Back to Home / #openttd / 2021 / 02 / Prev Day | Next Day
#openttd IRC Logs for 2021-02-22

---Logopened Mon Feb 22 00:00:49 2021
00:21-!-snail_UES_ [] has quit [Quit: snail_UES_]
00:29-!-Wuzzy [] has quit [Ping timeout: 480 seconds]
00:39-!-mirrorbird [~mirrorbir@2a00:801:446:6dfc:fadb:e310:6ed2:16d] has joined #openttd
00:39-!-mirrorbird is "mirrorbird" on #openttd
01:00-!-k-man [] has quit [Quit: WeeChat 3.0.1]
02:17-!-sla_ro|master [] has joined #openttd
02:17-!-sla_ro|master is "slamaster" on @#sla #openttd
02:51-!-NGC3982 [~henrik@] has quit [Remote host closed the connection]
03:05-!-Wolf01 [] has joined #openttd
03:05-!-Wolf01 is "Wolf01" on #openttd
03:09-!-mirrorbird [~mirrorbir@2a00:801:446:6dfc:fadb:e310:6ed2:16d] has quit [Ping timeout: 480 seconds]
03:23-!-heffer [] has quit [Quit: heffer]
03:24-!-heffer [] has joined #openttd
03:24-!-heffer is "Felix Kaechele" on #openttd
03:36-!-andythenorth [] has joined #openttd
03:36-!-andythenorth is "andythenorth" on #openttd
03:38-!-HerzogDeXtEr [] has joined #openttd
03:38-!-HerzogDeXtEr is "purple" on #openttd
04:14<TrueBrain>a bit sad that the OpenGL branch now has less impact on game performance :P
04:14<TrueBrain>it is still a huge difference
04:14<TrueBrain>just .. harder to measure :P
04:17<@peter1138>Stop making everything faster!
04:18<TrueBrain>I am sorry, we are just horrible people, we have to :P
04:19*andythenorth should try a nightly on the M1
04:19<TrueBrain>hmm, on SDL in VMWare I cannot move around with my rmb
04:19<TrueBrain>but I think that is VMWare acting up
04:20<@peter1138>You might need to fiddle with the mouse-scrolling option. Scrolling that relies on mouse warping generally doesn't work properly with remote/virtual desktops.
04:25<TrueBrain>yeah, I was told what to change, but I forgot :D
04:27<TrueBrain>owh, on Wayland I now finally have a top bar with SDL
04:27<TrueBrain>that is new
04:28<FLHerne>supermop_Home: Depends what vehicle type they're for, and/or what callback you used for what type :-/
04:28<FLHerne>grfspecs are crazy
04:28<FLHerne>(NML should hide that all from you)
04:49<TrueBrain>awh, macOS via VMWare doesn't give me an OpenGL context
04:56<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on pull request #7744: Draft Feature: OpenGL video driver
04:57<LordAro>TrueBrain: something i've been meaning to ask, does it fallback to non-opengl gracefully?
04:57<TrueBrain>michi_cc: I tried to break it, I could not. SDL works on Linux just fine for me, fallbacks etc work fine. Also tested Windows, couldn't find any issues. It is faster, and feels snappier, so yeah ...
04:57<TrueBrain>LordAro: absolutely
04:57<TrueBrain>see WSL2 in my last comment
04:57<TrueBrain>in general, our driver system is pretty solid
04:57<TrueBrain>it can just try multiple without any problems
04:57<LordAro>it would be irritating to have to reboot on linux after a graphics driver upgrade, like i do with certain other programs
04:58<TrueBrain>I see no reason why you should with OpenTTD :)
04:58<TrueBrain>michi_cc: code also looks good to me .. so maybe we can ask 2 or 3 people more to give it a spin, to make sure we don't have a regression .. but I think we should consider merging this :)
04:58-!-andythenorth [] has quit [Quit: andythenorth]
04:58<TrueBrain>so LordAro , if you are bored, please check out that PR :)
04:59<TrueBrain>the rendering difference between SDL and Win32 are insane btw
04:59<TrueBrain>no clue why they are this different
04:59<TrueBrain>especially as I run Win32 native, and SDL via VMWare :P
05:06<@DorpsGek>[OpenTTD/OpenTTD] orudge commented on pull request #8720: Remove remaining Windows 95 support
05:16-!-gelignite [] has joined #openttd
05:16-!-gelignite is "realname" on #llvm #openttd
05:16<TrueBrain> what(): Resource deadlock avoided
05:17<TrueBrain>how cute :)
05:20<@peter1138>Reboot after a driver update? How strange ;)
05:21<LordAro>that's arch for you
05:21<LordAro>only has one kernel
05:23<@DorpsGek>[OpenTTD/OpenTTD] LordAro commented on pull request #7744: Draft Feature: OpenGL video driver
05:24<TrueBrain>who cares about msys :P
05:24<LordAro>well it's the only compile environment i have while i'm in windows :p
05:28<TrueBrain>src/window.cpp:2992: void HandleMouseEvents(): Assertion `HasModalProgress() || IsLocalCompany()' failed. <- when I try to put gameloop in a thread and do drawing in the main thread
05:29<TrueBrain>while ... the code I currently have it either runs one thread or the other .. (basically made them fibers)
05:29<TrueBrain>interesting .. what is causing that :D
05:31<TrueBrain>pretty sure that instance of that assert does nothing for the code following
05:31<TrueBrain>but there are other places where it is important :D
05:43<TrueBrain>lol .. some details of our video-drivers are a bit weird :)
05:45<TrueBrain>hihi, some results are fun .. by accident, I now made that the game-tick wins from the draw-tick
05:45<TrueBrain>so if you FF
05:45<TrueBrain>your FPS drops to like 20 or so
05:46<TrueBrain>and the game is even faster :P
05:46<TrueBrain>a year takes ~4 seconds
05:46<LordAro>might actually have to fix the "money underflows" issue at this rate
05:47<TrueBrain>anyway, what is a bit shitty about our video drivers ... PollEvent can only be run if GameLoop is not running
05:47<TrueBrain>which is a rather odd thing to have
05:48<TrueBrain>mouse actions are processed immediately .. lets see what happens if I postpone it till the next draw frame
05:48<@peter1138>Lagface McLagface
05:49<FLHerne>TrueBrain: At this point, there needs to be a "sane FF" button that people can use to fast-forward their game and not arrive in the year 2451
05:49<TrueBrain>FLHerne: see my draft PR, yes :)
05:49<TrueBrain>haha, the game doesn't like it when you postpone mouse events .. rmb doesn't do anything :D
05:55<TrueBrain>SDL_WarpMouseInWindow triggers a new SDL_MOUSEMOTION
05:55<TrueBrain>so if the delta is not processed right there and then, it is invalidated soon after :D
06:15<TrueBrain>Wolf01: mind testing a custom build? . It is very rough, for example, Fast-Forward is shit. But I am curious if this resolves the tearing you see.
06:15<TrueBrain>You can easiest get this version with "git fetch gameloop-thread && git checkout FETCH_HEAD" , and build like normally
06:15<TrueBrain>LordAro: maybe you can give it a spin too?
06:15<TrueBrain>before I clean up the code and make it PR-able, I first would like to know if it solves anything :D
06:24<Wolf01>TrueBrain: yes, just aminute
06:25<TrueBrain>no rush, when ever you have a moment :)
06:27<Wolf01>How do I select that patch in VS?
06:27<TrueBrain>that I sadly cannot help you with .. I only know how to do it via CLI :(
06:27<TrueBrain>making a new checkout of my fork in another folder would work, I guess
06:32<Wolf01>Btw, I'll continue in the afternoon, got to go now
06:32<TrueBrain>no worries, tnx
06:33-!-Samu [] has joined #openttd
06:33-!-Samu is "realname" on #openttd
06:34<TrueBrain>hmm .. if I can get PollEvent to not influence the game-state, I can make it so the mouse stays responsive no matter what, I think
06:34<TrueBrain>by using a timed_mutex, I can just timeout if I don't get the lock in time, and draw the same frame as the last, but with the mouse moved
06:44<@peter1138>If only the OS could handle drawing the mouse itself.
06:52<Wolf01>Integrate cometcursors (that nasty malware) in OpenTTD :P
06:58<Wolf01>TrueBrain: in 5 minutes I tried with 30/60/144, it's almost the same on all, no lagging, micro jumps when moving the pointer even slowly
06:58<LordAro>yeah, no real change for me
06:59<Wolf01>If I move fast, the pointer almost disappear and I need to look for it... I think it's the lack of the ghosting
07:00<Wolf01>Maybe OpenTTD is too much optimised :P
07:05<TrueBrain>No lagging is an improvement not?
07:05<TrueBrain>And microjumps .. what does that mean for you? That you cannot move the mouse for 1 pixel, or?
07:11<TrueBrain>Wolf01: so basically, I am mostly interested if anything improved against master .. I cannot really get that from what you said, sorry :D
07:14<@DorpsGek>[OpenTTD/OpenTTD] spnda commented on issue #8687: CMake include error when building new project.
07:14<TrueBrain>okay, I can reproduce the blinking away from the cursor when I switch my screen to 60Hz
07:14<TrueBrain>so that helps
07:28<TrueBrain>well, in good news, with OpenGL that problem is totally gone
07:31<TrueBrain>okay, the problem has to do with the rect we send to Windows to repaint
07:31<TrueBrain>if I repaint the whole screen, the problem is gone
07:34<@DorpsGek>[OpenTTD/OpenTTD] glx22 commented on issue #8687: CMake include error when building new project.
07:36<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on issue #8687: CMake include error when building new project.
07:45<TrueBrain>funny, if you move at the exact right speed, the mouse is never visible to the eye :D
07:45<TrueBrain>this is pretty cool :P
07:45<TrueBrain>but Windows is having difficulties with that we say: only redraw this area
07:47<TrueBrain>I cannot see any performance issue between redrawing everything and only redrawing a partial
07:47<TrueBrain>issue = difference
07:48<Wolf01><TrueBrain> Wolf01: so basically, I am mostly interested if anything improved against master .. I cannot really get that from what you said, sorry :D <- at 30 there is an improvement, no more lag, in all the tests I can't see anymore tearing
07:49<TrueBrain>so the only remaining problem is the mouse not always showing up, right?
07:50<Wolf01>Yes, the jumps while moving
07:50<TrueBrain>okay, I found out what causes that, so that is good :)
07:52<Wolf01>Whan you want I'm available to test it :)
07:54<TrueBrain>we do this: 1) set region to draw, 2) draw, repeat 1) and 2). It feels like Windows is doing this: draw the image of the last round when 1) happens
07:55<TrueBrain>so basically, it updates the rect of the last frame
07:55<TrueBrain>instead of the upcoming
07:55<TrueBrain>that is how it feels
07:55<TrueBrain>as that would explain why you see it "jump" or at certain speeds, don't see the mouse at all
08:00-!-snail_UES_ [] has joined #openttd
08:00-!-snail_UES_ is "Jacopo Coletto" on #openttd
08:01-!-WormnestAndroid [~WormnestA@] has quit [Remote host closed the connection]
08:01-!-WormnestAndroid [~WormnestA@] has joined #openttd
08:01-!-WormnestAndroid is "WormnestAndroid" on #openttd
08:03-!-Wuzzy [] has joined #openttd
08:03-!-Wuzzy is "Wuzzy" on #openttd
08:03<TrueBrain>owh, when pausing it is easier to reproduce the problem btw :)
08:11<TrueBrain>Wolf01: can you try disabling full animation; does that solve the problem?
08:12<TrueBrain>nevermind, it doesn't
08:13<TrueBrain>it does have to do with palette animation, but full animation doesn't disable that :P
08:17<@peter1138>In theory OpenTTD is drawing an update which contains the moved mouse cursor. However it feels like it is drawing an update to remove the cursor, and then another update to draw it in the new location.
08:18<TrueBrain>the mouse is on every scene send to GDI
08:18<TrueBrain>but what is being redrawn by the OS seems to be the trouble
08:18<TrueBrain>but only during palette animation :D
08:19<@peter1138>Palette animation update happening without the mouse cursor somehow?
08:19<TrueBrain>I think this happens on every object that moves at a certain speed
08:20<TrueBrain>hmm., 32bpp-simple should not do palette animation .. hmm
08:20<TrueBrain>the win32 driver code is a bit weird
08:20<TrueBrain>it changes palette, marks whole screen dirty, while in the paint function already
08:20<TrueBrain>but okay, that is a problem for another day :D
08:23<TrueBrain>setting _cur_palette.count_dirty to zero causes the problem, lol
08:24<TrueBrain>lol, no, misleading as fuck .. keeping that non-zero repaints the screen completely
08:24<TrueBrain>which indeed solves the problem
08:24<TrueBrain>cause .. effect .. :D
08:25<TrueBrain>so the problem remains, doing an intersection with GetDCEx causes weirdness
08:27<TrueBrain>totally unrelated assert while testing something, w00p
08:28<TrueBrain>make 2 airports, make plane, create orders, click both airports, press delete
08:28<TrueBrain>ah, already happens when plane viewport is open
08:29<TrueBrain>lets see if that is my handywork :D
08:29-!-andythenorth [] has joined #openttd
08:29-!-andythenorth is "andythenorth" on #openttd
08:29<andythenorth>quite a nice read, on the community stuff also
08:33<LordAro>TrueBrain: just need to press delete with a hangar open
08:33<LordAro>i did touch some stuff with depots recently...
08:33<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain opened issue #8724: Pressing delete with open airplane crashes the game
08:33<TrueBrain>in that case, here, an issue ^^ for you to work on :P :D
08:34-!-andythenorth [] has quit [Quit: andythenorth]
08:34<TrueBrain>owh, closing the window is enough
08:35<LordAro>must be recent, no one else has reported it :)
08:35<TrueBrain>yeah, it is kinda game-breaking, so indeed :D
08:37<TrueBrain>okay, windows "jumping" issue also happens with things like aircrafts
08:37<TrueBrain>so it is not as much related to the mouse
08:37<TrueBrain>more that updating small regions acts up
08:38<@DorpsGek>[OpenTTD/OpenTTD] LordAro commented on issue #8724: Closing airport hanger window crashes the game
08:39<LordAro>i tested road vehicles & rail, what more could you need?
08:52<@DorpsGek>[OpenTTD/OpenTTD] LordAro opened pull request #8725: Fix #8724: Got wrong DepotID/StationID for airports, causing crash
08:54<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain approved pull request #8725: Fix #8724: Got wrong DepotID/StationID for airports, causing crash
09:11-!-GroovyNoodle [] has joined #openttd
09:11-!-GroovyNoodle is "" on #openttd
09:18<TrueBrain>tempted to just redraw the whole screen for windows
09:18<TrueBrain>doesn't seem to impact performance in any way
09:18<TrueBrain>and solves this whole issue :P
09:20<TrueBrain>from all observations it looks like there is a queue of what is marked "dirty" (or clipped, what-ever)
09:20<TrueBrain>but it always uses the last bmp
09:20<@DorpsGek>[OpenTTD/OpenTTD] LordAro merged pull request #8725: Fix #8724: Got wrong DepotID/StationID for airports, causing crash
09:20<TrueBrain>so while catching up with clipped regions, it is showing that of newer frames
09:20<@DorpsGek>[OpenTTD/OpenTTD] LordAro closed issue #8724: Closing airport hanger window crashes the game
09:20<TrueBrain>tnx LordAro for the quick fix
09:20<Wolf01>Probably, with monogame I found that is better redrawing the whole viewport instead of parts of it, video cards now are powerful enough to not even bother... and you can just prepare all the stuff to draw in the right order and send it to buffer at whole
09:23<TrueBrain>I just hate that I know something fishy is going on with the clipping, but I cannot put my fingers on it
09:24<TrueBrain>Wolf01: I pushed a new version to my branch; mind trying again?
09:24<TrueBrain>and LordAro too ofc, if you have the time :)
09:25<TrueBrain>I changed it so only the dirty rect is copied from buffer to screen, but the whole screen is marked as dirty
09:25<TrueBrain>(well, I don't decrease the clipping region, more accurate)
09:26<TrueBrain>always using 144Hz makes your eys hurt when you switch to 60Hz for a while :D You get spoiled real quick
09:31<TrueBrain>hmm .. not sure that sending only the dirty part to the screen is such a good idea ..
09:32<TrueBrain>yeah, no, that was a terrible choice
09:33<TrueBrain>pushed yet another version, sorry Wolf01
09:33<TrueBrain>I noticed that on-pause it was still terrible ... should be a lot better now
09:37<TrueBrain>in the meantime, lets play the game: how many wishlists are we on? :D
09:38<TrueBrain>if the page ever loads, I can tell ... :D
09:38<Wolf01>Title screen: 60=>jumping a lot more than before, 144=>it totally disappear when moving
09:38<Wolf01>Game: 60/144 not so much change
09:39<TrueBrain>you sure you have the latest? Sorry, that is a bit messy with 1 branch :)
09:39<Wolf01>Yup, I just pulled
09:40<TrueBrain>f363494 is latest
09:40<Wolf01>Wut... it says there are no changes
09:40<@peter1138>"3 friends want this game"
09:40<TrueBrain>Wolf01: it is a forced push, that might be related?
09:40<Wolf01>Still gameloop-thread?
09:41<TrueBrain>I noticed that what you can do with git in MSVC is a rather limited workflow :P
09:41<TrueBrain>"Wishlists 18,747"
09:42<Wolf01>Oh, wait, I told you the version shown in the title bar of the game
09:42<TrueBrain>those should be the same
09:42<TrueBrain>assuming it did not merge or something stupid :P
09:43<TrueBrain>otherwise, src/video/win32_v.cpp line 219 should read BitBlt(dc, 0, 0, _wnd.width, _wnd.height, dc2, 0, 0, SRCCOPY);
09:43<Wolf01>Maybe I did something weird by switching the branch at beginning, VS2019 is totally different
09:43-!-nielsm [] has joined #openttd
09:43-!-nielsm is "Niels Martin Hansen" on #openttd
09:43<Wolf01>It shows I have 2 edits
09:45<TrueBrain>euh, sorry, I meant src/video/win32_v.cpp line 364 should read BitBlt(dc, 0, 0, _wnd.width, _wnd.height, dc2, 0, 0, SRCCOPY);
09:45<TrueBrain>wrong line number :D
09:46<Wolf01>Nice, I see totally different ids in the local graph compared to github
09:47<TrueBrain>and this is why I use the CLI :D
09:47<TrueBrain>if you fetch the remote, you can often do a new branch checkout
09:47<TrueBrain>just to a random new name
09:47<TrueBrain>that should reset everything
09:48<Wolf01>Oooook deleted and re-extracted the branch, right id
09:49<TrueBrain>lets hope that fixes it a bit
09:49<TrueBrain>as I fixed what I could reproduce :D
09:51<Wolf01>Funny that it's still compiling and the output window is stuck
09:52<Wolf01>Maybe I broke it
09:52<Wolf01>WOW, it's perfect now at 144
09:52<Wolf01>Trying at 60 and 30
09:52<TrueBrain>hopefully also at 60 :)
09:54<Wolf01>Yup, perfect at 60 too
09:54<TrueBrain>okay, so at least I now know for sure where the problem comes from
09:54<TrueBrain>and that it has nothing to do with vsync :D
09:55<TrueBrain>not sure I like the solution, as really understanding it ... I do not
09:55<Wolf01>30 is really noticeable that is slower, but no artifacts
09:55-!-GroovyNoodle [] has quit [Quit: Leaving]
09:55<TrueBrain>yeah .. 30 is a bit too low for the commoner :)
09:55<TrueBrain>cool, tnx Wolf01
09:55<TrueBrain>hopefully this also fixes what LordAro was seeing
09:56<TrueBrain>but yeah, enjoying silky smooth movement is a joy to play with :)
09:58<TrueBrain>just ... DO NOT PRESS FAST FORWARD :P
10:00<TrueBrain>planes always look slightly better
10:00<TrueBrain>not good
10:00<TrueBrain>but better
10:00<TrueBrain>(fast planes, that is)
10:01<TrueBrain>their movement is still a bit jumpy, but that is because of pixels
10:11<LordAro>TrueBrain: yeah, much better :)
10:11<supermop_Home>here is a dumb question that i am sure is basic CS knowledge … can i take my 0-255 random number and instead make it two 0-16 random numbers?
10:12<LordAro>rand() % 16
10:12<TrueBrain>LordAro: awesmoe!
10:12<LordAro>oh, 2 numbers
10:12<TrueBrain>a = rand(); b1 = a >> 4; b2 = a % 0xF;
10:12<TrueBrain>I guess? :)
10:13<TrueBrain>euh, % = &
10:13<supermop_Home>like one switch tht picks what cars are on trailer, and one that picks what colors to recolor the cars
10:13<LordAro>what TB said
10:15-!-roadt__ [~roadt@] has quit [Ping timeout: 480 seconds]
10:16<FLHerne>Doesn't nml have a builtin for that?
10:16<supermop_Home>does it?
10:16-!-roadt__ [~roadt@] has joined #openttd
10:16-!-roadt__ is "roadt" on #openttd
10:16<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain opened pull request #8726: Change: [Win32] don't use clipping, but draw whole screen every frame
10:16<TrueBrain>Wolf01: one more time, I am sorry, but would you mind testing ^^ too?
10:17<FLHerne>getbits(value, first, amount) <- basically maps to what TrueBrain wrote
10:17<FLHerne>i.e. you'd do getbits(value, 0, 4) and getbits(value, 4, 4)
10:17<TrueBrain>it contains only the part about the win32 that is acting up, hopefully that fixes enough to already make it playable; but I would like to know if my other code fixes any more problems :)
10:17<FLHerne>Whether you think that's more readable or not is pretty arbitrary
10:21<supermop_Home>FLHerne do i need to write somewhere else that the bits im trying to get are random?
10:24<FLHerne>supermop_Home: ^ assuming that `value` is the random number you got somewhere, getbits() just takes some bits from it
10:24<FLHerne>So, yes
10:27<supermop_Home>like this?
10:32<Wolf01>TrueBrain: as soon as I learn how to test pull requests, yeah :P
10:32<TrueBrain>just in my fork, get the right branch :)
10:32<Wolf01>Ok then :P
10:33<FLHerne>supermop_Home: Where does `value` come from?
10:33<supermop_Home>….. i don't know
10:34-!-Wuzzy [] has quit [Ping timeout: 480 seconds]
10:34<@DorpsGek>[OpenTTD/OpenTTD] nielsmh updated pull request #8720: Remove remaining Windows 95 support
10:34-!-Wuzzy [] has joined #openttd
10:34-!-Wuzzy is "Wuzzy" on #openttd
10:35<supermop_Home>in other switches i use 'extra callback info'
10:35<Wolf01><TrueBrain> win32-clip-region <- is it private? I see many others but not that
10:36<FLHerne>supermop_Home: If you don't care about where the value comes from, you might as well use random_switch
10:36<FLHerne>Now that frosch fixed it, anyway :p
10:36<Wolf01>Strange, VS can't see shit :P
10:36<TrueBrain>fetch first?
10:37<TrueBrain>the MSVC GUI was weird as fuck
10:37<Wolf01>Yes, it's weird, a fetch on gameloop made win32-clip to appear
10:38<Wolf01>Hmmm, protected void __cdecl VideoDriver::Tick(void) not resolved
10:39<TrueBrain>make sure you have no pending changes?
10:39<supermop_Home>FLHerne the only part i really care about is that the switch to pick the cargo uses a different number than the one that picks the color
10:39<Wolf01>Nothing there, is updated to 3c2fb784
10:40<supermop_Home>so first one pick: car truck or hatchback, 2nd one pick red blue or green
10:40<FLHerne>supermop_Home: Then add `independent: my_other_switch;` to at least one of them (doc
10:41<supermop_Home>if the random value was just 1-3 i don't want only red cars, blue trucks, and green hatchbacks
10:41<FLHerne>I don't think extra_callback_info1 should work for you in road vehicles anyway...
10:41<TrueBrain>Wolf01: rebuild?
10:41<supermop_Home>FLHerne i use that for sprite stacks.. seems to work
10:41<@DorpsGek>[OpenTTD/OpenTTD] nielsmh commented on pull request #7235: Change: Non-rectangular sparse station catchment area
10:43-!-Flygon [~Flygon@2001:44b8:411e:4e00:e997:c254:e4e7:2471] has quit [Quit: A toaster's basically a soldering iron designed to toast bread]
10:44<Wolf01>Ok, maybe the first recompile didn't clean all
10:46-!-glx [] has joined #openttd
10:46-!-glx is "Loïc GUILLOUX" on #openttd
10:46-!-mode/#openttd [+v glx] by ChanServ
10:46<nielsm>hmm... known-bugs.txt references a number of things regarding video drivers that might not be relevant any longer
10:46<nielsm>or even testable
10:46<TrueBrain>a good problem to have :D
10:47<Wolf01>It seem to run fine, pointer is ok too at 60fps, FFWD still go fast
10:47<TrueBrain>is the other branch better, or equal?
10:48<TrueBrain>(subjective, btw)
10:49<Wolf01>Looks equal to me, maybe some more fps drops in FFWD
10:50<Wolf01>But may be map related (more/less snow and trees)
10:50<TrueBrain>yeah, and FFWD is hard to compare, because the other branch is a bit rough still
10:50<TrueBrain>k, cool, tnx :)
10:50<TrueBrain>always good to boil stuff to their essence :D
10:52<@DorpsGek>[OpenTTD/OpenTTD] spnda updated pull request #7955: WIP Feature: NewGRF Road Stops (Feature 14)
10:54<+glx>haha I broke everything using "openttd(2++)" dir name
10:54<supermop_Home>the 2cc recolor things at the bottom of this page:
10:55<supermop_Home>so i just write instead of PALETTE_CC_BLUE, "palette_2cc(0, 1)"
10:55<supermop_Home>for dark blue 1cc and pale green 2cc?
10:55<@DorpsGek>[OpenTTD/OpenTTD] nielsmh opened pull request #8727: Update known-bugs.txt
10:59<FLHerne>supermop_Home: You should be able to write that as palette_2cc(COLOUR_DARK_BLUE, COLOUR_PALE_GREEN)
10:59<TrueBrain>gratz glx :D
10:59<TrueBrain>now fix it :P
10:59<+glx>first thing to fix: a regex
11:00<FLHerne>supermop_Home: What you wrote also works, if you really want inscrutable numbers
11:05<+glx> string(REGEX REPLACE "^${CMAKE_SOURCE_DIR}/regression/" "${CMAKE_BINARY_DIR}/ai/" REGRESSION_BINARY_FILE "${REGRESSION_SOURCE_FILE}") <-- for some reason it doesn't like (2++) ;)
11:06<+glx>but it's easy to fix
11:07<@DorpsGek>[OpenTTD/OpenTTD] James103 commented on pull request #8727: Update known-bugs.txt
11:15<supermop_Home>FLHerne still not sure i get it:
11:20<FLHerne>supermop_Home: You should probably use TRIGGER_VEHICLE_NEW_LOAD instead of that bitmask
11:20*FLHerne keeps reading
11:23<FLHerne>I don't think line 34 works
11:26-!-Progman [] has joined #openttd
11:26-!-Progman is "Peter Henschel" on #openttd
11:26<Wolf01> uhm... maybe I could get rid of the entire VS stuff from my machine
11:28<supermop_Home>yeah idk about it
11:29<FLHerne>supermop_Home: If you use `switch_color_vehs_a()`, I think it should work with current NML
11:29<FLHerne>glx: ^ NML procedure call thing
11:29<supermop_Home>with empty ()?
11:30<supermop_Home>keep the 0x100 or no?
11:31<FLHerne>I can't help feeling there should be a better way to do that
11:31<+glx>you don't change the STORE_TEMP but as you call a switch it needs ()
11:33<+glx>in pre-procedure version I think it was implicit, but I made it explicit
11:36-!-Wormnest [~Wormnest@] has joined #openttd
11:36-!-Wormnest is "Wormnest" on #openttd
11:50-!-andythenorth [] has joined #openttd
11:50-!-andythenorth is "andythenorth" on #openttd
11:58<supermop_Home>well andy i was just going to release an update that fixes my messed up grain hoppers, with a little rally car easter egg.. and now i'm sucked into randomizing the colors of vehicles cargo on flatbeds
11:59<supermop_Home>as once there are little cars in the grf it makes no sense to not have drawn vehicles cargo sprites
11:59<Timberwolf>Heh. Road Vehicles does that. Trains I use company colour.
11:59<Timberwolf>But... have you thought about year-appropriate vehicles? :D
12:00<supermop_Home>Timberwolf no wrxs before 92
12:00<supermop_Home>before that you only get pickup trucks and hatchbacks
12:00<Timberwolf>Road Vehicles has a nice little cheat here, it uses Minis from 1959 through to 1990.
12:01<supermop_Home>and new minis from 98 onwards?
12:04<@DorpsGek>[OpenTTD/team] 2TallTyler opened issue #142: [en_US] Translator access request
12:06-!-gelignite [] has quit [Quit: Stay safe!]
12:06<@DorpsGek>[OpenTTD/team] LordAro commented on issue #142: [en_US] Translator access request
12:06<TrueBrain>you denied that, I hope? :P
12:07<LordAro>i saw no reason to :p
12:07<TrueBrain>FIND ONE! :D
12:07<TrueBrain>:D :D
12:14<+glx>add_custom_command is weird in quoting
12:15<+glx>from openttd(2++) it doesn't quote the path and compile fail because ( in commands
12:16<+glx>from openttd (2++) it quotes (because space) and everything works
12:16<+glx>and I still can't trigger the || thing
12:23-!-Tirili [] has joined #openttd
12:23-!-Tirili is "realname" on #openttd
12:32-!-Progman [] has quit [Remote host closed the connection]
12:32<Timberwolf>supermop_Home: MX-5, similar reasons (at TTD scale there's not really much difference between an NA and an ND)
12:34<@DorpsGek>[OpenTTD/OpenTTD] btzy commented on pull request #8715: Fix: Vehicle list windows did not update when this year's profit changed
12:40<+glx>TrueBrain: I think I now where the || come from
12:40<supermop_Home>oops left the headlights on on this sprite
12:48<+glx>ok solved the || issue
12:53<supermop_Home>getting car sprites in, plus fixing messed up hoppers would be enough things fixed that i could just about call this 1.0
12:53<supermop_Home>but my next sequential version is 0.6.8
12:54<supermop_Home>is it wrong that i want to release this as 0.6.9?
13:00<supermop_Home>Timberwolf nice title game
13:07<Timberwolf>Mostly hoping to inspire the people who are good at these things.
13:15-!-frosch123 [] has joined #openttd
13:15-!-frosch123 is "frosch" on #openttd
13:15-!-Tirili [] has quit [Remote host closed the connection]
13:16<@DorpsGek>[OpenTTD/OpenTTD] glx22 opened pull request #8728: Cmake fixes
13:19<Timberwolf>The maglev line is deadlockable if you add enough trains, needs a signal removed :/
13:19<+glx>oh and the funny part is || is supposed to trigger an error if command fails, but the command fails because cmd decided to pass || to our command
13:20<+glx>and of course it didn't trigger the error in this case
13:20-!-qwebirc6765 [~oftc-webi@2a00:f41:81a:200e:3d96:36a2:2140:c372] has joined #openttd
13:20-!-qwebirc6765 is "OFTC WebIRC Client" on #openttd
13:22-!-gelignite [] has joined #openttd
13:22-!-gelignite is "realname" on #llvm #openttd
13:23<@DorpsGek>[OpenTTD/OpenTTD] glx22 commented on issue #8687: CMake include error when building new project.
13:24<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on pull request #8728: Cmake fixes
13:24-!-jottyfan [] has joined #openttd
13:24-!-jottyfan is "jottyfan" on #openttd
13:25<+glx>hmm weird compile fails for all targets while it's ok locally
13:25-!-jottyfan [] has quit []
13:25<+glx>ok looks like my regexp fix is broken
13:29-!-qwebirc6765 [~oftc-webi@2a00:f41:81a:200e:3d96:36a2:2140:c372] has quit [Remote host closed the connection]
13:30<FLHerne>Did the 'fixed' version still have regexps? :p
13:33-!-peter1138 [] has left #openttd []
13:34<+glx>hmm files are copied, so regexp worked
13:43<+glx>oups <src>/regression/regression matches incorrectly in my regexp
13:43<LordAro>what if i call my OTTD checkout "regression" ? :p
13:43<Wolf01>Ha :D
13:43<+glx>yeah that's the issue
13:44<+glx>let's rewrite the regexp
13:44<LordAro>are there not path altering function in cmake? would be better to use those instead, no?
13:53<supermop_Home>well STORE_TEMP(switch_color_vehs_a(), 0x100), doesn't work
13:54<supermop_Home>nmlc says, "value to store must be an integer
13:55<+glx>hmm a call is an integer IIRC
13:58<+glx>unless there's a bug a call should mark it as a procedurre
13:59<FLHerne>glx: There seems to have been a merge conflict or something?
13:59<FLHerne>glx: You added it in 38070a878c, but the current version is back to ` if value.type() != Type.INTEGER:` in
14:00*FLHerne looks for where that happened
14:00<supermop_Home>nmlc has a problem with line 34 here:
14:01<@DorpsGek>[OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master
14:01<@DorpsGek> - Update: Translations from eints (by translators)
14:01<FLHerne>glx: Oh, you changed how that works in "simplify code and enforce ()", so that might not be relevant
14:07<FLHerne>supermop_Home: Can you upload the full source somewhere?
14:11<+michi_cc>TrueBrain: For OpenGL under VMware you need to use "-v cocoa-opengl:software" as a command argument. In general, using OpenGL with software fallback is pointless over using the normal cocoa driver.
14:11<TrueBrain>yeah ... so someone else should validate that the driver works and if this should gain priority over others :D
14:12<supermop_Home>FLHerne forum pm?
14:12<FLHerne>I guess?
14:12<FLHerne>glx might be interested too
14:13<+michi_cc>Indeed, some OSX validation would be sweet. I'd still tend to put cocoa-opengl below cocoa priority-wise.
14:14<TrueBrain>from what I got from others, that sounds like the right thing to do
14:14<TrueBrain>michi_cc: if you have time, I would love your opinion on
14:15<+michi_cc>TrueBrain: Let me get a bit to eat first, but I'll look at it afterwards.
14:18<supermop_Home>is there a way i can share it on my GitHub?
14:21<supermop_Home>FLHerne do you need the sprites too?
14:22<FLHerne>(might help to see if it actually works...)
14:22<supermop_Home>ok sent forum pm
14:27<FLHerne>I don't seem to have received one
14:28<FLHerne>oh, now I have
14:29<TrueBrain>now also spam in info@ because of Twitch?
14:29<TrueBrain>boy ...
14:30<TrueBrain>spam spam spammmmmmm
14:31<TrueBrain>michi_cc: btw, works surprisingly well, and should work on all platforms .. code still needs cleaning, ofc, but it also allows to slowly move more and more stuff out of the lock, I think :D Feels a lot less dirty, doing it this way :)
14:32<TrueBrain>need to test it with some extreme games to tune it a bit :)
14:37<FLHerne>supermop_Home: I don't get your problem, but I get different problems :p
14:38<supermop_Home>maybe my nml is old?
14:43<frosch123>TrueBrain: why twitch? pretty sure it is because of steam
14:44<supermop_Home>FLHerne any other issues you see i should fix?
14:45<supermop_Home>what if i change the color names to the color numbers? maybe that will held
14:47<@DorpsGek>[OpenTTD/nml] FLHerne opened pull request #192: Fix visual_effect
14:48<+michi_cc>TrueBrain: #8726 seems to fix the issue and causes no noticeable performance issues. It has a wrong commit message though, it should read "Fix #6319: ..." (Yes, that issue is not open, but was never resolved :D)
14:48<@DorpsGek>[OpenTTD/nml] FLHerne updated pull request #192: Fix visual_effect
14:49<TrueBrain>frosch123: because they mention Twitch and Youtube, so I guess one of the two :P
14:49<+michi_cc>Ha, was even closed by you :)
14:49<FLHerne>supermop_Home: The problems I have are ^ that bug, and that nmlc doesn't find enough bits to make the random_switches independent
14:49<TrueBrain>wait, we fix closed issues now, dammit :P
14:49<FLHerne>Which might be another bug, because there are only a couple of them...
14:49<+michi_cc>TrueBrain: I'll reopen it just so you can close it with a commit :P
14:49<FLHerne>If I temporarily comment out the `independent:` constraints, it compiles
14:50<FLHerne>supermop_Home: Oh, and I replaced spritegroup_w_s_c_vehs_d to ... _c on line 4174
14:50<TrueBrain>michi_cc: I am sure DorpsGek doesn't mind either way ;)
14:50<FLHerne>Because there is no _d
14:51<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain updated pull request #8726: Change: [Win32] don't use clipping, but draw whole screen every frame
14:51<TrueBrain>there, now fixes that ancient ticket :P
14:52<TrueBrain>I just hate that I do not understand why it fails :(
14:53<@DorpsGek>[OpenTTD/OpenTTD] michicc reopened issue #6319: Cursor: mousepointer flickers when moving the mouse
14:54<@DorpsGek>[OpenTTD/OpenTTD] michicc approved pull request #8726: Fix #6319: [Win32] don't use clipping; draw whole screen every frame
14:54<TrueBrain>you really reopened it :D Ghehe :D
14:56<TrueBrain>this also mostly addresses #7006, I think
14:57<TrueBrain>as in, the reports he refer to should be fixed now. Of course any change in fps will still be noticeable, but yeah
15:03-!-Progman [] has joined #openttd
15:03-!-Progman is "Peter Henschel" on #openttd
15:03<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on pull request #8726: Fix #6319: [Win32] don't use clipping; draw whole screen every frame
15:03<TrueBrain>fixing bugs reported in 2014, sweet.
15:04<TrueBrain>michi_cc: anything else I can do to help get OpenGL in "Ready to review" state?
15:04<+michi_cc>Merge #8726 so I can rebase? :D
15:05<TrueBrain>stupid macOS CI
15:05<TrueBrain>taking for-ever
15:05<TrueBrain>vcpkg had a cache-miss again
15:05<TrueBrain>I can't wait till we ditch those GitHub Actions
15:06<+michi_cc>#8076 also refers to the problem, just in a "I'm completely missing the point" comment and not the issue proper.
15:06<TrueBrain>the dude tried hard to get it to work, but it feels like it is now doing everything in a single action, and none of it particularly well :(
15:07<TrueBrain>the original report is interesting ..
15:07<TrueBrain>sadly, the chances of reproducing that are next to zero :P
15:09<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on issue #8076: Odd Cursor Glitch When Moving Through Windows and Game
15:09<TrueBrain>lets ask if with all the latest changes, that problem still exists
15:12<TrueBrain>20 minutes in for macOS CI, and it is at "f" of the files to compile
15:12<TrueBrain>this is like watching paint dry
15:13<+michi_cc>andythenorth: Is still an issue or was that fixed during the course of the cocoa video driver refactor?
15:14<TrueBrain>the lovely toolbar issue :P
15:14<TrueBrain>that looked odd as fuck :P
15:15<+michi_cc>It might have been fixed by my change of the used cursor tracking API.
15:16<TrueBrain>I did not hear andythenorth complain about it lately .. and he normally always mentions it at least when testing macOS builds :P
15:16<andythenorth>I think you imagine that
15:16<andythenorth>I never play full screen :P
15:17<TrueBrain>it feels like the macOS runners in the GitHub farm are slow lately
15:17<@DorpsGek>[OpenTTD/OpenTTD] andythenorth commented on issue #8038: Fullscreen mode in MacOSX is almost unusable
15:17<@DorpsGek>[OpenTTD/OpenTTD] andythenorth closed issue #8038: Fullscreen mode in MacOSX is almost unusable
15:18<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain merged pull request #8726: Fix #6319: [Win32] don't use clipping; draw whole screen every frame
15:18<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain closed issue #6319: Cursor: mousepointer flickers when moving the mouse
15:18<TrueBrain>there you go michi_cc :)
15:23<@DorpsGek>[OpenTTD/nml] frosch123 updated pull request #173: Fix: Access to persistent storage of towns
15:25<frosch123>ah, forgot black again
15:25<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on issue #8019: AppImage support
15:25<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain closed issue #8019: AppImage support
15:26<TrueBrain>again?! :P
15:26<TrueBrain>I never do that
15:26<TrueBrain>nevah!!!! :D
15:31<@DorpsGek>[OpenTTD/nml] frosch123 updated pull request #173: Fix: Access to persistent storage of towns
15:31<frosch123>FLHerne: i think you need to rebase 192 first. it also contains your previous PR
15:33<frosch123>TrueBrain: it's probably a problem of old people
15:33<TrueBrain>Ugh ... getting old sucks
15:35<Timberwolf>You get wild eyebrows though, which is not to be underestimated.
15:36<supermop_Home>FLHerne i still get that error about line 4162 even if i comment out the 'independent' parts of the switches
15:36<andythenorth>Timberwolf you got any standard values for cargo aging, different vehicle types?
15:36*andythenorth curious what other grfs are doing
15:37<Timberwolf>Er... let me check my spreadsheet.
15:37<@DorpsGek>[OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver
15:38<+michi_cc>Let me add a comment on what the latest state is and what specifically needs some testing.
15:39<Timberwolf>205 for express services, 222 for long distance express, 165 for suburban, 120 for tubes (and 185 as standard)
15:39<Timberwolf>Preservation railways start at 74 (!) but increase to 5*185 if the loco is old enough. That's balanced by max speed of 25mph though.
15:39*andythenorth was wondering about standardising between grfs, was suggested in forum
15:40<andythenorth>it's a flawed mechanic, because of wildly differing map sizes, but eh
15:40<andythenorth>I have 35 for metro, 56 for high-capacity pax, and 185 for everything else
15:40<Timberwolf>In terms of anything I've done, I'd suggest "find the values which actually work, let Timberwolf know so he can fix his broken set"
15:41<andythenorth>there are no correct values :P
15:41<andythenorth>it depends on map size, and preferred play style :P
15:41<andythenorth>very 'oof'
15:42<Timberwolf>I've been pleasantly surprised by how well most of my gameplay mechanics work - I've tweaked things a bit as a result of playing, but every so often I get an, "oh, that actually worked out rather well"
15:42<andythenorth>good :)
15:42<andythenorth>most things I add, I expect to delete a few years later :P
15:43<Timberwolf>Well, I'm only 6 months in on the trains :)
15:43<andythenorth>time yet
15:43<supermop_Home>hmm changing all the color names to their numbers also does not work
15:46<@DorpsGek>[OpenTTD/OpenTTD] michicc commented on pull request #7744: Draft Feature: OpenGL video driver
15:46<+michi_cc>Calling everybody with an AMD GPU to test point 2 from my last comment on the OpenGL PR.
15:47<+michi_cc>Otherwise, I'd declare it to not be a draft anymore.
15:50<TrueBrain>michi_cc: did you fix the MSYS error btw?
15:50<+michi_cc>Yes, I think so :)
15:51<supermop_Home>FLHerne my nml was too old
15:51<supermop_Home>compiled now with independent bits commented out
15:54<TrueBrain>michi_cc: the only thing that comes to mind when reviewing, is that we might want to split cocoa and win32 in different files too
15:54<TrueBrain>like 1 class per file approach
15:54<TrueBrain>but .. nothing that cannot be done later too
15:55<+michi_cc>Maintainer pushes are allowed :D
15:55<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain approved pull request #7744: Draft Feature: OpenGL video driver
15:55<TrueBrain>I might some day
15:56<TrueBrain>up to you if you want to wait for feedback from AMD GPU users :)
15:56<+michi_cc>Eddi|zuHause: Didn't you have an AMD video card or did I misremember that?
15:57<+michi_cc>I'd love if you could test my second point in the comment
15:57<+michi_cc>milek7_: You, too, if you have time.
15:58<Eddi|zuHause>i tested that before, yes
15:58<+michi_cc>The GL_CLIENT_STORAGE_BIT is new if you uncomment the #define. It might or might not improve performance.
16:02<Eddi|zuHause>try poking me tomorrow
16:03<Eddi|zuHause>then i can probably test it
16:06-!-jottyfan [] has joined #openttd
16:06-!-jottyfan is "jottyfan" on #openttd
16:06-!-jottyfan [] has quit []
16:07-!-jottyfan is "jottyfan" on #openttd
16:07-!-jottyfan [] has joined #openttd
16:07-!-jottyfan [] has quit []
16:13<+michi_cc>TrueBrain: If I merge that now I'm hiding for the rest of the week :D GPU drivers are broken enough that the crash chance for at least somebody is 1000%.
16:13<+michi_cc>TrueBrain: Re-enabling persistent mapping on AMD is easy enough as an extra commit.
16:15<TrueBrain>It has to happen some day... :p
16:15<TrueBrain>And nobody tests nighties :D
16:16<@DorpsGek>[OpenTTD/OpenTTD] michicc merged pull request #7744: Feature: OpenGL video driver
16:16<+michi_cc>TrueBrain made me :)
16:17<nielsm>guess I'll have to do some rebasing of the win9x cleansing now
16:17<frosch123>nielsm: adding support for minigl?
16:19<milek7_>michi_cc: I'm not sure what you want me to test?
16:19<milek7_>as you noted I tested GL_CLIENT_STORAGE_BIT previously, did you change anything that could influence the result?
16:20<+michi_cc>I would like a confirmation of that just in case some other unrelated change had some unintended consequences.
16:20<milek7_>ok, will check later
16:22<+michi_cc>There's a "#define GL_MAP_PERSISTENT_AMD" at the top of src/video/opengl.cpp. If you remove the comment signs, it should use persistent buffer mapping even on AMD. I'm interesting if a) performance in okay and b) if there's any advantage over not using persistent mapping.
16:26<nielsm>smoke test for sdl2-opengl on intel gpu on fedora 33 gnome is fine at least..
16:26<@DorpsGek>[OpenTTD/OpenTTD] glx22 updated pull request #8728: Cmake fixes
16:27<+glx>redone without regexp
16:27<LordAro>now you have 98 problems!
16:32<@DorpsGek>[OpenTTD/OpenTTD] nielsmh updated pull request #8720: Remove remaining Windows 95 support
16:32<nielsm>okay just one merge conflict that needed fixing
16:34<supermop_Home>ok now to test these sprites i just need to get a whole vehicle chain going in steeltown
16:35<andythenorth>there should be a savegame :P
16:36<@DorpsGek>[OpenTTD/OpenTTD] nielsmh updated pull request #8720: Remove remaining Windows 95 support
16:36<andythenorth>supermop_Home DMed you
16:36<andythenorth>dunno if it will work
16:37<andythenorth>local openttd build, probably refuses to load in your client
16:39<@DorpsGek>[OpenTTD/OpenTTD] spnda updated pull request #7955: WIP Feature: NewGRF Road Stops (Feature 14)
16:41<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on issue #7006: Hardware mouse cursor
16:41<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain closed issue #7006: Hardware mouse cursor
16:44<TrueBrain>always takes a bit of time to type out my thought-process :D
16:46<TrueBrain>saying "no" takes more time than saying yes :P
16:47<TrueBrain>glx: if you fix my comment, we can merge your PR :)
16:51<@DorpsGek>[OpenTTD/OpenTTD] michicc opened pull request #8729: Fix 8706c36f: Change RELEASE code, too.
16:51<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain approved pull request #8729: Fix 8706c36f: Change RELEASE code, too.
16:54<frosch123>meh, opengl does not work for me :/
16:54-!-nielsm [] has quit [Ping timeout: 480 seconds]
16:54<LordAro>file a bug!
16:54<frosch123>all glgetstring return nulltpr
16:54<+michi_cc>TrueBrain: Do you want to create the biggest shitstorm ever? :D
16:54<frosch123>internet says that happens when not properly initialised
16:55<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain commented on pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack
16:55<TrueBrain>michi_cc: lolz :D Let me look at what you did ...
16:55<TrueBrain>and michi_cc , we have the first 1000% chance user ;)
16:56-!-Samu [] has quit [Quit: Leaving]
16:56<LordAro>opengl working here with msys (and 980Ti)
16:56<+michi_cc>frosch123: SDL is supposed to do that for us.
16:57<TrueBrain>michi_cc: why would that branch give a shitstorm? :)
16:58<+michi_cc>"Change", and one that doesn't even look that much better.
16:58<TrueBrain>besides it appears to be doing nothing currently .. "smooth" is added as parameter, but no callers give it :P
16:58<supermop_Home>andythenorth ideas for a grf name?
16:58<andythenorth>supermop_Home 'grf.grf'
16:58<+michi_cc>TrueBrain: Oops, looks like I didn't commit/push the remaining changes :)
16:58<frosch123>supermop_Home: 4 letters
16:59<TrueBrain>michi_cc: :D
16:59<andythenorth>oh livestream lolz
16:59<andythenorth>or gruph.grf
16:59<andythenorth>we need to call them gruphs now
16:59<andythenorth>frosch123 have you seen rocket gronk btw?
16:59<+michi_cc>TrueBrain: It should be obvious what to change to see the effect, though.
16:59<TrueBrain>it is :)
17:00<frosch123>another place suggested to make the sdlcontext current, i inserted a SDL_GL_MakeCurrent, which succeeds without error, but still no version numbers
17:02<+glx># For MSVC CMake runs this script from a batch file using || to detect errors,
17:02<+glx># depending on source path it may quote args, and cause cmd to not understand ||
17:02<+glx># and pass it as argument to ourself.
17:02<+glx>TrueBrain: ^^ enough ?
17:03<TrueBrain>very much so yes
17:03<TrueBrain>a lovely facepalm for future-us :D
17:03<frosch123>andythenorth: yes, we should play it as closing scene to the next livestream
17:04<frosch123>andythenorth: music though :/
17:04<andythenorth>oh yes, copyright problems
17:04<@DorpsGek>[OpenTTD/OpenTTD] glx22 updated pull request #8728: Cmake fixes
17:04<+glx>anyway cmake and quoting is kind of silly
17:05<@DorpsGek>[OpenTTD/OpenTTD] TrueBrain approved pull request #8728: Cmake fixes
17:05<+glx>"-DINCLUDES_SOURCE_FILE=D:/developpement/GitHub/glx22/OpenTTD - cmake error/src/script/api/" <-- seriously, who would quote like that ?
17:06<+glx>but I can understand why add_custom_command does it that way
17:07<TrueBrain>okay, OpenTTD in browser is now a lot more enjoyable to play on a higher framerate .. that is nice
17:07<TrueBrain>frosch123: what distro (and version) are you on?
17:08<frosch123>debian stable
17:08<TrueBrain>remind me what name stable carries?
17:08<frosch123>x11 backend, no wayland
17:08<frosch123>that's what i wanted to avoid to look up :p
17:09<TrueBrain>lsb_release -c
17:09<TrueBrain>or -a, and pick the right line
17:09<TrueBrain>LibSDL 2.0.9 .. a bit old, but not that old
17:09<TrueBrain>thought, it might be related, but seems unlikely
17:10<+michi_cc>Some mesa or whatever component that is not installed by default?
17:10<frosch123>well, factorio works :p
17:10<frosch123> <- glxinfo is also pretty happy
17:10<+glx>yeah a free hl
17:11<+glx>I love opengl discussions
17:11<+michi_cc>At least OTTD will fall back to non-gl video drivers.
17:11<frosch123>yes, i can confirm that works :)
17:11<milek7_>oh no, nvidia driver
17:11<frosch123>yep, non-free :p
17:12<frosch123>i bought that card for factorio only
17:12<TrueBrain>frosch123: if you use SDL_GL_GetProcAddress("glGetString") to get the proc
17:12<TrueBrain>does that change anything?
17:15<FLHerne>frosch123: I did rebase it, that's what the second push was
17:15<FLHerne>Github seems unhappy
17:16<@DorpsGek>[OpenTTD/OpenTTD] glx22 commented on pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack
17:16<@DorpsGek>[OpenTTD/OpenTTD] michicc approved pull request #8727: Update known-bugs.txt
17:16<@DorpsGek>[OpenTTD/OpenTTD] michicc merged pull request #8729: Fix 8706c36f: Change RELEASE code, too.
17:17<milek7_>michi_cc: why it does request compat-profile by default?
17:17<FLHerne>Oh, I know what happened
17:17<+michi_cc>It does not on Win32 and OSX. For SDL, ask SDL :)
17:17<@DorpsGek>[OpenTTD/OpenTTD] glx22 closed issue #8687: CMake include error when building new project.
17:17<@DorpsGek>[OpenTTD/OpenTTD] glx22 merged pull request #8728: Cmake fixes
17:18<@DorpsGek>[OpenTTD/nml] FLHerne updated pull request #192: Fix visual_effect
17:18<frosch123>TrueBrain: yay, that does work
17:18<frosch123>ugly function pointer casts though :)
17:18<TrueBrain>michi_cc: seems you need to wrap glGetString the same as glGetStringi :)
17:19<TrueBrain>SDL uses it to dynamically load gl extensions, it seems
17:19<+michi_cc>If glGetString needs to be wrapped, I would expect everything needs to be wrapped, as that is OGL 1.0 or so.
17:20<TrueBrain> <- lol @ hardcoded vendor names :)
17:23<+michi_cc>frosch123: Does it display something then? It would feel really strange if only that singular call needs to be dynamically loaded.
17:24-!-mirrorbird [~mirrorbir@2a00:801:446:6dfc:fadb:e310:6ed2:16d] has joined #openttd
17:24-!-mirrorbird is "mirrorbird" on #openttd
17:24<frosch123>i have to replace all getstring first
17:25<frosch123>when i only faked the version, it failed on no-textures next
17:26<frosch123>but that's something for tomorrow to try
17:27<+michi_cc>Modern Linux OpenGL should use a central dispatch table that all calls (no matter if linked dynamically or statically) go through. The non-free NVidia driver might bypass that table somehow.
17:28<+michi_cc>And yes, technically we could dynamically load all functions, its just annoying to code :)
17:28<frosch123>glGetIntegerv would be next :p
17:28<TrueBrain>SDL does it in their codebase, it is absolutely annoying to read :P
17:28<TrueBrain>but SDL loads everything dynamic, so for them it makes sense I guess :)
17:29<TrueBrain>I can try and see what the non-free nvidia driver does for me tomorrow
17:29<TrueBrain>might give one more datapoint :)
17:29<frosch123>hmm, can i ld-preload something?
17:30<frosch123>oh great, openttd links libwayland-egl
17:30<frosch123>now i am confused
17:30<milek7_>do you have libglvnd installed?
17:32<milek7_>>its just annoying to code :)
17:32<milek7_>I think using autogenerated loaders is nice
17:32<TrueBrain>Hmm, I should check the linux-generic build tomorrow, see if that has blown up
17:33<frosch123>i have a pkgconfig file for libglvnd, but can't find the lib
17:36<@DorpsGek>[OpenTTD/OpenTTD] glx22 updated pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack
17:37<frosch123>yay, LD_PRELOAD worked :p
17:37<frosch123>now i see stuff
17:39-!-moll [] has quit [Ping timeout: 480 seconds]
17:40<frosch123>in my testcase non-opengl is at 7ms for video output, opengl is at 0.06ms
17:41<frosch123>which results in 3x the simulation rate
17:41<TrueBrain>Ain't that something :D
17:41<frosch123>i wanted to buy a new cpu thoug
17:41<TrueBrain>So how many game months go by per second?
17:41<+glx>compiling current master (debug build as always)
17:41<frosch123>i never had a mainboard so old (11 years), i am scared it blows up somewhen
17:42<TrueBrain>Haha, that is old
17:42-!-moll [] has joined #openttd
17:42-!-moll is "Andri Möll" on #openttd
17:42<+glx>and pressed start
17:42<+glx>let's see
17:43-!-k-man [] has joined #openttd
17:43-!-k-man is "Jason Lewis" on #debian-au #debian-next #debian #oftc #openttd #virt
17:44<+michi_cc>Okay, GL is officially annoying. OSX defines PFNGLGETSTRINGPROC, Windows does not. Debian Buster doesn't either, but who knows what other Linuxes/Unixes do.
17:45<TrueBrain>Revert? :D
17:46<+glx>wow FFWD in debug build exists
17:46-!-jjavaholic [~jjavaholi@2a00:23c6:9508:e601:8d11:5425:637b:40f2] has joined #openttd
17:46-!-jjavaholic is "realname" on #openttd
17:46<TrueBrain>I have not tried Release yet ...
17:47<+glx>I get above 20x
17:48<TrueBrain>Happy days :)
17:48<+glx>well mouse is still laggy with modal windows
17:48<TrueBrain>Yeah .. I have plans for that :D
17:49<TrueBrain>As that is solvable now :)
17:49<+glx>ah yes drawing is independant of input now
17:50-!-andythenorth [] has quit [Quit: andythenorth]
17:50<TrueBrain>Right, time to zzzzzz .. dream of sweet smooth mouse movement :)
17:52<@DorpsGek>[OpenTTD/nml] frosch123 approved pull request #192: Fix visual_effect
17:52-!-frosch123 [] has quit [Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn]
17:53<@DorpsGek>[OpenTTD/nml] FLHerne merged pull request #192: Fix visual_effect
18:01<milek7_>michi_cc: tbh, I feel trying to support opengl <2.0 is.. unreasonable ;P
18:02<+michi_cc>It does not support OpenGL < 2.0. What it does support is the equivalent of OpenGL 3.2 build up from the needed extensions.
18:03<+michi_cc>We only need a select few parts of 3.2.
18:03<milek7_>yeah but it tries to use extensions all the way down to 1.3
18:03<milek7_>like, selecting between glActiveTexture vs glActiveTextureARB
18:04<LordAro>mm, 5000fps on release mode
18:04<+michi_cc>On Windows, anything above 1.1 has to be dynamically linked anyway, so that if doesn't hurt.
18:04-!-HerzogDeXtEr [] has quit [Read error: Connection reset by peer]
18:04<LordAro>not even an empty map
18:07-!-Wolf01 [] has quit [Quit: Once again the world is quick to bury me.]
18:16-!-Tirili [] has joined #openttd
18:16-!-Tirili is "realname" on #openttd
18:25<@DorpsGek>[OpenTTD/OpenTTD] LordAro approved pull request #8727: Update known-bugs.txt
18:25<@DorpsGek>[OpenTTD/OpenTTD] LordAro merged pull request #8727: Update known-bugs.txt
18:26<@DorpsGek>[OpenTTD/OpenTTD] michicc opened pull request #8730: Codechange: [OpenGL] Load all OpenGL functions dynamically.
18:26<@DorpsGek>[OpenTTD/OpenTTD] LordAro approved pull request #8720: Remove remaining Windows 95 support
18:27<+michi_cc>Will Linux fail? Probably yes.
18:30<@DorpsGek>[OpenTTD/OpenTTD] LordAro approved pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack
18:40<milek7_>michi_cc: alternatively:
18:46<@DorpsGek>[OpenTTD/OpenTTD] stormcone opened issue #8731: OpenGL font rendering
18:46-!-sla_ro|master [] has quit []
18:49<milek7_>(as system ogl headers are not used at all, I removed cmake opengl searching and linking)
18:49<milek7_>(and for win32 it also needs your win32_v.cpp patch)
18:53-!-jottyfan [] has joined #openttd
18:53-!-jottyfan is "jottyfan" on #openttd
18:53-!-glx [] has quit []
18:54-!-jottyfan [] has quit []
18:59<@DorpsGek>[OpenTTD/OpenTTD] Milek7 commented on pull request #8730: Codechange: [OpenGL] Load all OpenGL functions dynamically.
19:02<milek7_>and now persistent mapping makes practically no difference
19:24-!-mirrorbird [~mirrorbir@2a00:801:446:6dfc:fadb:e310:6ed2:16d] has quit [Quit: Leaving]
19:24-!-hotgirl55 [~hunkofmon@2a00:801:446:6dfc:fadb:e310:6ed2:16d] has joined #openttd
19:24-!-hotgirl55 is "hunkofmonk" on #openttd
19:33-!-Progman [] has quit [Remote host closed the connection]
19:56-!-gelignite [] has quit [Quit: Stay safe!]
20:01-!-Flygon [~Flygon@2001:44b8:411e:4e00:a80f:60fd:5a8c:b680] has joined #openttd
20:01-!-Flygon is "Flygon" on #openttd
20:14-!-Tirili [] has quit [Remote host closed the connection]
20:30-!-WormnestAndroid [~WormnestA@] has quit [Remote host closed the connection]
20:32-!-WormnestAndroid [~WormnestA@] has joined #openttd
20:32-!-WormnestAndroid is "WormnestAndroid" on #openttd
20:37<FLHerne>hotgirl55: Hi
20:37<FLHerne>(this channel is usually a bit quiet during EU nighttime)
20:38<FLHerne>supermop_Home: I looked at it a bit more, the probabilities in switch_color_vehs_a are just too complicated for the number of random bits available
20:39-!-WormnestAndroid [~WormnestA@] has quit [Remote host closed the connection]
20:39<FLHerne>If you make them all `1` (i.e. equal probabilities) it works
20:41-!-WormnestAndroid [~WormnestA@] has joined #openttd
20:41-!-WormnestAndroid is "WormnestAndroid" on #openttd
20:42<FLHerne>As it is, the chances are n/41, so NML needs 6 random bits, and the other one needs 4, while there are only 8 available
20:42<FLHerne>Making it even is a straight 1/16 chance, so it only takes 4 bits and is possible to implement
20:45*FLHerne sleeps
20:59-!-Wormnest [~Wormnest@] has quit [Quit: Leaving]
21:42<@DorpsGek>[OpenTTD/OpenTTD] 2TallTyler commented on issue #8686: Unable to change townnames in Scenario editor after pr8566
21:54<supermop_Home>FLHerne thanks!
22:23-!-D-HUND [~debdog@2a00:79c0:603:ab00:7a24:afff:fe8a:d04d] has joined #openttd
22:23-!-D-HUND is "Wowbagger" on #openttd
22:27-!-debdog [~debdog@2a00:79c0:62d:2b00:7a24:afff:fe8a:d04d] has quit [Ping timeout: 480 seconds]
22:35-!-hotgirl55 [~hunkofmon@2a00:801:446:6dfc:fadb:e310:6ed2:16d] has quit []
22:35-!-muffindrake2 [] has joined #openttd
22:35-!-muffindrake2 is "muffindrake" on #debian-next #openttd
22:37-!-muffindrake1 [] has quit [Ping timeout: 480 seconds]
23:46<@DorpsGek>[OpenTTD/OpenTTD] 2TallTyler opened pull request #8732: Change: Improve graph period markings
---Logclosed Tue Feb 23 00:00:51 2021