ScummVM IRC Log - 13 Jul 2017

[Back to Index]

[00:04] <ScummBot_> Port build status changed with 906c6de3: Success: master-ios7, master-mingw-w32, master-openpandora, master-android_mips, master-android_x86, master-ios, master-wii, master-debian-x86, master-mingw-w64-cplusplus11, master-webos, master-osx_intel, master-psp2, master-gp2x, master-dingux, master-debian-x86-clang, master-gamecube, master-gp2xwiz, master-android_arm,
[00:04] <ScummBot_> master-mingw-w64, master-deb
[00:04] <ScummBot_> ian-x86_64
[00:12] <-- SylvainTV left irc: Read error: Connection reset by peer
[00:31] --> travis-ci joined #scummvm.
[00:31] <travis-ci> scummvm/scummvm#3795 (master - 256b275 : Paul Gilbert): The build has errored.
[00:31] <travis-ci> Change view : https://github.com/scummvm/scummvm/compare/34969d50b87a...256b27581c73
[00:31] <travis-ci> Build details : https://travis-ci.org/scummvm/scummvm/builds/253016897
[00:31] travis-ci (travis-ci@ec2-54-157-231-228.compute-1.amazonaws.com) left #scummvm.
[00:31] <-- frankyboy_ left irc: Remote host closed the connection
[01:11] <-- girafe left irc: Read error: Connection reset by peer
[01:16] <-- criezy left irc: Ping timeout: 240 seconds
[01:17] --> Strangerke_ joined #scummvm.
[01:18] --> criezy joined #scummvm.
[01:18] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[01:19] <-- Strangerke left irc: Ping timeout: 240 seconds
[01:19] Nick change: Strangerke_ -> Strangerke
[01:34] --> Joefish_ joined #scummvm.
[01:34] #scummvm: mode change '+v Joefish_' by ChanServ!ChanServ@services.
[01:34] <-- Joefish left irc: Ping timeout: 258 seconds
[01:58] <-- Dominus left irc: Ping timeout: 260 seconds
[01:58] --> Dominus joined #scummvm.
[02:09] <-- Joefish_ left irc: Ping timeout: 276 seconds
[02:10] <-- dreammaster left irc:
[02:25] <tsoliman> snover: how were you able to debug the fullscreen/window switching on a mac? I am having a weird issue
[02:26] <tsoliman> in other words: is it just lots of "debug()" or did you actually use a debugger?
[02:30] <tsoliman> I misread "{14} All open tickets last commented on by someone other than you" as "by someone older than you"
[02:32] <snover> tsoliman: debug which part? the problem was clear, the backend code was always deleting and recreating windows instead of persisting the window and just reinitialising the rendering surface.
[02:33] <tsoliman> oh ok .. there's an issue and I don't know if it is macOS specific
[02:33] <snover> the rest of it was just figuring out what the correct incantations were for different platforms which were having some issues updating the surface correctly, and bgK did most of that work
[02:33] --> Axy joined #scummvm.
[02:33] <-- Axy left irc: Changing host
[02:33] --> Axy joined #scummvm.
[02:34] <tsoliman> it's the whole "start a game from the launcher" then "go fullscreen" then "return to launcher" .. if the game was exactly the same size as the launcher .. stuff gets weird .. (the launcher window size assumes the game's size but contains the launcher scaling
[02:35] <snover> uch. gross.
[02:35] <tsoliman> if the launcher is set to 2x, it happens to be "close enough" to Torin's forced resolution
[02:36] <tsoliman> but make the launcher 3x and Torin makes it smaller
[02:37] <tsoliman> torin does this .. COMI does this
[02:37] <snover> if you want to look into it you will probably need to use debug calls. definitely dont try to breakpoint in this code or windowserver will be locked up until you can either ssh in and kill the debugserver process, or force-reboot
[02:37] <tsoliman> ok cool I will use debug()
[02:38] <snover> a lot of the backend code seems more complicated than it ought to be, but i am definitely no expert here and i do not know what kinds of crazy stuff needs to be handled
[02:40] <tsoliman> it seems to be game-specific? like some games just say "don't scale me" or something?
[02:40] <snover> yes, the initGraphics function has a force normal1x parameter
[02:41] <snover> there are also assumptions in parts of backend code like that 320x200 shall be considered the default resolution
[02:52] <-- Cheeseness left irc: Quit: Leaving.
[02:56] --> Cheeseness joined #scummvm.
[03:02] <tsoliman> stupid C++ question: `a = b = c;` means `a = c; b = c;` right?

[03:34] <snover> tsoliman: Its Complicated". i would ask someone with more C++ knowledge to verify what i say here, but at least, with `a=b=c` the value assigned to a has to be the type of c converted to type b, so it is not the same as `a=c;b=c;` depending upon the types.
[03:42] <snover> also, = is not a sequence point, so while you can rely on the right-to-left associativity, you cant rely on side-effects applying in a specific order (e.g. `a[i] = i++` is UB)
[03:48] <snover> i had this same question a little while ago and still dont feel like i have a very satisfactory answer. i think it also even depends upon whether the side-effect-producing operators are user-defined operator functions, since arguments to a function must have been evaluated before invoking the function, so any side-effects must have taken place before the operator function is called. i think.
[03:54] <-- Lightkey left irc: Ping timeout: 258 seconds
[04:07] --> Lightkey joined #scummvm.
[04:11] <-- Axy left irc: Ping timeout: 260 seconds
[04:51] <snover> omg. omg omg omg omg omg. wat.
[04:51] <snover> this phant1 palette bug is being caused by something that i swear to god i had checked already days and days ago
[04:52] <snover> i must have teleported directly to the room to check if it fixed the problem or not!
[04:52] <snover> since teleporting in the original game is also broken!
[04:57] <snover> so the only reason this room is not glitched out all the time in the original interpreter is because, of all things, the P button in the interface is redrawn when switching to the new room for some reason, and this P buttons palette happens to have a used entry at index 255, and the colour is slightly different than what some other thing had set it to previously. games are not allowed to change 255 in the hardware (it is forced
[04:57] <snover> to white) but this room relies on this palette invalidation of an unusable index to fix the glitch.
[05:01] <snover> i had noticed that in KQ7, palettes were being invalidated by this sort of writing to index 255, and since it is not actually valid to change index 255, i just stopped copying it over to reduce the number of unnecessary palette invalidations in that game&
[05:05] <snover> smh.
[05:14] <snover> this doesnt seem to fix the basement glitch but at least it looks like the engine is still getting palettes from that bogus rat view in the basement so there probably isnt an error in calclists after all (phew!)
[05:26] <-- eriktorbjorn left irc: Quit: Leaving.
[05:52] <snover> basement glitch is looking like an identical-priority sorting failure.
[05:55] --> waltervn joined #scummvm.
[05:55] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[05:57] <waltervn> morning
[05:58] <snover> hey waltervn. just getting ready to sign off here, as soon as i hopefully finalise my understanding of this last bug.
[06:08] <snover> actually, i think i wont be quite getting there. close though. forcing the bad view to sort below the background pic ensures that the palette ends up being correct on most frames, but there are a couple frames where the bad view is submitted, but not the background pic which is normally overriding its palette changes on every frame in ssci.
[06:13] <snover> guess i will figure it out tomorrow.
[06:26] <-- waltervn left irc: Read error: Connection reset by peer
[06:28] <-- freanux left irc: Remote host closed the connection
[06:33] <-- TMM left irc: Quit: Ex-Chat
[06:57] --> Strangerke|work joined #scummvm.
[06:57] <Strangerke|work> hi guys
[06:59] <Strangerke|work> If it interests someone, there's a "cheap" Dylan Dog on ebay: http://www.befr.ebay.be/itm/Retrogame-PC-Dylan-Dog-attraverso-lo-specchio-/162588215165?ssPageName=ADME:B:SS:BE:1120
[07:01] <Strangerke|work> If someone is looking for Fascination CD, I also found one for a reasonable price, pm me
[07:02] <P2E> "cheap" indeed
[07:02] <P2E> gonna have to pass at that price
[07:02] <P2E> one of these days my museum will exist and I can afford to put together an exhibit or something
[07:03] <Strangerke|work> P2E: usually it's over 300eur
[07:03] <Strangerke|work> (maybe it'll end over 300, as it's a bid)
[07:04] <P2E> I won't be one of them, but I'll watch it =]
[07:04] <Strangerke|work> :)
[07:06] --> freanux joined #scummvm.
[07:15] --> GitHub18 joined #scummvm.
[07:15] <GitHub18> [scummvm] criezy pushed 1 new commit to master: https://git.io/vQ9EK
[07:15] <GitHub18> scummvm/master b9358b9 wreednumero2: I18N: Update translation (German)...
[07:15] GitHub18 (GitHub18@192.30.252.40) left #scummvm.
[07:20] --> waltervn joined #scummvm.
[07:20] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[07:40] --> Joefish joined #scummvm.
[07:40] #scummvm: mode change '+v Joefish' by ChanServ!ChanServ@services.
[07:55] --> girafe joined #scummvm.
[08:18] <-- LittleToonCat left irc: Remote host closed the connection
[08:18] <-- snover left irc: Ping timeout: 248 seconds
[08:29] --> jamm joined #scummvm.
[08:29] <-- jamm left irc: Changing host
[08:29] --> jamm joined #scummvm.
[08:31] --> snover joined #scummvm.
[08:31] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[08:34] <-- jamm left irc: Ping timeout: 260 seconds
[09:01] <-- girafe left irc: Read error: Connection reset by peer
[09:08] --> criezy|Work joined #scummvm.
[09:08] #scummvm: mode change '+o criezy|Work' by ChanServ!ChanServ@services.
[09:09] <-- snover left irc: Ping timeout: 260 seconds
[09:12] --> omer_mor_ joined #scummvm.
[09:12] <-- kurtwr left irc: Read error: Connection reset by peer
[09:14] --> snover joined #scummvm.
[09:14] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[09:14] <-- omer_mor left irc: Ping timeout: 260 seconds
[09:16] <-- waltervn left irc: Ping timeout: 258 seconds
[09:21] <-- ST1 left irc: Ping timeout: 240 seconds
[10:05] --> t0by joined #scummvm.
[10:05] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[10:44] --> ajax16384 joined #scummvm.
[10:44] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[10:53] --> jamm joined #scummvm.
[10:53] <-- jamm left irc: Changing host
[10:53] --> jamm joined #scummvm.
[10:57] <-- jamm left irc: Ping timeout: 240 seconds
[11:14] --> user9 joined #scummvm.
[11:29] <-- exmensa left irc: Ping timeout: 240 seconds
[11:39] --> exmensa joined #scummvm.
[12:42] --> Axy joined #scummvm.
[12:42] <-- Axy left irc: Changing host
[12:42] --> Axy joined #scummvm.
[12:49] --> deep-book-gk_ joined #scummvm.
[12:49] deep-book-gk_ (1wm_su@94.242.252.58) left #scummvm.
[12:55] --> jamm joined #scummvm.
[13:12] --> Littleboy joined #scummvm.
[13:12] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[13:57] --> ST joined #scummvm.
[13:57] #scummvm: mode change '+o ST' by ChanServ!ChanServ@services.
[14:00] --> kurtwr joined #scummvm.
[14:10] --> waltervn joined #scummvm.
[14:10] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[15:11] --> Henke37 joined #scummvm.
[15:44] --> Begasus joined #scummvm.
[16:02] --> Begas_VBox joined #scummvm.
[16:13] <-- Strangerke|work left irc: Quit: Page closed
[16:18] --> LittleToonCat joined #scummvm.
[16:20] <-- ajax16384 left irc: Read error: Connection reset by peer
[16:27] <_sev_> yes!
[16:27] --> GitHub141 joined #scummvm.
[16:27] <GitHub141> [scummvm] sev- closed pull request #968: SLUDGE: Add Sludge engine code (master...newWIP1-1) https://git.io/vQ6WJ
[16:27] GitHub141 (GitHub141@192.30.252.42) left #scummvm.
[16:27] --> GitHub146 joined #scummvm.
[16:27] <GitHub146> [scummvm] sev- pushed 118 new commits to master: https://git.io/vQH2R
[16:27] <GitHub146> scummvm/master 94439e2 yinsimei: SLUDGE: Add first classes in scummvm to detect .slg game file
[16:27] <GitHub146> scummvm/master 219044a yinsimei: SLUDGE: Add sludge files and make it compile
[16:27] <GitHub146> scummvm/master c4a5af2 yinsimei: SLUDGE: Add namespace
[16:27] GitHub146 (GitHub146@192.30.252.45) left #scummvm.
[16:31] --> WooShell joined #scummvm.
[16:35] <WooShell> meow =^.^=
[16:36] <wjp> Simei: congratulations :-)
[16:37] <Simei> :D
[16:37] <Simei> Thanks!
[16:40] --> Strangerke_ joined #scummvm.
[16:42] <-- Strangerke left irc: Ping timeout: 268 seconds
[16:42] Nick change: Strangerke_ -> Strangerke
[16:45] --> GitHub146 joined #scummvm.
[16:45] <GitHub146> [scummvm] sev- pushed 2 new commits to master: https://git.io/vQHr3
[16:45] <GitHub146> scummvm/master e36f3ad Eugene Sandulenko: SLUDGE: int32_t -> int32
[16:45] <GitHub146> scummvm/master 50f1204 Eugene Sandulenko: SLUDGE: Fix line endings
[16:45] GitHub146 (GitHub146@192.30.252.45) left #scummvm.
[16:48] <t0by> Yay!
[16:49] --> GitHub190 joined #scummvm.
[16:49] <GitHub190> [scummvm] sev- pushed 2 new commits to master: https://git.io/vQHrV
[16:49] <GitHub190> scummvm/master 2b0cbd3 Eugene Sandulenko: SLUDGE: Fix warning
[16:49] <GitHub190> scummvm/master 469d177 Eugene Sandulenko: SLUDGE: Fix more warnings
[16:49] GitHub190 (GitHub190@192.30.252.40) left #scummvm.
[16:50] --> m_kiewitz joined #scummvm.
[16:50] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[16:58] <ScummBot_> Port build status changed with 94ddf526: Failure: master-psp, master-mingw-w32
[17:07] <_sev_> that has been addressed
[17:07] <_sev_> (hopefully)
[17:17] <-- jamm left irc: Ping timeout: 240 seconds
[17:20] --> Farmboy0 joined #scummvm.
[17:20] <-- Farmboy0 left irc: Changing host
[17:20] --> Farmboy0 joined #scummvm.
[17:29] --> Strangerke_ joined #scummvm.
[17:31] <-- Strangerke left irc: Ping timeout: 255 seconds
[17:31] Nick change: Strangerke_ -> Strangerke
[17:46] <ScummBot_> Port build status changed with 94ddf526: Failure: master-amigaos4
[17:49] <-- criezy|Work left irc: Quit: Page closed
[17:50] --> ajax16384 joined #scummvm.
[17:50] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[17:53] --> ny00123 joined #scummvm.
[17:54] <Simei> what is this one about? How can I know it?
[17:56] <wjp> still the int32_t one
[17:56] <snover> Simei: http://buildbot.scummvm.org/console?category=changes;category=master;branch=master the buildbot has not caught up yet to _sev_s commits
[17:59] <Simei> Ah, ok, I see
[17:59] <Simei> Thanks
[18:03] <ScummBot_> Port build status changed with 94ddf526: Failure: master-wii
[18:09] <snover> well, i really was hopeful that there was no calclists issue, but now i think i am flip-flopping back to the idea that were still slightly off somewhere. these frames in the basement where the screen flickers have only view 3171 in the draw list, whereas in ssci, pic 20200 is also in the draw list.
[18:10] <-- Strangerke left irc: Ping timeout: 240 seconds
[18:21] <ScummBot_> Port build status changed with 94ddf526: Failure: master-mingw-w64-cplusplus11
[18:27] <m_kiewitz> uh oh
[18:28] --> girafe joined #scummvm.
[18:56] <ScummBot_> Port build status changed with 94ddf526: Failure: master-psp2
[19:19] --> Strangerke joined #scummvm.
[19:19] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[19:39] <ScummBot_> Port build status changed with 94ddf526: Failure: master-gamecube
[19:49] <_sev_> only 3 ports to do
[20:05] <-- Strangerke left irc: Ping timeout: 255 seconds
[20:08] <ScummBot_> Port build status changed with 94ddf526: Failure: master-mingw-w64
[20:08] <_sev_> let's see
[20:09] Nick change: _sev_ -> _sev
[20:12] --> Strangerke joined #scummvm.
[20:22] --> omer_mor joined #scummvm.
[20:23] <-- Begasus left irc: Ping timeout: 240 seconds
[20:24] <-- omer_mor_ left irc: Ping timeout: 246 seconds
[20:36] --> Begasus joined #scummvm.
[20:41] --> GitHub142 joined #scummvm.
[20:41] <GitHub142> [scummvm] sev- pushed 1 new commit to master: https://git.io/vQQmk
[20:41] <GitHub142> scummvm/master 8fde046 Eugene Sandulenko: SLUDGE: Make UTF-8 code portable
[20:41] GitHub142 (GitHub142@192.30.252.40) left #scummvm.
[20:43] <-- Strangerke left irc: Ping timeout: 240 seconds
[20:45] --> omer_mor_ joined #scummvm.
[20:45] --> Strangerke joined #scummvm.
[20:47] <-- omer_mor left irc: Ping timeout: 240 seconds
[20:50] <ScummBot_> Port build status changed with 469d1777: Success: master-psp, master-mingw-w32, master-amigaos4, master-wii, master-mingw-w64-cplusplus11, master-gamecube, master-mingw-w64
[20:53] <-- ajax16384 left irc: Read error: Connection reset by peer
[21:06] <_sev> \o/
[21:06] --> TMM joined #scummvm.
[21:06] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[21:06] <snover> nice work Simei, _sev!
[21:06] <Simei> :D
[21:06] <Simei> That's great
[21:09] <snover> ha ha HA FML ok. the remaining basement bug is *still* a broken priorities bug
[21:09] <-- Axy left irc: Ping timeout: 240 seconds
[21:09] <snover> i forgot that Plane::calcLists used a different algorithm that checked only the priority & creation order, which would fail the same as ScreenItem::operator< when the creation order is inverted from what happens most of the time in SSCI
[21:09] <ScummBot_> Port build status changed with 8fde0463: Success: master-psp2
[21:11] <snover> depending upon how memory is set up in SSCI, the same problem may or may not occur
[21:12] <snover> phantasmagoria sure has a lot of these screen item ordering bugs.
[21:14] <TMM> I say we just merge dosbox
[21:14] <TMM> fuck this shit
[21:14] Action: TMM runs
[21:14] <snover> hey, i already suggested that!
[21:16] <snover> now, i guess i need to decide whether to change the code in Plane::calcLists to use ScreenItem::operator< or not, and then write a script patch to stop the game from doing broken things like this.
[21:16] <TMM> :D
[21:18] <snover> the difference between what is done in Plane::calcLists and what is in ScreenItem::operator< is that the latter one also sorts on position y+z, not just on priority & creation order
[21:19] <wjp> argh, again that < ?
[21:19] <snover> yup.
[21:19] Action: wjp sighs :/
[21:20] <TMM> I suggest we bend some sierra person over a barrel until they give us sources
[21:20] <TMM> in the long term this is clearly the way forward
[21:21] <snover> that wouldnt help here unfortunately. the problem here is that SSCI does a totally bogus check when comparing screen items with identical priorities
[21:23] <snover> it compares memory handle IDs, which can be reused, so it is not deterministic and you can make things sort in the wrong order just by getting unlucky and reusing a handle ID that is below another screen item's handle ID
[21:25] <snover> we do something a little smarter than that, and give all screen items monotonically increasing creation IDs, which at least provides a guarantee that an item created later sorts higher than an item created earlier, and is usually what the game scripts expect to happen.
[21:26] <m_kiewitz> snover: so it just works by accident?
[21:26] <snover> but, rarely, like in this room, the game only works correctly if a newer item sorts lower than an older item.
[21:26] <snover> m_kiewitz: yep!
[21:29] --> GitHub112 joined #scummvm.
[21:29] <GitHub112> [scummvm] yinsimei pushed 2 new commits to master: https://git.io/vQQWU
[21:29] <GitHub112> scummvm/master b60aa5e Simei Yin: SLUDGE: fix warnings
[21:29] <GitHub112> scummvm/master cce0aac Simei Yin: SLUDGE: get random number
[21:29] GitHub112 (GitHub112@192.30.252.42) left #scummvm.
[21:33] <snover> so, i would really like to use the same comparison code in calcLists, since it is more correct, but i am concerned that it might start breaking games that expect a screen items y+z value to not influence draw order.
[21:37] <-- ny00123 left irc: Quit: Leaving
[21:38] --> Strangerke_ joined #scummvm.
[21:39] <snover> oh, well, the good news is that doing that actually breaks rendering fairly spectacularly, so i guess i have the answer to my dilemma.
[21:41] <-- Strangerke left irc: Ping timeout: 260 seconds
[21:41] Nick change: Strangerke_ -> Strangerke
[21:51] --> Axy joined #scummvm.
[21:51] <-- Axy left irc: Changing host
[21:51] --> Axy joined #scummvm.
[21:58] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[21:59] <-- Begas_VBox left irc: Ping timeout: 246 seconds
[21:59] <-- Begasus left irc: Ping timeout: 255 seconds
[22:11] --> Begasus joined #scummvm.
[22:17] <snover> getting close to bug 10000
[22:17] <snover> 43 to go!
[22:21] <-- Farmboy0 left irc: Read error: Connection reset by peer
[22:26] <-- 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.

[22:38] <snover> m_kiewitz: what are your thoughts if I were to suggest turning ScriptPatcher into ResourcePatcher so it could be used to patch anything? Does this sound possible/good?
[22:39] <snover> I haven't looked into it at all, just thinking about how I might fix one of these bugs caused by a palette issue inside a view resource
[22:41] <snover> I don't know if there are enough of these cases to make the effort for this
[22:41] <m_kiewitz> that makes not really sense
[22:42] <m_kiewitz> i guess an additional resource patcher could make sense in a way
[22:42] <m_kiewitz> the script patcher is fully designed to patch executeable code
[22:43] <m_kiewitz> sure, you could use the very basic logic behind it on all resources, but then why not create a separate resource patcher for that purpose
[22:43] <m_kiewitz> still what exactly is wrong with the view?
[22:44] <m_kiewitz> also how often do we need this? I patch in "dual" views for dual mode, but that's even extending the view
[22:44] <m_kiewitz> so maybe special code to fix exactly that view could make more sense
[22:45] <snover> On phone, one moment please...
[22:45] <m_kiewitz> np
[22:47] <snover> https://bugs.scummvm.org/ticket/9788 the game adds a view with top priority so its palette overwrites the palette of the background, turning black into light brown
[22:48] <snover> This causes a 3 frame flash of bad palette until the game accidentally fixes it by resubmitting palettes of the dirty rects
[22:49] <m_kiewitz> wait, the problem is top priority?
[22:49] <m_kiewitz> ah btw. that sounds like the gabriel knight 1 issue that i see all the time
[22:50] <m_kiewitz> some part of the background changes from real black to really dark grey and then back again
[22:50] <snover> Embedded palettes are submitted in the draw order, back to front
[22:50] <m_kiewitz> so if the view wasn't top priority it would work?
[22:50] <m_kiewitz> is the view actually needed at all?
[22:51] <m_kiewitz> why not fix the script code that adds it?!
[22:51] <m_kiewitz> (I have no idea about this specific issue, so if I ask something dumb, just say so :P)
[22:51] <snover> The priority can't be lower than the background or it would draw under the background
[22:52] <m_kiewitz> so there is only that single view?
[22:52] <snover> There is the ego, the chest, and the background, iirc.
[22:52] <m_kiewitz> so if you give that view the lowest priority (one above background), would that fix it?
[22:53] <-- waltervn left irc: Quit: Leaving
[22:53] <m_kiewitz> and what is that view anyway? is it actually neded in the scene?
[22:53] <m_kiewitz> ah chest box
[22:54] <-- Axy left irc: Read error: Connection reset by peer
[22:54] --> Axy joined #scummvm.
[22:54] <-- Axy left irc: Changing host
[22:54] --> Axy joined #scummvm.
[22:54] <m_kiewitz> where is that chest btw? I can't see any chest on the supplied screenshot
[22:54] <snover> Another object would need to have a higher priority, be visible in a dirty rect, and have the correct palette to override the bad value in the chest palette
[22:54] <snover> It is at the bottom right
[22:55] <-- Axy left irc: Read error: Connection reset by peer
[22:55] <snover> At the right edge above the carpet
[22:55] <m_kiewitz> hard to see
[22:56] <snover> It is a silly non interactive view that does nothing of consequence
[22:56] <snover> So one option I mentioned in the ticket is to just never draw it
[22:56] <m_kiewitz> and there is no way to disable the palette altogether via script code?
[22:57] <m_kiewitz> i mean the palette submit of the view palette
[22:57] <snover> Nope, the script cannot control ignoring palettes of a screen item
[22:58] <m_kiewitz> hmm
[22:58] <m_kiewitz> that other glitch is caused by different MemIDs?
[22:58] <m_kiewitz> what's exaclty the difference for that one?
[22:59] <m_kiewitz> did SSCI reuse those?
[22:59] <m_kiewitz> and we don't?
[22:59] <snover> See the comment in screen_item32.h in operator< for details
[23:00] <snover> In the basement the bad view is not visible, so there is no reason to even think about drawing it
[23:01] --> Mia joined #scummvm.
[23:01] <-- Mia left irc: Changing host
[23:01] --> Mia joined #scummvm.
[23:01] <m_kiewitz> hmmm, is there no way to reimplement Sierra's MemID reuse algo?
[23:02] <m_kiewitz> what exactly did they use? the pointer itself? or did they assign a separate ID of some sort?
[23:02] <snover> No, it is impossible and a bad idea since it randomly breaks in ssci too
[23:02] <m_kiewitz> urgh
[23:02] <m_kiewitz> and that one is the only known case for now?
[23:03] <m_kiewitz> that doesn't work right?
[23:03] <snover> No, there are many broken places in ssci which are fixed in scummvm by doing what we are doing now with the monotonic insert id
[23:03] <m_kiewitz> yes, i meant the only known issue in scummVM of course
[23:04] <m_kiewitz> i already read the comment stating various things that break
[23:04] <m_kiewitz> (in SSCI)
[23:04] <m_kiewitz> for the latter I think one could remove that specific view to fix it, right?
[23:04] <snover> OK, sorry, I misunderstood the question :)
[23:04] <m_kiewitz> and the other one, well, for now I would fix that view manually by code
[23:05] <snover> Yes my plan is to remove the view from the script if that is possible
[23:05] <m_kiewitz> the view unpacking code of sci16 makes changing up views in such a way easy
[23:06] <snover> I don't know if it is some thing that only shows up in chapter 7 or if it is only broken in chapter 7 but draws in earlier chapters on screen
[23:06] <m_kiewitz> for sci32 i guess it's going to be a bit more hack-y, idk
[23:07] <m_kiewitz> phantasmagoria 1 was only released on x86, right?
[23:07] <m_kiewitz> that at least makes it a bit easier, no bye swapping or stuff like that
[23:07] <snover> No, 68k Mac exists
[23:07] <m_kiewitz> oh, then our wiki is incorrect
[23:08] <m_kiewitz> does anyone own that specific version?
[23:08] <snover> (No Mac version of anything is supported in SCI32)
[23:09] <m_kiewitz> well for now :p
[23:09] <m_kiewitz> but that makes it easier for now I guess
[23:09] <m_kiewitz> still the localized versions will have to be checked
[23:10] <m_kiewitz> is color 0 or color 255 broken?
[23:10] <m_kiewitz> it looks like color 0
[23:10] <snover> Hacking in a fix for the palette is trivial, it's just that I was thinking if there are other times where resources need patching that a generalised mechanism would be better
[23:10] <m_kiewitz> i can't think of many resources that are getting patched
[23:10] <snover> It is 138 or something, I'll have to check my notes when I am back at my desk
[23:11] <snover> I think I saw some codename iceman hack when I was looking earlier
[23:11] <m_kiewitz> i think i had to change up a few line draws for dithering being disabled to work properly
[23:12] <m_kiewitz> but changing the drawing code is way easier, you can't really just use signatures for that
[23:12] <m_kiewitz> i mean in a way you could, but it would get way more complicated
[23:13] <m_kiewitz> so for view patching or background image patching, a completely different patcher would make way more sense
[23:13] <m_kiewitz> like a patch that basically says: "change color X of palette for view Y"
[23:13] <m_kiewitz> or "change loop count of view Y"
[23:14] <m_kiewitz> well i don't really see a point, because resource patching is quite rare, especially compared to script patching
[23:14] <m_kiewitz> in case there would be way more issues like this...
[23:15] <m_kiewitz> or for picture drawing "do not draw lines in this way when color X is selected"
[23:16] <snover> bbiab, I'll check logs when I return
[23:16] <m_kiewitz> oh and that qfg3 demo view patch is actually an ugly hack, which shouldn't be there in the first place
[23:23] <-- t0by left irc: Ping timeout: 260 seconds
[23:40] <-- girafe left irc: Quit: Leaving
[23:43] --> Littleboy joined #scummvm.
[23:43] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[23:49] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[00:00] --- Fri Jul 14 2017