ScummVM IRC Log - 13 Aug 2016

[Back to Index]

[00:00] <dreammaster> About the warning disabling? I saw it in the logs.
[00:01] <dreammaster> Since I do happen to have MSVC, I can add it in and test it :)
[00:07] <-- jamm left irc: Read error: Connection reset by peer
[00:07] --> GitHub53 joined #scummvm.
[00:07] <GitHub53> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6uqg
[00:07] <GitHub53> scummvm/master bc9abfa Paul Gilbert: TITANIC: Implemented CArboretumGate class
[00:07] GitHub53 (GitHub53@192.30.252.40) left #scummvm.
[00:09] <Trixar_za> You know, I'm watching these Al Lowe interviews that m_kiewitz posted one link to and it's interesting as hell.
[00:11] <snover> dreammaster: thanks, let me know how it goes
[00:18] --> GitHub52 joined #scummvm.
[00:18] <GitHub52> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6um2
[00:18] <GitHub52> scummvm/master add058a Paul Gilbert: CREATE_PROJECT: Add warning C4373 to ignore list for SCI engine
[00:18] GitHub52 (GitHub52@192.30.252.40) left #scummvm.
[00:18] <dreammaster> Looks good
[00:21] --> jamm joined #scummvm.
[00:22] <dreammaster> Actually, if there's no objection, I'd like to add C4125 to the list as well. Compiling the engine gives a bunch of spurious decimal digit terminates octal escape sequence warnings
[00:27] <dreammaster> Hmmm... actually, I take that back. It may be somewhat legitimate. GUIO_GAMEOPTIONS9 at least is non-valid octal, since the digits should only be 0-7.
[00:28] <dreammaster> Should be be defining it as "\060" instead of "\058"?
[00:40] <snover> i do believe that would be the next octal
[00:42] <snover> any particular reason you wanted to limit C4373 to SCI engine?
[00:42] <snover> dreammaster: ^
[00:43] <dreammaster> Since it doesn't seem like I'm missing the obvious, I'll go ahead and change the option.
[00:43] <dreammaster> Not particularly; just a general aversion to completely disabling warnings for the entire solution without reason. Since sci is the only one exhibiting the problem, I figure better to simply constrain the disablement to it for now.
[00:44] <snover> nobody else wants to be a const-loving rebel? :)
[00:44] <dreammaster> Rebel without a const.
[00:45] <snover> it seems unfortunate to me that there is not a way of opting translation units into immutable-by-default mode
[00:45] <snover> #pragma rust
[00:46] --> GitHub19 joined #scummvm.
[00:46] <GitHub19> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6uYF
[00:46] <GitHub19> scummvm/master 9e0a6cd Paul Gilbert: COMMON: Fix octal for GUIO_GAMEOPTIONS9
[00:46] GitHub19 (GitHub19@192.30.252.46) left #scummvm.
[00:48] <snover> also, why are these even octal to begin with
[00:48] <dreammaster> That in itself is a very good question
[00:50] --> Strangerke_ joined #scummvm.
[00:52] <snover> embarrassingly, i think i was the one that dumbed that line up, so thanks for fixing it for me
[00:52] <snover> im only slightly better at writing code than a monkey
[00:52] <-- Strangerke left irc: Ping timeout: 240 seconds
[00:52] Nick change: Strangerke_ -> Strangerke
[00:53] <dreammaster> Speaking of create_project issues in general; seems like it's been a while since I rebuilt the create_project executable. Anyone know/remember why LordHoto changed libpng.lib to libpng16.lib for Visual Studio? It was a comm on Mar 9.. 5510fdc7e47d0b2a1bd734bef812cdf819a2a4c2
[01:17] <dreammaster> Guess it's not important. Replacing my existing libs folder with a fresh one from wiki.scummvm.org works
[01:25] <-- Dominus left irc: Ping timeout: 244 seconds
[01:25] --> Dominus joined #scummvm.
[01:25] <-- Dominus left irc: Changing host
[01:25] --> Dominus joined #scummvm.
[01:30] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[02:14] --> Littleboy joined #scummvm.
[02:14] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[02:19] <-- jamm left irc: Read error: Connection reset by peer
[02:19] --> jamm joined #scummvm.
[02:20] <-- snover left irc: Quit: Leaving.
[02:26] <-- jamm left irc: Read error: Connection reset by peer
[02:27] --> jamm joined #scummvm.
[02:43] --> GitHub136 joined #scummvm.
[02:43] <GitHub136> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6uZ1
[02:43] <GitHub136> scummvm/master 4f8c31e Paul Gilbert: TITANIC: Implemented CAnnounce and CActButton classes
[02:43] GitHub136 (GitHub136@192.30.252.42) left #scummvm.
[02:48] <-- jamm left irc: Read error: Connection reset by peer
[02:49] --> jamm joined #scummvm.
[03:53] --> GitHub74 joined #scummvm.
[03:53] <GitHub74> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6uch
[03:53] <GitHub74> scummvm/master 6a9923e Paul Gilbert: TITANIC: Implemented some game object classes
[03:53] GitHub74 (GitHub74@192.30.252.41) left #scummvm.
[03:56] <-- jamm left irc: Read error: Connection reset by peer
[03:57] --> jamm joined #scummvm.
[04:02] --> Poly-C joined #scummvm.
[04:04] <-- Polynomial-C left irc: Ping timeout: 250 seconds
[04:27] <-- dreammaster left irc:
[06:01] <-- jamm left irc: Read error: Connection reset by peer
[06:01] --> jamm joined #scummvm.
[06:02] --> ny00123 joined #scummvm.
[06:09] --> rootfather joined #scummvm.
[06:12] <-- rootfather left irc: Changing host
[06:12] --> rootfather joined #scummvm.
[06:12] #scummvm: mode change '+o rootfather' by ChanServ!ChanServ@services.
[06:12] <-- rootfather left irc: Changing host
[06:12] --> rootfather joined #scummvm.
[06:12] #scummvm: mode change '+o rootfather' by herbert.freenode.net
[06:59] <-- rootfather left irc: Quit: Page closed
[07:27] --> t0by joined #scummvm.
[07:27] #scummvm: mode change '+v t0by' by ChanServ!ChanServ@services.
[07:45] --> waltervn joined #scummvm.
[07:45] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[07:45] <waltervn> morning
[07:47] <-- Littleboy left irc: Read error: Connection reset by peer
[08:21] --> jammm joined #scummvm.
[08:21] <-- jamm left irc: Read error: Connection reset by peer
[08:30] --> m_kiewitz joined #scummvm.
[08:30] <-- m_kiewitz left irc: Changing host
[08:30] --> m_kiewitz joined #scummvm.
[08:30] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[08:35] <m_kiewitz> haha I can't believe it. Amazon France had some pricing error yesterday for the James Bond Collection and sold them for 44 EUR
[08:36] <m_kiewitz> Yesterday I ordered it, then it was cancelled. Then I ordered it a second time, was cancelled again. But amazon still kept offering it at that price point.
[08:36] <m_kiewitz> Then I tried a third time and today I got a shipping e-mail :p :p :p
[08:43] <waltervn> We should ask the person of that SQ5 bug report what backend he's using
[08:46] <waltervn> I only tried SDL1 though, m_kiewitz are you on SDL2?
[08:46] <m_kiewitz> I'm using SDL1
[08:47] <m_kiewitz> very interesting - Space Quest game commentaries with some of the Space Quest creators https://www.youtube.com/watch?v=qjDQ7Nrse4E
[08:55] <m_kiewitz> waltervn: could the error actually happen only on another backend? I seriously doubt that
[08:56] <m_kiewitz> well the mouse click may not come through correctly of course
[08:56] <waltervn> it might if for some reason the mouse won't reach x=319 in full screen mode
[08:56] <m_kiewitz> yes
[08:57] <waltervn> m_kiewitz: any idea how to convince scummvm to use SDL2?
[08:57] <m_kiewitz> for windows you need to specify to create SDL2 project files
[08:57] <m_kiewitz> that worked for me, but I wanted to go back to SDL
[08:57] <m_kiewitz> 1
[08:57] <waltervn> I'm using msys2
[09:00] <m_kiewitz> i tried full screen mode and it works for me as well
[09:08] <waltervn> I'll check SDL2 just to make sure..
[09:09] <-- Lightkey left irc: Ping timeout: 258 seconds
[09:11] <waltervn> SDL2 works fine for me as well
[09:19] --> jammmm joined #scummvm.
[09:19] <-- jammm left irc: Read error: Connection reset by peer
[09:21] --> Lightkey joined #scummvm.
[09:33] <-- Cruel` left irc: Quit: Leaving
[09:51] --> frankyboy_ joined #scummvm.
[09:53] <-- jammmm left irc: Quit: Leaving
[09:54] --> jamm joined #scummvm.
[09:57] --> GitHub9 joined #scummvm.
[09:57] <GitHub9> [scummvm] wjp pushed 2 new commits to master: https://git.io/v6uu7
[09:57] <GitHub9> scummvm/master 4d12080 Ori Avtalion: ALL: Don't use 'defined' in macro definitions...
[09:57] <GitHub9> scummvm/master 68a15c1 Willem Jan Palenstijn: Merge pull request #790 from salty-horse/clang_fixes...
[09:57] GitHub9 (GitHub9@192.30.252.40) left #scummvm.
[09:57] --> GitHub74 joined #scummvm.
[09:57] <GitHub74> [scummvm] wjp closed pull request #790: ALL: Don't use 'defined' in macro definitions (master...clang_fixes) https://git.io/vKyRF
[09:57] GitHub74 (GitHub74@192.30.252.41) left #scummvm.
[09:58] --> ajax16384 joined #scummvm.
[09:58] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[09:59] <wjp> so let's see if buildbot likes that patch :-)
[10:00] <m_kiewitz> waltervn: https://sourceforge.net/p/scummvm/bugs/7181/
[10:00] <m_kiewitz> it seems 3x is broken in full screen?!?!
[10:01] <waltervn> maybe because of rounding the x coordinate?
[10:01] <m_kiewitz> yeah 3x is in fact broken, mouse is broken
[10:01] <wjp> interesting
[10:01] <m_kiewitz> seems to be a system issue, not SCI issue
[10:02] <waltervn> sorry, I hadn't read the new post yet
[10:04] <waltervn> it seems to work for me in SDL2
[10:05] <m_kiewitz> o_O
[10:05] <m_kiewitz> how is that possible? So actually SDL1 breaks it?
[10:05] <waltervn> that has a different full screen implementation than SDL1 does though
[10:05] <waltervn> let me go back to SDL1
[10:07] <m_kiewitz> 1:29:15 https://www.youtube.com/watch?v=qjDQ7Nrse4E lol, Space Quest Inside Copy + commentary
[10:08] <waltervn> yes, I see the problem in SDL1
[10:09] <m_kiewitz> snover: ^^ you should watch this lol
[10:09] <m_kiewitz> waltervn: yikes, so is it our fault or SDL1's fault?
[10:09] <waltervn> no idea, but SDL1 fullscreen has other problems for me as well, since it changes the resolution
[10:10] <waltervn> when I quit scummvm, my IRC window has been resized etc...
[10:10] <waltervn> not fun
[10:13] <m_kiewitz> lol
[10:18] --> jammm joined #scummvm.
[10:18] <-- jamm left irc: Read error: Connection reset by peer
[10:19] <-- jammm left irc: Client Quit
[10:20] --> jamm joined #scummvm.
[10:41] --> GitHub25 joined #scummvm.
[10:41] <GitHub25> [scummvm] sev- pushed 2 new commits to master: https://git.io/v6uz5
[10:41] <GitHub25> scummvm/master fb5c38c Eugene Sandulenko: DIRECTOR: Lingo: Implement charToNum, added stubs for constrainH/V functions
[10:41] <GitHub25> scummvm/master 963826c Eugene Sandulenko: DIRECTOR: Lingo: Simplified Datum constants pushing
[10:41] GitHub25 (GitHub25@192.30.252.41) left #scummvm.
[10:51] --> jammm joined #scummvm.
[10:51] <-- jamm left irc: Read error: Connection reset by peer
[10:51] --> GitHub188 joined #scummvm.
[10:51] <GitHub188> [scummvm] sev- pushed 1 new commit to master: https://git.io/v6ugO
[10:51] <GitHub188> scummvm/master dc2a456 Eugene Sandulenko: DIRECTOR: Lingo: Implemented integerp function and added stubs for 'the free*'
[10:51] GitHub188 (GitHub188@192.30.252.34) left #scummvm.
[10:56] --> Henke37 joined #scummvm.
[11:04] <m_kiewitz> ah info on freddy pharkas dancing balls lol, around 15:00 https://www.youtube.com/watch?v=vHaVrLJhRME
[11:06] --> WooShell joined #scummvm.
[11:06] <WooShell> good meowning =^.^=
[11:19] <m_kiewitz> wjp: any idea why we do all those weird cursor type thing detections?
[11:19] <m_kiewitz> for lsl5 it detects cursor type SCI0 early, which is why the pseudo mouse doesn't work
[11:20] <m_kiewitz> couldn't we simply force it for SCI1.1 and otherwise also check for PseudoMouse class?
[11:36] <-- frankyboy_ left irc: Remote host closed the connection
[11:45] <m_kiewitz> wjp/waltervn: it seems even Leisure Suit Larry 1 demo (almost the first SCI1 game) sets the event type differently than qfg2 demo (which will then cause keydown + direction bits to be set)
[11:48] <-- jammm left irc: Quit: Leaving
[11:48] --> jamm joined #scummvm.
[11:51] --> criezy joined #scummvm.
[11:51] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[11:51] <m_kiewitz> anyone owns King's Quest 5 floppy 0.000.055 / 0.000.051?
[12:07] --> jammm joined #scummvm.
[12:07] <-- jamm left irc: Ping timeout: 265 seconds
[12:27] <-- t0by left irc: Remote host closed the connection
[12:27] --> t0by joined #scummvm.
[12:27] #scummvm: mode change '+v t0by' by ChanServ!ChanServ@services.
[12:41] --> GitHub67 joined #scummvm.
[12:41] <GitHub67> [scummvm] sev- pushed 1 new commit to master: https://git.io/v6uVt
[12:41] <GitHub67> scummvm/master 409772e Eugene Sandulenko: DIRECTOR: Lingo: Added stub for 'label' function
[12:41] GitHub67 (GitHub67@192.30.252.46) left #scummvm.
[12:43] --> jamm joined #scummvm.
[12:44] <-- jammm left irc: Ping timeout: 265 seconds
[12:58] --> GitHub67 joined #scummvm.
[12:58] <GitHub67> [scummvm] m-kiewitz pushed 1 new commit to master: https://git.io/v6uVi
[12:58] <GitHub67> scummvm/master c270b30 Martin Kiewitz: SCI: Fix pseudo mouse in various SCI1 games like e.g. Larry5...
[12:58] GitHub67 (GitHub67@192.30.252.45) left #scummvm.
[12:58] <m_kiewitz> waltervn: i hope this does it. Someone will have to check SCI32, but I think until now we also set it for SCI32.
[13:07] <m_kiewitz> _sev: can we disable the widget <> warning? I'm getting it almost all the time
[13:08] <m_kiewitz> e.g. WARNING: Widget <> has clipping area y < 0 (-14)!
[13:15] --> dreammaster joined #scummvm.
[13:15] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[13:16] <m_kiewitz> damn, youtube.com is down for me :(
[13:16] <dreammaster> Oh my god. Where will you get your funny cat videos now? ;)
[13:17] <m_kiewitz> i was listening to funny Space Quest commentaries :(
[13:17] <dreammaster> Ah. That's somewhat more important, indeed
[13:22] --> snover joined #scummvm.
[13:22] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[13:23] <m_kiewitz> snover: can you please check what kMapKeyToDir does in sci32?
[13:23] <snover> sure
[13:23] <m_kiewitz> if it also adds bitflag "direction" (40h) to event type or if it sets that explicitly
[13:31] <snover> the type appears to be explicitly set to 10h
[13:32] <snover> not a flag
[13:32] <m_kiewitz> o_O 10h
[13:33] <m_kiewitz> 10h was menustart in sci16
[13:35] <snover> thats gone as far as i can tell
[13:36] <snover> menus are built as sci scripts now
[13:36] <m_kiewitz> this has to get checked in depth, was there keyboard control available for the mouse cursor in some games?
[13:40] <snover> i dont *think* so, but i am not sure
[13:41] <m_kiewitz> gk1 didn't, I just checked
[13:52] --> GitHub32 joined #scummvm.
[13:52] <GitHub32> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6urd
[13:52] <GitHub32> scummvm/master 5c64e27 Paul Gilbert: TITANIC: Implemented several NPC related game classes
[13:52] GitHub32 (GitHub32@192.30.252.45) left #scummvm.
[14:01] <-- Deledrius left irc: Quit: App.Exit
[14:02] --> Deledrius joined #scummvm.
[14:05] --> jammm joined #scummvm.
[14:05] <-- jamm left irc: Read error: Connection reset by peer
[14:12] <_sev> m_kiewitz: sure, I can do it
[14:12] --> GitHub127 joined #scummvm.
[14:12] <GitHub127> [scummvm] sev- pushed 2 new commits to master: https://git.io/v6uoi
[14:12] <GitHub127> scummvm/master 02f0692 Retro-Junk: FULLPIPE: Use proper Surface coordinates
[14:12] <GitHub127> scummvm/master b4f418a Retro-Junk: FULLPIPE: Fix wrong member name in InteractionController::handleInteraction
[14:12] GitHub127 (GitHub127@192.30.252.40) left #scummvm.
[14:27] <-- criezy left irc: Quit: criezy
[14:34] <-- jammm left irc: Quit: Leaving
[14:35] --> jamm joined #scummvm.
[14:38] --> Vampire0 joined #scummvm.
[14:41] <waltervn> m_kiewitz: seems to be working now, thanks for looking into this!
[14:44] <waltervn> I'm getting some "casts away qualifiers" warnings in sci/graphics/video32.cpp
[14:45] --> rootfather joined #scummvm.
[14:45] <-- rootfather left irc: Changing host
[14:45] --> rootfather joined #scummvm.
[14:45] #scummvm: mode change '+o rootfather' by ChanServ!ChanServ@services.
[14:45] <-- rootfather left irc: Changing host
[14:45] --> rootfather joined #scummvm.
[14:45] #scummvm: mode change '+o rootfather' by herbert.freenode.net
[14:48] <snover> hey sci folks, does this seem reasonable? am i missing anything obvious? https://gist.github.com/csnover/5443479b469948f52c4b2d0f2e3757d0
[14:56] <snover> (cc wjp :))
[14:58] <-- jamm left irc: Read error: Connection reset by peer
[14:59] --> jamm joined #scummvm.
[15:00] <-- m_kiewitz left irc: Ping timeout: 244 seconds
[15:01] <-- Vampire0 left irc: Quit: Verlassend
[15:19] --> Vampire0 joined #scummvm.
[15:37] <snover> hum. this is going to invalidate save games.
[15:40] <-- waltervn left irc: Read error: Connection reset by peer
[15:41] <-- ajax16384 left irc: Read error: Connection reset by peer
[15:43] <wjp> is it?
[15:47] <snover> im trying to make it not so, SegmentObjTableEntrySyncer needs to do more than it is doing in that patch
[15:48] <snover> for some reason when i try loading an earlier save game it spins for a while trying to reconstruct the stack and then fails
[15:54] --> waltervn joined #scummvm.
[15:54] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[15:55] <-- dreammaster left irc:
[15:58] <-- ny00123 left irc: Quit: Leaving
[15:58] --> ny00123 joined #scummvm.
[15:59] --> m_kiewitz joined #scummvm.
[15:59] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[16:02] <-- waltervn left irc: Quit: Leaving
[16:11] <wjp> ah, yes, Syncer would indeed need work
[16:29] --> GitHub97 joined #scummvm.
[16:29] <GitHub97> [scummvm] sev- pushed 1 new commit to master: https://git.io/v6u1l
[16:29] <GitHub97> scummvm/master 2092422 Retro-Junk: FULLPIPE: Fix incorrect array element indirection in MGM::recalcOffsets
[16:29] GitHub97 (GitHub97@192.30.252.45) left #scummvm.
[16:30] <snover> i dont understand what is going on with this crazy thing
[16:43] <snover> doesnt even *get* to SegmentObjTableEntrySyncer
[16:45] --> GitHub3 joined #scummvm.
[16:45] <GitHub3> [scummvm] sev- closed pull request #803: SCUMM: Detection for dutch airport demo (master...detection_fixes) https://git.io/v6n56
[16:45] GitHub3 (GitHub3@192.30.252.46) left #scummvm.
[16:45] --> GitHub42 joined #scummvm.
[16:45] <GitHub42> [scummvm] sev- pushed 2 new commits to master: https://git.io/v6uMc
[16:45] <GitHub42> scummvm/master b9561b4 Ben Castricum: SCUMM: Add dutch airport demo
[16:45] <GitHub42> scummvm/master 6193f40 Eugene Sandulenko: Merge pull request #803 from BenCastricum/detection_fixes...
[16:45] GitHub42 (GitHub42@192.30.252.34) left #scummvm.
[16:45] <snover> hm. with this other save i have, it does, but then somehow it thinks there is an 8MB string to unserialise
[16:47] --> GitHub102 joined #scummvm.
[16:47] <GitHub102> [scummvm] sev- closed pull request #800: Add snapcraft.yaml so you can `snapcraft build` (master...master) https://git.io/v6UXH
[16:47] GitHub102 (GitHub102@192.30.252.46) left #scummvm.
[16:47] --> GitHub56 joined #scummvm.
[16:47] <GitHub56> [scummvm] sev- pushed 1 new commit to master: https://git.io/v6uM8
[16:47] <GitHub56> scummvm/master fe256bb spads-spads: BUILD: Add snapcraft.yaml so you can `snapcraft build` (#800)
[16:47] GitHub56 (GitHub56@192.30.252.40) left #scummvm.
[16:48] --> Strangerke_ joined #scummvm.
[16:49] --> Littleboy joined #scummvm.
[16:49] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[16:51] <-- Strangerke left irc: Ping timeout: 264 seconds
[16:51] Nick change: Strangerke_ -> Strangerke
[16:57] --> Cruel` joined #scummvm.
[17:18] <-- Axy left irc: Ping timeout: 244 seconds
[17:22] <snover> wjp, or m_kiewitz, or anyone: https://gist.github.com/csnover/5443479b469948f52c4b2d0f2e3757d0 do you see something in here that would break the old saves?
[17:22] <m_kiewitz> just try to restore old saves, then create a new one and try to load it with the old code
[17:22] <snover> well, that is the problem& restoring an old save doesnt work
[17:23] <snover> it spins for a bit
[17:23] <snover> then there is an assertion failure
[17:24] <snover> it seems fairly obvious that something is under-reading or over-reading data from the save game, causing garbage values to be read later on
[17:24] <snover> later on in the restore process*
[17:26] <snover> since it gobbles up a few hundred megs of memory and thinks it is restoring 8MB long strings
[17:31] <wjp> mysterious
[17:31] <wjp> your old save doesn't also accidentally have version 37?
[17:32] <wjp> (but from a previous commit that also increased the version or something)
[17:33] <snover> nope
[17:33] <snover> checked that
[17:34] <snover> well, and even tried the save game from https://sourceforge.net/p/scummvm/bugs/7181
[17:49] <snover> this is so annoying. i really want to fix this today because games are super-crashy now that kListEachElementDo is working properly. I could just re-dereference the list after each loop but that doesnt solve this underlying problem.
[17:54] --> Mia joined #scummvm.
[18:01] <m_kiewitz> nooooo R2D2 is dead :(
[18:02] <m_kiewitz> https://www.theguardian.com/film/2016/aug/13/kenny-baker-r2-d2-dies-star-wars
[18:04] <wjp> snover: on top of which branch are you applying that segment patch?
[18:04] <snover> my working branch
[18:04] <wjp> hm, doesn't apply here
[18:04] <snover> allow me to repush that.
[18:05] <snover> its already committed to the branch so you should be able to just pull and go
[18:06] <snover> hmm. hold on a second, something different is happening
[18:12] <snover> never mind. i still have no idea what is happening.
[18:17] <wjp> huh
[18:25] <m_kiewitz> snover: entry.data was SciArray before, correct?
[18:25] <snover> it was T, now it is T*
[18:27] <wjp> ah
[18:27] <wjp> wait
[18:28] <wjp> the bitmap table stored by version 36 is now being read as a normal array
[18:30] <wjp> while version 36 stored it in a custom way (that now matches the new version 37?)
[18:31] <m_kiewitz> snover: sci16 saves still work?
[18:32] <m_kiewitz> i think we can live with sci32 saves not working anymore
[18:32] <snover> ye, i am worried about the sci16 ones, they werent working either
[18:32] <snover> wjp: some part of my brain thinks it knows what you are talking about
[18:32] <wjp> however it presumably wouldn't explain SQ5 saves breaking?
[18:33] <m_kiewitz> yes, there has to be an issue with SegmentTable then
[18:35] <snover> wjp: ok. i got what you are saying about bitmap saving now.
[18:36] <-- jamm left irc: Ping timeout: 258 seconds
[18:37] --> jamm joined #scummvm.
[18:37] <snover> so i am not worried about version 36 saves, actually i will just not even increment the save versioning for now
[18:38] <m_kiewitz> agreed, even if you change the bitmap format, i wouldn't care about that, because sci32 is wip anyway
[18:38] <m_kiewitz> i can't figure out what's going wrong with the Segment table
[18:38] <wjp> void syncWithSerializer(Common::Serializer &s, SciBitmap *&obj) can go, by the way
[18:39] <snover> wjp: yeah i just deleted that :)
[18:39] <wjp> k
[18:39] <m_kiewitz> how did it work before? it saved every entry of the table separately?
[18:39] <m_kiewitz> always starting with "entry.next_free"?
[18:39] <m_kiewitz> or did entry.data point to the full table?
[18:40] <m_kiewitz> but the current code doesn't go through table entries
[18:40] <snover> entry.data pointed to whatever kind of object it is, Node, List, SciString, etc.
[18:40] <snover> well, pointed is wrong
[18:40] <m_kiewitz> so it saved:
[18:40] <m_kiewitz> entry.next_free
[18:40] <snover> entry.data *was* whatever object it is
[18:40] <m_kiewitz> entry.data (first entry)
[18:40] <m_kiewitz> entry.next_free
[18:40] <m_kiewitz> entry.data (second entry)
[18:40] <m_kiewitz> ?
[18:41] <m_kiewitz> yeah, but was entry.data the full table or just 1 single entry?
[18:41] <snover> just one object
[18:41] <m_kiewitz> so it saved next_free multiple times per entry?
[18:41] <snover> for each entry in the segment table
[18:41] <m_kiewitz> hmmm
[18:43] <snover> Entry objects are nodes in a linked list
[18:43] <m_kiewitz> and for empty entries it saved an empty entry?
[18:44] <snover> i believe it would have recorded whatever data had been associated with that Entry, regardless of whether or not the entry was in use.
[18:54] --> omer_mor joined #scummvm.
[18:57] <wjp> one problem is that the version < 37 restore code never sets .data == nullptr
[18:57] <wjp> while allocEntry() does assume that for unused entries
[18:58] <wjp> (the assert in there)
[19:00] <snover> do you end up most of the time crashing on the assert for reconstructStack?
[19:01] <wjp> on 'working', yes, but on adfd718fcae2da7734c7a6f8844d0ddf4a06e4fa, no
[19:01] --> dreammaster joined #scummvm.
[19:01] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[19:01] <snover> ok. yeah.
[19:02] <wjp> oh, a later commit seems to do a wrong version check on ResMan
[19:02] Action: wjp was looking at adfd718fcae2da7734c7a6f8844d0ddf4a06e4fa only
[19:03] <snover> yeah, definitely just look at that one only
[19:03] <snover> the multi-disc audio is not fixed. :)
[19:03] <wjp> so the crash in adfd718 is caused by this assert in allocEntry
[19:03] <wjp> which is caused by not clearing .data of free entries
[19:04] <wjp> the reconstructStack crash _seems_ to be the '< 36' check being wrong in ResourceManager::sync in the multi-disc commit
[19:06] <snover> Unknown SegmentObj type 855638016! mhmm.
[19:06] <omer_mor> Hi
[19:06] <snover> hi omer_mor
[19:06] <omer_mor> What do you think of http://wiki.scummvm.org/index.php/Quest_for_Glory#NEC_PC-9801_Version ?
[19:06] <omer_mor> I added it this week
[19:06] <omer_mor> Do you think scummvm should support these color modes?
[19:08] --> GitHub110 joined #scummvm.
[19:08] <GitHub110> [scummvm] sev- pushed 2 new commits to master: https://git.io/v6uHP
[19:08] <GitHub110> scummvm/master e585453 Eugene Sandulenko: FULLPIPE: Put behavior debug output to a separate channel
[19:08] <GitHub110> scummvm/master 0fdab36 Eugene Sandulenko: COMMON: Introduce --debug-channels-only command line flag....
[19:08] GitHub110 (GitHub110@192.30.252.46) left #scummvm.
[19:10] --> GitHub90 joined #scummvm.
[19:10] <GitHub90> [scummvm] dreammaster pushed 2 new commits to master: https://git.io/v6uHy
[19:10] <GitHub90> scummvm/master bbc9e07 Paul Gilbert: DEVTOOLS: Add Barbot's action frame ranges to create_titanic
[19:10] <GitHub90> scummvm/master 786f2ca Paul Gilbert: TITANIC: Implementing CBarbot game object class
[19:10] GitHub90 (GitHub90@192.30.252.46) left #scummvm.
[19:14] <wjp> You mean currently? I don't think so
[19:14] <wjp> Or are you asking if we'll accept it if you add it?
[19:14] <wjp> I'm not sure what you mean by "should", I suppose
[19:15] <omer_mor> I'm asking if you think scummvm should (eventually) support the pc-98 SCI graphics
[19:15] <omer_mor> the 8-color 640x480 dithered mode is unique
[19:15] <omer_mor> actually 640x400
[19:15] <wjp> is that a roundabout way of asking someone to write it?
[19:16] Action: t0by hides
[19:17] <omer_mor> wjp: No really. More of a roundabout way of raising awareness to this SCI oddity.
[19:20] <omer_mor> Actually when I look closely at the pixels, they are all double-height
[19:20] <omer_mor> So it's dithered to 640x200 8 color
[19:21] <omer_mor> The Japanese font however (not shown in the screenshot) is using the full resolution.
[19:24] <omer_mor> I also read that the PC-98 that supported 16 colors, could choose those colors out of 4096. The screenshot seem to support that. I wonder if the palette selection is encoded in the SCI resources anywhere.
[19:25] <omer_mor> How do Amiga / Atari-ST SCI0 games choose their 16 color palette?
[19:26] <omer_mor> Is it hardcoded in the interpreter or chosen by the game?
[19:27] <m_kiewitz> afaik hardcoded
[19:28] <m_kiewitz> and i doubt that sierra added some fancy color selection for pc9801
[19:36] <wjp> snover: any new insights?
[19:38] <snover> wjp: no i took a break, just getting back into it now
[19:39] <wjp> same here :-)
[19:40] <wjp> things seem ok here at adfd718 with the assert in allocEntry disabled, by the way
[19:40] <wjp> (other than the memleaks of course)
[19:41] --> frankyboy_ joined #scummvm.
[19:41] <snover> the complained-about entry memory is allocated by the SegmentObjTableEntrySyncer
[19:41] <snover> not sure how it then becomes a free slot for allocEntry to reuse
[19:42] <wjp> it would if it's in the free list
[19:42] <wjp> free-but-allocated slots get saved/loaded
[19:42] <snover> annoying.
[19:43] <wjp> yes
[19:43] <snover> shall we just delete and then reallocate the memory in allocEntry, or shall we walk the list in saveload and not allocate at all?
[19:44] <snover> i feel like the second one is a better idea, then the assert can remain and catch if any runtime stuff does bad stuff with memory
[19:44] <wjp> I think you have to allocate since you need to do the sync
[19:44] <snover> could call s.skip or whatever
[19:44] <wjp> although you can of course use a dummy for the sync
[19:44] <snover> yeah, or that
[19:44] <wjp> the annoying thing is that SegmentObjTableEntrySyncer doesn't know enough to decide if it's free
[19:46] <wjp> (since it's free if next_free == idx, but the EntrySyncer doesn't know idx)
[19:46] <wjp> but can easily enough add idx as an argument to DefaultSyncer and SegmentObjTableEntrySyncer I suppose
[19:47] <snover> hmm.
[19:48] <wjp> um, I meant, free if next_free != idx
[19:52] <snover> DefaultSyncer and SegmentObjTableEntrySyncer are only used for arrays, is this correct from what you see?
[19:54] <wjp> yes
[20:02] <-- Littleboy left irc: Read error: Connection reset by peer
[20:02] --> Littleboy joined #scummvm.
[20:02] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[20:10] <snover> hum.
[20:10] <snover> well that seems like it may work!
[20:12] <snover> sweet.
[20:14] <snover> and shaved a whole 3kB off the save ;)
[20:17] <wjp> what did you change in saving?
[20:17] <snover> free segment objects are omitted from the save
[20:18] <wjp> oh, you mean relative to current master, not relative to the version in adfd718
[20:18] <-- ny00123 left irc: Quit: Leaving
[20:19] <snover> i passed the index down to SegmentObjTableEntrySyncer and then `hasData = (entry.next_free == index)`
[20:19] <snover> for version 36 and earlier
[20:19] <snover> pull and show 5f487d79739f3e6283aeac9991b1ef43d62e5acd
[20:21] <wjp> nice
[20:26] <snover> now lets see how much memory is being leaked&
[20:29] <snover> seems like every time i run valgrind i get another uninitialised read in SciMusic&
[20:30] <snover> good reminder to add that MIDI work to the TODO list
[20:30] <snover> yay, no identified leaks.
[20:30] <snover> tentatively this seems ready to land. did you want me to do a PR so some more testing can be done first though?
[20:33] <snover> i suppose it would get more testing if it landed since it seems there are folks using master scummvm releases
[20:35] <wjp> I'm personally fine with this SegmentObjTable commit going in without a PR
[20:35] <wjp> self-contained, and already discussed and tested here now
[20:37] --> Axy joined #scummvm.
[20:39] <-- Mia left irc: Ping timeout: 258 seconds
[20:42] <snover> hold onto your butts.
[20:42] --> GitHub43 joined #scummvm.
[20:42] <GitHub43> [scummvm] csnover pushed 1 new commit to master: https://git.io/v6udS
[20:42] <GitHub43> scummvm/master 741ac22 Colin Snover: SCI: Fix pointer invalidation caused by array storage moves...
[20:42] GitHub43 (GitHub43@192.30.252.40) left #scummvm.
[20:43] <m_kiewitz> i thought you didn't want to increase the version?
[20:44] <snover> since it affects SCI16 save games it seemed prudent to bump the version for anyone running SCI16 on master
[20:45] <snover> if it was only a change to SCI32 games then i probably would have left it at 36
[20:45] <snover> otherwise it would have just invalidated everyones version 36 SCI16 saves
[20:45] <m_kiewitz> ah, so no way to keep the same saved game format...
[20:46] <wjp> no clean way at least
[20:46] <snover> right. on the upside, now your SCI saves will be up to 7% smaller!
[20:46] <snover> ;P
[20:48] <snover> phew, buildbot is running a little behind today
[20:48] <snover> so much new code! :)
[20:54] --> waltervn joined #scummvm.
[20:54] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[20:56] --> Strangerke_ joined #scummvm.
[20:57] <snover> actually, what the heck is going on with buildbot
[20:58] <snover> its been building one revision for two hours
[20:59] <-- Strangerke left irc: Ping timeout: 250 seconds
[20:59] Nick change: Strangerke_ -> Strangerke
[21:00] <m_kiewitz> im getting 1>d:\git-scummvm\engines\sci\engine\savegame.cpp(229) : warning C4018: '<' : signed/unsigned mismatch
[21:01] <snover> oops. yeah. sorry about that.
[21:01] <snover> seems i still have a couple of flags more that need to be added to create_project for xcode.
[21:02] <m_kiewitz> but that one actually is a correct warning, "i" should be uint, not int
[21:02] <snover> yes.
[21:05] <-- frankyboy_ left irc: Read error: Connection reset by peer
[21:06] --> frankyboy_ joined #scummvm.
[21:09] <snover> not that any of these things will get anywhere near the signed maximum.
[21:09] <snover> im fixing it.
[21:17] <m_kiewitz> well the saved game could also be corrupt, you never know
[21:17] <m_kiewitz> i think we even should check all of that code at some point for buffer overflows etc.
[21:17] <m_kiewitz> my new saved game code checks for quite a lot of things already
[21:18] <m_kiewitz> because well someone could offer saved games on the internet, that use such errors to execute virus/trojan code
[21:18] <m_kiewitz> the current code literally crashes on all sorts of issues, like for example even when just a script got patched in some way
[21:20] <snover> this code uses -1 for invalid index
[21:22] <m_kiewitz> argh
[21:23] <snover> you probably have a memory leak if you have 2 billion active memory handles.
[21:23] <snover> :)
[21:23] <-- frankyboy_ left irc: Read error: Connection reset by peer
[21:24] --> frankyboy_ joined #scummvm.
[21:25] --> GitHub35 joined #scummvm.
[21:25] <GitHub35> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6uNC
[21:25] <GitHub35> scummvm/master 3c4c605 Paul Gilbert: TITANIC: Added CBarbot CMovieEndMsg event handler
[21:25] GitHub35 (GitHub35@192.30.252.45) left #scummvm.
[21:29] --> GitHub96 joined #scummvm.
[21:29] <GitHub96> [scummvm] csnover pushed 1 new commit to master: https://git.io/v6uNu
[21:29] <GitHub96> scummvm/master 22c7be2 Colin Snover: SCI: Fix sign comparison warning
[21:29] GitHub96 (GitHub96@192.30.252.45) left #scummvm.
[21:29] <-- t0by left irc: Quit: Bye!
[21:31] --> Strangerke_ joined #scummvm.
[21:33] <-- Strangerke left irc: Ping timeout: 244 seconds
[21:33] Nick change: Strangerke_ -> Strangerke
[21:34] --> GitHub81 joined #scummvm.
[21:34] <GitHub81> [scummvm] csnover pushed 2 new commits to master: https://git.io/v6uN1
[21:34] <GitHub81> scummvm/master beba350 Colin Snover: SCI32: Fix cast warnings
[21:34] <GitHub81> scummvm/master 97b3437 Colin Snover: CREATE_PROJECT: Fix warning flags in Xcode generator...
[21:34] GitHub81 (GitHub81@192.30.252.41) left #scummvm.
[21:37] <snover> can buildbot use ccache? seems like that might be useful.
[21:38] <wjp> I think we should already be using that for a fair number of toolchains
[21:39] <wjp> if not all
[21:39] <snover> i just dont remember it taking so long to run builds as it is taking today
[21:40] <-- WooShell left irc: Quit: Zu gotdy od mpy nrmy stpimf. Zu drvpmf zrsmd aogy jrt iq pt viy jrt yp yjr htpimf.
[21:41] <wjp> _sev touched a few headers in common earlier
[21:41] <snover> i suppose that would do it.
[21:46] <snover> wjp: did you review 804? (if not, you dont have to now, just wondering if you had)
[21:48] --> Vampire0_ joined #scummvm.
[21:49] <wjp> hadn't yet
[21:49] <wjp> interesting diff; I guess the main thing code-wise is actually cursor32 support?
[21:50] <snover> yes
[21:50] <-- Vampire0 left irc: Ping timeout: 265 seconds
[21:53] <snover> could probably use better names for some of that, please do make suggestions if you can
[21:56] <-- jamm left irc: Ping timeout: 252 seconds
[21:58] --> jamm joined #scummvm.
[21:59] <m_kiewitz> snover: why are you changing code inside GfxScreen?
[22:00] <m_kiewitz> and why did you change SCI_SHAKE_DIRECTION_VERTICAL to something else? In that code there seems to be an error, because you don't check for vertical both times, but one time for vertical and the other time for horizontal
[22:03] <m_kiewitz> and why is there an include for the header of frameout.h now inside GfxScreen? That doesn't make sense to me
[22:07] <m_kiewitz> is it also necessary to render the cursor always by ourselves now? Because there are SCUMMVM system functions for that
[22:12] <m_kiewitz> it also seems that according to the code GK1 on Macintosh only supports hires mode and atm that is not implemented in the same manner
[22:12] <m_kiewitz> i don't own it, I guess clone2727 owns it.
[22:23] <-- frankyboy_ left irc: Remote host closed the connection
[22:25] <m_kiewitz> and why is an non existing object now ignored inside kSetNowSeen?
[22:26] <m_kiewitz> in sci32 that results in a fatal error
[22:29] --> GitHub54 joined #scummvm.
[22:29] <GitHub54> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/v6uxP
[22:29] <GitHub54> scummvm/master e5c100c Paul Gilbert: TITANIC: Implemented CBellBot class
[22:29] GitHub54 (GitHub54@192.30.252.41) left #scummvm.
[22:30] --> Strangerke_ joined #scummvm.
[22:33] <-- Strangerke left irc: Ping timeout: 260 seconds
[22:33] Nick change: Strangerke_ -> Strangerke
[22:44] <-- waltervn left irc: Read error: Connection reset by peer
[23:00] <snover> boy, lots of questions
[23:04] <snover> ugh, im being called away. ill answer as soon as i get back. feel free to comment on the code itself on github too :)
[23:11] <-- jamm left irc: Ping timeout: 250 seconds
[23:11] --> jamm joined #scummvm.
[23:17] <snover> > why are you changing code inside GfxScreen? >> getting rid of the SCI32 stuff, looks like I rushed it too much. oops. thats why i get code review! :)
[23:23] <snover> > is it also necessary to render the cursor always by ourselves now? Because there are SCUMMVM system functions for that >> id have to try again to be sure but i think there was some problem when trying to restrict the cursor, it would be rendered outside the restricted area and then bounced back in
[23:23] <snover> that is what my memory is telling me anyway
[23:23] <snover> > GK1 on Macintosh only supports hires mode and atm that is not implemented in the same manner >> unfortunately i do not have access to any mac versions of any of the games, so they probably dont work for a lot of different reasons
[23:26] <snover> > why is an non existing object now ignored inside kSetNowSeen? >> because thats how it works in SCI32? :) if the screenitem is not found it returns false, otherwise it sets the nsrect and returns true
[23:28] <snover> as far as why changing SCI_SHAKE_DIRECTION_VERTICAL, just getting rid of #defines where i can, since they dont interact with the type system
[23:28] <snover> and arent namespaced, etc etc.
[23:31] <snover> oh, the weirdo changes to GfxScreen came from before i decided to just go all the way and implement the remaining SCI32 features needed to separate it, so thats why that include got there.
[23:32] <snover> (not the kShakeHorizontal thing, that was just some brain damage :))
[23:44] <m_kiewitz> snover: GK1 should already work, it "worked" before. I can't test it of course
[23:45] <m_kiewitz> and sure, it would bounce back with ScummVM system, but then I guess ScummVM system should be improved
[23:46] <m_kiewitz> because now we are basically implementing stuff that already exists in ScummVM system and could be implemented in a better way on certain platforms for performance and that would be skipped by us
[23:47] <m_kiewitz> my sq6 disassembly does:
[23:47] <m_kiewitz> cseg01:00054899 call ScreenItemList__Search1
[23:47] <m_kiewitz> and then does this
[23:47] <m_kiewitz> cseg01:000548B7 push inst_MsgMgr
[23:47] <m_kiewitz> cseg01:000548BD call MsgMgr__Fatal
[23:47] <m_kiewitz> when it's not found
[23:48] <m_kiewitz> but maybe thats the wrong code, idk
[23:49] <m_kiewitz> oh right about GK1 Mac - as I said - clone2727 implemented it that way, so I guess Mac didn't support low-res, which would make sense.
[23:49] <m_kiewitz> i think it would make sense to just implement it the same way until we can actually test it again
[23:52] <m_kiewitz> it would be really weird in case the Mac version supported low-res + hires
[23:53] <m_kiewitz> ah, there it is - Mac seems to be hires only
[23:53] <m_kiewitz> https://www.youtube.com/watch?v=W49FgsXhE7Y
[00:00] --- Sun Aug 14 2016