_sev> that must be removed in favour of our image/png.h
[20:05] <_sev> and these:
[20:05] <_sev> C++ engines/sludge/transition.o
[20:05] <_sev> In file included from engines/sludge/transition.cpp:6:
[20:05] <_sev> In file included from engines/sludge/backdrop.h:2:
[20:05] <_sev> engines/sludge/graphics.h:13:10: fatal error: 'GLee.h' file not found
[20:05] <_sev> #include "GLee.h"
[20:05] <_sev> ^
[20:05] <_sev> 1 error generated.
[20:05] <Simei> Ah, ok. Why did it work for me
[20:06] <_sev> and these
[20:06] <_sev> engines/sludge/sound_openal.cpp:23:10: fatal error: 'AL/alure.h' file not found
[20:06] <_sev> probably because you have those libs installed in your default include path
[20:07] <Simei> How can I make it as if I don't have those libs ?
[20:07] <t0by> "AL/alure.h" can be ifdef'd out anyway, no?
[20:07] <_sev> just remove those includes
[20:07] <wjp> it's probably better to just remove all non-scummvm #includes
[20:08] <t0by> ROFL
[20:08] <t0by> https://github.com/yinsimei/scummvm/commit/9ed8bd943b58be6d1fb278b55d09bfc72befbede#diff-a8c529ab73ecb310b0d371361cf578e0R133
[20:08] <t0by> Wait, that's no longer funny in light of recent events.
[20:10] <Simei> Ok, I will ifdef'd them and make another commit
[20:10] <t0by> Uh, no
[20:10] <t0by> amend this one
[20:10] <t0by> we don't want non-compiling commits, remember?
[20:11] <Simei> Yes, it's want i meant to do
[20:12] <t0by> You might as well do as wjp said and strip those out. We are going to replace all SDL and OpenGL primitives with OSystem calls (the OSystem is our very own patented, ultraportable abstraction layer)
[20:12] <wjp> s/patented/Free/ :-)
[20:13] <t0by> "Patented as in beer".
[20:13] <Simei> So it's all the non-scummvm and non-system #includes to remove?
[20:13] <_sev> Simei: yes, and make it compilabel again after that
[20:13] <logix> my inner eye just pictured t0by like this: http://tinyurl.com/c6gj666
[20:13] <_sev> it will reveal more non-portable function calls
[20:13] <Simei> Ok, I see
[20:14] <wjp> engine code can't use any libraries/headers from outside of ScummVM (with the exception of a few basic C headers if required)
[20:14] <t0by> logix: this is how t0by's inner eye pictures t0by: https://media.giphy.com/media/yr7n0u3qzO9nG/giphy.gif
[20:14] <wjp> since inside engines you can't make any assumption on the presence/absence of these libraries
[20:14] <Simei> Ok, I note it
[20:15] --> ajax16384 joined #scummvm.
[20:15] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[20:15] <wjp> ScummVM comes with a large set of interfaces for things like I/O, graphics, as well as image/video/audio codecs
[20:15] <wjp> speaking of which, the webm thing will be interesting
[20:16] <t0by> don't we do matroska already?
[20:16] <t0by> oh, no, that's theora.
[20:16] <wjp> not as far as I know
[20:17] <Simei> what is "that webm thing"?
[20:17] <wjp> you have a subdirectory 'webm' in your tree
[20:17] <t0by> opensludge/source/Engine/libwebm
[20:17] <wjp> or libwebm yes
[20:18] <t0by> I wouldn't concern too much with that for now.
[20:18] <Simei> Ok, so I keep them there at present
[20:18] <t0by> *I wouldn't concern myself too much with that for now
[20:19] exg (giaquie1@unaffiliated/exg) left #scummvm ("fading...").
[20:20] <wjp> yeah, something for later
[20:24] <Simei> I see
[20:24] --> TMM joined #scummvm.
[20:24] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[20:25] <wjp> (It's presumably for videos, which are probably not a priority. But take whatever I say about sludge with a grain of salt, since I know nothing about it)
[20:27] <t0by> wjp: it's only used in playMovie from movie.cpp, in turn called only from the builtin playMovie, so...
[20:28] <t0by> I think it's pretty safe to say you're right.
[20:31] <Simei> So can I presume only the headers in scummsys.h are legal?
[20:31] <-- Harekiet left irc: Ping timeout: 240 seconds
[20:32] <wjp> well, the ones in there you don't need to include yourself anymore
[20:33] <Simei> Ok, so all other stuffs not from the engine itself
[20:33] <Simei> I remove them
[20:33] <TMM> Simei, ah, you're the GSoC student?
[20:35] <wjp> but you may have to add new #include "common/scummsys.h" lines
[20:35] <Simei> Yes, I've been accepted this year by GSoC
[20:36] <Simei> wjp: Ok, I get it
[20:43] --> Harekiet joined #scummvm.
[21:01] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[21:05] <-- ajax16384 left irc: Read error: Connection reset by peer
[21:22] <TMM> Simei, congratulations! And good luck
[21:23] --> girafe joined #scummvm.
[21:30] <snover> m_kiewitz: in the areas where rects are scaled (and so may have different rounding behaviour), the mulinc and mulru helper functions are used to do this correctly. otherwise, rects coming in from game scripts via kernel calls are converted from inclusive rects in the game scripts to exclusive rects in the kernel.
[21:30] <-- ny00123 left irc: Quit: Leaving
[21:35] <TMM> snover, I had a look at that Krandom codebase, it seems that swapping that out for a manually seedable prng wouldn't be terrible difficult, although it would touch osystem. Would you be interested in me taking a stab at that?
[21:37] <snover> TMM: the OSystem PRNG is seedable :) Common::RandomSource::setSeed
[21:37] <TMM> snover, actually, never mind, Common::RandomSource() actually has a setSeed(), I must be on crack
[21:37] <snover> but weirdly i never see any call to try to try to set seed in my copy of lsl7
[21:38] <TMM> I'll just hook something up and see if it fixes the dice game
[21:38] <TMM> or of it changes the behavior of the dice game at all
[21:38] <snover> if you would fix the dice game so i dont have to, i would be grateful :)
[21:38] <TMM> I can try
[21:38] <TMM> If it's actually somewhere in the guts of sci I probably won't be able to in any reasonable amount of time
[21:39] <TMM> but if it's stuff on the periphery like this I should be able to be somewhat useful :)
[21:40] <TMM> If it was something else that's unimplemented I'd have seen different debug messages right? Or do I need to enable some other debug flag in scummvm to see that?
[21:41] <snover> generally, unimplemented kernel calls blow up.
[21:42] <Simei> TMM: Thanks!
[21:42] <snover> it doesnt mean that there isnt an incorrectly-implemented kcall though :|
[21:42] <m_kiewitz> TMM: i wouldnt be surprised in case the mini game uses for example temp variables to store a state or something like that
[21:43] <m_kiewitz> in sci16 at least temp variables were not set to 0
[21:43] <m_kiewitz> but we do that, in fact we actually set a special value so that we can detect when uninitialized temps are used
[21:43] <m_kiewitz> anyway in such a case it would work in ssci but not with ScummVM
[21:43] <m_kiewitz> no idea if temp space gets reset in sci32, snover probably knows that
[21:44] <TMM> Yeah, maybe. The weird thing is that the opponent will think you are bluffing 100% of the time. I imagine that there's a certain amount of weighted chance she thinks you're bluffing
[21:44] <TMM> I wonder if they picked some kind of 'safe' seed in the prng arrived at by empirical study to balance that
[21:44] <TMM> al lowe did like his 'clever hacks'
[21:45] <snover> TMM: well, again, i dont see any call to set the seed in my game
[21:45] <snover> and it is broken
[21:45] <snover> it seems like the game does not know what the dice values are
[21:45] <TMM> the dice game also doesn't work for you?
[21:46] <snover> right. the opponent always challenges.
[21:46] <snover> and challenging the opponent seems to always fail even if their bet was wrong
[21:46] <snover> at first i thought that i just didnt understand the rules of the game :)
[21:46] <TMM> I'll give the sci debugger a go
[21:46] <TMM> sounds like fun
[21:47] <TMM> this *is* the sort of bug I'm usually pretty good with
[21:47] <snover> i wonder about this workaround { GID_LSL7, -1, 64017, 0, "oFlags", "clear", NULL, 0, { WORKAROUND_FAKE, 0 } }, // demo version, when it starts, and whenever the player chooses to go to the "Strip Liar's Dice" mini game
[21:47] <m_kiewitz> uh oh
[21:48] <TMM> oh, that looks... dangerous if I understand the workaround correctly
[21:48] <m_kiewitz> TMM: start ScummVM with parameters
[21:48] <m_kiewitz> --debugflags=ScriptPatcher,Workarounds
[21:48] <m_kiewitz> and then play that game
[21:48] <snover> i need to run; i will be back in a couple hours or so.
[21:49] <snover> the cheese maker is the only lsl7 script patch
[21:49] <TMM> I'll just try some things, I quite *like* the liar's dice minigame
[21:49] <TMM> as the only human on the planet apparently
[21:50] <m_kiewitz> maybe ssci only works because they use an uninitialized variable (so pseudo RNG)
[21:58] <-- Polynomial-C left irc: Remote host closed the connection
[22:06] <-- girafe left irc: Read error: Connection reset by peer
[22:09] <t0by> Simei: still awake? Compiling yet?
[22:09] <TMM> Is it possible to open the sci debugger on the controlling tty on Linux?
[22:09] <TMM> rather than inside the scummvm window?
[22:14] <m_kiewitz> TMM: i don't think so, but I'm on windows so what do I know :P
[22:15] <TMM> bleh
[22:15] <TMM> I like being able to copy/paste :)
[22:16] <wjp> TMM: yes it is
[22:16] <wjp> ./configure --enable-text-console
[22:16] <wjp> debugger is pretty much unusable without that
[22:17] <m_kiewitz> wjp: dont tell me that also works on windows
[22:17] <wjp> not sure; with libreadline it might work
[22:18] <TMM> wjp, thank god
[22:18] <m_kiewitz> I always wanted to disable the animation of the debug window, because that slows me down like crazy
[22:18] <Simei> t0by: yes, still awake. Still stubbing. I'll put the github link here when it's good
[22:18] <wjp> that oFlags::clear() function makes no sense
[22:19] <t0by> Aaaah
[22:19] <t0by> now it clicked
[22:19] <t0by> Simei: sorry, but you *have* fixed the compilation issues, haven't you?
[22:19] <t0by> Have you committed and pushed that?
[22:21] <Simei> _sev, t0by: with all GLee.h remove, I get variable types like GLubyte, GLuint exposed
[22:21] <_sev> yes, and those do not belong to ScummVM in any case
[22:22] <wjp> m_kiewitz: if you want we can have a look at getting this to work for windows some time. (Maybe for the msys terminal if not for the standard windows console window)
[22:22] Action: t0by squints
[22:22] <t0by> I'm sorry, "exposed" how?
[22:22] <Simei> _sev, t0by: So I put all the lines concerning these variables into #if 0 #endif ?
[22:22] <_sev> Simei: yes
[22:22] <m_kiewitz> wjp: that would be very helpful. well if at least that debug window animation could be turned off for single stepping alone
[22:23] <Simei> t0by: Because I didn't pay attention to the illegal headers
[22:23] <TMM> wjp, it should at least already work for cygwin builds I imagine
[22:23] <m_kiewitz> what i also never understood why we print out some stuff in the additional window and other stuff inside the GUI debug window
[22:23] Action: t0by double squints
[22:23] <wjp> m_kiewitz: personally I don't really see the point in improving the our internal debugger window too much, because it'll never be as useful as an external one
[22:23] <Simei> I just fix the reference compiling errors with them
[22:23] <wjp> TMM: cygwin? *shudder*
[22:23] <m_kiewitz> for example object state, wouldn't it be useful to get this out in both windows?
[22:24] <TMM> wjp, for debugging? who cares :P
[22:24] <m_kiewitz> yeah well, if the single stepping worked inside the text window, I wouldn't care about the graphical animation anymore
[22:24] <t0by> Simei, _sev: the plan is still that Simei is gonna make it compile by stupidliy ifdef'ing all GL- and SDL-specific stuff, commit, share for review and *then* move on, right?
[22:24] <wjp> m_kiewitz: I have no idea what happens with debugging output currently, since I never use the internal debugger window
[22:24] <t0by> Also, what do we mean by "stubbing", here?
[22:24] <_sev> t0by: yes
[22:24] <Simei> I only dealt with gl function then. With headers removed now, there are also varaibles to be removed
[22:25] <Simei> yes, i think so
[22:25] <wjp> m_kiewitz: for me everything appears in the external terminal, so I have no idea what is shown where otherwise
[22:25] <t0by> I call "stubbing" writing a "dumb" implementation of an interface or taking an existing implementation and ripping its effectful part out.
[22:25] <m_kiewitz> wjp: ah i see, well it's a mess. Some gets on the text window, some on the internal one, some on both
[22:25] <Simei> stubbed*
[22:25] <_sev> body #endif
[22:25] <m_kiewitz> and it's of course a nightmare when you want to copy things out.
[22:25] <_sev> grrr
[22:25] <t0by> s/ripping/ifdef'ing/
[22:25] <wjp> m_kiewitz: which build system and/or environment do you use?
[22:26] <_sev> t0by: warning("function FOO was called"); #if 0 // real function body ... #endif
[22:26] <t0by> Yes.
[22:26] <m_kiewitz> wjp: visual c++, windows xp / 7
[22:26] <t0by> _sev: Simei should be doing that *only as much as needed* to make it compile thoyugh, right? (That probably means all the GL stuff and the like, no?)
[22:27] <_sev> yes
[22:27] <t0by> i.e. should not start stubbing out stuff at random?
[22:27] <_sev> because then it will help with gradually fixing it and watching the regressions
[22:27] <t0by> It shouldn't be an awful lot of stuff, should it?
[22:29] <_sev> normally not
[22:29] <_sev> you come down to the low level functions, and that is where it goes
[22:29] <_sev> Simei: and if you have too many things, just comment out those files from module.mk, so you could continue later
[22:31] <Simei> _sev: Ok, I see, it should be more effective with files like graphics
[22:31] <t0by> Well, you start pruning the tree at the top, so to speak.
[22:32] <t0by> Simei: I'm sorry, if you have a second could you please commit on a throwaway branch and share your current status? I'd please like to have a look at one thing.
[22:33] <Simei> Ok, I'll do it
[22:37] <Simei> https://github.com/yinsimei/scummvm/tree/throw_out_branch
[22:37] <Simei> t0by: still half way, not compiled at all
[22:37] <t0by> yeah yeah
[22:40] <t0by> https://github.com/yinsimei/scummvm/commit/dc01698ef929bf80fdd6b9dd6ef1bc160051a369#diff-a8c529ab73ecb310b0d371361cf578e0R64
[22:40] <t0by> _sev: is this too fine grained?
[22:40] --> dreammaster joined #scummvm.
[22:40] <t0by> I would be tempted to stub out the whole body on that low level stuff
[22:40] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[22:43] <-- dreammaster left irc: Client Quit
[22:43] <t0by> In fact, being an axe-welding caveman, I would probably be tempted to comment out stuff like *openal.cpp and *sdl.cpp from modules.mk altogether and stub them out from the client code that calls them.
[22:44] --> dreammaster joined #scummvm.
[22:44] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[22:46] <t0by> Simei: my *very personal* opinion is that if we are trying to make it *compile* you shouldn't probably be cherry-picking lines inside stuff like *openal.cpp
[22:46] <t0by> As _sev said a few lines above, warning("function FOO was called"); #if 0 // real function body ... #endif
[22:46] <t0by> But check with him
[22:47] <t0by> I think something quicker and dirtier is in order.
[22:47] <Simei> ok, so there are like "low fuctions"
[22:47] <Simei> we don't care
[22:48] <t0by> sound_openal.cpp seems like an OpenAL-specific implementation of an interface, so low-level by definition, but my point is that... we are just trying to make it compile, not to make actual sense.
[22:48] <t0by> We want it to compile, not compile-and-not-crash-horribly
[22:49] <Simei> Ok, I understand
[22:49] <t0by> That's how I interpret it, at least.
[22:49] <t0by> _sev, agree?
[22:50] <Simei> _sev, t0by: what about the ALL_ALLOW or not thing
[22:50] <t0by> What's with ALL_ALLOW?
[22:50] <Simei> _sev, t0by: this time, we allow things like fopen or not?
[22:52] <t0by> Uh, I think we decided to put those inside an ifdef as well?
[22:52] <Simei> You were talking about it. I tried to read a bit the history, but what's the conclusion?
[22:52] <Simei> Ok, I get it
[22:52] <t0by> Yes, I'm not super sure myself.
[22:53] <t0by> Simei: just in case put those inside #ifdef ALLOW_FOPEN or something like that
[22:53] <t0by> instead of #ifdef 0
[22:53] <t0by> (same effect until you actually define an ALLOW_FOPEN)
[22:54] <Simei> Ok, that seems reasonable
[22:54] <t0by> But don't overthink it.
[22:55] <t0by> do this quick & dirty, let's have a compiling commit, then let's have some actual fun.
[22:55] <t0by> remember that as _sev said you can actually comment out whole modules from modules.mk and stub out the calls in the client code.
[22:56] <t0by> (10 bucks says that if you do this with enough modules you'll be left with just the interpreter and builtins.cpp with everything stubbed, since basically that's the entry point for everything)
[22:57] <-- tsoliman left irc: Ping timeout: 268 seconds
[23:03] <-- waltervn left irc: Quit: Leaving
[23:04] <Simei> Yes, that makes sense
[23:09] --> Strangerke_ joined #scummvm.
[23:11] <-- Strangerke left irc: Ping timeout: 245 seconds
[23:11] Nick change: Strangerke_ -> Strangerke
[23:19] --> tsoliman joined #scummvm.
[23:19] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[23:25] <snover> TMM: did you fix it yet? ;)
[23:25] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:26] <t0by> I'm crawling to bed
[23:26] <t0by> nite
[23:38] <TMM> snover, I attempted to go to sleep
[23:38] <TMM> snover, I seem to have failed, so maybe I'll look at it some more now :P
[23:38] <snover> then what happened?
[23:38] <t0by> So did I
[23:38] <t0by> I think we have broken PM
[23:38] <t0by> TMM: what firmware are you running?
[23:39] <TMM> t0by, some beta I got off of usenet a couple years ago
[23:39] <t0by> Oh, that must be it then.
[23:40] <TMM> It was SUPPOSED to have all the latest features
[23:40] <TMM> but it turned out there was no upgrade path to the final release and I can't seem to downgrade
[23:40] <TMM> I've been in a GPL dispute with the intelligent designer ;)
[23:43] <-- exmensa_ left irc:
[23:44] <t0by> Let me try once more
[23:44] <TMM> I'm applying the 'do something else for a while' workaround
[23:44] <t0by> Good luck
[23:45] <t0by> I'll try the all-powerful incantation:
[23:46] <t0by> yo, nite folks
[23:46] <-- t0by left irc: Quit: t0by
[23:47] --> exmensa joined #scummvm.
[23:58] --> Stormkpr joined #scummvm.
[23:59] <-- Stormkeeper left irc: Ping timeout: 268 seconds
[00:00] --- Fri May 26 2017