ScummVM IRC Log - 01 Jun 2017

[Back to Index]

[00:01] --> omer_mor joined #scummvm.
[00:03] <-- OmerMor left irc: Ping timeout: 240 seconds
[00:21] --> VonGuardian joined #scummvm.
[00:27] <-- criezy left irc: Ping timeout: 240 seconds
[00:30] --> criezy joined #scummvm.
[00:30] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[00:39] <-- Mia left irc: Read error: Connection reset by peer
[00:57] <-- SylvainTV left irc: Read error: Connection reset by peer
[01:12] <-- Vampire0 left irc: Ping timeout: 240 seconds
[01:17] --> Vampire0 joined #scummvm.
[01:39] --> DominusExult joined #scummvm.
[01:39] <-- Dominus left irc: Ping timeout: 260 seconds
[01:39] Nick change: DominusExult -> Dominus
[01:58] --> Joefish_ joined #scummvm.
[01:58] #scummvm: mode change '+v Joefish_' by ChanServ!ChanServ@services.
[01:58] <-- Joefish left irc: Ping timeout: 260 seconds
[02:00] <-- dreammaster left irc:
[03:07] --> GitHub149 joined #scummvm.
[03:07] <GitHub149> [scummvm] csnover pushed 1 new commit to master: https://git.io/vHR5C
[03:07] <GitHub149> scummvm/master 03ba3e4 Colin Snover: SCI32: Return CD speed in kPlatform...
[03:07] GitHub149 (GitHub149@192.30.252.45) left #scummvm.
[03:41] <-- VonGuardian left irc: Quit: Connection closed for inactivity
[03:44] <tsoliman> I just realized that savegames have a _saveDate that's a String with DD.MM.YYYY .. I was thinking of switching it to YYYY.MM.DD (to be less ambiguous for regions that don't use DD/MM like the US) .. is there any interest in this? What happens to old saves, should they parsed and flipped or left alone? Is any of this a good or bad idea?
[04:17] <tsoliman> actually this is much simpler than I thought
[04:19] <snover> tsoliman: the string isnt saved, it is created by call to SaveStateDescriptor::setSaveDate
[04:20] <tsoliman> yep .. that's what I found .. it parses the header and calls that
[04:21] <snover> of course my response to this is to use ICU, but i expect this will not get much traction since were still using legacy encoding for text :)
[04:24] <tsoliman> https://github.com/tsoliman/scummvm/commit/4e6babdba3ce7157c89ecac27ab867ae8ed58271
[04:24] <snover> if you want to flip the order you might as well go full ISO-8601 and change those dots to dashes
[04:25] <tsoliman> I really want to, honestly
[04:28] <snover> and while you are at it maybe you can fix the metadata text drawing off the bottom edge of the background rect in modern theme? :)
[04:32] <tsoliman> ?
[04:32] <tsoliman> I don't see it
[04:38] <snover> tsoliman: https://zetafleet.com/i/592f9a2e7e7c1.jpg no?
[04:38] <tsoliman> wow .. how did I not see that before?
[04:39] <tsoliman> it's as if there's a blank line above "Date: ___"
[04:40] <snover> Im also not really sure why date and time have separate lines
[04:40] <snover> Saved on: 1234-56-78 90:12
[04:53] <tsoliman> is 1234-56-78T90:12 too ISO?
[04:59] <snover> yes. this is supposed to be for humans, not machines.
[05:05] --> GitHub186 joined #scummvm.
[05:05] <GitHub186> [scummvm] tsoliman opened pull request #956: ENGINES: Display savestate date as YYYY-MM-DD (master...savestate-date-format1) https://git.io/vHRpG
[05:05] GitHub186 (GitHub186@192.30.252.42) left #scummvm.
[05:06] <tsoliman> I am honestly not sure if this should be tagged GUI or ENGINES
[05:07] <snover> GUI, of course, since it is not the one you picked :)
[05:07] <tsoliman> d'oh
[05:08] <snover> i am out of here. have a good night!
[05:09] <tsoliman> you too
[05:16] <-- Lightkey left irc: Ping timeout: 272 seconds
[05:22] <-- kurtwr2 left irc:
[05:25] --> kurtwr joined #scummvm.
[05:28] --> Lightkey joined #scummvm.
[05:54] --> waltervn joined #scummvm.
[05:54] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[06:00] <waltervn> morning
[06:15] <-- TAS_2012v left irc: Ping timeout: 260 seconds
[06:15] --> TAS-2012v joined #scummvm.
[06:15] Nick change: TAS-2012v -> TAS_2012v
[06:38] <cpasjuste> hi !
[06:41] <TMM> hi
[07:05] <-- Cheeseness left irc: Remote host closed the connection
[07:07] --> Cheeseness joined #scummvm.
[07:25] <-- TMM left irc: Quit: Ex-Chat
[07:30] --> TAS-2012v joined #scummvm.
[07:30] <-- TAS_2012v left irc: Ping timeout: 260 seconds
[07:30] Nick change: TAS-2012v -> TAS_2012v
[08:01] --> kurtwr2 joined #scummvm.
[08:03] <-- kurtwr left irc: Ping timeout: 240 seconds
[08:24] <-- LittleToonCat left irc: Remote host closed the connection
[08:33] --> _sev_ joined #scummvm.
[08:33] #scummvm: mode change '+o _sev_' by ChanServ!ChanServ@services.
[08:34] <-- _sev left irc: Ping timeout: 240 seconds
[08:36] <-- Harekiet left irc: Ping timeout: 245 seconds
[08:40] --> criezy|Work joined #scummvm.
[08:40] #scummvm: mode change '+o criezy|Work' by ChanServ!ChanServ@services.
[08:45] --> _sev|work joined #scummvm.
[08:45] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[08:48] --> Harekiet joined #scummvm.
[08:48] --> TMM joined #scummvm.
[08:48] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[09:12] <-- Strangerke_ left irc: Ping timeout: 260 seconds
[09:21] --> Strangerke_ joined #scummvm.
[10:42] --> Mia joined #scummvm.
[10:42] <-- Mia left irc: Changing host
[10:42] --> Mia joined #scummvm.
[10:47] <-- Strangerke|work left irc: Ping timeout: 260 seconds
[11:01] <-- TMM left irc: Quit: Ex-Chat
[11:52] <-- Tomaz^W left irc:
[12:00] --> Littleboy joined #scummvm.
[12:00] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[12:18] --> Strangerke|work joined #scummvm.
[12:31] <-- ST1 left irc: Ping timeout: 260 seconds
[12:41] <-- antlarr left irc: Remote host closed the connection
[12:41] --> antlarr joined #scummvm.
[13:55] --> TMM joined #scummvm.
[13:55] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[13:57] Nick change: Joefish_ -> Joefish
[14:16] --> rootfather joined #scummvm.
[14:27] --> ajax16384 joined #scummvm.
[14:27] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[14:46] <-- TMM left irc: Quit: Ex-Chat
[15:07] --> Henke37 joined #scummvm.
[15:13] --> LittleToonCat joined #scummvm.
[15:21] <Simei> Hi, everyone :)
[15:42] <Strangerke|work> G'day Simei
[15:43] <Simei> Thanks! You too. :)
[15:53] --> ny00123 joined #scummvm.
[15:59] <-- _sev|work left irc: Quit: This computer has gone to sleep
[16:03] --> m_kiewitz joined #scummvm.
[16:03] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[16:12] --> WooShell joined #scummvm.
[16:13] <WooShell> meow =^.^=
[16:14] <-- ajax16384 left irc: Read error: Connection reset by peer
[16:19] --> Farmboy0 joined #scummvm.
[16:19] <-- Farmboy0 left irc: Changing host
[16:19] --> Farmboy0 joined #scummvm.
[17:04] <-- criezy|Work left irc: Quit: Page closed
[17:17] --> girafe joined #scummvm.
[17:18] <Simei> _sev: Are there any examples of scaling and making mirror symmetry of surfaces in other engines?
[17:25] <-- Harekiet left irc: Ping timeout: 255 seconds
[17:30] <wjp> many engines do some kind of sprite mirroring, but I think most do that manually
[17:35] --> Harekiet joined #scummvm.
[17:35] <wjp> what are your requirements exactly for scaling/mirroring surfaces?
[17:35] <Simei> Ok, I think what they do here is just reverse the x coordinates of reading them from the texture.
[17:35] <Simei> _sev, wjp: Also how to we deal with shaders in scummvm? In sludge, every time something displayed, they pass "primary color" and "secondary color" to a shader.
[17:36] <wjp> you're going to have to determine what the scalers do, and reproduce the effect in software
[17:36] <wjp> shaders*
[17:37] <Simei> wjp, it's a function that has a bool "mirror" and a float "scale" that I want to make work
[17:37] <Simei> Ok, so the shaders are for later
[17:38] <Simei> I think generally the function recalculate its position and change its size then display it
[17:38] <wjp> you might start from TransparentSurface::scaleT
[17:38] <wjp> (in graphics/transparent_surface.*)
[17:38] <wjp> it does arbitrary scaling, and in theory also mirroring, but that isn't enabled
[17:39] <wjp> if you don't need an alpha channel, TransparentSurface will be very slow though
[17:39] <Simei> Ok, I will search about it. Thanks!
[17:48] --> SylvainTV joined #scummvm.
[17:48] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[17:56] --> ajax16384 joined #scummvm.
[17:56] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[18:03] <-- rootfather left irc: Quit: Heb' auf deine erste Scherbe, Hinter den Wolken wartet das Licht. Denn du brauchst deine Erde, doch die Erde braucht dich nicht.
[18:40] --> Strangerke joined #scummvm.
[18:42] <-- Strangerke_ left irc: Ping timeout: 260 seconds
[18:42] Nick change: Strangerke -> Strangerke_
[18:52] <m_kiewitz> wjp / snover: oh wow, I'm lucky today. Will get the Spanish version of Gabriel Knight 1 for 2.75 EUR only :P
[18:53] <m_kiewitz> and the Spanish version of Torin's Passage for 2 EUR or so. heh
[18:58] <m_kiewitz> :(
[19:13] <-- EmeraldM2 left irc: Read error: Connection reset by peer
[19:37] <wjp> nice :-)
[20:00] --> TMM joined #scummvm.
[20:00] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[20:06] <-- borosky left irc:
[20:08] <TMM> YOU'RE ALL SCUMM!
[20:08] <TMM> :D
[20:08] --> borosky joined #scummvm.
[20:38] <-- ajax16384 left irc: Read error: Connection reset by peer
[20:42] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[20:56] --> DJWillis joined #scummvm.
[20:56] #scummvm: mode change '+o DJWillis' by ChanServ!ChanServ@services.
[21:35] <snover> wjp, m_kiewitz: I have been struggling with what to do about resource locking and could use your thoughts. SSCI has two different mechanisms for locking: there is both a disposable flag set on memory handles in the memory manager, and a lock counter on memory handles. kLock sets/clears the disposable flag, it does not change the lock counter; getting a resource from the resource manager with lock=true changes the lock counter.
[21:35] <snover> Calling kUnload *seems to* (I am not 100% sure on this) dispose of memory regardless of whether it was locked or disposable. So I dont think that we can just use a lock counter because with kLock, it seems that some games like Lighthouse will call to unlock the same resource more than once, which right now ends up removing locks that were made by e.g. Audio32.
[21:35] <snover> But, Im uncertain if it is safe to change this. So if you guys can give your opinions about this situation, I would appreciate it. At the moment I am thinking that I should implement the flag lock and fix kLock to use that; the flag (which is part of the memory handle attributes) seems to be the only thing that is saved in GK2/SCI3 (and only for View/Pic/Wave/Sync/Audio resources); the lock counter seems to be ignored/reset.
[21:37] <snover> sorry about the wall 'o text.
[21:40] <snover> and i imagine you might need more background/context, so let me know about that, though i will be a bit slow to respond this evening (i have been meaning to say this for 2 or 3 days now, so i just finally had to send it even though i dont have free time to do anything else yet with it)
[21:40] <snover> (until tomorrow)
[21:40] <snover> (or after you are in bed)
[21:40] <wjp> can the GC play a role here?
[21:42] <wjp> but I'll be away until Sunday evening, so won't really be able to contribute meaningfully until then
[21:42] <snover> part of the problem is that the resource manager in scummvm does not go through segmanager, but in ssci it goes through the memorymgr (segmanager equivalent)
[21:43] <snover> the LRU in ResourceManager is what is currently responsible for clearing away unlocked resources when a new request to load a resource is made
[21:43] <wjp> I think the script-based locking causing potential problems in SCI16 was part of the reason for a GC (although that was back in the freesci days before I was involved, so not entirely sure of the motivations)
[21:46] <snover> so, the GC handles references to memory within the VM, but (a) resources are not guaranteed to actually be held by any reference in the VM, since they are loaded and used by kernel calls, and (b) thinking ahead to C++ SSCI used by Shivers 2, the GC becomes irrelevant, but the ResourceManager is still a thing that the game uses
[21:49] <snover> for a little more background, one of the three problems here is that in GK2, kLock seems to be the call which is responsible for audio resources to be loaded into memory, and the audio resources *have* to be loaded into memory in order for kDoAudioWaitForPlay to return the expected value (0) in the grotto (https://bugs.scummvm.org/ticket/9675)
[21:50] <snover> (or at least, we need to be able to flag that the resource was requested to be loaded into memory, even if it is not)
[21:50] <snover> the second problem is that Lighthouse seems to call too many times to unlock resources via kLock which breaks the lock counter
[21:51] <snover> the third problem is the one i just mentioned about ResourceManager needing to be usable in the absence of the SCI VM
[21:52] <snover> (oh, and with problem 1, the locked-ness of the audio resource needs to be stored in the save game, since the game does not call kLock again when the save game is restored)
[21:54] <wjp> fun fun fun
[21:54] <wjp> 2 seems to imply the need for some kind of reference tracking
[21:55] <-- girafe left irc: Read error: Connection reset by peer
[21:55] <wjp> or maybe separate lock counters for (unreliable) script locks and (reliable) internal engine locks
[21:58] <wjp> is it easy to "fake" the audio resource status needed for KDoAudioWaitForPlay ?
[21:58] <wjp> with some shadow bookkeeping, I mean
[22:02] <snover> This is where I have been getting stuxj
[22:02] <snover> Stuck *
[22:04] <snover> Audio 32 channels are restored via soundcmd
[22:05] <snover> As the BG music is a sfx
[22:06] <snover> Audio32 gets one lock for itself, the script klock gives it two locks
[22:06] <snover> Two locks total*
[22:08] <snover> So right now I think I'm returning that the channel is loaded in memory if there are 2 locks on the resource
[22:08] <snover> 2+ locks*
[22:11] <-- Korak left irc: Ping timeout: 268 seconds
[22:12] <snover> There are just a ton of possible edge cases, like a resource can be played simultaneously in multiple channels, though I don't know that this actually happens
[22:14] <snover> kDoAudioPreload also makes channels in memory, but I only see that call in hoyle5
[22:18] <-- Farmboy0 left irc: Remote host closed the connection

[22:26] <-- ny00123 left irc: Quit: Leaving
[22:50] <-- waltervn left irc: Quit: Leaving
[23:22] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:35] <-- 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.
[23:53] <-- Mia left irc: Read error: Connection reset by peer
[00:00] --- Fri Jun 2 2017