ScummVM IRC Log - 05 Jun 2017

[Back to Index]

[00:01] <-- LePhilousophe left irc: Remote host closed the connection
[00:02] --> LePhilousophe joined #scummvm.
[00:18] --> GitHub180 joined #scummvm.
[00:18] <GitHub180> [scummvm] dreammaster pushed 2 new commits to master: https://git.io/vHaux
[00:18] <GitHub180> scummvm/master 032cd3f Paul Gilbert: TITANIC: Simplify DVector fn1
[00:18] <GitHub180> scummvm/master 39ffcf7 Paul Gilbert: TITANIC: Simplifications in lockMarker2
[00:18] GitHub180 (GitHub180@192.30.252.41) left #scummvm.
[00:22] <-- 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.
[01:04] <-- Strangerke left irc: Ping timeout: 255 seconds
[01:05] <-- SylvainTV left irc: Read error: Connection reset by peer
[01:11] <-- Begasus left irc: Quit: Ex-Chat
[01:13] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[01:44] --> DominusExult joined #scummvm.
[01:47] <-- Dominus left irc: Ping timeout: 260 seconds
[01:47] Nick change: DominusExult -> Dominus
[01:50] --> Joefish_ joined #scummvm.
[01:50] #scummvm: mode change '+v Joefish_' by ChanServ!ChanServ@services.
[01:54] <-- Joefish left irc: Ping timeout: 255 seconds
[03:12] <-- dreammaster left irc:
[04:00] --> exmensa joined #scummvm.
[04:16] <snover> oh, what the what. SSCI didnt even have a lock counter in SCI16?
[04:20] <snover> i just assumed since it was in SCI32 that it was in SCI16 too, but it appears to be not so
[04:21] <snover> i looked through the old freesci source and its been using a counter ever since the first implementation
[05:03] <logix> is that just a case of an additional opcode in freesci or plain luck that no sci game claims two locks before releasing them?
[05:04] <logix> i.e., all official games work but you can easily create one that works in the original interpreter but not in scummvm (or vice versa)
[05:05] <snover> the most likely outcome would be transient memory leaks that disappear when scummvm is restarted (by a game script expecting one unlock call will undo all locks)
[05:07] <snover> the alternative (making more calls to unlock resources than there were lock calls) i would expect to have appeared in a more explosive manner earlier than now
[05:07] <-- Lightkey left irc: Ping timeout: 255 seconds
[05:08] <snover> but its possible that such an error would be masked either by the LRU cache or just by dumb luck in an undetected use-after-free situation
[05:10] <snover> on the other hand its also possible that some games call lock N times + unlock N times, which could have caused a use-after-free in SSCI but is handled correctly by ScummVM because of the lock counter
[05:10] <snover> its all very nebulous.
[05:13] <logix> wait - what I meant up there was not a script claiming two locks and releasing two locks but a script claiming two locks and releasing only one (from freesci/scumm's POV)
[05:14] <snover> yeah. in that case there would still be 1 lock left on the resource, which means the memory used by that resource would never be reclaimed
[05:15] <snover> (where never is until scummvm is restarted since the lock state of resources is not saved)
[05:17] <snover> (and not saving the lock states is just another problem&)
[05:20] --> Lightkey joined #scummvm.
[05:20] <snover> i feel like i must be missing some key insight, since its difficult to imagine how all of this would not have blown up something at some point, unless the LRU cache has just done an amazing job of masking resource lock defects.
[05:22] <snover> i looked through open SCI tickets in the bug tracker and https://bugs.scummvm.org/ticket/7073 is the only one that raises an eyebrow, though i can just as easily believe that it is a defect in the MIDI code
[05:25] <snover> anyway. i am working on some solution for GK2 that is less obtrusive than changing all of the lock code.
[05:26] <snover> but maybe i will look tomorrow if there are any mystery worksforme tickets that sound like lock problems.
[05:40] <logix> snover: ah, ok, I assumed there might have been something along the lines of lock(); lock(); unlock(); and then in some other place "while existlock {};"
[05:41] <logix> s/assumed/thought/
[05:41] <logix> but "just" memory leaks isn't quite that bad (in terms of working games)
[06:22] --> Begasus joined #scummvm.
[06:27] --> Begas_VBox joined #scummvm.
[06:50] Lightkey (~Darklock@p200300764C7B171722CF30FFFE083718.dip0.t-ipconnect.de) got netsplit.
[06:52] FleetAdmiral[m] (fleetadmir@gateway/shell/matrix.org/x-jwzrhnoqnkkmarxv) got netsplit.
[06:52] harekiet[m] (harekietha@gateway/shell/matrix.org/x-qkzyzfwivmtpceud) got netsplit.
[06:52] Lightkey (~Darklock@p200300764C7B171722CF30FFFE083718.dip0.t-ipconnect.de) returned to #scummvm.
[06:52] <-- tsoliman left irc: Ping timeout: 259 seconds
[06:52] --> tsoliman joined #scummvm.
[06:52] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[06:55] harekiet[m] --> (harekietha@gateway/shell/matrix.org/x-hfhxgqavnyytqtfj) joined #scummvm.
[06:55] FleetAdmiral[m] --> (fleetadmir@gateway/shell/matrix.org/x-mzmbmepblzjorecr) joined #scummvm.
[07:07] <-- ced117 left irc: Ping timeout: 255 seconds
[07:08] --> ced117 joined #scummvm.
[07:09] --> t0by joined #scummvm.
[07:10] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[07:10] --> Mia joined #scummvm.
[07:11] <-- Mia left irc: Changing host
[07:11] --> Mia joined #scummvm.
[07:18] <t0by> o/
[07:34] <_sev> \o
[07:38] <cpasjuste> hi
[07:38] <Begasus> https://0x0.st/60s.png
[07:39] <Begasus> one of the guys on Haiku running Sam & Max :)
[07:42] <logix> scummvm used here / to play the sam and max game / cuz I like retro
[07:42] <t0by> lol
[07:54] --> m_kiewitz joined #scummvm.
[07:54] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[07:54] --> cpasjuste_ joined #scummvm.
[07:56] <-- cpasjuste_ left irc: Client Quit
[08:07] <TMM> well, I have convinced myself that the actual decoding of the pixeldata is all correct but that my page swapping is subtly wrong somewhere
[08:07] <TMM> so, that's fun
[08:11] <wjp> what do you mean by page swapping in this context?
[08:13] <logix> my guess is a double buffered graphics surface
[08:18] <TMM> yeah
[08:18] --> ajax16384 joined #scummvm.
[08:18] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[08:18] <TMM> it appears that I made some mistake in the opcode for copying from the previous frame
[08:18] <TMM> everything is in the right place, but that's not working properly it seems
[08:22] <-- LittleToonCat left irc: Remote host closed the connection
[08:53] <-- TMM left irc: Quit: Ex-Chat
[09:13] --> Strangerke joined #scummvm.
[09:13] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[09:14] <Strangerke> hi guys
[09:20] <criezy> Good morning
[09:37] --> WooShell joined #scummvm.
[10:11] --> rootfather joined #scummvm.
[10:11] #scummvm: mode change '+o rootfather' by ChanServ!ChanServ@services.
[10:26] <WooShell> good meowning =^.^=
[10:45] <-- t0by left irc: Quit: t0by
[11:19] <-- Mia left irc: Ping timeout: 260 seconds
[11:43] --> h00ligan joined #scummvm.
[11:43] #scummvm: mode change '+o h00ligan' by ChanServ!ChanServ@services.
[11:44] <-- ajax16384 left irc: Ping timeout: 240 seconds
[11:48] --> LyTchet joined #scummvm.
[11:49] <-- LyTchet left irc: Client Quit
[12:07] --> SylvainTV joined #scummvm.
[12:07] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[12:12] --> waltervn joined #scummvm.
[12:12] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[12:16] <-- waltervn left irc: Client Quit
[12:17] --> waltervn joined #scummvm.
[12:17] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[12:30] --> Strangerke_ joined #scummvm.
[12:33] <-- Strangerke left irc: Ping timeout: 268 seconds
[12:33] Nick change: Strangerke_ -> Strangerke
[12:40] --> GitHub93 joined #scummvm.
[12:40] <GitHub93> [scummvm] criezy opened pull request #959: GUI: Separate bevel and shadow effect when extending widget rect (master...shadowhunter) https://git.io/vHVks
[12:40] GitHub93 (GitHub93@192.30.252.42) left #scummvm.
[13:36] <-- antlarr left irc: Remote host closed the connection
[13:39] --> antlarr joined #scummvm.
[13:59] --> Littleboy joined #scummvm.
[13:59] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[14:13] <-- waltervn left irc: Quit: Leaving
[14:24] Nick change: Joefish_ -> Joefish
[14:31] <Joefish> hi. I'm trying to draw simply a line on the screen but it doesn't show https://github.com/Joefish/scummvm/blob/supernova/engines/supernova/supernova.cpp#L79 . Would be great if someone can point out my probably obvious mistake
[14:38] --> Polynomial-C joined #scummvm.
[14:39] <Strangerke> I don't think you initialized your palette
[14:40] <Strangerke> Hum...
[14:40] <Strangerke> I'm in the middle of the REing of an opcode, give me 5 minutes
[15:03] --> Henke37 joined #scummvm.
[15:04] <Joefish> Strangerke: you're right.. I assumed that Engine defines a default palette and didn't load my own. works now. thanks :)
[15:06] <Strangerke> rah, happy. https://pastebin.com/9bLBfLj3
[15:07] <Strangerke> so, well, sorry, it took me a bit more than 5 minutes O:^)
[15:49] <-- h00ligan left irc: Quit: Leaving
[16:00] <harekiet[m]> wooop
[16:01] <Harekiet> hey that does actually work, just not the other way around :)
[16:14] <Simei> Hi, everyone :)
[16:18] --> ajax16384 joined #scummvm.
[16:18] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[16:21] <wjp> shadowhunter criezy :-)
[16:21] <wjp> Hi Simei
[16:22] --> Mia joined #scummvm.
[16:22] <-- Mia left irc: Changing host
[16:22] --> Mia joined #scummvm.
[16:22] --> ny00123 joined #scummvm.
[16:24] <Simei> :)
[16:32] <snover> wjp: did you see my wall of text from yesterday and have any thoughts about where i might be going wrong with my expectation that the locking code should be causing more problems than it seems to be?
[16:33] <wjp> Which type/class of resource are you expecting trouble with?
[16:34] <wjp> Ideally (maybe) things wouldn't depend on scripts at all
[16:34] <snover> all of them
[16:34] <wjp> And I think e.g. views are always loaded on-demand anyway
[16:37] <wjp> I think I'd prefer script calls like kLock to be treated more as hints than requirements wherever possible
[16:37] <snover> in the SCI32 renderer, the views/pics are loaded on demand; in SCI16 they are loaded by e.g. GfxView::initData and stored for the duration of the GfxView instance
[16:38] <snover> (which looks like it could be a while because they go into GfxCache)
[16:40] --> LittleToonCat joined #scummvm.
[16:40] <wjp> (back in 30 mins or so)
[16:40] <snover> ok
[16:52] --> Axy joined #scummvm.
[16:52] <-- Axy left irc: Changing host
[16:52] --> Axy joined #scummvm.
[16:55] <-- Mia left irc: Ping timeout: 246 seconds
[16:59] --> t0by joined #scummvm.
[16:59] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[17:06] --> TMM joined #scummvm.
[17:06] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[17:07] Action: TMM throws some root beer at LeChuck
[17:08] <Strangerke> hey TMM :)
[17:09] <Strangerke> TMM: It's still a bit too large to be displayed easily, but I think I only have 2 opcodes left in GL1 :)
[17:13] --> Strangerke_ joined #scummvm.
[17:15] <-- Strangerke left irc: Ping timeout: 260 seconds
[17:15] Nick change: Strangerke_ -> Strangerke
[17:19] <TMM> Strangerke, sweet
[17:19] <TMM> Strangerke, I'm still struggling with this video format 0x10
[17:19] <TMM> Strangerke, it seems that everything is in the right place, but somehow copying from the previous frame seems broken
[17:22] <-- Strangerke left irc: Ping timeout: 246 seconds
[17:38] --> waltervn joined #scummvm.
[17:38] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[17:41] --> Strangerke joined #scummvm.
[17:43] --> girafe joined #scummvm.
[17:48] <TMM> Hmm, something just occured to me
[17:48] <TMM> I'm doing the RE of 0x10 from the 32bit player
[17:48] <TMM> but the 32bit player is actually not capable of displaying king200.mve either
[17:52] <TMM> disassembly of the relevant functions looks basically the same though
[17:52] <TMM> weird
[18:04] <-- Begasus left irc: Ping timeout: 255 seconds
[18:16] --> Begasus joined #scummvm.
[18:33] <-- t0by left irc: Read error: Connection reset by peer
[18:35] --> Strangerke_ joined #scummvm.
[18:36] --> frankyboy_ joined #scummvm.
[18:38] <-- Strangerke left irc: Ping timeout: 246 seconds
[18:38] Nick change: Strangerke_ -> Strangerke
[18:58] harekiet[m] (harekietha@gateway/shell/matrix.org/x-hfhxgqavnyytqtfj) left #scummvm ("User left").
[19:11] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[19:12] <Strangerke> TMM: Sorry, super bad connection
[19:12] <TMM> Strangerke, still stuck on something of that 0x10 video frame format
[19:13] <Strangerke> for the logo? :/
[19:13] <TMM> Strangerke, yeah
[19:13] <Strangerke> stupid logo :(
[19:28] --> omer_mor joined #scummvm.
[19:28] <-- omer_mor_ left irc: Ping timeout: 255 seconds
[19:45] <TMM> Strangerke, I think everything is mostly working correctly but I think I'm doing something wrong with the buffer flipping, or failing that, I'm misidentifying what I think is a compiler optimization for some logic I'm not implementing
[19:47] <Strangerke> :/
[19:47] <Strangerke> On my side, I just finished the last grouping in IDA. GL1 is now fully RE and readable in IDA. :P
[19:48] <TMM> niiiice
[19:48] <TMM> does it make sense?
[19:49] <-- rootfather left irc: Ping timeout: 240 seconds
[19:49] <Strangerke> well, you have access to the C file :)
[19:49] <Strangerke> but yes, clearly
[19:52] <-- peterkohaut left irc: Ping timeout: 260 seconds
[19:56] <Strangerke> I hesitate a bit to start GL2
[19:56] --> omer_mor_ joined #scummvm.
[19:57] <Strangerke> I'm not sure it's useful at this point
[19:57] <-- omer_mor left irc: Ping timeout: 255 seconds
[20:00] --> ec2-user2 joined #scummvm.
[20:01] Nick change: ec2-user2 -> peterkohaut
[20:02] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[20:03] <-- chadj left irc: Ping timeout: 268 seconds
[20:10] --> chadj joined #scummvm.
[20:14] <TMM> Strangerke, I think we should have everything now then to start the first part of the game minus the intro logo
[20:15] <Strangerke> So, would it be more productive to start an engine?
[20:15] <TMM> Strangerke, maybe, but I'm soooo cloooose :D
[20:15] <wanwan> so, it's all done now?
[20:16] <TMM> wanwan, well, we can play all the gameplay related videos, load the images, and we have debug symbols, and Strangerke implemented the first large hardcoded logic block
[20:16] <Strangerke> movie format almost done, 1/4 of the hardcoded logic REed
[20:16] <wanwan> cool
[20:17] <TMM> the engine engine outside of the hardcoded logic is super trivial, so it's looking quite good
[20:17] <Strangerke> the hardcoded logic is simple, but a bit long
[20:18] <Strangerke> 1300 lines for group 1... it's clearly not a disaster
[20:18] <TMM> that's not so bad
[20:18] <TMM> movie player code is ~700 lines
[20:23] <TMM> Strangerke wins :D
[20:31] <Strangerke> bah. hardcoded logic is simple, video player isn't
[20:33] --> Mia joined #scummvm.
[20:33] <-- Mia left irc: Changing host
[20:33] --> Mia joined #scummvm.
[20:35] <-- Axy left irc: Ping timeout: 255 seconds
[20:36] --> omer_mor joined #scummvm.
[20:39] <-- omer_mor_ left irc: Ping timeout: 255 seconds
[20:53] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[21:11] <-- ajax16384 left irc: Read error: Connection reset by peer
[21:16] --> t0by joined #scummvm.
[21:16] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[21:21] <TMM> I thought I had a breakthrough, but it turns out that the intro video in the demo is my already known format :P
[21:25] --> Begas_VBox joined #scummvm.
[21:25] <-- girafe left irc: Read error: Connection reset by peer
[21:25] <-- waltervn left irc: Quit: Leaving
[21:38] <-- Mia left irc: Ping timeout: 255 seconds
[21:50] <TMM> I don't get it, this loop seems to be implemented twice, but it jmps between the two implementations so it functions as just one loop
[21:50] <TMM> I.. think
[21:50] <TMM> why would the compiler do this
[21:51] <TMM> it's in both the 32bit as well as the 16bit version
[21:51] <TMM> so maybe it's not just a compiler quirk
[21:51] <TMM> I'm probably just missing something
[21:56] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[21:56] <wjp> want a second pair of eyes?
[22:00] <t0by> Wow, plastic surgery is really getting out of hand.
[22:01] <wjp> and that's just the start; they also come with a brain attached
[22:01] <t0by> Eeek, cyborgs are upon us.
[22:01] <t0by> Good night all :)
[22:01] <wjp> night
[22:05] <Lightkey> twitch.tv/gogcom The Last Express nau
[22:05] <-- t0by left irc: Quit: t0by
[22:11] <Lightkey> "so someone didn't want to pay a ticket, haha" "oh wait, that's me? oh crip crap"
[22:13] <-- 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.
[22:13] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[22:19] <-- Begasus left irc: Quit: Ex-Chat
[22:19] <-- frankyboy_ left irc: Remote host closed the connection
[22:19] <TMM> wjp, probably. I'll write the relevant bits of the loop in assembly. I'd appreciate you taking a look at it.
[22:30] <-- ny00123 left irc: Quit: Leaving
[23:40] --> dreammaster joined #scummvm.
[23:40] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[00:00] --- Tue Jun 6 2017