[Back to Index]

[00:31] --> _sev_ joined #scummvm.
[00:31] #scummvm: mode change '+o _sev_' by ChanServ!ChanServ@services.
[00:31] <Lightkey> funny, wanted to take a look what I had on DotEmu before it closes at the end of the month and it took my browser with it while loading
[00:31] <-- _sev left irc: Ping timeout: 260 seconds
[00:32] <Lightkey> full page video background probably explains it
[00:38] <Simei> t0by: Sorry, fallen asleep just now. :p
[00:41] <snover> i think i need another pair of eyes on this. i cannot for the life of me see how the original interpreter succeeds here but we dont
[00:41] <Simei> t0by: Yes, that's why I thought slg files should be pushed into git as well
[00:45] <Simei> t0by: I'll do the memory check. But I added an output for png load failure of PNGDecoder, it doesn't print it. I will look into il later.
[00:49] --> OmerMor joined #scummvm.
[00:49] <-- omer_mor_ left irc: Ping timeout: 260 seconds
[00:52] <snover> OH.
[00:52] <snover> THERE YOU ARE.
[00:52] <snover> way over in kLock.
[00:54] <snover> code was added so that when an audio resource is locked it gets copied into memory.
[01:14] <snover> and now i think i shall collide once again with the resource manager&
[01:22] <snover> the resource manager which does not seem to be recording resource locks in save games?&
[01:38] <snover> hmmmm.
[01:39] <snover> serialisation of locks looks like it may have shown up somewhere after SQ6 but before/at Torin
[01:40] --> DominusExult joined #scummvm.
[01:41] <snover> i wonder if this has anything to do with Torins save games sometimes missing correct palette restoration
[01:43] <-- Dominus left irc: Ping timeout: 260 seconds
[01:43] Nick change: DominusExult -> Dominus
[02:04] --> GitHub74 joined #scummvm.
[02:04] <GitHub74> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vHcpf
[02:04] <GitHub74> scummvm/master a4fca68 Paul Gilbert: TITANIC: Fix placement of stars in photo starfield
[02:04] GitHub74 (GitHub74@192.30.252.45) left #scummvm.
[02:25] --> GitHub140 joined #scummvm.
[02:25] <GitHub140> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vHcpE
[02:25] <GitHub140> scummvm/master 77703ed Paul Gilbert: TITANIC: Renamed CStarControlSub7 to CStarMarkers
[02:25] GitHub140 (GitHub140@192.30.252.41) left #scummvm.
[03:05] --> GitHub40 joined #scummvm.
[03:05] <GitHub40> [scummvm] criezy pushed 1 new commit to master: https://git.io/vHch0
[03:05] <GitHub40> scummvm/master a614638 Thierry Crozat: I18N: Regenerate translations data file
[03:05] GitHub40 (GitHub40@192.30.252.42) left #scummvm.
[03:43] --> GitHub143 joined #scummvm.
[03:43] <GitHub143> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vHcjQ
[03:43] <GitHub143> scummvm/master 4358712 Paul Gilbert: TITANIC: Fix copying star positions selected in photo
[03:43] GitHub143 (GitHub143@192.30.252.40) left #scummvm.
[03:44] <tsoliman> snover: are you opposed to me marking all the SCI32 "Blinding you with SCIence" games as ADGF_TESTING (instead of UNSTABLE)?
[03:46] <tsoliman> nevermind .. I thought QFG4 was on the list :(
[03:46] <tsoliman> they are already TESTING
[03:53] <snover> yeah, that one is not ready.
[04:03] --> GitHub40 joined #scummvm.
[04:03] <GitHub40> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vHCe4
[04:03] <GitHub40> scummvm/master 496839d Paul Gilbert: TITANIC: Merge incorrectly duplicated field in CPhotoCrosshairs
[04:03] GitHub40 (GitHub40@192.30.252.45) left #scummvm.
[04:13] <-- dreammaster left irc:
[04:50] --> omer_mor_ joined #scummvm.
[04:54] <-- OmerMor left irc: Ping timeout: 260 seconds
[05:17] <-- Mia left irc: Read error: Connection reset by peer
[05:19] --> Mia joined #scummvm.
[05:19] <-- Mia left irc: Changing host
[05:19] --> Mia joined #scummvm.
[05:21] <-- Lightkey left irc: Ping timeout: 245 seconds
[05:33] --> Lightkey joined #scummvm.
[06:00] --> Strangerke_ joined #scummvm.
[06:02] <-- Strangerke left irc: Ping timeout: 246 seconds
[06:02] Nick change: Strangerke_ -> Strangerke
[06:03] <Simei> Morning!
[06:05] <Simei> _sev_, t0by: I've pushed the slg into the git that you can test. Also, I find a code formatting shortcut that I didn't found. I did a mass code formatting and pushed it. :)
[06:10] <Simei> So, I think there won't be mess in the future commits :)
[06:10] <Simei> Sorry for last unreadable commit, t0by :)
[06:21] <Simei> _sev_, t0by: The memory position for png file reading is good, both at 2878.
[06:38] <-- girafe left irc: Read error: Connection reset by peer
[06:46] <-- Mia left irc: Read error: Connection reset by peer
[06:47] --> m_kiewitz joined #scummvm.
[06:47] <-- m_kiewitz left irc: Changing host
[06:47] --> m_kiewitz joined #scummvm.
[06:47] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[06:48] --> Mia joined #scummvm.
[06:48] <-- Mia left irc: Changing host
[06:48] --> Mia joined #scummvm.

[06:50] <m_kiewitz> https://www.amazon.co.uk/gp/offer-listing/B004E10QCI/ref=dp_olp_all_mbc?ie=UTF8&condition=all
[06:53] <m_kiewitz> wjp, DrMcCoy etc: ^^^
[06:55] <DrMcCoy> I'd kill myself before I'd buy a Blu-ray
[06:56] <Lightkey> I was about to ask, does it use BD+? :-P
[06:56] <Harekiet> seems cheap though :)
[07:05] --> girafe joined #scummvm.
[07:05] <Simei> _sev, t0by: I have a question concerning, endian in file reading, in sludge, there is an operation dealing with big endian and little endian to read float, but only with float. So do I need to pass my initial file function to something like SeekableReadStreamEndian ? How do I pass the file and the endian option at the same time ?
[07:08] <logix> m_kiewitz: holy phaser set to stun, batman - I don't even have a BD player or drive and I'm about to buy it for that price
[07:22] <_sev_> Simei: SeekableReadStreamEndian is used when you have same file in both BE and LE formats, e.g. it happens sometimes with Mac/Win releases
[07:22] <_sev_> Simei: so, that class lets you avoid writing countless if (platform = Mac) { use BE } else { use LE }
[07:23] <_sev_> when you know about format of a specific field and it is always like that, you'd better use explicit methods which specify the endianness
[07:26] <Simei> _sev_: Ok, I see, so I keep the original methods that deal with it then. Thanks!
[07:31] <Simei> _sev_, t0by: Another problem, I believe the png is correctly loaded with PNGDecoder using the code:
[07:33] --> waltervn joined #scummvm.
[07:33] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[07:33] <Simei> https://www.irccloud.com/pastebin/niuizwiv/
[07:33] <Simei> The display in a while loop with :
[07:34] <Simei> https://www.irccloud.com/pastebin/94OI24MH/
[07:35] <Simei> _sev_, t0by: Did I miss somthing anywhere? It's just a sum up of what I did.
[07:36] <Simei> Also, with "initGraphics(640, 480, false);" before the loop
[07:45] <-- LittleToonCat left irc: Remote host closed the connection
[07:45] <_sev_> Simei: yes, that is correct, except one important thing
[07:45] <_sev_> after updateScreen() you need to call g_system->delayMillis(10);
[07:45] <_sev_> so you yield the CPU
[07:57] <Simei> _sev_: That makes sense, but still nothing appears. Do you have any idea about it?
[07:58] <_sev_> Simei: how did you initialize screen?
[07:58] <_sev_> e.g. what are the parameters to the initGraphics() call?
[07:59] <_sev_> is it true color or 8-bit?
[07:59] <Simei> 640 480 false
[08:00] <_sev_> then it is 8-bit
[08:00] <_sev_> what about the original?
[08:02] <Simei> Sorry, but how to see it in code?
[08:02] <_sev_> no idea. it's Sludge
[08:02] <_sev_> is it palettized or truecolor?
[08:03] <_sev_> and in what format is its graphics? Does it have palettes or only 16 or 32-bit pixels?
[08:04] <Simei> The texture is a table of uint8
[08:05] <Simei> It has a palette with un signed char as attributes
[08:05] <_sev_> okay, then it is 8-bit
[08:05] <_sev_> then you initialize the screen properly
[08:05] <Simei> But not used for background
[08:05] <Simei> Ah, ok
[08:05] <_sev_> but perhaps you're missing to set the palette in ScummVM
[08:06] <_sev_> which makes every color [0, 0, 0]
[08:07] <Simei> So I should set palette first.
[08:07] <Simei> Let me see where is it.
[08:22] <Simei> _sev_, t0by: It appears! That's a good start for today! :DD
[08:23] <-- girafe left irc: Read error: Connection reset by peer
[08:23] <_sev_> \o/
[08:23] <Simei> I'll push it and leave now. See you later today. :D
[08:23] <Simei> Thanks a lot for the help!!
[08:27] <-- Harekiet left irc: Ping timeout: 258 seconds
[08:30] <Simei> _sev_, t0by: https://github.com/yinsimei/scummvm/commits/newWIP1-1
[08:30] <wjp> first visible pixels? Hurray :-)
[08:30] <Simei> XD
[08:40] <Deledrius> Wow, that Trek set is Region-Free, too.
[08:47] --> t0by joined #scummvm.
[08:47] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[08:50] --> Harekiet joined #scummvm.
[08:52] <t0by> Simei: great news!
[08:52] <t0by> Now it doesn't segfault for me anymore
[08:52] <t0by> No idea what you changed
[08:52] <t0by> let me see
[08:54] <t0by> Oh, right, the palette
[08:54] <t0by> I'm very puzzled
[08:55] <t0by> Last night it wasn't loading correctly.
[08:55] <t0by> Simei: at this point consider moving the body of loadPng inside its caller
[08:56] <t0by> In particular since it does more than loading, you are mucking with globals, etc
[08:56] <t0by> Simei: great job!
[08:56] <t0by> Hum
[09:04] <t0by> I'm *super* confused
[09:04] <t0by> 9e5287b does load cleanly for me this morning.
[09:05] <t0by> I literally went to sleep and came back now, haven't touched anything
[09:08] <TMM> #magic
[09:08] <TMM> We all know that ScummVM is a magical piece of software! :D
[09:09] <wjp> t0by: was it segfaulting, or failing that assertion that you copy-pasted?
[09:10] <t0by> wjp: it was failing to load the PNG. As a result, the surface was not initialized and passing surface.getPixels() to a drawing method would make ev erything explode.
[09:10] <t0by> Simei: btw remember to check for that onw
[09:10] <t0by> Now it's loading.
[09:11] <t0by> I feel super silly.
[09:11] <wjp> sure you weren't in the wrong directory again? ;-)
[09:11] <t0by> No, thsi time I checked :P
[09:11] <wjp> (by the way, a failed assertion is not a segfault)
[09:12] <t0by> It's not, but the cause was still a nullptr being passed around.
[09:13] <t0by> (Did I say segfault? I did. I meant failing assert(), sorry.)
[09:15] <t0by> And yes, it's the same exact slg
[09:16] <t0by> ?_?
[09:20] <wjp> ah well, I'm sure there'll be a lot more testing of sludge in the future to catch any non-reproducible issues again
[09:21] <t0by> Apparently, my grandmother was right about a good night's sleep being able to fix anything.
[09:36] --> _sev|work joined #scummvm.
[09:36] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[10:11] <-- TMM left irc: Quit: Ex-Chat
[10:46] <-- Harekiet left irc: Read error: Connection reset by peer
[10:57] --> Harekiet joined #scummvm.
[11:36] <t0by> Simei: I was looking at your commit history and it's rather wacky
[11:37] <t0by> I took the liberty of rebasing your WIP branch: https://github.com/tobiatesan/scummvm/tree/rebase_WIP
[11:37] <t0by> I tried to screenrecord it, but apparently my X is broken
[11:37] <t0by> ...and I can't be bothered to do that again
[11:38] <t0by> Simei: Note how I split your "add namespace" commit into two commits, one that actually adds the namespace and one that does everything else you did in that ocmmit
[11:38] <t0by> Note also how I squashed your "add namespace for file.h" commit into "add namespace"
[11:47] <t0by> Simei: or how I put adding of debug channels and detection stuff into their own commit.
[11:50] <t0by> Simei: I'm sorry I failed miserably at screenrecording it, next time we'll go through the process together
[12:15] --> Joefish joined #scummvm.
[12:15] #scummvm: mode change '+v Joefish' by ChanServ!ChanServ@services.
[12:19] <t0by> Joefish: hello
[12:20] --> TMM joined #scummvm.
[12:20] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[12:22] <Joefish> hey t0by
[13:21] --> Littleboy joined #scummvm.
[13:21] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[13:25] <-- TMM left irc: Read error: Connection reset by peer
[13:27] --> TMM joined #scummvm.
[13:27] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[13:41] <Strangerke|work> Oh god, dotemu store is closing! :(
[13:48] <t0by> Strangerke|work: what are the consequences?
[13:48] <_sev|work> no surprise
[13:49] <Strangerke|work> t0by: You bought games on dotemu? Download quickly
[13:49] <t0by> Never.
[13:50] <t0by> But what are the possible consequences for the retrogaming and community and economy, as a whole?
[13:50] <Strangerke|work> And it seems they plan to sell those games or steam, or something... Damn, I hate Steam so much!
[13:50] <t0by> lol
[13:51] <Strangerke|work> t0by: afaik, it was the only place to buy tony tough, for example
[13:52] <t0by> OH NO
[13:52] <t0by> That's a legitimate catastrophe, then.
[13:52] <t0by> Strangerke|work: who owns that anyway nowadays?
[13:52] <t0by> IIRC Nayma made one game.
[13:53] <Strangerke|work> still the same
[13:53] <t0by> Oh, well, good news is -- it's on Steam.
[13:53] <Strangerke|work> You call it a good news?
[13:53] <Strangerke|work> Being obliged to have a working connection and a working steam server to be able to do anything?
[13:54] <t0by> Honestly, I don't know the first thing about those sellers, you'll have to pry the CD from my cold dead hands.
[13:54] <t0by> Oh.
[13:54] <t0by> Then it's, I suppose, bad news.
[13:54] <Strangerke|work> so the day Steam closes his shop, you'll have a bunch of useless encrypted bull shit for the price you paid? Is that really a good news?
[13:54] <t0by> Whaaaaaaaat.
[13:55] Action: t0by checks Tony Tough CD is still where he left it.
[13:56] <t0by> An elegant weapon, for a more civilized time.
[13:56] --> _sev|work_ joined #scummvm.
[13:56] <-- _sev|work_ left irc: Changing host
[13:56] --> _sev|work_ joined #scummvm.
[13:56] #scummvm: mode change '+o _sev|work_' by ChanServ!ChanServ@services.
[13:57] <logix> Strangerke|work: afaik with gog you download a cd image of the game you bought and that's it, no online checks or anythign alike, right? did (well, "does") it work the same way with dotemu?
[13:57] <Strangerke|work> yes
[13:57] <-- _sev|work left irc: Read error: Connection reset by peer
[13:58] <logix> (I'm a "physical medium" person, I've never used any of these stores)
[13:58] <t0by> ^
[13:59] <logix> I suppose steam is better than nothing, but that's not a very high standard...
[14:01] <t0by> Oh
[14:03] <t0by> Does create_project need any special help in recognizing a new engine?
[14:09] <t0by> Uh, no, it only lists directories in engines/... :|
[14:11] <t0by> O_o
[14:11] Action: t0by curses in italian
[14:14] --> Begasus joined #scummvm.
[14:15] <t0by> ah
[14:16] <criezy> Sadly Tony Tough is not available on GoG, only on Steam. And the Gobliins pack on GoG is English only while the one on Dotemu used to be multilingual.
[14:16] <criezy> As far as I know these are the only two consequences for us of Dotemu closing.
[14:16] <criezy> (this is still a bit sad)
[14:18] <Strangerke|work> criezy: gobliiins is multilingual
[14:18] <criezy> On GoG?
[14:18] <t0by> criezy: do you have any idea why my create_project.exe might be oblivious to engines/sludge ?
[14:19] <Mataniko> t0by i had some issues with line breaks/extra line at the end of configure.engine
[14:19] <Mataniko> make a copy of an existing one and just edit it
[14:20] <t0by> ah
[14:20] <criezy> Strangerke|work: I just checked since I have it in my GoG library and they only have the English version
[14:20] <t0by> extra line missing indeed
[14:20] <t0by> thanks
[14:21] <criezy> t0by: if you have further questions on create_projects I might not be the best person to ask.
[14:21] <criezy> I did have a quick look a few months back when I wanted to build ScummVM for my iPad, but I don't otherwise use it.
[14:21] <t0by> Well, I saw your name in the log and... it worked.
[14:21] <t0by> Thanks.
[14:23] <Strangerke|work> criezy: You mean that it's detected as english only by ScummVM, really?
[14:25] --> Begas_VBox joined #scummvm.
[14:26] <Deledrius> FYI for the collectors: it looks like Al Lowe is selling his personal copy of LSL collection www.ebay.com/itm/182595224128
[14:29] <criezy> Strangerke|work: just tried with a fresh download and it detected as the multilingual game (EFIGS). But if I select French, I still get the game in English...
[14:32] <t0by> _sev_, Simei: I can confirm Sludge doesn't build on Win32.
[14:33] <t0by> That's due to Windows-specific stuff inside #ifdef _WIN32 that got broken
[14:33] <t0by> and, frankly, I would not bother with it, since we'll eventually going to replace all of it with osystem stuff.
[14:34] <Strangerke|work> criezy: it's weird
[14:34] <t0by> Unless somebody really cares about using the One True IDE for working on it...
[14:34] <Strangerke|work> true IDE? => MSVC?
[14:34] Action: Strangerke|work hides
[14:35] <t0by> Clearly.
[14:47] <t0by> In fact we might as well zap all those #ifdef WIN32 rightaway
[14:47] <t0by> nothing good can come out of them, I guess
[14:47] <t0by> I
[14:47] <t0by> I'll have a look later
[14:48] <t0by> It's file-related stuff or... worse.
[15:02] --> Henke37 joined #scummvm.
[15:07] <-- _sev|work_ left irc: Quit: This computer has gone to sleep
[15:09] --> _sev|work joined #scummvm.
[15:09] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[15:21] <-- _sev|work left irc: Read error: Connection reset by peer
[15:21] --> _sev|work joined #scummvm.
[15:21] <-- _sev|work left irc: Changing host
[15:21] --> _sev|work joined #scummvm.
[15:21] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[15:27] <-- _sev|work left irc: Ping timeout: 245 seconds
[15:30] <-- Begas_VBox left irc: Read error: Connection reset by peer
[15:32] <-- Begasus left irc: Ping timeout: 240 seconds
[15:38] --> _sev|work joined #scummvm.
[15:38] <-- _sev|work left irc: Changing host
[15:38] --> _sev|work joined #scummvm.
[15:38] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[15:44] --> ny00123 joined #scummvm.
[15:52] <t0by> Deledrius: that's... sad.
[15:55] <t0by> Simei, _sev_: apparently this is enough to make it build under Windows: https://github.com/tobiatesan/scummvm/commit/f1e5199d337e84f54a9d66cb57e33c843ac97212
[15:56] <t0by> I'll split and polish it later
[16:02] <t0by> Simei, _sev_: notice the missing return statements
[16:02] <t0by> gcc isn't bothered, but MSVC is.
[16:05] <Simei> t0by: Thanks for the effort in making it compile! That's great.
[16:06] <Simei> I know I have a messy commit history. Do we need to rebase a lot to clear them ?
[16:07] --> LittleToonCat joined #scummvm.
[16:09] <t0by> Simei: I've done the job for you this time around.
[16:09] <t0by> Next time we have to polish up the history a bit (it's normal, it happens) we'll do it together, ok?
[16:10] <t0by> Simei: again, great work with the background!
[16:10] <t0by> Do you plan on loading the bitmaps next?
[16:11] <t0by> wjp, csnover: btw I have the vague suspicion that stuff is tga with RLE.
[16:11] <Simei> Thanks! You mean the history or the windows? The history you've cleared it? Or we'll get to arrange them together?
[16:11] --> Joefish_ joined #scummvm.
[16:11] #scummvm: mode change '+v Joefish_' by ChanServ!ChanServ@services.
[16:11] <t0by> o_O
[16:11] <-- Joefish left irc: Ping timeout: 245 seconds
[16:12] <t0by> Simei: https://github.com/tobiatesan/scummvm/tree/rebase_WIP
[16:12] <t0by> ^ i polished up your history a bit
[16:12] <t0by> https://github.com/tobiatesan/scummvm/commit/f1e5199d337e84f54a9d66cb57e33c843ac97212
[16:12] <t0by> ^ these are required to make it build on windows
[16:12] <t0by> not super important in itlse,f but those missing return calls *might* be the reason why _sev_'s clang complains
[16:13] <Simei> Ok, that must have a lot of work. Thank you very much!
[16:13] <t0by> Nah, just a matter of zapping #ifdefs away until it builds
[16:13] <t0by> Simei: so, back to the important stuff. What are you planning to tackle next?
[16:14] <Simei> Haha

[16:14] <t0by> Great!
[16:14] <t0by> Have you had a look at the one in prince?
[16:14] <Simei> As we have a way to see if it works now
[16:15] <Simei> Then try to refactor the things
[16:15] <snover> t0by: im not sure which stuff you are talking about exactly?
[16:15] <t0by> The thing with the stuff, no?
[16:15] <Simei> _sev_, t0by : When will we begin to objectify the code ?
[16:16] --> cpasjuste_ joined #scummvm.
[16:16] <t0by> snover: sorry, I'm broken. I meant to highlight dreammaster.
[16:16] <Simei> Because I'm always working on with the global variables
[16:16] <snover> thats a funny way to spell dreammaster
[16:17] <t0by> Simei: that's tolerable so far. I would wait until we get to the point we can *at least* display characters (implying; where the interpreter doesn't segfault at us).
[16:19] <Simei> _sev_, t0by : I've looked the code in prince. But still not very clear about the palettes the pixelFormat thing. Are there some docs that I can read about them?
[16:19] <t0by> Simei: eh, better ask.
[16:20] <Simei> Or in sludge it's just always 8 bit int stuff
[16:21] <Simei> _sev_, t0by: So, I'll get back here when I get stuck with a specific thing.
[16:21] Action: t0by nods
[16:21] <-- _sev|work left irc: Quit: This computer has gone to sleep
[16:22] <Simei> _sev_, t0by : Another thing is that how to stretch surface in scummvm ?
[16:22] <Simei> And is it possible to resize screen window?
[16:22] --> WooShell joined #scummvm.
[16:24] <WooShell> meow =^.^=
[16:24] <t0by> Simei: TransparentSurface has a bunch of scaling methods
[16:24] <Simei> Hello, WooShell :)
[16:25] <t0by> PLain old Surface, I don't think so...
[16:26] <Simei> Ah, ok. So we are generally using the transparentSurface instead?
[16:26] <t0by> Simei: initGraphics(height, width, ...)
[16:26] <t0by> Simei: good question.
[16:26] <t0by> Simei: why do you care right now, though?
[16:26] <t0by> not that caring is bad
[16:26] <t0by> just legitimately curious
[16:26] <t0by> you need that for backdrops as well?
[16:28] --> Huerbine joined #scummvm.
[16:28] --> ajax16384 joined #scummvm.
[16:28] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[16:29] <-- WooShell left irc: Read error: Connection reset by peer
[16:29] <wjp> t0by: tga RLE is quite different
[16:29] <Simei> That's not very important, but I think in sludge they resize the window to have a bigger background without changing the scale of it. And stretch the background at the same time. _sev_ t0by
[16:29] <-- t0by left irc: Remote host closed the connection
[16:29] --> dreammaster joined #scummvm.
[16:29] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[16:29] <Simei> Or maybe I'm wrong about it
[16:31] --> t0by joined #scummvm.
[16:31] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[16:32] <Simei> Not very important for now.
[16:32] <t0by> Simei: I think you should diregard that for now
[16:33] <t0by> (Or possibly forever, I'm not sure if that should be the engine's concern)
[16:34] <wjp> depends on if it's scaling everything or just the background
[16:34] <snover> m_kiewitz: yes, the 'day 6 police beignet timer issue' patch applies on script 230 offset 9730.
[16:34] <wjp> scaling everything is probably indeed better left to the scummvm backend
[16:35] <m_kiewitz> maybe there are multiple issues and i missed one
[16:35] <t0by> wjp: I think Simei means scaling everything, if I'm understanding correctly.
[16:37] <t0by> somaen: that you know of is there a good reason *not* to use TransparentSurface for everything?
[16:37] <t0by> Besides performance.
[16:37] <t0by> I can't remember if there was some caveat
[16:38] <t0by> I'm going afk for a little while, see you in an hour
[16:39] <t0by> Simei, context: somaen was my mentor during GSoC 2013. I worked on adding rotoscale to TransparentSurface, which he in turn had ported from the BS2.5 engine for Wintermute. Then it ended up being grafted onto Graphics::
[16:40] <t0by> I *think* you could get away with using TransparentSurface exclusively.
[16:40] <t0by> But I wouldn't concern myself with it just yet.
[16:40] <t0by> ttyl
[16:40] <-- t0by left irc: Quit: t0by
[16:41] <Simei> t0by: See you. But I get a little confused. That means it's better to use TransparentSurface or not ? :x
[16:42] <Simei> _sev_, t0by: Anyway, I'm get started on image decoder. :)
[16:43] --> t0by joined #scummvm.
[16:43] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[16:43] <t0by> Simei: TL;DR: there is probably no harm in doing so. But don't until you have a good reason to, if nothing else because it's more complicated.
[16:44] <t0by> Particularly, I don't think you should be scaling backgrounds, if I understand what you are saying.
[16:44] <t0by> Simei: better?
[16:45] <Simei> t0by: Ok, I see. :)
[16:45] <t0by> Yay :)
[16:45] <t0by> Ttyl
[16:45] <-- t0by left irc: Remote host closed the connection
[17:00] <m_kiewitz> snover: it seems somthing is breaking the seconds counter
[17:00] <snover> breaking in what way?
[17:00] <-- cpasjuste_ left irc: Ping timeout: 260 seconds
[17:01] <m_kiewitz> hmm, i just left gabriel alone and it took around 20 seconds or so
[17:02] <m_kiewitz> the patch definitely worked fine when I wrote it and it's waiting for a specific amount of seconds
[17:03] <m_kiewitz> wth is happening now
[17:04] <m_kiewitz> police officer returns within 10 seconds or so when I go for the door
[17:04] <m_kiewitz> this makes no sense at all
[17:06] <m_kiewitz> maybe the ingame timer is not working properly anymore
[17:06] <m_kiewitz> it should take around 23 seconds in any case whatever you do
[17:06] <m_kiewitz> will have to debug this
[17:07] <m_kiewitz> what's weird is that police officer being asleep is set to 42 seconds
[17:07] <m_kiewitz> and that seems to work?!?!
[17:08] <m_kiewitz> or maybe the counter goes faster in some way, but not that fast so you can still make it
[17:13] <m_kiewitz> snover: you need to look into this, i have to assume that seconds counter in gk1 is broken somehow
[17:13] <m_kiewitz> it sets 23 seconds correctly
[17:13] <m_kiewitz> and waits for 23 seconds when I don't do anything
[17:13] <m_kiewitz> but when I walk over to the door, it triggers after around 10 seconds already
[17:14] <m_kiewitz> maybe something going wrong with the speed throttler?
[17:16] <m_kiewitz> ah, i think i played through this a few months ago before saved game compatibility got broken
[17:16] <m_kiewitz> and i would have gotten stuck as well, but i wasn't
[17:16] <m_kiewitz> so I guess something in the meantime broke it
[17:17] <m_kiewitz> my saved games are from august 2016
[17:18] <wjp> the seconds cue works via GetTime(1)
[17:18] --> Farmboy0 joined #scummvm.
[17:18] <-- Farmboy0 left irc: Changing host
[17:18] --> Farmboy0 joined #scummvm.
[17:19] <wjp> that just gets its time from the system
[17:19] --> girafe joined #scummvm.
[17:20] <m_kiewitz> hmm, then what is going wrong?
[17:21] <m_kiewitz> how can it be different depending on if you do something or not
[17:21] <wjp> something else is cueing it maybe?
[17:21] <m_kiewitz> checking
[17:23] <m_kiewitz> waiting makes it get cue'd by rm230::doit, daySixBeignet::doit
[17:25] <m_kiewitz> walking to door makes it get cue'd by the same
[17:25] <m_kiewitz> is the instruction pointer shown somewhere within the backtrace?
[17:26] <m_kiewitz> so that i can verify that it gets cued by the exact same code?
[17:26] <m_kiewitz> wjp ^^
[17:27] <wjp> should be in there
[17:29] <m_kiewitz> "pc"?
[17:29] <wjp> yeah
[17:29] <wjp> he's already walking back despite daySixBeignet still being at 14 or something
[17:29] <wjp> daySixBeignet::seconds
[17:29] <m_kiewitz> what is happening...
[17:31] <m_kiewitz> daySixBeignet::doit pc 6:622
[17:32] <m_kiewitz> simply waiting gets me daySixBeignet::doit pc 6:646 aha
[17:33] <m_kiewitz> it triggers on cycles???
[17:34] <m_kiewitz> but cycles are not set, at least not by me
[17:34] <m_kiewitz> ah maybe something sets additional cycles when you walk through the little thingie
[17:35] <m_kiewitz> wjp: your debugger changes are ready?
[17:35] <wjp> yes
[17:35] <m_kiewitz> why not push it?
[17:36] <wjp> need testing
[17:36] <m_kiewitz> ah, it's sInGateWithPermission::changeState
[17:36] <m_kiewitz> that one is setting cycles on top
[17:37] <m_kiewitz> weird that it worked before
[17:37] <m_kiewitz> well, on master there would be tons of testing :P
[17:37] <wjp> so that was intended to reset the timer I guess?
[17:37] <m_kiewitz> it's not like you could break anything for the user
[17:38] <m_kiewitz> not exactly resetting
[17:38] <m_kiewitz> it sets cycles to 0xC8
[17:38] <snover> im testing&
[17:38] <snover> (&the new debugging code)
[17:38] <m_kiewitz> where the original call sets them to 0xDC
[17:38] <m_kiewitz> this is possibly just an overall timer
[17:39] <m_kiewitz> anyway, should be trivial to fix
[17:39] <m_kiewitz> and i could slow down that beignet guy too
[17:39] <m_kiewitz> it seems sierra "optimized" it because the view is so large, so they set him to be as fast as possible
[17:52] --> omer_mor joined #scummvm.
[17:54] <-- omer_mor_ left irc: Ping timeout: 240 seconds
[18:04] --> SylvainTV joined #scummvm.
[18:04] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[18:08] <m_kiewitz> snover: we could in theory fix those timer issues by simply watch for writes to the cycles selector, and when it's >10, just set seconds instead and do a div 20 or whatever before doing so
[18:08] <m_kiewitz> *watching
[18:13] <snover> i would expect that to work kind of poorly as a universal rule, since they probably just increased the cycle count as computers got progressively faster
[18:13] <snover> but, i dont know for sure. fixing cycles in the transitions code seemed to work OK.
[18:14] <-- LittleToonCat left irc: Remote host closed the connection
[18:14] <m_kiewitz> we would have to of course change calculation based on SCI version
[18:14] <m_kiewitz> but right, the risk is too high that it won't work properly in some case
[18:15] --> LittleToonCat joined #scummvm.
[18:24] --> Begasus joined #scummvm.
[18:50] --> Begas_VBox joined #scummvm.
[19:04] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[19:06] --> m_kiewitz joined #scummvm.
[19:06] <-- m_kiewitz left irc: Changing host
[19:06] --> m_kiewitz joined #scummvm.
[19:06] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[19:11] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[19:22] <m_kiewitz> snover: more and more bugs are coming in.
[19:22] --> GitHub122 joined #scummvm.
[19:22] <GitHub122> [scummvm] m-kiewitz pushed 1 new commit to master: https://git.io/vHWIA
[19:22] <GitHub122> scummvm/master 04b5c3a Martin Kiewitz: SCI32: GK1: Add second beignet timer issue script patch...
[19:22] GitHub122 (GitHub122@192.30.252.34) left #scummvm.
[19:22] <m_kiewitz> at some point we will have to jump ship ;-)
[19:22] <snover> im pumping out the water as fast as i can :)
[19:23] <snover> MIDI is the #1 broken thing
[19:24] <-- Begasus left irc: Quit: Ex-Chat
[19:33] <somaen> TransparentSurface is _slow_
[19:33] <-- somaen left irc: Changing host
[19:33] --> somaen joined #scummvm.
[19:33] #scummvm: mode change '+o somaen' by ChanServ!ChanServ@services.
[19:34] <somaen> But, as long as you avoid unneccessarily using the transparency functions when not needed, it's probably OK
[19:34] <Stormkeeper> Hmm...
[19:36] <m_kiewitz> wjp: is it possible that for closed bugs a RSS update is sent out?
[19:37] <wjp> no idea
[19:37] <wjp> should it be, or should it not be?
[19:40] <Stormkeeper> Quick question guys... what's required to be able to add a translation to your translation DB?
[19:42] <Strangerke|work> Stormkeeper: taslking to criezy is certainly the best idea, he's coordinating the translation efforts
[19:42] <Strangerke|work> -s
[19:42] Action: Stormkeeper nods
[19:43] <Stormkeeper> Admittedly, I don't have a lot of knowledge of the language I'd translate into, but I can do some research, and at least get the ball rolling, if nothing else... thanks for the info :)
[19:44] <Strangerke|work> oh, are you sure you'll manage to reach a decent quality if you dont' know the target language?
[19:44] <criezy> Stormkeeper: what language do you have in mind?
[19:44] <Stormkeeper> Irish
[19:44] <Strangerke|work> (and is the target Klingon?)
[19:44] <Strangerke|work> wooow :)
[19:44] <criezy> We have https://translations.scummvm.org where new language can be added and exising language can be worked on.
[19:44] <Stormkeeper> I managed to translate about 14% of OpenXCom, using what I knew, and a little bit of research and creativity
[19:45] <logix> Strangerke|work: hey, I also guessed klingon :)
[19:45] <Stormkeeper> XD
[19:45] <Strangerke|work> :)
[19:46] <Stormkeeper> I am signed up to that website criezy, but I'm afraid I couldn't locate where to add a language... not a good start on my part, admittedly XD
[19:46] <criezy> Isin't Irish using it's own alphabet though?
[19:46] <logix> you're thinking of elvish
[19:47] <Stormkeeper> It's mostly standard, apart from accented vowels, and a few letters missing
[19:47] <logix> seriously, I thought it was mostly roman letters
[19:47] <Stormkeeper> It is
[19:47] <Simei> _sev_, t0by: Hello, I have a problem concerning palette. In sludge, they have a structure like palette, but it is only used for pasting texts onto the screen. When loading images, I can't see anything like that. What on earth is a palette ? Is it necessary for image displaying? So in this case, what should I do to get a palette ?
[19:47] <Stormkeeper> At least, it's how I was taught in primary school, all those years ago
[19:47] <logix> Stormkeeper: sorry, yeah, I was basically pressing return when I saw your explanation
[19:47] <Stormkeeper> np :)
[19:48] <m_kiewitz> logix: have you gotten Star Trek TOS?
[19:48] <_sev_> Simei: ugh
[19:49] <criezy> Stormkeeper: on that website we have two level of registered users. The officially recognized translators can create new languages and save changes to existing ones. The default register users can only make suggestions.
[19:49] <Stormkeeper> With the exception of the extra 5 vowels (which have accents - called fadas - over them), Irish is pretty much the same as English in terms of characters
[19:49] Action: Stormkeeper nods
[19:49] <_sev_> Simei: please read this: https://en.wikipedia.org/wiki/Indexed_color
[19:49] <criezy> So we would need to upgrade you to official translator if you want to start a Irish translation.
[19:49] <Stormkeeper> Sure
[19:49] <snover> criezy: when you have a moment would you be able to update http://wiki.scummvm.org/index.php/HOWTO-Translate_ScummVM_GUI with this information, and get rid of the old information?
[19:50] <wjp> are you sure it's a good idea to translate to a language you're not fluent in?
[19:50] <criezy> Yes, that might be an issue...
[19:50] <logix> m_kiewitz: no, I hesitated - and the offer seems to be gone now :(
[19:50] <wjp> getting translations to sound natural is tricky even if you _are_ fluent in both languages
[19:51] <Simei> _sev_: Thanks, I'll read it
[19:51] <Stormkeeper> That is true wjp
[19:52] <Stormkeeper> Irish isn't widely spoken in Ireland, unfortunately... although it is getting somewhat of a revival
[19:52] <m_kiewitz> logix: well yeah, they even sold the Clint Eastwood Collection for 2 pounds (which contains more than 10 movies on blu-ray). it seems it was a pricing mistake
[19:52] <m_kiewitz> according to my info though items were shipped
[19:52] <m_kiewitz> i personally misse the Eastwood collection, I already own all of Star Trek on DVD+Bluray
[19:52] <m_kiewitz> *missed
[19:53] <Stormkeeper> I was hoping to do my little part to help spur it on, and perhaps encourage someone better than me to get involved... although I admit, it's a bit of an idealist approach
[19:53] <logix> m_kiewitz: ah, I was about to ask which of the two you had ordered (or both)
[19:53] <m_kiewitz> I own all of them already, so I didn't order anything
[19:53] <m_kiewitz> they also sold Star Trek Enterprise on blu-ray for almost nothing (the last TV series)
[19:55] <Simei> _sev_: So if I understand well, it's in the palette that we have colors, in textures, it's just indexes pointing to that color
[19:57] <Simei> https://www.irccloud.com/pastebin/HO1o9ysW/
[19:57] <_sev_> Simei: yes, for palettized techures
[19:58] <Simei> _sev_: they have something like this "makeColour", Is this what I need ?
[19:58] <_sev_> Simei: https://github.com/scummvm/scummvm/blob/master/graphics/pixelformat.h#L169
[19:59] <Simei> Ah, ok, they are same. So I build a palette with it ?
[19:59] <_sev_> err, no
[19:59] <Simei> Sorry
[20:00] <_sev_> Simei: that function converts color components to 16-bit color
[20:00] <_sev_> Simei: you need to make sure what is the color format of the sludge
[20:00] <Simei> So, it's the inverse direction
[20:01] <Simei> https://www.irccloud.com/pastebin/YW3MqTqD/
[20:01] <Simei> https://www.irccloud.com/pastebin/ensXlLIu/
[20:01] <Simei> They stock these values into the byte arrays
[20:03] <_sev_> in order to understand the graphics format you need to look into graphics initialisation
[20:04] <_sev_> Simei: if (SDL_SetVideoMode(realWinWidth, realWinHeight, 32, videoflags) == 0) {
[20:05] <_sev_> Simei: what's the format there?
[20:08] <Simei> Sorry, I don't understand. That means from this setGraphicsWindow(...), we can see the graphic format?
[20:09] <_sev_> yes, you can see the format from the call I quoted above
[20:11] <_sev_> Simei: got any clue?
[20:11] <Simei> Sorry, I don't know much about sdl neither
[20:12] <Simei> SDL_Surface?
[20:12] <_sev_> and how to find it out?
[20:13] <_sev_> Simei: please follow this URL: http://bfy.tw/C49k
[20:14] <Simei> Ok
[20:15] <Simei> it's SDL_Surface the type of framebuffer surface that they return
[20:16] <_sev_> but what is the requested format?
[20:16] <_sev_> see the first link in the results
[20:17] <Simei> this one ? https://www.libsdl.org/release/SDL-1.2.15/docs/html/sdlsetvideomode.html
[20:17] <_sev_> yes
[20:18] <Simei> requested format means input parameters ?
[20:19] <_sev_> Simei: https://en.wikipedia.org/wiki/Color_depth
[20:19] <_sev_> Simei: I think that should be a required read for you
[20:20] <Simei> Ah, ok, bpp
[20:22] <_sev_> so, as you see, your initial idea that it is 8bpp is not correct
[20:23] <Simei> Yes, it's 32 bits per pixel
[20:23] <Simei> I see
[20:25] <Simei> So, that means we should init graphics with a different PixelFormat?
[20:25] <_sev_> yep, like here: https://github.com/scummvm/scummvm/blob/master/engines/sword25/sword25.cpp#L99
[20:26] <Simei> And all the surfaces should be changed ?
[20:27] <_sev_> not necessary
[20:29] <Simei> In sword25, I see they also used the same format to convert png
[20:29] <_sev_> yes
[20:30] <_sev_> you could convert the surface before blitting
[20:30] <_sev_> Simei: like this: https://github.com/scummvm/scummvm/blob/master/graphics/surface.h#L250
[20:35] <-- dreammaster left irc: Ping timeout: 260 seconds
[20:35] <Simei> Ok, I see.
[20:36] <Simei> So I get the png displayed it's because both are assumed as 8bpp
[20:37] <Simei> But what I load there byte by byte don't work, it's because the format is not correct ? It's not only related to palette?
[20:38] Nick change: Joefish_ -> Joefish
[20:45] <wjp> what exactly doesn't work?
[20:46] <Simei> simply, in sludge, they separate "png loading" and "other image loading"
[20:46] --> Harekiet2 joined #scummvm.
[20:47] <-- Harekiet left irc: Ping timeout: 255 seconds
[20:47] <Simei> And I was able to get the png displayed using the existing PNGDecoder
[20:47] <Simei> ANd I'm trying to write the one for other images
[20:48] <Simei> I get the data loaded into a byte * and tried to use copyToRect to show it just like what I did with png
[20:49] <Simei> But, it doesn't show, I think there is a number of reasons for it
[20:50] --> t0by joined #scummvm.
[20:50] <-- t0by left irc: Changing host
[20:50] --> t0by joined #scummvm.
[20:50] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[20:50] <t0by> Hello
[20:51] <t0by> I've see there's been quite a lot of activity in the backlog
[20:51] <t0by> I'll be with you in ~30 mins
[20:52] <m_kiewitz> wjp/snover: that sq4 cd floppy issue happens because kScriptID sets acc to diskDrive
[20:52] <m_kiewitz> is that supposed to happen?
[20:52] <wjp> m_kiewitz: which one is that?
[20:52] <m_kiewitz> https://bugs.scummvm.org/ticket/9812
[20:53] --> dreammaster joined #scummvm.
[20:53] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[20:53] <m_kiewitz> the script then pushes acc and calls ScriptID again
[20:53] <m_kiewitz> i wonder if it's a script bug, or us setting incorrect acc
[20:53] <m_kiewitz> would original ssci error out in such a case?
[20:54] <wjp> I've never understood the name kScriptID
[20:54] <wjp> because as far as I know it returns an export entry?
[20:54] <wjp> which script is this?
[20:55] <m_kiewitz> ah wait, it seems it's actually a script bug
[20:55] <m_kiewitz> well ScriptID returns driveCloseUp actually
[20:55] <wjp> DisposeScript(ScriptID(152))
[20:55] <m_kiewitz> and thats get pushed to DisposeScript
[20:55] <m_kiewitz> exactly
[20:55] <wjp> that looks rather wrong indeed
[20:56] <wjp> 151 also does this once
[20:56] <m_kiewitz> what did ssci do in such a case? nothing I guess?
[20:56] <m_kiewitz> in that case i could simply ignore that call
[20:56] <m_kiewitz> for what is 151 used?
[20:57] <m_kiewitz> ah the fight
[20:57] <wjp> DisposeScript looks up the argument in a script map, so it's probably just treating some random object offset as a script number
[20:57] <wjp> DisposeScript returns (without an error) if it can't find the script in the map
[20:58] <wjp> (this is in KQ5 which IDB I happened to have open)
[20:59] <m_kiewitz> and kScriptID is supposed to return what exactly?
[21:00] <m_kiewitz> wouldn't it make sense for it to return a script map offset?
[21:02] <wjp> ScriptID returns an entry from the script's export table
[21:02] <m_kiewitz> and when there are no exports?
[21:03] <m_kiewitz> it's pretty weird
[21:03] <m_kiewitz> and where do you get that offset into the script map from?
[21:04] <wjp> kq5 errors out if the index is >= the number of exports it seems
[21:04] <wjp> which offset?
[21:04] <m_kiewitz> ah
[21:04] <m_kiewitz> kScriptID is for getting export offsets?
[21:04] <wjp> yes
[21:04] <wjp> crazy name
[21:04] <m_kiewitz> ah now it makes sense
[21:04] <m_kiewitz> yes
[21:05] <m_kiewitz> i guess the developer behind that code thought it would return the script offset
[21:05] <wjp> the 1 argument version returns the first export
[21:05] <m_kiewitz> yes
[21:05] <m_kiewitz> ScriptID(152d) if I see it correctly
[21:05] <m_kiewitz> so I guess that would cause an error?
[21:06] <m_kiewitz> someone should check in original interpreter
[21:06] <m_kiewitz> ah wait
[21:07] <m_kiewitz> gregs engine does this here
[21:07] <m_kiewitz> uint16 nExports = READ_LE_UINT16(ptr);
[21:07] <m_kiewitz> if (nobject <= nExports)
[21:07] <m_kiewitz> outObj = READ_LE_UINT16(ptr + (nobject << 2) + 2);
[21:07] <m_kiewitz> return (sciScript *)heap2Ptr(hScr);
[21:07] <m_kiewitz> so it almost seems that asking for anything above export count will actually return offset to script?
[21:07] <wjp> ignoring outObj?
[21:08] <m_kiewitz> ah right
[21:08] <m_kiewitz> yes, nvm
[21:08] <m_kiewitz> so it won't error out, at least greg's engine didn't
[21:08] <wjp> kq5 has a PError
[21:08] <m_kiewitz> let me check in sci1.1
[21:10] <wjp> looks the same in sq4
[21:12] <m_kiewitz> hmm ok
[21:12] <m_kiewitz> so we should really verify that this happens in SSCI just to make absolutely sure
[21:12] <m_kiewitz> and then we can ignore those calls
[21:13] <m_kiewitz> its weird, i can't find PError calls inside kq6 kScriptID
[21:14] <wjp> in _GetDispatchAddrInHeap
[21:15] <wjp> two calls down
[21:16] <m_kiewitz> ah down there, missed it
[21:18] Nick change: Harekiet2 -> Harekiet
[21:20] <m_kiewitz> ah wait, it probably won't crash because it's called with the scriptid, which is valid
[21:20] <m_kiewitz> and kDisposeScript will just ignore the bad call
[21:20] <m_kiewitz> anyway, trying right now
[21:20] <m_kiewitz> have to wait for count down
[21:21] <m_kiewitz> yes, no crash
[21:21] <wjp> scripts 151 and 152 both have exports, so that PError shouldn't trigger
[21:21] <t0by> Here we are
[21:22] <t0by> Simei: I've given a quick glance at the backlog -- are you still stuck?
[21:25] <m_kiewitz> wjp: there are already workarounds in place, just with room 150 instead of 900
[21:26] <Simei> t0by, _sev: Advanced quite a bit. I get other images shown, but the color are not right.
[21:26] <t0by> Oooh, great! Oooh, bad!
[21:26] <t0by> Simei: "not right" how?
[21:26] <t0by> I smell wrong byte order.
[21:26] <m_kiewitz> wjp: really weird, room number is actually 150
[21:27] <Simei> t0by, _sev_: black to red, purple to yellow, yellow to purple :/
[21:28] <_sev_> t0by: you're wrong with the pixelformat
[21:28] <_sev_> Simei: that was for you
[21:28] <_sev_> Simei: try to play with the shift parameters
[21:28] <t0by> That was for me too.
[21:29] <t0by> "Black to red" is hardly wrong byte order
[21:29] <t0by> (But wrong format, possibly)
[21:29] <_sev_> t0by: it is when you mix alpha and red channels
[21:29] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[21:29] <t0by> Oh, right.
[21:31] <t0by> Simei: care to show the code?
[21:31] <_sev_> i'd recommend commit and push
[21:31] <t0by> Yes, that kind of show the code.
[21:32] <Simei> Ok, I'll push it.
[21:32] <t0by> (To be honest if you need help debugging it's generally pretty useless to sprunge portions of code when you can share your whole tree)
[21:33] <m_kiewitz> wjp: ah it makes sense, 900 is the game over room
[21:33] <wjp> aha
[21:38] <Simei> https://github.com/yinsimei/scummvm/commit/60223a625e0a904bbe15233b678ca01276021b01
[21:38] <m_kiewitz> wjp: can you die during the fight?
[21:38] <wjp> never played SQ4... :/
[21:38] <m_kiewitz> and if so, will it cause the same error in room 900?
[21:38] <Simei> t0by, _sev_: https://github.com/yinsimei/scummvm/commit/60223a625e0a904bbe15233b678ca01276021b01
[21:38] <m_kiewitz> i played through it, but i don't have proper saved games during the end
[21:39] <wjp> m_kiewitz: I'm thinking just turn this room number into -1 in the workaround table?
[21:39] <Simei> There are still some reformatting in backdrop.cpp
[21:39] <t0by> Sweet, thanks
[21:39] <t0by> Let's see
[21:39] <Simei> I'll fix that later :/
[21:39] <m_kiewitz> yes, I did that for the disk menu workarounds
[21:39] <m_kiewitz> but the fight one I would rather keep as it is unless it can happen in room 900 too
[21:39] <wjp> the other conditions are pretty strict, so I don't see how this could cause false positives
[21:39] <t0by> Simei: sure, don't be shy to commit, you can always fix that later
[21:40] <wjp> m_kiewitz: both scripts only have a single kDisposeScript call AFAICT
[21:40] <t0by> Ew
[21:40] <t0by> it *would* help if it weren't 100000 useless lines, though.
[21:40] <t0by> Simei: PNGs work fine, right?
[21:41] <t0by> it's just the TGAs that are wrong, right?
[21:41] <Simei> Yes
[21:42] <Simei> Only this part: https://github.com/yinsimei/scummvm/commit/60223a625e0a904bbe15233b678ca01276021b01#diff-bcf30428c7a9654b3c357622b7f42f14
[21:42] <Simei> bytearray.cpp
[21:43] <wjp> t0by: I still don't see how this is TGA
[21:43] <t0by> wjp: entirely possible it's not, you're right
[21:44] <t0by> I'm basing this on the fact that the dev env takes TGAs and PNGs
[21:44] <Simei> It was tga, encoded in some way now. :/
[21:44] --> GitHub114 joined #scummvm.
[21:44] <GitHub114> [scummvm] m-kiewitz pushed 1 new commit to master: https://git.io/vHWZj
[21:44] <GitHub114> scummvm/master 266f68c Martin Kiewitz: SCI: SQ4CD: Remove room restriction of existing workarounds for end...
[21:44] GitHub114 (GitHub114@192.30.252.42) left #scummvm.
[21:45] <t0by> Ah
[21:45] <t0by> A ssec
[21:46] <Simei> By the way, I switched the r and a channel, got another colour look, still not very right.
[21:46] <wjp> try some more permutations :-)
[21:46] <t0by> lol
[21:47] <t0by> The format you are passing is RGBA, isn't it?
[21:47] <t0by> Graphics::PixelFormat(4, 8, 8, 8, 8, 24, 16, 8, 0)
[21:48] <Simei> Yes
[21:49] <t0by> But your assignment doesn't look very RGBA...
[21:49] <t0by> Let me see if it's just that
[21:50] <wjp> you can look up in graphics/pixelformat.h what exactly those values mean
[21:50] <m_kiewitz> wjp: can you please add "resolution" to "Open SCI tickets by last modified date"?
[21:50] <t0by> wjp: they mean RGBA :)
[21:50] <t0by> Simei: this is where "jump to/peek definition" or similar comes handy
[21:51] <t0by> (Come on, compile, you stupid machine)
[21:51] <m_kiewitz> wjp: and maybe owner too
[21:52] <wjp> m_kiewitz: for _open_ tickets?
[21:52] <m_kiewitz> yes, sometimes I set myself as the owner when I'm working on one
[21:52] <Simei> Hmmm, Ok. I'll try to see the definition. :/
[21:53] <wjp> looking up documentation is really no reason for a ":/" I hope :-)
[21:53] <-- Farmboy0 left irc: Remote host closed the connection
[21:54] <Simei> True :) Type too much ":/" today
[21:55] <t0by> Oh
[21:55] --> Strangerke_ joined #scummvm.
[21:55] <t0by> Well
[21:55] <t0by> 1. *Always* have the definition of any function you use handy
[21:55] <t0by> (This is where an IDE can help you navigate an unfamiliar code base and remind you of the definitions)
[21:56] <t0by> Simei: so, what is the byte order you specified in your PixelFormat?
[21:56] <t0by> (and, for bonus points, how many bytes is a pixel?)
[21:57] <Simei> 4 bytes for a pixel here, just discussed it with _sev_
[21:57] <Simei> :)
[21:57] <-- ny00123 left irc: Quit: Leaving
[21:57] <-- Strangerke left irc: Ping timeout: 240 seconds
[21:57] Nick change: Strangerke_ -> Strangerke
[21:58] <Simei> that's r g b a, I think ?
[21:58] <t0by> Good.
[21:58] <t0by> And what does your assignment look like?
[21:58] <wjp> m_kiewitz: do you have permissions to edit the report yourself?
[21:59] <wjp> (out of curiosity)
[21:59] <m_kiewitz> ah it seems i actually can
[21:59] <Simei> t0by: argb
[22:01] <t0by> Simei: or is that bgra?
[22:02] <t0by> (wjp brought this up, and he did well)
[22:02] <Simei> How to see this thing?
[22:02] <Simei> I will look into that shift bit thign
[22:04] <wjp> yes, that's the right thing to look at
[22:09] <Simei> So I put r g b a in PixelFormat, is it also r g b a the order that I should put into the array ?
[22:11] <-- |Cable| left irc: Ping timeout: 240 seconds
[22:12] <-- ajax16384 left irc: Read error: Connection reset by peer
[22:13] <m_kiewitz> snover: how were you able to get audio going again in gk1?
[22:17] <Simei> t0by, _sev_: It's good now :) I've reversed it.
[22:17] <t0by> Yay!
[22:17] <Simei> :D
[22:18] <t0by> wjp points out to me that this is, in fact, ABGR, because... well, you put A first, then B, then...
[22:18] <wjp> well, ABGR memory order, RGBA word order
[22:19] <Simei> that's true. didn't realize it, haha
[22:19] <wjp> (gotta love LE sometimes)
[22:21] <t0by> wjp: thanks for clarifying.
[22:22] <t0by> Simei: so yay, great progress!
[22:22] <t0by> Good job!
[22:23] <t0by> Now, where does this leave us.
[22:23] <Simei> Haha, made a lot of stupid mistakes. Thanks a lot! _sev_, t0by, wjp
[22:23] <t0by> Who doesn't.
[22:23] <Simei> rearrange the things I guess?
[22:23] <Simei> fix the push first
[22:23] --> |Cable| joined #scummvm.
[22:24] <t0by> Oh
[22:24] <Simei> stretch image?
[22:24] <t0by> By the way I see what you mean
[22:24] <t0by> by exactly that.
[22:25] <t0by> You mean that original sludge has a fixed viewport of 640x480, and if the backdrop was not provided in the correct dimension it stretches it to fit the viewport, right?
[22:25] <t0by> Have you isolated the lines in opensludge where it does so?
[22:26] <Simei> Let me see
[22:27] <snover> t0by: sprunge. you nerd.
[22:27] <snover> m_kiewitz: sorry, i dont understand your question about getting audio going again in GK1. could you be more specific?
[22:28] <t0by> Simei: I'm not sure it does by default, though?
[22:28] <m_kiewitz> snover: bug about setting gk1 to subtitles only and then not getting audio in game
[22:28] <snover> oh
[22:28] <t0by> Simei: http://imgur.com/a/BrJKl
[22:29] <t0by> is there a special switch perhaps
[22:29] <Simei> there is a resizeBackdrop function
[22:29] <Simei> can be called if "reserve is set false"
[22:29] <snover> m_kiewitz: it was a sequence of steps that i didnt try to turn into a minimal reproduction. something like, click the voice button so it says on, then turn the sound volume back up, then click the voice button twice so it goes off and on again
[22:29] <t0by> o_O
[22:30] <Simei> set to false by default
[22:30] <t0by> Simei: have you seen the screenshot?
[22:30] <Simei> https://github.com/opensludge/opensludge/blob/master/source/Engine/backdrop.cpp#L275
[22:31] <Simei> t0by: Yes, I see, What is that?
[22:31] <t0by> That is VerbCoin with room.tga replaced with a resized file.
[22:31] <t0by> So it doesn't scale automatically, it seems...
[22:31] <t0by> ah, wait, there are the floor and z-buffer thingies
[22:31] <t0by> a second
[22:32] <t0by> Uuh, setScale() is interesting.
[22:32] <snover> m_kiewitz: also, for all query-based reports, if you want to show extra columns after running the report, open the columns at the top-left and check extra ones
[22:33] <t0by> setScale() only applies to characters, it seems...
[22:35] <Simei> https://github.com/opensludge/opensludge/blob/master/source/Engine/backdrop.cpp#L1032
[22:35] <t0by> Simei: it appears that the backdrop has a fixed dimension *but* you scale the whole viewport with everything it contains - e.g. to set fullscreen
[22:35] <Simei> There is a resizebackdrop() called here when it is loading
[22:35] <t0by> ^^
[22:36] <t0by> ^
[22:36] <t0by> I have just tried screwing it up in every possible way but no, there isn't a way to scale the backdrop in-game, it seems.
[22:36] <Simei> is this only reallocation of memory?
[22:36] <t0by> If you can find it I'd be happy to be wrong.
[22:37] <t0by> Simei: here's something fun to try
[22:37] <t0by> alter VerbCoin.slp with
[22:37] <t0by> width=1920
[22:37] <t0by> height=1280
[22:37] <Simei> But the welcome image is not streched as it is with sludge engine
[22:38] <t0by> you'll see that the whole viewport is scaled to fit your monitor
[22:38] <t0by> Simei: I've noticed that, but I suspect the *whole viewport* is being scaled.
[22:38] <t0by> let me try
[22:39] <Simei> Ah, ok, so that's different from stretching the image
[22:39] <t0by> Simei: something super funny to try: alter Welcome.slp with
[22:39] <t0by> width=1024
[22:39] <t0by> height=768
[22:39] <t0by> you'll see something familiar.
[22:40] <t0by> I *suspect* there is no such thing as in-game backdrop scaling. Until we can find evidence there is (and frankly it would be pretty useless and inefficient), I'd assume there is no such thing.
[22:40] <t0by> Scaling of the whole viewport (e.g. for fullscreen) is alaready handled by scummvm backends.
[22:41] <Simei> Ah, ok, I see
[22:41] <Simei> It's just init the graphics with the value read in game data
[22:42] <-- waltervn left irc: Quit: Leaving
[22:42] <Simei> So, there is actually no streching
[22:42] <t0by> So, I *think* that's safe to disregard and I'll keep thinking that until we find a counterexample :)
[22:43] <t0by> I was thinking you could tidy up things a *little bit* and then set out to load characters and sprite banks?
[22:43] <t0by> Let's hear what _sev_ has to say.
[22:44] <Simei> Ok, that sounds good.
[22:44] <t0by> This won't be easy, but...
[22:44] <t0by> Or at least not *as* easy.
[22:45] <t0by> for the time being you could handle animations as static images, i.e. use only the first frame to display characters.
[22:45] <t0by> A great result would be having the interpreter not blowing up.
[22:45] <Simei> Or texts ?
[22:45] <t0by> Eh, I'd work on characters first.
[22:45] <Simei> Ok
[22:46] <t0by> The point is - the problem of uninitialized sprite banks is making everything blow up in our faces if you try loading anything that has characters (and thus text), so we might as well attack that.
[22:47] <Simei> Ok, that sounds reasonable.
[22:47] <t0by> Simei: it it will probably require a bit of heavy-handedness and creativity on your part. You probably want to stub some stuff first to get to a point where stuff doesn't break.
[22:48] <t0by> When we have characters we can tackle input and essentially we have VerbCoin running...
[22:48] <t0by> (well, minus the audio)
[22:49] <Simei> Hmm, ok, so to fix the segfault first
[22:49] <t0by> Simei: by the way -- things are about to get quite hairy. Make sure you can use the tools you have - i.e. the VCS, the debugger, etc.
[22:50] <t0by> Half an hour spent on reading docs and experimenting might save you half a week later on.
[22:51] <t0by> Simei: I know I sound like your grandmother, but I have been in your very shoes a few years ago.
[22:51] <Simei> Ok, I'll read some tutos about them. So that's git cola, valgrind, dbg for now
[22:52] <t0by> Simei: about git -- *please* make sure you understand you know how it works and what it allows you to do. The gui, if any, you choose to use is irrelevant.
[22:53] <Simei> I understand, what you said helps
[22:53] <Simei> a lot
[22:53] <Simei> Ok, i note that
[22:53] <t0by> In the future you'll probably start needing to reorder commits, split them, test them individually, squash them.
[22:54] <t0by> Simei: so, good job,thank you. I can't wait to have characters on screen :)
[22:54] <t0by> I'll go to bed it you don't need me.
[22:55] <t0by> (You'll have realized, by now, that you're never alone in here anyway)
[22:55] <Simei> No, I need to thank you all, it's good for now. Good night.:)
[22:55] <t0by> Good night then!
[22:55] <-- t0by left irc: Quit: t0by
[22:58] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:05] --> m_kiewitz joined #scummvm.
[23:05] <-- m_kiewitz left irc: Changing host
[23:05] --> m_kiewitz joined #scummvm.
[23:05] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[23:24] <snover> oh thank god.
[23:25] <snover> restoring a save game at the end of torin has the same palette problems in the original interpreter. one less bug of my own to fix.
[23:39] <m_kiewitz> is there already a bug about it?
[23:40] <snover> i dont think so, pretty sure it was just something that i had stumbled upon myself
[23:42] <m_kiewitz> please make a bug report on it, maybe we can fix it at some point
[23:42] <m_kiewitz> including saved game
[23:42] <m_kiewitz> have you played through torin successfully?
[23:42] <m_kiewitz> you can upload your saves onto ScummVM FTP
[23:42] <m_kiewitz> i uploaded quite a few saves of other games today
[23:42] <m_kiewitz> always helpful
[23:46] <m_kiewitz> anyway got 2 go, bye
[23:46] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:52] <-- Huerbine 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.
[00:00] --- Tue May 30 2017