Log for #openttd on 15th January 2021:
Times are UTC Toggle Colours
00:12:15  <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8574: Feature: Ships stop to be passed by another ship
00:12:35  *** Wormnest has quit IRC
00:27:31  *** SmatZ has quit IRC
00:28:11  *** SmatZ has joined #openttd
00:34:26  *** gelignite has quit IRC
00:37:02  *** kevin has joined #openttd
00:37:36  *** kevin is now known as Guest10484
00:40:00  *** Guest10484 has left #openttd
00:41:38  *** kevin_ has joined #openttd
01:16:17  *** kevin_ has quit IRC
01:33:49  *** Lejving_ has joined #openttd
01:36:56  *** Lejving has quit IRC
02:01:20  *** aliasbkilgrinhu[m] has joined #openttd
02:09:29  *** Flygon has joined #openttd
02:52:45  *** orudge has quit IRC
02:52:46  *** orudge has joined #openttd
03:08:45  *** debdog has joined #openttd
03:12:08  *** D-HUND has quit IRC
03:25:02  *** heffer has quit IRC
03:25:24  *** heffer has joined #openttd
03:51:11  *** glx has quit IRC
03:58:53  *** berndj has quit IRC
03:59:27  *** berndj has joined #openttd
04:37:17  *** SpComb has quit IRC
04:37:19  *** SpComb has joined #openttd
04:42:41  *** supermop_Home_ has quit IRC
05:54:05  *** Exec has quit IRC
05:54:05  *** Exec has joined #openttd
06:03:38  *** Webster has joined #openttd
06:46:30  *** Gustavo6046_ has joined #openttd
06:48:23  *** Gustavo6046_ has joined #openttd
06:50:41  *** Gustavo6046 has quit IRC
06:50:41  *** Gustavo6046_ is now known as Gustavo6046
07:05:09  *** andythenorth has joined #openttd
07:11:30  *** Smedles has joined #openttd
07:21:41  *** sla_ro|master has joined #openttd
07:32:13  *** azubieta has quit IRC
07:32:38  *** azubieta has joined #openttd
08:58:25  *** DasPoseidon has joined #openttd
09:03:18  *** snail_UES_ has quit IRC
09:48:06  <TrueBrain> orudge: and, figured out a way to make the CA happy? :D How are you looking time-wise for the OSX website changes? Anything I can help with or do for you?
10:05:52  *** andythenorth has left #openttd
10:07:51  *** TinoDidriksen has quit IRC
10:08:02  *** TinoDidriksen has joined #openttd
10:08:36  *** TinoDidriksen is now known as Guest10515
10:11:15  <TrueBrain> haha, found a nice "bug" with the 75% screen detection :D It now spans for 75% over my 2 screens by default :P
10:11:36  <TrueBrain> high chance this is due to the fact I am running X11 on Windows, and it isn't aware I have 2 screens :P
10:11:57  <LordAro> oh yeah, i've seen that before with xforwarding stuff
10:12:58  <TrueBrain> would be weird if it was not :)
10:13:04  *** Guest10515 is now known as TinoDidriksen
10:13:23  <TrueBrain> so I guess it thinks I have a 32:9 screen :D
10:17:14  <TrueBrain> was doing some random profiling of OpenTTD ... in fast-forward on an empty map, most time is still spend in drawing the screen :P
10:17:46  <TrueBrain> despite it only drew like once every 3 in-game months :D
10:18:55  <TrueBrain> I assume ReadSprite is cached, so this might mean I just need to run it longer :)
10:33:20  <TrueBrain> yeah, indeed, it takes more CPU time than landscape generation :P
10:33:21  <TrueBrain> nice :D
10:33:45  <TrueBrain> funny how expensive PaletteAnimation is :D No shocker :P
10:34:38  <LordAro> need that opengl driver :)
10:34:45  <TrueBrain> it is why I was profiling :)
10:35:03  <TrueBrain> DrawString is also not a small one
10:35:15  <TrueBrain> drawing the status bar every time also has somewhat of an impact
10:36:02  <TrueBrain> DrawViewport, after caching of everything, 3% CPU. DrawString, 6% CPU
10:36:14  <TrueBrain> (PaletteAnimation: 45%)
10:36:58  <TrueBrain> lol, over 50% of the time of DrawString is spend in DrawLayoutLine :)
10:37:11  <TrueBrain> funny :D
10:37:52  <TrueBrain> which draws strings btw .. did not expect that by reading the name :)
10:39:04  <TrueBrain> anyway ... let's see if I can understand any of this OpenGL stuff ..
10:40:45  <TrueBrain> PFNWGLCREATECONTEXTATTRIBSARBPROC  <- these names, lol
10:43:48  <orudge> TrueBrain: bank account is set up at last. Not looked at website stuff yet but maybe today. Will try to phone CA.
10:44:20  <orudge> Might need to set up a £1/month VoIP line if they're being awkward about it
10:44:27  <TrueBrain> orudge: if there is anything we can do to help, just let us know :) Especially the website stuff, is a bit blocking for a beta release I would say. Not that we are in a real rush with that, but .. if we can help, that is not a problem :)
10:44:46  <orudge> Shouldn't be hard
10:45:31  <TrueBrain> oeh, and bank account you say .. so Steam access is also getting closer :D Nice :)
10:50:12  <TrueBrain> boy, understanding what the OpenGL PR does while remembering how SDL OpenGL worked .. this will be fun :P
10:53:39  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #8563: Prepare for 1.11.0-beta1 release
10:57:51  <TrueBrain> dbg: [driver] OpenGL driver: NVIDIA Corporation (1.4 (4.6.0 NVIDIA 457.51))
10:57:57  <TrueBrain> well, at least I have a valid OpenGL context :)
11:00:58  <TrueBrain> sadly ... the version is a bit old :P
11:11:01  <DorpsGek> [OpenTTD/OpenTTD] ldpl commented on pull request #8574: Feature: Ships yield when passing through one another
11:13:44  <TrueBrain> dbg: [driver] SDL2: Couldn't allocate OpenGL context: Single element texture formats not supported
11:13:45  <TrueBrain> awwwwhhhh
11:24:37  <orudge> TrueBrain: so, fun fact - both Safari and Firefox, ARM64 native builds, report "Intel" on the Mac
11:24:56  <orudge> and the updated ua detection JS doesn't pick up the CPU on Mac at all
11:25:05  <orudge> Apparently Apple do this to avoid 'breaking compatibility'
11:25:34  <orudge> Apparently there's a way of detecting it by probing WebGL stuff, but that seems... a bit wrong to me
11:25:53  <LordAro> user agent intensifies
11:26:11  <orudge> Maybe at some point Apple will switch the user agent
11:26:23  <orudge> but for the moment, we're out of luck there I think
11:26:32  <orudge> So we can either leave it as it is, or push for a Universal build instead
11:33:57  <LordAro> leave as is, i would say
11:40:19  *** Flygon has quit IRC
11:42:20  *** Flygon has joined #openttd
11:43:28  *** Samu has joined #openttd
11:49:18  <_dp_> should I use enum classes for new prs or just stick to plain enums?
11:50:14  <LordAro> don't think we've used enum classes anywhere yet
11:50:23  <_dp_> there are 2 xD
11:50:30  <LordAro> well.
11:50:37  <_dp_> ElementFloat and MidiSysexMessage
11:54:17  <TrueBrain> orudge: that is just stupid :P But not surprising ..
11:55:49  <orudge> TrueBrain: I can push my Universal branch, I know it will use more bandwidth but it might be less confusing for end users, or we can hope they are able to read and pick the right installer :)
11:56:15  <orudge> Do we have any stats on what downloads are most popular, how many people download the Mac build just now anyway, etc?
12:00:59  <TrueBrain> not sure it matters; if people download the wrong one hurts more ;) As I said earlier, it is not about not wanting universal, it is about it being the wrong solution to a problem :) So we tried to start from the proper angle. But if Apple is doing its hardest to make that impossible .. ugh
12:01:17  <TrueBrain> one thing in all the years of doing IT stuff I noticed: you have the rest of the world, which is not always easy but doable .. and than you have Apple :P
12:01:24  <TrueBrain> we used to say that about Microsoft :D
12:03:45  <TrueBrain> owh, the WebGL solutions you mention, are comparing hardware of known Apples .. that is a horrible idea :D
12:04:30  <TrueBrain> who ever wrote that stack overflow should be ashamed!
12:06:11  <TrueBrain> (well, not really, but you get what I mean :D)
12:26:01  *** DasPoseidon has quit IRC
12:28:10  *** DasPoseidon has joined #openttd
12:30:39  <TrueBrain> oef, I forgot it takes 5 minutes to build vcpkg :P Taking for ever if you are waiting for it :D
12:32:08  *** DasPoseidon has quit IRC
12:35:09  <_dp_> hm, a random idea that's impossible without GS industry production control: have industries issue quests to increase production
12:35:19  <_dp_> e.g. distribute goods to 5 different cities
12:40:02  * _dp_ misread 1.11 changelog and thought it has full production control somehow
12:41:13  <TrueBrain> orudge: I myself was thinking about something like this:
12:41:23  <TrueBrain> the diff looks a bit weird, but it basically just adds a job :)
12:41:42  <TrueBrain> I am still testing it, but it is a bit slow to test :D
12:44:52  *** DasPoseidon has joined #openttd
12:49:19  <orudge> OK, I was thinking about how it could run as 2 jobs, but hadn't got as far as writing code :)
12:49:29  <orudge> Can test and see how it works
12:50:45  <TrueBrain> yeah, I thought: I can annoy you with it and that it possibly doesn't work at all, or I can try it myself :P
12:50:54  <TrueBrain> it is still running to see if it works at all
12:50:59  <TrueBrain> but you will have to test if signing works :D
12:54:22  <TrueBrain> meh, it seems to fail on a lazy-thing I tried, but it takes me 10+ minutes to get back to that same point :D
12:54:23  <TrueBrain> haha
12:55:31  *** WormnestAndroid has quit IRC
12:55:44  *** WormnestAndroid has joined #openttd
13:13:36  <DorpsGek> [OpenTTD/OpenTTD] Eddi-z updated pull request #8444: Feature: Close adjacent level crossings as if they were one large crossing
13:13:54  <DorpsGek> [OpenTTD/OpenTTD] Eddi-z updated pull request #8444: Feature: Close adjacent level crossings as if they were one large crossing
13:17:49  <TrueBrain> that moment a step is doing something, but you don't know what, and it is taking a lot lnoger than it should :P Lol
13:18:45  <TrueBrain> I have no clue if this is because it fails to sign or something, or that it is actually broken
13:18:56  <TrueBrain> CPack: - Run preinstall target for: openttd
13:18:58  <TrueBrain> was hanging
13:19:09  <TrueBrain> orudge: did you test without secrets available, if all steps worked?
13:19:31  <TrueBrain> (just to avoid me going down the wrong rabbithole :D)
13:26:47  <DorpsGek> [OpenTTD/OpenTTD] Eddi-z updated pull request #8556: Feature: Allow diagonal tracks on level crossings
13:34:16  <orudge> TrueBrain: yes, and it should do
13:34:24  <TrueBrain> hmm
13:34:27  <orudge> You do get an error in the "importing certificate" stage but it's set up to not fail on error
13:34:37  <TrueBrain> cpack just hangs for ever and ever :
13:34:42  <TrueBrain> but I think I know what I can do about that :D
13:35:18  <TrueBrain> just not as elegant as what I have now .. hmm
13:35:32  <TrueBrain> always this balance between effort and result :D
13:36:31  <TrueBrain> cpack just freeze up, by the looks of it
13:36:32  <orudge> You're not actually running make after cmake
13:36:35  <orudge> will cpack work in that state?
13:36:41  <TrueBrain> it should, according to docs
13:36:49  <orudge> I can try it out on my own machine
13:37:28  <orudge> Try a cpack -v?
13:37:31  <orudge> -V even
13:38:31  <TrueBrain> okay, it is cpack doing a full recompile on a single core
13:38:34  <TrueBrain> that is not what it is suppose to do
13:38:40  <TrueBrain> it only does that for certain targets
13:38:40  <TrueBrain> lolz
13:38:50  <orudge> You could copy over all the state of the "build" folder from one of the platforms
13:38:57  <orudge> then just cpack on top of that
13:38:59  <orudge> not really ideal
13:39:05  <orudge> but might do the job
13:39:20  <orudge> (well, lipo then cpack, of course)
13:40:22  <TrueBrain> yeah, that was what I wanted to try, only ro realise artifacts have to be as few files as possible, so yet-another-zip :)
13:40:27  <TrueBrain> it is just a bit odd :P
13:40:33  <TrueBrain> should I be surprised? Guess not
13:41:15  <orudge> I might at some point try to get the Windows Store (appx) generation going in Azure Pipelines too. That's just the binary files from the regular installer wrapped up and signed, though.
13:41:58  <TrueBrain> can't it be done in GitHub Actions? :D
13:47:52  *** azubieta has quit IRC
13:48:09  *** azubieta has joined #openttd
13:48:31  <orudge> TrueBrain: eh, that's what I meant, yes
13:48:46  <TrueBrain> :D
13:49:30  <TrueBrain> Steam, Windows Store, Windows signed, Mac signed .. can 2021 be the year OpenTTD gets modern? :P
13:50:25  <Eddi|zuHause> what about OpenGL? :p
13:54:06  <orudge> Vulkan surely these days
13:54:17  <orudge> and Metal on the Mac, of course
13:54:20  <orudge> because Apple has to be different
13:54:28  <Xaroth> Isn't that a trend with Apple
13:57:31  <TrueBrain> right, I finally got SDL2 working on Windows .. took long enough
13:57:38  <TrueBrain> orudge: if you want to build us a PR for Vulkan, sure :P
13:58:39  *** snail_UES_ has joined #openttd
13:59:58  <TrueBrain> okay, OpenGL context created and everything .. now for the drawing itself ...
14:02:34  *** Flygon has quit IRC
14:02:34  *** nielsm has joined #openttd
14:06:47  <TrueBrain> funny, michi_cc added a NULL instead of a nullptr :P Not judging, just ... well .. judging :P
14:07:35  <orudge> TrueBrain: :D
14:07:48  <orudge> Phoned up about the code signing certificate, got somebody on a very noisy line that was quite hard to understand
14:07:55  <orudge> I think hopefully they are going to take another look at the link I sent
14:07:56  <orudge> but we will see
14:08:19  <TrueBrain> pam pam pammmmm
14:08:22  <TrueBrain> and Steam? :D :D
14:08:36  <orudge> Not even started with that yet, was waiting for a bank account :P
14:08:39  <orudge> but now that that's in place...
14:08:41  <LordAro> TrueBrain: the branch definitely predates NULL -> nullptr, not surprising one got missed :p
14:09:01  *** Timberwolf has quit IRC
14:09:05  *** Timberwolf has joined #openttd
14:09:39  <TrueBrain> found a way to crash everything, which is nice :)
14:12:25  <Eddi|zuHause> sometimes you're just in the mood for that :p
14:13:49  <TrueBrain> not sure these are bugs or intended behaviour :D
14:14:30  <Eddi|zuHause> "undocumented behaviour" :p
14:14:49  <TrueBrain> I am now just hacking around the code a bit in order to get something workable :P
14:14:49  *** glx has joined #openttd
14:14:49  *** ChanServ sets mode: +v glx
14:17:12  <TrueBrain> okay, it renders in an GL context now
14:17:14  <TrueBrain> that was easy
14:17:29  <TrueBrain> just need to talk to michi_cc to understand why the 40bpp-anim crashes on me :D
14:22:04  <TrueBrain> funny how much faster OpenGL is
14:22:12  <TrueBrain> from ~0.2ms per video frame to 0.00
14:22:17  <TrueBrain> it doesn't even show up in the performance thing
14:24:18  <Eddi|zuHause> anywhere i can try that?
14:24:33  <Eddi|zuHause> because display stuff is like 10ms over here
14:24:44  <TrueBrain> for Windows only, atm
14:24:54  <Eddi|zuHause> right, not helping then :)
14:26:01  <TrueBrain> and I have a VERY ugly implementation on SDL2
14:26:06  <TrueBrain> needs a bit more work before that becomes sane
14:27:25  <TrueBrain> SDL2 seems to cap the max-speed on FF more than Win32, which is interesting
14:27:27  <TrueBrain> but not for now
14:29:23  <Eddi|zuHause> well, with 10ms per frame just on displaying i'm barely getting by on normal speed, not even thinking about FF
14:31:59  <TrueBrain> I don't really get the win32 code :P It tracks a "dirty_rect" but never uses it, it seems to me
14:33:04  <TrueBrain> owh, wait, there
14:35:24  <LordAro> lol
14:36:10  <Samu> i found a bug with vehicle profits in the list of vehicles
14:36:20  <Samu> they're not being updated in real time, they look paused
14:36:23  <orudge> No confirmation e-mail from Steam, do we already have an account I wonder?
14:36:40  <TrueBrain> orudge: I know I do, as I tried that process, but I think I used truebrain@
14:37:21  <TrueBrain> yeah, I did
14:40:13  <Samu> profit this year is paused
14:43:04  <TrueBrain> ah, the OpenGL backend needs a resize event to work, oops .. missed that part :D
14:52:29  <DorpsGek> [OpenTTD/OpenTTD] SamuXarick opened issue #8575: Profit for each vehicle in vehicle lists aren't being updated regularly
14:52:58  <TrueBrain> okay, so it no longer crashes with animation, but it now just hangs
14:53:00  <TrueBrain> improvements!
14:54:23  <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #8575: Profit for each vehicle in vehicle lists aren't being updated regularly
14:54:28  <Samu> damn, getting the version of openttd is the boring part
14:55:08  <LordAro> Samu: first line of `./openttd -v`
14:55:37  <Eddi|zuHause> from crashing to hanging is definitely a step upwards :p
14:57:39  <DorpsGek> [OpenTTD/OpenTTD] ldpl opened pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window
14:58:04  <TrueBrain> I find some many odd things I don't understand .. but learning how the win32 driver works while doing this, is a bit lot to take in :D Getting there ...
14:58:57  <DorpsGek> [OpenTTD/OpenTTD] SamuXarick commented on issue #8575: Profit for each vehicle in vehicle lists aren't being updated regularly
15:00:06  <TrueBrain> orudge: my approach is not going to work, at all, so we can forget about that :P
15:00:17  <TrueBrain> orudge: the minor thing I don't like about the approach you took, is that there is a lot of duplication going on
15:00:27  <TrueBrain> but for  now I don't see an easy way around that :)
15:00:44  <TrueBrain> but tl;dr, cpack needs all the dependencies available too, like all vcpkg stuff, build-host, .....
15:01:10  <TrueBrain> I do have some minor comments on your branch, but if you PR it, I can fill those in :)
15:01:22  <DorpsGek> [OpenTTD/OpenTTD] James103 commented on pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window
15:03:15  <orudge> TrueBrain: yes, that is a bit unsatisfactory for me too
15:03:18  <orudge> I'll do a PR though
15:03:40  <TrueBrain> I assumed cpack would do a lot less than it does :)
15:04:22  <DorpsGek> [OpenTTD/OpenTTD] ldpl updated pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window
15:05:00  *** Der_Herr[m] has left #openttd
15:06:30  <Samu> autoreplace/autorenew is incredibly huge
15:06:34  <Samu> the code
15:06:45  <TrueBrain> that's what she said
15:06:49  <TrueBrain> giggedy
15:07:49  <glx> hey at least autoreplace/autorenew is easier to read than yapf :)
15:07:53  <DorpsGek> [OpenTTD/OpenTTD] orudge opened pull request #8577: Feature: Build macOS build as a universal binary
15:08:13  <Samu> I think yapf was easier
15:08:47  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
15:09:11  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
15:09:22  <TrueBrain> can I comment in peace ffs? :P
15:09:38  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
15:09:45  <orudge> Sorry
15:09:55  <orudge> I realised I'd made the change regarding the trailing slash yesterday but hadn't commited it
15:10:01  <orudge> then pushed without actually adding the change
15:10:04  <Samu> so, i'm changing the order some stuff is executed in autoreplace, because of profit_last_year on a group / vehicle not being synced
15:10:06  <orudge> Brain not working today :P
15:10:39  <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8574: Feature: Ships yield when passing through one another
15:10:49  *** supermop_Home has joined #openttd
15:10:54  <Samu> but... given the huge amount of code involving autoreplace, I am not sure I'm breaking something else
15:11:14  <Xaroth> Better test thoroughly then.
15:12:16  <Samu> it involves cargoes, engines, refits, capacities, it's a whole new level of complexity for me
15:12:21  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8577: Feature: Build macOS build as a universal binary
15:12:24  <Samu> and groups too
15:12:30  <Samu> which is the part im touching
15:14:00  <supermop_Home> hi
15:14:05  <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #8575: Profit for each vehicle in vehicle lists aren't being updated regularly
15:16:45  <TrueBrain> hmm ... not sure I was really using GL to draw surface :P Ghehe, ironic
15:26:40  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
15:27:57  <TrueBrain> orudge: "make package" on line 462 should become "make" to I think?
15:28:02  <TrueBrain> (in the x64 build job)
15:28:17  <orudge> Oh yeah
15:28:18  <orudge> Thanks
15:28:35  <TrueBrain> otherwise it looks good to me
15:28:43  <TrueBrain> if you can test it before we merge, that would be nice :D
15:28:51  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
15:28:58  <orudge> Just doing so now :)
15:29:34  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8577: Feature: Build macOS build as a universal binary
15:38:16  <TrueBrain> michi_cc: I have a GL context working, but blittering so far is a black screen :D Pretty sure I am doing something silly, so I will figure that out. Anyway, the SDL driver needs a similar refactor as you did with win32 for this, and the win32-opengl extension will look a lot like the sdl-opengl one :P
15:38:26  <TrueBrain> not sure yet what we want to do with that ...
15:38:31  <TrueBrain> either way, your PR needs a rebase first :D
15:40:39  <milek7_> curious how is performance with that opengl blitter from forums
15:40:53  <milek7_> but code-quality wise it needs lot of work
15:49:31  <DorpsGek> [OpenTTD/OpenTTD] orudge dismissed a review for pull request #8577: Feature: Build macOS build as a universal binary
15:49:34  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
15:50:58  *** Wormnest has joined #openttd
15:58:57  *** frosch123 has joined #openttd
16:00:54  <Eddi|zuHause> what exactly happened to GetSlopeZ?
16:01:34  <LordAro> still exists?
16:01:53  <Eddi|zuHause> my grep came up pretty empty
16:02:09  <LordAro> oh, GetTileZ ?
16:02:12  <LordAro> or is that different
16:04:55  *** iSoSyS has joined #openttd
16:05:13  <frosch123> z positions distinguish height levels and world/pixel coordinates, since more-height-levels
16:05:19  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
16:05:29  <frosch123> so, whenever you have a 10 year old patch, check for PixelZ instead of Z
16:06:14  <Eddi|zuHause> right, needs some investigation what exactly was meant
16:10:16  *** iSoSyS has quit IRC
16:22:49  <DorpsGek> [OpenTTD/OpenTTD] SamuXarick commented on pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui
16:24:41  *** Wolf01 has joined #openttd
16:25:00  <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8577: Feature: Build macOS build as a universal binary
16:25:22  <orudge> Third time lucky, I hope
16:27:16  <Samu> who has the patience to 8577
16:27:30  <Samu> oops 7918, my bad
16:42:01  *** Lejving_ has quit IRC
17:06:59  <Eddi|zuHause> well... almost? :p
17:09:58  *** tokai has joined #openttd
17:09:58  *** ChanServ sets mode: +v tokai
17:12:16  <supermop_Home> looks good eddi
17:14:54  <Eddi|zuHause> i think i missed something in the sprite loading code
17:15:45  <Timberwolf> The question mark and the Chaney Jubilee, the two favourite sights of anyone who does anything related to sprites. :)
17:16:57  *** tokai|noir has quit IRC
17:17:13  *** _2TallTyler has joined #openttd
17:18:54  <Eddi|zuHause> gnarf... now i need to patch nforenum to accept the new action5 type...
17:20:13  <Timberwolf> nml branch when? ;)
17:20:32  *** andythenorth has joined #openttd
17:20:48  <Timberwolf> I'm interested to see what horror adapting my abuse of what the crossing sprites allows turns up for diagonal tracks.
17:21:15  <Eddi|zuHause> there won't be support for railtypes for now
17:22:33  <Eddi|zuHause> hmz... i must have messed with nforenum data files before, but i don't remember how
17:28:07  <frosch123> there is 5 years of stuff left to add to nforenum
17:28:35  <frosch123> noone deleted it by accident :)
17:28:47  <Eddi|zuHause> my checkout is from 2017
17:28:55  <Eddi|zuHause> i'm assuming the repo moved? :p
17:29:39  <supermop_Home> where did we land on crossing sprites coming from roads vs railtypes?
17:30:03  <supermop_Home> all my roads have a crossing sprite but it doesn't get used
17:30:05  <Eddi|zuHause> there's a road sprite and a rail overlay
17:30:20  <supermop_Home> I might need to look into that
17:30:34  <andythenorth> yo
17:30:40  <supermop_Home> yo andythenorth
17:30:58  * andythenorth so broken
17:31:13  <Eddi|zuHause> i'm not entirely understanding it either, but currently i'm trying to fit this old code into the case where the railtype has no overlay defined
17:31:33  <supermop_Home> oof I just told my friend in LA i'd mail them my extra copy of this Peggy gou record... but now I see its going for over  on discogs
17:32:03  <supermop_Home> maybe I can back out of it
17:32:32  *** Progman has joined #openttd
17:36:00  <supermop_Home> would I draw new diag crossings in unspooled?
17:37:36  *** gelignite has joined #openttd
17:37:48  *** Yoshi_DE has quit IRC
17:40:51  <Eddi|zuHause> > make
17:40:53  <Eddi|zuHause> Makefile:55: *** Ungültige Syntax in der Bedingung.  Schluss.
17:40:54  <Eddi|zuHause> what?
17:41:10  <Eddi|zuHause> (grfcodec, clean checkout)
17:44:02  <LordAro> Eddi|zuHause: works for me. something weird in your path somewhere?
17:44:29  *** TrueBrain has quit IRC
17:44:32  *** TrueBrain has joined #openttd
17:44:34  <LordAro> oh, line55 is that line
17:44:58  <LordAro> wouldn't have thought there could be anything invalid in there...
17:46:38  <DorpsGek> [OpenTTD/grfcodec] Eddi-z opened issue #11: Compilation fails on clean checkout
17:49:02  <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened pull request #8578: Fix: Stopped ships shouldn't block depots
17:49:40  <Eddi|zuHause> the cmake PR seems to build
17:49:49  <Eddi|zuHause> with a huge pile of warnings
17:52:17  <andythenorth> did we merge anything exciting?
17:52:25  * andythenorth has been inattentive
17:52:50  <DorpsGek> [OpenTTD/grfcodec] LordAro commented on issue #11: Compilation fails on clean checkout
18:01:35  <DorpsGek> [OpenTTD/grfcodec] Eddi-z commented on issue #11: Compilation fails on clean checkout
18:11:01  <DorpsGek> [OpenTTD/OpenTTD] orudge commented on pull request #8577: Feature: Build macOS build as a universal binary
18:21:47  <DorpsGek> [OpenTTD/grfcodec] Eddi-z commented on pull request #10: Switch to CMake
18:24:02  <Samu> who's free?
18:24:49  <andythenorth> odd question
18:27:19  *** Progman has quit IRC
18:29:11  <Samu> I have a somewhat time consuming request: #7918
18:30:19  <Samu> glx, are you free?
18:38:27  <Eddi|zuHause> getting closer?
18:38:45  <Eddi|zuHause> i think the png file is for windows palette :p
18:38:55  <Eddi|zuHause> anyone got a conversion script?
18:43:40  <orudge> TrueBrain: macosx-universal.dmg is already "taken" in the web site description by various ancient versions. Are there actually pages on the site that use them? Maybe the new build could just be renamed "macosx.dmg" (or morr accurately these days "macos"?)
18:44:13  *** m1cr0man has quit IRC
18:44:26  <LordAro> macos-universal.dmg ?
18:44:37  *** m1cr0man has joined #openttd
18:49:46  <orudge> That's possible too of course
18:52:14  <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8578: Fix: Stopped ships shouldn't block depots
18:57:01  *** Hobbyboy has quit IRC
18:57:07  *** Hobbyboy has joined #openttd
18:58:12  <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8578: Fix: Stopped ships shouldn't block depots
18:59:46  <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8578: Fix: Stopped ships shouldn't block depots
19:05:12  <andythenorth> well
19:05:22  <andythenorth> I did a livery spec, did I mention it yet? :P
19:10:06  <frosch123> did people like it?
19:10:41  <frosch123> was there a result from the discussion? or did you press the reset button?
19:11:25  <_2TallTyler> I liked it, but don't know enough about what happens behind the veil of NML to fully understand
19:17:06  *** Compu has joined #openttd
19:24:19  <andythenorth> frosch123 too early to tell :)
19:24:25  <andythenorth> was very...crickets
19:24:46  <Eddi|zuHause> hm... weird... i'm fairly sure i converted the palette of the png file, but it still displays pink ingame
19:32:16  <andythenorth> has the grf recompiled correctly?
19:33:50  <Eddi|zuHause> i tried with -B, and the timestamp seems right
19:35:25  *** Heiki has quit IRC
19:35:29  *** Heiki has joined #openttd
19:36:50  <andythenorth> where is the png?
19:40:43  <DorpsGek> [OpenTTD/OpenTTD] Eddi-z updated pull request #8556: Feature: Allow diagonal tracks on level crossings
19:41:17  <Eddi|zuHause> there
19:42:13  <andythenorth> ?
19:42:35  <michi_cc> orudge: For your macOS Universal builds, how are the libs handled? Are the libs static or is vcpkg building universal libs? If not, where are those stitched together?
19:42:56  <Eddi|zuHause> andythenorth: yes
19:43:22  <michi_cc> TrueBrain: I think the pushed OpenGL patch version is missing a performance counter somewhere. I've fixed that locally already, but I was tinkering with some other stuff first.
19:43:44  *** Markk_ has quit IRC
19:43:52  *** Markk_ has joined #openttd
19:44:08  <DorpsGek> [OpenTTD/OpenTTD] Eddi-z commented on pull request #8556: Feature: Allow diagonal tracks on level crossings
19:44:14  <andythenorth> Eddi|zuHause the colour table in the png misses one entry
19:45:10  <andythenorth> hmm no
19:45:17  <Eddi|zuHause> hm... i picked the palette from another .png from media/baseset, and clicked the button for "don't remove unused entries"
19:45:18  <andythenorth> my DOS palette in photoshop is broken
19:45:24  <andythenorth> that's going to be 'interesting'
19:47:04  <DorpsGek> [OpenTTD/OpenTTD] Eddi-z commented on pull request #8556: Feature: Allow diagonal tracks on level crossings
19:49:39  <michi_cc> TrueBrain: Catching up on backlog: What do you think most of the recent "cleanup OSX" PRs come from? And that one is soo much worse so I will have a few more PRs to go on that :)
19:53:39  <andythenorth> Eddi|zuHause afaict the palette in that image is correct
19:54:05  <Eddi|zuHause> then maybe something is wrong in the drawing code?
19:54:11  <Eddi|zuHause> it's all ancient
19:58:33  <orudge> michi_cc: vcpkg produces static libs, which makes life easier
19:58:46  <Eddi|zuHause> drawing code starts here:
20:12:00  <michi_cc> orudge: Indeed. Interestingly, vcpkg on windows is defaulting to shared libs, so whoever did the macos stuff for vcpkg had different experiences :D
20:13:08  <michi_cc> Oh, and if your review that macos font patch, the universal binary is reduced by 2xfreetype :)
20:18:45  <TrueBrain> orudge: macosx might be taken, but indeed, we name the binaries macos already, so there should be no conflict :) Do you also make a PR for website? :D
20:20:35  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8577: Feature: Build macOS build as a universal binary
20:22:15  <orudge> TrueBrain: will do!
20:22:41  <TrueBrain> macos-universal is fine by me btw; I like the postfixes, as that makes it easier for the us in 5 years :P
20:22:55  <orudge> michi_cc: I gave it a test earlier and it seems to work for me. Will give it a proper review when I can
20:23:14  <TrueBrain> michi_cc: why do you disable threading for draw with OpenGL?
20:24:22  <michi_cc> TrueBrain: You can't use a single OpenGL context on multiple threads. You'd have to create extra context, and the user space time saving is just not worth it.
20:24:36  <TrueBrain> fair :)
20:24:57  <TrueBrain> hmm .. I now have most things working, it just keeps showing a black screen :P Pretty sure I am forgetting something here :D
20:25:09  <michi_cc> Did you do the equivalent to SwapBuffers?
20:25:30  <TrueBrain> OpenGLBackend::Get()->Paint();
20:25:30  <TrueBrain> 	SDL_GL_SwapWindow(_sdl_window);
20:25:59  <michi_cc> Which blitter are you trying?
20:26:21  <TrueBrain> any
20:26:26  <TrueBrain> they all work, as in, they no longer crash
20:26:33  <TrueBrain> took me a while to figure out how you did 40bpp :D
20:26:44  <TrueBrain> animator buffer was tricky :D
20:26:57  <TrueBrain> (more than one place that can allocate it, depending on detected settings)
20:27:03  <DorpsGek> [OpenTTD/OpenTTD] orudge merged pull request #8577: Feature: Build macOS build as a universal binary
20:27:08  <michi_cc> Is it possible that SDL needs some kind of dirty marking?
20:29:26  <TrueBrain> possibly, so I left it enabled for now
20:31:21  <milek7_> what enabled? are you sure that old code is not interrfering with gl?
20:32:49  <michi_cc> Well, if you push it somewhere, I can have a look for anything obvious.
20:33:22  <TrueBrain> btw, michi_cc , I really like what you did with the video driver
20:33:27  <TrueBrain> a lot of things make a lot more sense
20:33:33  <TrueBrain> like putting simple stuff as locks in a function
20:33:40  <TrueBrain> that alone makes it less: WTF?! :D
20:34:10  <michi_cc> I'm ever so slowly untangling the macos subdriver mess as well.
20:36:23  <TrueBrain> honestly, my current approach has just been: hack the stuff in there; I should just start separating the SDL driver too in a frontend and backend driver, so I know it is less likely I missed something :D
20:36:32  <TrueBrain> took me like 30 minutes to find out I wasn't calling Resize() for example :P
20:36:45  <TrueBrain> (which did more than just resizing .. it has to be called at least once :D)
20:37:37  <Eddi|zuHause> i can has preview label on #8556?
20:39:09  <michi_cc> cheese applied
20:43:08  *** Smedles_ has joined #openttd
20:44:18  <TrueBrain> weird, the win32-opengl can open maximized, but with a buffer allocated I indicated with "-r"
20:44:27  <TrueBrain> well, a bug to trace for another day .. writes it down in his mind
20:46:18  *** Smedles has quit IRC
20:46:46  *** frosch123 has quit IRC
20:47:22  <Samu> I have 8 PR's open
20:47:33  <Samu> must reduce to 5
20:47:35  <Samu> lol
20:51:45  <Samu> dang, there's nobody else with more open PR's than me
20:51:56  <Samu> michicc is at 6
20:54:54  <Eddi|zuHause> so, it looks right in the preview
20:55:00  <Eddi|zuHause> no clue what's happening locally
20:56:24  <TrueBrain> michi_cc: funny, because you open the GL context in Windows as 3.2, you get that assigned .. if you don't, you get a higher version :P Not sure if that matters :D
20:56:36  <TrueBrain> later on you detect what parts we need I noticed
20:57:43  <michi_cc> Yes and no. My baseline is OpenGL 3.2 because that is mostly what you get on macOS. There are a few selective things I use if available, like the persistent buffer mapping.
20:58:04  <TrueBrain> you also have checks for 2.0 etc?
20:58:08  <TrueBrain> or can those be cleaned up?
20:58:26  <michi_cc> We don't technically need all of 3.2.
20:59:00  <TrueBrain> it felt so double: both request a minimum version, but fallback to any if none found, and check for the features we need :)
20:59:12  <TrueBrain> but okay, that is not important now; just something that caught my eye :)
20:59:32  <michi_cc> I think I said that somewhere already, but some of that is written the way it is for me to learn more about OpenGL. Canonical way would be to just stick GLEW in.
21:00:08  <TrueBrain> honestly, I am more than fine with this. Just trying to understand what I see, learning from your learnings :D
21:00:14  <michi_cc> When I started with it, we were lower on the OSX front and some version/gpus there are only 2.1 + extensions.
21:00:33  <Samu> Eddi|zuHause, the graphics don't seem to match opengfx
21:00:47  <TrueBrain> I wouldn't mind if we just get what-ever OpenGL is available and check for features we want/need :)
21:00:52  <michi_cc> Also, old Intel GPUs driver are bad at new OpenGL versions. Current drivers are a lot better in that regards.
21:01:32  <TrueBrain> the persistent buffers feature toggle makes for a bit weird code, but .. after you understand it, it is fine :)
21:01:43  <michi_cc> Getting a 3.2 _core_ context by default is also mostly about me getting notified if I try to use legacy stuff from the compatibility profile.
21:01:54  <TrueBrain> ghehe, makes sense :)
21:02:29  <michi_cc> If you don't specify anything, you'll often get a compatiblity context. And I wasn't in the mood to test all the versions down from 4.6 to whatever.
21:04:20  <LordAro> Samu: ..did you expect them to?
21:04:41  <Samu> yesn't
21:04:43  <milek7_> this fallback doesn't make much sense really
21:05:12  <milek7_> if ogl 3 core context is not provided it is better to just stop and don't use opengl at all
21:07:42  <milek7_> better to use software render than to get gl2 context and crash
21:10:01  <milek7_> ah it does feature checks later
21:10:07  <milek7_> then maybe it's fine
21:13:19  <TrueBrain> michi_cc: meh; traced it up to: dst_ptr is filled, dirty_rect is set correctly, I can draw manually on the GL context ... some glue is missing :D Tomorrow or so I will just rewrite the SDL driver in the same format as the Win32 driver, that should rule out any stupidity on my side :D
21:13:33  <TrueBrain> just something is not sending the buffer to the screen, it seems :P
21:13:35  <michi_cc> milek7_: I selected what to use (including appropriate feature tests), along as that is the most limiting target we have.
21:13:44  *** Wormnest has quit IRC
21:15:00  <michi_cc> TrueBrain: Hmm, what else could be missing... Doing calls to OpenGLBackend::Get()->UpdatePalette?
21:15:43  <TrueBrain> I copied Paint(), so I assume that works as expected, but did not validate :D
21:16:41  <TrueBrain> 0 times count_dirty is > 0 :D
21:16:41  <michi_cc> It should, unless SDL is doing something stupid.
21:18:28  <TrueBrain> hahahahaha
21:18:33  <michi_cc> I suspect SDL might set/change some OpenGL state during init I'm ignoring right now.
21:18:35  <TrueBrain> yeah ... having no palette doesnt' work :D
21:18:37  <TrueBrain> so I got an image now :P
21:18:55  <TrueBrain> no ... I was just being an silly lazy bastard :P
21:18:59  <michi_cc> Party!
21:19:29  <TrueBrain> all kind of weird glitches, but that is completely because of all the hacks I added :D
21:19:57  *** crem has quit IRC
21:20:02  <milek7_> michi_cc: personally I wouldn't bother with <3.2
21:20:06  <TrueBrain> but yeah, I didn't "return" in CheckPaletteAnim(), causing count_dirty to always be zero when I was reading it :)
21:20:09  *** crem has joined #openttd
21:20:15  <TrueBrain> you solved that correctly in win32 .. when I copy this, it will be fine
21:20:36  <milek7_> because this is rather hard to test
21:21:23  <TrueBrain> michi_cc: very cool; and it is clearly faster too :)
21:21:31  <TrueBrain> not as fast as native win32, but ... yeah ..
21:21:35  <TrueBrain> SDL on Windows is slow anyway
21:22:50  <TrueBrain> I hope to clean this up this weekend :D
21:23:10  <TrueBrain> I like that you separated the frontend from the backend (in the videodrivers) with this :)
21:23:14  <TrueBrain> feels very clean
21:32:48  *** WormnestAndroid has quit IRC
21:33:03  *** WormnestAndroid has joined #openttd
21:40:30  *** gelignite has quit IRC
21:53:58  *** Samu has quit IRC
22:04:49  <DorpsGek> [OpenTTD/OpenTTD] LC-Zorg commented on pull request #8578: Fix: Stopped ships shouldn't block depots
22:09:18  *** _2TallTyler has quit IRC
22:10:47  <andythenorth> comment caused a rage quit?
22:13:43  <supermop_Home> don't ships already try to space out a bit?
22:13:59  <supermop_Home> i thought since a year or so ago?
22:14:04  <Eddi|zuHause> i think that was a patch?
22:15:54  <supermop_Home> hmm i guess so
22:19:30  <supermop_Home> obviously i do not use ships enough
22:23:05  <Eddi|zuHause> i feel like this old drawing code is overly convoluted
22:24:23  *** tokai|noir has joined #openttd
22:24:23  *** ChanServ sets mode: +v tokai|noir
22:31:16  *** tokai has quit IRC
22:32:52  <DorpsGek> [OpenTTD/OpenTTD] LC-Zorg commented on pull request #8574: Feature: Ships yield when passing through one another
22:33:09  *** andythenorth has quit IRC
22:47:25  *** tonyfinn[m] has quit IRC
22:47:31  *** tonyfinn[m] has joined #openttd
22:53:19  *** nielsm has quit IRC
22:53:53  *** Progman has joined #openttd
22:56:40  *** andythenorth has joined #openttd
23:06:52  *** Wolf01 has quit IRC
23:16:15  <TrueBrain> michi_cc: draw the mouse with OpenGL or let SDL handle it, any preference?
23:16:51  <michi_cc> OpenGL will be a tiny bit faster because it does not need to "undraw".
23:17:16  <TrueBrain> (I now had 2 cursors, one dragging behind the other .. SDL is clearly faster with mouse movement)
23:18:14  *** twpol[m] has joined #openttd
23:18:28  <michi_cc> Are you/Is SDL enabling double buffering?
23:18:48  <TrueBrain> how to find out? I did, but I removed that statement as I saw no difference
23:18:54  <TrueBrain> I assumed it would have a double-buffer given I have to swap :)
23:19:34  <michi_cc> No, swap has not really anything to do with double buffering nowadays in the time of composition window managers.
23:19:50  <TrueBrain> shows when I last did anything with GL :D
23:20:12  <michi_cc> For OTTD you should not double buffer, but the swap is still necessary to tell the driver/OS what to do when.
23:20:52  <michi_cc> Well, if you're doing a 3D game that should have as high FPS as it can, double buffering is still necessary
23:21:03  <TrueBrain> hmm, mouse movement doesn't work, something is pushing back :D So right-click doesn't work, which is funny :)
23:21:38  <TrueBrain> video-output, ~0.03ms when news is up, 0.00ms otherwise
23:21:43  <TrueBrain> this is shitty fast :P
23:22:13  <michi_cc> Video driver Paint needs a "PerformanceMeasurer framerate(PFE_VIDEO);" at front, I think the PR is missing it for win32.
23:22:21  <TrueBrain> I like that full animation doesn't matter, on/off
23:23:31  <TrueBrain> something is vsyncing, and something is holding back FF because of it
23:23:41  <TrueBrain> at least, it is waaayyyyyyy to close to 140Hz :)
23:23:56  <TrueBrain> 144Hz, sorry
23:24:11  <TrueBrain> Monitor: 144Hz, frame rate: 144 frames/s on FF
23:24:11  <TrueBrain> :D
23:24:54  <michi_cc> You might get more speedby uncommenting NO_GL_BUFFER_SYNC at the top of opengl.cpp. You get more glitches, too, though :)
23:25:10  <TrueBrain> I mostly wonder why the Win32 doesn't vsync like this
23:25:37  <michi_cc> Because I'm explicitly disabling vsync in the context.
23:26:09  <TrueBrain> via WGL, I see
23:26:21  <michi_cc> Though I am assuming the screen compositor simply ignores updates faster than the refresh rate.
23:28:23  <TrueBrain> hmm .. if I do a palette update, do I also have to call Paint() in OpenGL?
23:29:16  <TrueBrain> in Win32 you do this in the Paint() function, but if there is no part marked dirty, WM_PAINT is not emitted, is there? So it would not palette animate?
23:35:20  <TrueBrain> owh, and michi_cc , do we also want to port this to SDL1? Seems trivial, but .. just more work :P
23:35:38  <michi_cc> MainLoop calls VideoDriver_Win32Base::CheckPaletteAnim every frame, which InvalidateRect the window if count_dirty > 0.
23:36:03  <TrueBrain> michi_cc: ah! I see :)
23:36:15  <TrueBrain> some of these things are hiding in different places :D
23:37:48  <michi_cc> If you want to follow it all, reading the PR commit by commit will show you exactly how I went from the start over each step (pure 8bpp, also 32bpp, optimize this, optimize that, cursor stuff and etc) up to the 40bpp blitter.
23:38:27  <TrueBrain> okay, mouse on Windows SDL2 is just broken, that is not my fault :D
23:41:14  <TrueBrain> michi_cc: any chance you can rebase this PR? :D
23:41:39  <TrueBrain> (not that it fixes my mouse issue, but before I rework the SDL driver .. the recent changes are conflicts :P)
23:42:31  <michi_cc> Probably :) I'll do that tomorrow.
23:42:38  <TrueBrain> cool :)
23:42:42  <michi_cc> Or more like today :P
23:42:52  <TrueBrain> haha, yes, it is late .. holy crap
23:43:07  <TrueBrain> right, seems SDL_WarpMouseInWindow is re-emitting an event on Windows
23:43:16  <TrueBrain> so it is compensating the change constantly :D
23:45:49  <TrueBrain> but, nobody should be using SDL for Windows, so there isn't a real problem :P
23:47:57  *** twpol[m] is now known as JamesRoss[m]
23:50:04  <TrueBrain> I can disable vsync on SDL too
23:52:10  <TrueBrain> Windows does something else than SDL with, I tihnk, SwapBuffers, as in: on SDL the PerformanceMeasurer reads 0ms, where Windows does ~0.14ms (on FF), but SDL reaches lower frames/s
23:52:13  <TrueBrain> bit weird :P
23:52:38  <TrueBrain> euh, no, it is just unfair way of measuring ... on SDL it also counts frames there was nothing to draw :P
23:53:10  <TrueBrain> there we go ... cheating SDL driver ..
23:54:26  <TrueBrain> michi_cc: shall I also move all those global variables inside the instance when reworking this?
23:55:07  <michi_cc> I'm not going to stop you :)
23:55:25  <TrueBrain> is that wanted, the question is :P
23:56:04  <TrueBrain> hmm .. do we just remove SDL1, or are we going to support it ...
23:56:06  <TrueBrain> tricky question :D
23:58:16  <TrueBrain> Eddi|zuHause: <- in case you want to test, I am not stopping you :D It "should work", but it is still really rough
23:58:32  <TrueBrain> use -vsdl -b40bpp-anim for the most enjoyment

Powered by YARRSTE version: svn-trunk