ScummVM IRC Log - 25 Oct 2016

[Back to Index]

[00:04] --> ruskie joined #scummvm.
[00:18] --> somaen joined #scummvm.
[00:18] <-- somaen left irc: Changing host
[00:18] --> somaen joined #scummvm.
[00:18] #scummvm: mode change '+o somaen' by ChanServ!ChanServ@services.
[00:30] --> GitHub62 joined #scummvm.
[00:30] <GitHub62> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vPxjc
[00:30] <GitHub62> scummvm/master 1c370c9 Paul Gilbert: TITANIC: Add conversation text to the PET
[00:30] GitHub62 (GitHub62@192.30.252.34) left #scummvm.
[00:36] <-- SylvainTV left irc: Read error: Connection reset by peer
[01:04] <-- Polynomial-C left irc: Ping timeout: 260 seconds
[01:07] <-- cody56 left irc: Ping timeout: 244 seconds
[01:20] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[01:41] <-- Dominus left irc: Ping timeout: 252 seconds
[01:43] --> Dominus joined #scummvm.
[01:43] <-- Dominus left irc: Changing host
[01:43] --> Dominus joined #scummvm.
[01:46] --> Vampire0_ joined #scummvm.
[01:50] <-- Vampire0 left irc: Ping timeout: 260 seconds
[02:24] <-- snover left irc: Quit: Leaving.
[02:52] <-- dreammaster left irc:
[03:00] --> Polynomial-C joined #scummvm.
[03:15] <-- ignalina left irc: Quit: ignalina
[04:03] <-- Polynomial-C left irc: Ping timeout: 256 seconds
[04:11] <-- Littleboy left irc: Read error: Connection reset by peer
[05:16] <-- LittleToonCat left irc: Remote host closed the connection
[05:43] --> m_kiewitz joined #scummvm.
[05:43] <-- m_kiewitz left irc: Changing host
[05:43] --> m_kiewitz joined #scummvm.
[05:43] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[06:00] --> Polynomial-C joined #scummvm.
[06:00] <-- Strangerke left irc: Ping timeout: 276 seconds
[06:04] <-- TMM left irc: Quit: Ex-Chat
[06:24] <-- wanwan left irc: Ping timeout: 260 seconds
[06:36] --> wanwan joined #scummvm.
[06:41] --> niska joined #scummvm.
[06:42] <-- niska` left irc: Ping timeout: 265 seconds
[06:42] <-- Cheeseness left irc: Ping timeout: 265 seconds
[06:42] --> Cheeseness joined #scummvm.
[07:35] <-- girafe left irc: Quit: Leaving
[07:58] <rootfather> hey folks
[07:59] <rootfather> is it okay that I submit my next changes as PRs, because I want to take part in the Github Hackathon and get the free t-shirt? :P
[08:01] <Harekiet> whoever commits the most lines gets a tshirt?
[08:02] <rootfather> no, you get a free shirt if you open 4 PRs until oct 31st
[08:03] <Harekiet> hehe nice
[08:04] <rootfather> I don't know if it works if you are a member of the organization you are commiting too, but worth a try I think :P
[08:04] <-- Cheeseness left irc: Ping timeout: 250 seconds
[08:05] <Harekiet> Can't pass on a free tshirt :)
[08:05] <fuzzie> it's pretty normal to submit PRs to your own organizations, to do a review process
[08:06] <fuzzie> and the rules say 'any' :p
[08:10] <_sev> hehehe
[08:14] <fuzzie> hm
[08:15] <fuzzie> I just added a comment to a PR and clicked the big green button and it gives me a 'review' :o
[08:43] <-- Lightkey left irc: Ping timeout: 260 seconds
[08:56] --> Lightkey joined #scummvm.
[09:01] <-- Harekiet left irc: Ping timeout: 245 seconds
[09:01] --> Cheeseness joined #scummvm.
[09:08] --> Harekiet joined #scummvm.
[09:17] --> kurtwr2 joined #scummvm.
[09:19] <-- kurtwr left irc: Ping timeout: 268 seconds
[09:44] <madmoose> rootfather: damn, I made like 8 or 9 commits yesterday, they could have been PRs...
[10:24] --> dos11 joined #scummvm.
[10:24] --> Henke37 joined #scummvm.
[10:24] <-- dos1 left irc: Read error: Connection reset by peer
[10:25] Nick change: dos11 -> dos1
[10:49] --> ajax16384 joined #scummvm.
[10:49] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[10:49] <-- ajax16384 left irc: Remote host closed the connection
[10:51] --> ajax16384 joined #scummvm.
[10:51] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[11:22] --> ignalina joined #scummvm.
[12:12] --> omer_mor_ joined #scummvm.
[12:14] <-- omer_mor left irc: Ping timeout: 250 seconds
[12:29] <-- ajax16384 left irc: Read error: Connection reset by peer
[12:39] <-- Harekiet left irc: Ping timeout: 250 seconds
[12:56] --> Harekiet joined #scummvm.
[13:07] --> te_lanus joined #scummvm.
[13:28] --> Littleboy joined #scummvm.
[13:28] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[14:04] --> omer_mor joined #scummvm.
[14:05] <-- omer_mor_ left irc: Ping timeout: 260 seconds
[14:18] <-- Paaris left irc: Ping timeout: 246 seconds
[14:20] --> Paaris joined #scummvm.
[14:22] --> NuSuey joined #scummvm.
[14:25] <-- _sev left irc: Quit: Leaving
[14:27] --> ajax16384 joined #scummvm.
[14:27] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[14:31] <-- ajax16384 left irc: Read error: Connection reset by peer
[14:38] --> _sev joined #scummvm.
[14:38] #scummvm: mode change '+o _sev' by ChanServ!ChanServ@services.
[14:40] <-- DrMcCoy left irc: Quit: He's dead, Jim
[14:44] --> snover joined #scummvm.
[14:44] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[14:51] --> uruk-hai joined #scummvm.
[14:52] #scummvm: mode change '+o uruk-hai' by ChanServ!ChanServ@services.
[14:55] --> DrMcCoy joined #scummvm.
[14:55] <-- DrMcCoy left irc: Changing host
[14:55] --> DrMcCoy joined #scummvm.
[14:55] #scummvm: mode change '+o DrMcCoy' by ChanServ!ChanServ@services.
[14:56] --> Guest80247 joined #scummvm.
[15:14] --> LittleToonCat joined #scummvm.
[15:28] <snover> wjp: interestingly, that fix for the error in the sq6 cutscene also seems to make the animations on the rama main menu work properly
[15:28] <snover> i thought that that flag wasnt used any more in sci3 for some reason
[15:29] <m_kiewitz> \o/
[15:52] <wjp> sorry, which fix was that?
[15:54] <-- DrMcCoy left irc: Quit: He's dead, Jim
[15:54] <snover> c9ad2062db80fc9b6339bebdcd70f58269e38bf6
[15:54] --> DrMcCoy joined #scummvm.
[15:54] #scummvm: mode change '+o DrMcCoy' by ChanServ!ChanServ@services.
[15:55] <wjp> ah, that one
[15:55] <wjp> cool :-)
[15:55] <wjp> so not just the one bug after all :-)
[16:02] <snover> yeah!
[16:06] <-- dtcrshr left irc: Read error: Connection reset by peer
[16:15] <-- uruk-hai left irc: Read error: Connection reset by peer
[16:18] <-- omer_mor left irc: Quit: Leaving
[16:19] --> omer_mor joined #scummvm.
[16:20] <-- omer_mor left irc: Client Quit
[16:21] --> omer_mor joined #scummvm.
[16:29] <-- ignalina left irc: Quit: ignalina
[16:31] --> ny00123 joined #scummvm.
[16:33] --> dtcrshr joined #scummvm.
[16:47] --> Wishmaster2 joined #scummvm.
[16:48] <-- DrMcCoy left irc: Quit: He's dead, Jim
[16:48] --> waltervn joined #scummvm.
[16:48] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[16:49] --> DrMcCoy joined #scummvm.
[16:49] #scummvm: mode change '+o DrMcCoy' by ChanServ!ChanServ@services.
[16:51] <-- Wishmaster2 left irc: Ping timeout: 260 seconds
[16:57] --> ignalina joined #scummvm.
[16:58] --> Wishmaster2 joined #scummvm.
[17:13] --> WooShell joined #scummvm.
[17:13] <WooShell> meow =^.^=
[17:15] <_sev> woo!
[17:20] <madmoose> yooo
[17:20] <m_kiewitz> moo!

[17:34] <omer_mor> I downloaded a win32 daily today to test kirben's fix to Hebrew simon. Scummvm crashed because of what I think is a newer SDL2.dl.
[17:35] <-- bgK left irc: Ping timeout: 250 seconds
[17:35] <omer_mor> I think it's SDL2 because after installilng an older daily I still got crashes, and I found that installing a previous version keeps the newer SDL2.dll
[17:36] <omer_mor> removing that dll and re-installing the older daily fixed the problem
[17:36] <omer_mor> _sev: ^^^
[17:36] --> ajax16384 joined #scummvm.
[17:36] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[17:37] <omer_mor> so I suspect that latest SDL2.dll bundled with the win32 daily is incompatible with scummvm.
[17:37] --> bgK joined #scummvm.
[17:37] #scummvm: mode change '+o bgK' by ChanServ!ChanServ@services.
[17:37] <omer_mor> I can help test if needed
[17:37] <_sev> omer_mor: Kirben is building the builds
[17:38] <_sev> eventually we make the buildbot doing that completely
[17:38] <_sev> you may also consider downloading nightly builds from the buildbot
[17:39] <omer_mor> _sev: I don't believe kirben is manually building the nightly on a daily basis, right?
[17:39] <_sev> omer_mor: he does
[17:39] <omer_mor> btw - the "good" sdl2 is 2.0.4.0 and the "bad" is 2.0.5.0
[17:40] <_sev> that's weird. as the previous bad version was 2.0.4 and we recommended to use 2.0.3
[17:41] <omer_mor> _sev: unbelievable! he does that for several years now? daily??
[17:41] <_sev> omer_mor: yes, he does
[17:41] <_sev> he was insisting on this procedure
[17:42] <omer_mor> 2.0.4.0 is from a 2016/10/16 nightly and 2.0.5.0 is from today's.
[17:43] <omer_mor> I just checked the buildbot build and it's not bundled with SDL2 at all.
[17:44] --> GitHub18 joined #scummvm.
[17:44] <GitHub18> [scummvm] sev- closed pull request #854: ANDROIDSDL: Changed logic for back button & getting sdcard storage location (master...master) https://git.io/vPbiY
[17:44] GitHub18 (GitHub18@192.30.252.40) left #scummvm.
[17:44] --> GitHub90 joined #scummvm.
[17:44] <GitHub90> [scummvm] sev- pushed 3 new commits to master: https://git.io/vPhQ8
[17:44] <GitHub90> scummvm/master a0c2fca lubomyr: ANDROIDSDL: back button re-mapped to F13 keycode. F13 keycode assigned to call scummvm in-game menu. CONTROL key now can be uses with games
[17:44] <GitHub90> scummvm/master 55f3771 lubomyr: ANDROIDSDL: default storage-sdcard directory location getting direct from libSDL wrapper
[17:44] <GitHub90> scummvm/master 97bc65e Eugene Sandulenko: Merge pull request #854 from lubomyr/master...
[17:44] GitHub90 (GitHub90@192.30.252.34) left #scummvm.
[17:44] <omer_mor> copying the 2.0.4.0 SDL2 over the latest nightly fixes the crash
[17:45] <m_kiewitz> so 2.0.5.0 has different bugs? oh my
[17:46] <m_kiewitz> 2.0.4.0 is totally broken on Windows
[17:49] <omer_mor> I didn't notice any issues with 2.0.4.0 so far
[17:49] <omer_mor> 2.0.5.0 crashes on load, just after trying to start the game
[17:50] <omer_mor> so the latest win32 nightly is totally unusable
[17:50] <snover> backtrace?
[17:51] <omer_mor> I'll try to get one. I think I'll to build it locally for that, right?
[17:53] <snover> im not sure what needs to happen in windows. i would assume that the nightly builds would be built with debug symbols but i dont know that either
[17:55] <snover> omer_mor: by the way, i tried signing up for an account on sciprogramming since you asked; the forum is set up to require manual verification of new accounts and mine still has not been verified in several days
[17:55] <snover> so im just going to make you keep reporting stuff there ok ;)
[17:56] <omer_mor> thanks for signing up! you didn't respond to me asking you to report the bug, so I figured you just didn't want the hassle of reporting it and did it for you :)
[17:57] <omer_mor> I'll ping the board owner about your acccount. is it 'snover' ?
[17:57] <snover> well, i dont want the hassle too&
[17:57] <snover> :)
[17:58] <snover> i guess i can just report through github too
[17:58] <omer_mor> I actually reported both in the forum and on github
[17:59] <omer_mor> did a batch of reports for a few bugs/feature requests
[17:59] <omer_mor> https://github.com/icefallgames/SCICompanion/issues
[17:59] <omer_mor> if you have some more let me know or report them
[18:01] <omer_mor> (started building scummvm, will probably take a few minutes)
[18:01] <omer_mor> snover: this is what I wrote in your bug (https://github.com/icefallgames/SCICompanion/issues/5) :
[18:02] <omer_mor> The scummvm team found that SCI32 changed its handling of zero-offset exports: scummvm/scummvm@55222ec
[18:02] <omer_mor> SV and SciCompanion are both affected by this change.
[18:02] <omer_mor> You can see this in Space Quest 6, where in rm500.sc sEndChapter1::changeState(6) is calling __proc64909_0 (originally ShakePlane) but the 64909 script has "export64909_0 = $0" (in SV) and the procedure is not detected and decompiled.
[18:02] <snover> yes, i saw it
[18:02] <omer_mor> snover: Am I correct?
[18:02] <omer_mor> I'm not 100% understanding the issue, which is why I preferred you to report it in the first place :)
[18:05] <snover> as far as i know
[18:05] <snover> yes
[18:05] <m_kiewitz> omer_mor: maybe we could patch SV
[18:06] <-- DrMcCoy left irc: Quit: He's dead, Jim
[18:06] <omer_mor> m_kiewitz: how?
[18:07] <m_kiewitz> well we can disassemble SV too
[18:07] --> DrMcCoy joined #scummvm.
[18:07] #scummvm: mode change '+o DrMcCoy' by ChanServ!ChanServ@services.
[18:07] <omer_mor> m_kiewitz: disassemble?
[18:07] <omer_mor> oh
[18:07] <omer_mor> :)
[18:07] <-- bgK left irc: Ping timeout: 245 seconds
[18:08] <snover> without support for CSC its not likely that id run into any new problems with SC since it cant decompile anything else
[18:08] <snover> though who knows
[18:08] <omer_mor> m_kiewitz: that would be great. SV is a very useful tool.
[18:08] <omer_mor> snover: CSC?
[18:09] <snover> the format used by sci3
[18:09] <omer_mor> snover: have you considered adding CSC support via PR?
[18:09] <snover> no, i have my own engine to work on :)
[18:10] <omer_mor> snover: don't you use sci companion during your development? won't it help?
[18:10] <snover> i also dont have any system set up that i can use to run vs2015
[18:11] <-- ScottT1 left irc: Ping timeout: 260 seconds
[18:20] <omer_mor> snover: ok...
[18:21] <omer_mor> I've built scummvm but it seems to be using SDL.dll and not SDL2.dll
[18:21] <omer_mor> is there something I need to do?
[18:24] --> bgK joined #scummvm.
[18:24] #scummvm: mode change '+o bgK' by ChanServ!ChanServ@services.
[18:25] <-- NuSuey left irc: Quit: Connection closed for inactivity
[18:26] <omer_mor> seems like VS is using the libraries I copied from http://www.scummvm.org/frs/build/scummvm_libs_2015.zip
[18:27] <omer_mor> however these contain an older SDL2
[18:27] <omer_mor> is there a newer libs package that matches the latest nightly?
[18:33] <-- bgK left irc: Ping timeout: 256 seconds
[18:38] --> bgK joined #scummvm.
[18:38] #scummvm: mode change '+o bgK' by ChanServ!ChanServ@services.
[18:47] <-- bgK left irc: Ping timeout: 260 seconds
[18:48] --> LyTchet joined #scummvm.
[18:48] <snover> wjp: oh my god i think i just figured out https://bugs.scummvm.org/ticket/9582
[18:49] --> bgK joined #scummvm.
[18:49] #scummvm: mode change '+o bgK' by ChanServ!ChanServ@services.
[18:49] <snover> https://zetafleet.com/i/580fa93ea472f.png whats wrong with this string?
[18:50] <snover> (which is used in conjunction with strchr to find the decimal value of a hex string)
[18:52] <_sev> there are two zeros?
[18:52] <snover> yup.
[18:53] <snover> this bug goes back to at least sci1.1
[18:55] --> girafe joined #scummvm.
[18:58] <snover> it is in kq6
[18:59] <snover> not in sq4&
[18:59] <snover> not in ecoquest cd&
[18:59] <-- ny00123 left irc: Read error: Connection reset by peer
[19:00] <snover> not in sq5&hm
[19:01] <snover> is in lsl6 lores
[19:02] <snover> m_kiewitz: can you (or anyone) help me track down which sci1.1 games have this bug?
[19:02] <m_kiewitz> sure
[19:02] <m_kiewitz> but what did it cause in sci1.1?
[19:03] --> Strangerke joined #scummvm.
[19:03] <snover> if someone uses a hex escape sequence in a message, it treats a-f as 11-16 instead of 10-15
[19:03] <m_kiewitz> argh
[19:03] <m_kiewitz> although i wonder if any sci1.1 actually did this
[19:03] <m_kiewitz> i think there are no open bug reports atm
[19:04] --> ny00123 joined #scummvm.
[19:04] <-- LyTchet left irc:
[19:05] <m_kiewitz> hmm i can't find that string inside qfg3 demo
[19:06] <m_kiewitz> it's inside the data segment, I guess?
[19:07] <snover> yeah
[19:07] <m_kiewitz> hmm non in pq1 sci
[19:08] <m_kiewitz> no in qfg3
[19:15] <snover> i have to run i will be back in a while. thanks for checking :)
[19:16] <-- snover left irc: Quit: Leaving.
[19:18] --> TMM joined #scummvm.
[19:18] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[19:18] <-- girafe left irc: Ping timeout: 250 seconds
[19:19] <-- ajax16384 left irc: Read error: Connection reset by peer
[19:19] --> salty-horse joined #scummvm.
[19:19] #scummvm: mode change '+o salty-horse' by ChanServ!ChanServ@services.
[19:21] <-- Wishmaster2 left irc: Read error: Connection reset by peer
[19:22] <TMM> hi
[19:23] <omer_mor> _sev: any idea how I can build scummvm on my machine w/latest SDL2?
[19:24] --> ScottT joined #scummvm.
[19:24] #scummvm: mode change '+o ScottT' by ChanServ!ChanServ@services.
[19:24] <omer_mor> the scummvm_lib_2015 package contains an old one
[19:24] <_sev> omer_mor: by following the build instructions?
[19:24] <omer_mor> _sev: I did
[19:25] <omer_mor> it tells me to get the scummvm_lib package which I did
[19:25] <omer_mor> but it's an old package
[19:25] <omer_mor> _sev: I'm just trying to help find the reason new SDL2 crashes on start
[19:26] <_sev> ok, let me replace it
[19:26] <_sev> with an updated version
[19:27] --> criezy joined #scummvm.
[19:27] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[19:27] <_sev> omer_mor: please redownload it. The new package is 60MB
[19:29] <_sev> but I am not sure that that is not 2.0.4
[19:30] <-- Guest80247 left irc: Read error: Connection reset by peer
[19:30] <omer_mor> it's 2.0.4
[19:49] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[19:49] <Strangerke> hi guys
[19:50] <-- waltervn left irc: Quit: Leaving
[19:51] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[19:56] --> girafe joined #scummvm.
[20:06] <-- ignalina left irc: Quit: ignalina
[20:07] --> SylvainTV joined #scummvm.
[20:07] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[20:12] <rootfather> omer_mor then you should download sdl 2.0.5 from libsdl.org
[20:13] Nick change: rootfather -> rootfather|AFK
[20:13] Nick change: rootfather|AFK -> rootfather
[20:13] --> m_kiewitz joined #scummvm.
[20:13] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[20:14] <rootfather> 11 years ago, around this time, I first discovered ScummVM btw
[20:14] <rootfather> what a great journey since then
[20:14] <rootfather> guys, I thank you so much.
[20:15] <omer_mor> rootfather: runtime binaries or development libraries?
[20:16] <rootfather> I think for compilation you need the dev libraries
[20:17] <rootfather> any plans to contact Tobis87 in the forums to get something like http://forum.scummvm.org/viewtopic.php?t=14110 in the codebas?
[20:17] <rootfather> codebase?
[20:22] <rootfather> and another thing... IIRC, residualvm checks the integrity of the game files on the first startup, right?
[20:22] <rootfather> maybe we should implement something like this with extensive checksums for all datafiles in scummvm too?
[20:27] <wjp> snover: oh wow. Broken hex parsing in SSCI?
[20:28] <rootfather> TMM, do you have any plans to create a 64bit build for windows
[20:29] <TMM> rootfather, if people care about it, sure
[20:29] <TMM> I was lead to believe people generally didn't care about it
[20:29] <rootfather> well, I care :P
[20:30] <rootfather> and in http://forum.scummvm.org/viewtopic.php?t=14156 there is at least one person that cares
[20:30] <rootfather> and 64bit is the future, so maybe we should provide official builds for it
[20:32] <TMM> If I can test it properly I'll do it for 1.10 probably
[20:32] <rootfather> count me in for testing
[20:32] <rootfather> running Windows 10 64bit myself
[20:33] <TMM> Is there really a way to run win64 without wow?
[20:33] <TMM> I didn't think microsoft gave that option
[20:34] <TMM> Maybe I'm misunderstanding something
[20:35] <TMM> So on most linuxes you can run 64-bit only, is that possible on windows now?
[20:35] <rootfather> hm, I have trouble understanding your question
[20:35] <rootfather> on 64bit windows, 64bit apps run native, while 32bit applications run through the 32bit abstraction/emulation layer
[20:35] <rootfather> to be honest, I don't understand your question
[20:36] <TMM> so, on Linux when you install a 64bit distro out of the box you get the ability to run 32bit binaries but you get 0 libraries or runtime support until you install it
[20:36] <TMM> it is my understanding that the 32bit userland required to run 32bit applications on windows 64bit is non-optional
[20:36] <TMM> ie: the gigabytes requires to run 32bit libraries has already been paid on every windows installation, 32bit or 64bit
[20:37] <TMM> I don't think there's any *runtime* cost for running 32bit apps on a 64bit os, at least not on Linux
[20:38] <rootfather> yeah, running 32bit binaries on win 64 cause some overhad
[20:38] <rootfather> overhead
[20:38] <rootfather> I don't know how linux handles this
[20:38] <TMM> what is this overhead?
[20:38] --> cody56 joined #scummvm.
[20:39] <TMM> If I can get it tested properly I'll still do it, but I don't really understand what the benefit is
[20:41] <rootfather> I think there is some overhead if you run 32bit windows applications on 64bit windows caused by the WoW64 compatibility layer
[20:41] <rootfather> the benefit is that 64bit systems are the "future", and in my opinion, the earlier we achieve stable 64bit builds, the better
[20:42] <TMM> I think this overhead only exists if you run WoW64 applications on itanic
[20:43] <rootfather> to my understanding, you have some overhead as soon as you run 32bit apps on Win64
[20:44] <TMM> yes, but this overhead is the same as running the 64bit version in the first place
[20:44] <TMM> you need to pad datastructures sent into the kernel, so the 32bit app uses more ram than the 32bit app on a 32bit windows
[20:45] <TMM> but the same as that same app running in 64bit windows natively
[20:46] <TMM> I think you're trading potential overhead for guaranteed overhead
[20:46] <rootfather> okay, I have to admit that this is beyond my knowledge of the architectures
[20:48] <TMM> I am like 99% sure that this is the case
[20:49] <TMM> and we'll have to update the migration.bat file
[20:50] <TMM> because if someone upgrades from 32bit 1.5.0 to 64bit 1.10 the savegames will stay in c:/program files (x86)
[20:50] <TMM> whereas migration.bat won't look there
[20:51] <rootfather> maybe we should get rid of migration.bat...
[20:51] <rootfather> we made the changes a long time ago, and I doubt anybody will upgrade from 1.5.0 to 1.10
[20:52] <TMM> it does seem that if we do make a 64bit build that will not get a migration.bat
[20:53] <somaen> Well
[20:53] <somaen> As far as I know, there are upsides and downsides wrt 32-bit binaries vs 64-bit binaries
[20:54] <somaen> 64-bit binaries have to pay the additional cost of the increased pointer size, where most apps rarely need that much memory
[20:54] <TMM> what are the downsides? because I couldn't think of any, and googling hasn't really revealed any
[20:54] <TMM> (of 32bit apps on a 64bit os, on x86)
[20:54] <somaen> They do gain the additional 3 registries though
[20:55] <somaen> So, you can in practice go slower running a 64-bit app on a 64-bit os for some cases
[20:55] <TMM> well, on linux there's the X86 'architecture' that is specifically for that
[20:55] <TMM> but I don't think windows has that
[20:56] <somaen> There's WoW64
[20:56] <TMM> X32, I think
[20:56] <TMM> no, you miss my meaning. You can run a 32bit app in long mode, getting access to the extra regs without using the extra pointer space
[20:56] <TMM> but the application needs to be compiled for it
[20:56] <TMM> this won't just magically happen for any x86_32 app
[20:57] <TMM> and it requires runtime and kernel support
[20:57] <TMM> neither of which exist on win64, afaik
[20:57] <somaen> Any additional costs do seem to be itanium specific: https://msdn.microsoft.com/en-us/library/windows/desktop/aa384219(v=vs.85).aspx
[20:58] <somaen> You can't run a 32 bit app with extra regs, that would be an app using the 64-bit architecture
[20:58] <rootfather> so there is currently no benefit in using 64bit application?
[20:59] <somaen> @rootfather: It all depends on what you plan to run
[20:59] <TMM> rootfather, well, you get to use a couple extra registers which can help
[20:59] <TMM> but if you don't use more than 2GB of ram... not really
[21:00] <somaen> I do remember the CounterStrike server example that Valve used, where simply recompiling made it 20% faster
[21:00] <TMM> I really, really doubt that scummvm falls in that category though
[21:00] <somaen> Presumably because of the registers.
[21:00] <TMM> (what would a 20% faster scummvm even mean on a desktop pc :))
[21:00] <somaen> The one bonus you DO get going 64-bit, is that you can make some further assumptions about availability of SSE and stuff
[21:01] <somaen> @TMM: A lot for Wintermute
[21:01] <somaen> :P
[21:01] <somaen> I seriously need to optimize the rendering there, some day.
[21:03] <TMM> na'h just wait for x86_128
[21:03] --> ignalina joined #scummvm.
[21:08] <somaen> @TMM: What I gathered above was that you said that software compiled for x86 (not x64 or x86_64 or whatever the proper term is) can access the extra registers.
[21:08] <somaen> As far as I know, that's not true.
[21:09] <TMM> somaen, no, but Linux has a special subarch called (I believe) x32 that is specifically designed to do that
[21:09] <TMM> somaen, and, as far as I know, windows does not have that
[21:09] <somaen> Well, you would presumably need to recompile everything
[21:09] <TMM> somaen, https://en.wikipedia.org/wiki/X32_ABI
[21:09] <somaen> Which Windows software doesn't lend itself as easily to, given it's binary nature in most cases
[21:10] <somaen> https://en.wikipedia.org/wiki/X86-64#OPMODES
[21:10] <somaen> But X32 does seem to give the best of both worlds for anything that doesn't need the additional memory
[21:13] <somaen> And, yes, sorry about the confusion then.
[21:13] <TMM> anyway,I'll probably still do the win64 build
[21:14] <TMM> but I'm not so sure it'll actually matter
[21:16] <-- ignalina left irc: Quit: ignalina
[21:17] --> GitHub191 joined #scummvm.
[21:17] <GitHub191> [scummvm] wjp pushed 1 new commit to master: https://git.io/vPjRv
[21:17] <GitHub191> scummvm/master b4515d0 Willem Jan Palenstijn: KYRA: (LOL) Fix buffer overflow in _lastOverridePalFile...
[21:17] GitHub191 (GitHub191@192.30.252.41) left #scummvm.
[21:23] --> ignalina joined #scummvm.
[21:26] --> snover joined #scummvm.
[21:26] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[21:29] <snover> m_kiewitz: any others that you found with the bad escape code?
[21:29] <snover> right now i know only of LSL6 and KQ6
[21:29] <m_kiewitz> i need to look through the others, moment
[21:29] <snover> wjp: yep :)
[21:31] <m_kiewitz> these (LOL) comments look so weird, i always think it's literally lol (laughing out loud) for a second

[21:33] <m_kiewitz> snover: i can't find any "0123456789" string in for example the qfg3 at all
[21:33] <snover> m_kiewitz: i can verify that it is not in qfg3
[21:33] <m_kiewitz> so I guess they added a different algo on this?
[21:34] <m_kiewitz> or maybe didn't support that originally at all?
[21:34] <-- ny00123 left irc: Quit: Leaving
[21:35] <snover> it seems there are 5 different versions of messages so it is definitely possible that they added in the escape mode later. (i mean, it is clear that they did, as LSL6 and KQ6 and all of SCI32 have this broken thing)
[21:35] <m_kiewitz> Mixed Up Mother Goose SCI1.1 doesnt have it either
[21:36] <m_kiewitz> you checked kq6 cd only, I guess?
[21:36] <snover> ah, yes. both of those are CD versions.
[21:36] <snover> that i checked.
[21:37] <m_kiewitz> kq6 German floppy doesn't have it
[21:38] <m_kiewitz> Pepper's Adventure in Time (demo) has it
[21:38] <m_kiewitz> you checked sq4 cd?
[21:38] <m_kiewitz> because that's following pepper
[21:38] <snover> im pretty sure i did
[21:39] <snover> let me double check
[21:39] <snover> yeah, it is no in SQ4CD
[21:39] <m_kiewitz> eco quest follows that one, and you said it also doesn't have it
[21:39] <snover> nor in SQ5
[21:39] <m_kiewitz> i don't have "Hoyle Classic" (SCI1.1)
[21:39] <m_kiewitz> so can't check that
[21:41] <m_kiewitz> ecoquest 2 doesn't have it either
[21:42] <m_kiewitz> pepper's adventure in time (full game) has it too
[21:42] <snover> woo, random. :)
[21:43] <m_kiewitz> Laura Bow 2 doesn't have it
[21:43] <m_kiewitz> (CD)
[21:44] <m_kiewitz> gabriel Knight 1 demo has it
[21:45] <snover> so, uh. it appears that the dos version of hoyle classic card games does not have it but the windows version does.
[21:45] <m_kiewitz> ok, missing atm are: Freddy Pharkas (demo+game), "Hoyle classic" SCi1.1, "Hoyle Classic Card Games" (maybe the same as Hoyle Classic)
[21:46] <m_kiewitz> police quest 4 demo, slater & charlie, qfg4 demo
[21:47] <m_kiewitz> Slater & Charlie has it
[21:48] <m_kiewitz> Freddy Pharkas CD has it
[21:48] <snover> so given the randomness of this bug, i guess a better question is whether the sci1.1 games *without* it had a different way of doing hex literals
[21:48] <m_kiewitz> Freddy Pharkas floppy ha it
[21:48] <snover> or if this was just a feature that was introduced with a bug and they never fixed the bug
[21:48] <m_kiewitz> well either that or they couldn't do them at all
[21:49] <snover> right
[21:49] <snover> if they had no hex literals then we can just reimplement the bug and everything is fine
[21:50] <snover> do you have an idb of one of those games without it? the bad function is MassageText
[21:51] <m_kiewitz> all 3 freddy pharkas demos have it
[21:51] <m_kiewitz> i got idbs for almost every game
[21:51] <m_kiewitz> which kernel call is it?
[21:52] <m_kiewitz> so wait, which games have it again?
[21:52] <m_kiewitz> kq6 CD only, freddy pharkas (all), slater & charlie, gabriel knight 1 demo, pepper's adventure in time
[21:53] <snover> LSL6 CD and Hoyle Classic Card Games (windows only)
[21:53] <m_kiewitz> larry 6
[21:53] <snover> and all of SCI32 :)
[21:53] <m_kiewitz> well of course
[21:53] <m_kiewitz> and i bet pq4 demo + qfg4 demo have it too
[21:53] <m_kiewitz> haven't checked them, can't find them right now
[21:54] <m_kiewitz> which kernel call is it again?
[21:54] <m_kiewitz> i don't have debug symbols for the sci16 games
[21:55] <m_kiewitz> ah wait, i can check using the kq6 cd idb
[21:55] <snover> it is kMessage, the specific function is MassageText
[21:56] <m_kiewitz> which sub-op is it?
[21:56] <snover> 1, getnext
[21:57] <m_kiewitz> i just figured that out
[21:59] <m_kiewitz> hmm did those other games actually have a kMessage kernel call
[22:00] <snover> interesting question. there is a string "Msg %d: %d %d %d %d not found" in this call, and that string exists in the binary, but ida does not seem to find a reference to it
[22:00] <snover> (im just looking at sq5 here)
[22:01] <m_kiewitz> hmm i think i found it but the code looks completely different
[22:02] <m_kiewitz> it seems they added more opcodes at the same time
[22:02] <snover> its normally in the same function as the one that gets rid of the stage directions
[22:04] <snover> i gotta run again for a bit, ill be back
[22:05] <snover> let me know what you find
[22:05] <-- WooShell left irc: Quit: Walking upside down in the sky, between the satellites passing by. Gliding along the black rainbow, I fly away with my shadow. Scratching the moon like a DJ, the night follows its odyssey.
[22:07] <m_kiewitz> snover: this is the routine http://pastebin.com/eRbx7NiF
[22:07] <m_kiewitz> it's completely different from the one that is used by the buggy interpreters
[22:07] <m_kiewitz> too sleepy to look through it
[22:08] <m_kiewitz> but it checks for '0' and letters, so I guess maybe it's doing the same thing
[22:36] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[22:37] --> dreammaster joined #scummvm.
[22:38] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[22:43] --> GitHub144 joined #scummvm.
[22:43] <GitHub144> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vPjKA
[22:43] <GitHub144> scummvm/master 20092fb Paul Gilbert: TITANIC: Implement hide counting for mouse cursor hide/show
[22:43] GitHub144 (GitHub144@192.30.252.34) left #scummvm.
[22:48] <-- ignalina left irc: Quit: ignalina
[22:52] --> ignalina joined #scummvm.
[23:04] <snover> m_kiewitz: those letter checks are for the stage directions, looks like the whole thing is only the stage directions code and there is no hex escape code
[23:16] --> GitHub103 joined #scummvm.
[23:16] <GitHub103> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vPjXW
[23:16] <GitHub103> scummvm/master 44bcade Paul Gilbert: TITANIC: Keep cursor disabled for entire Doorbot intro
[23:16] GitHub103 (GitHub103@192.30.252.40) left #scummvm.
[23:20] <-- ced117_ left irc: Ping timeout: 260 seconds
[23:21] --> ced117_ joined #scummvm.
[23:22] <-- criezy left irc: Quit: criezy
[23:23] <-- salty-horse left irc: Remote host closed the connection
[23:24] <snover> unfortunately now a new question is raised, which is when did they start allowing numbers in stage directions? since there is some comment in scummvm message.cpp that claims that that was only a sci32 thing, but here they are in this subroutine
[23:26] <snover> unfortunately i cant follow these jumps since they are using a different addressing scheme than the location names
[23:27] --> GitHub150 joined #scummvm.
[23:27] <GitHub150> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vPj1i
[23:27] <GitHub150> scummvm/master 8ec1f77 Paul Gilbert: TITANIC: Fix setting active NPC
[23:27] GitHub150 (GitHub150@192.30.252.46) left #scummvm.
[23:43] <-- ced117_ left irc: Ping timeout: 260 seconds
[23:44] --> ced117_ joined #scummvm.
[23:49] <snover> awh man, why was that an expiring pastebin :(
[23:51] --> GitHub178 joined #scummvm.
[23:51] <GitHub178> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vPjD5
[23:51] <GitHub178> scummvm/master 6c54d5e Paul Gilbert: TITANIC: Fix dragging Photograph
[23:51] GitHub178 (GitHub178@192.30.252.41) left #scummvm.
[23:51] --> GitHub159 joined #scummvm.
[23:52] <GitHub159> [scummvm] sev- pushed 1 new commit to master: https://git.io/vPjyv
[23:52] <GitHub159> scummvm/master 5649036 Eugene Sandulenko: FULLPIPE: Fix if statement. Thanks to PVS-Studio
[23:52] GitHub159 (GitHub159@192.30.252.41) left #scummvm.
[23:57] <snover> oh well. i found it without too much trouble.
[00:00] --- Wed Oct 26 2016