Config
Log for #openttd on 11th December 2020:
Times are UTC Toggle Colours
00:17:07  *** nielsm has quit IRC
00:27:17  <milek7> https://github.com/emscripten-core/emscripten/pull/12990/commits/90dd4d4c6b1cedec338ff5b375fffca93700f7bc#diff-5d76639a22445df4ff34f4df3f48af0bbc827a2818e340bf445068384bdc39cd
00:27:52  <milek7> much uglier now, but doesn't use configure; faster and works on windows
01:13:19  *** Delzur has quit IRC
01:56:53  *** glx_ has joined #openttd
01:56:53  *** glx is now known as Guest8087
01:56:53  *** glx_ is now known as glx
02:03:28  *** Guest8087 has quit IRC
02:06:39  *** Flygon has joined #openttd
03:46:04  *** D-HUND has joined #openttd
03:49:28  *** debdog has quit IRC
03:55:34  *** glx has quit IRC
03:55:41  *** Wormnest has quit IRC
04:29:59  *** supermop_Home has quit IRC
05:32:13  *** keoz has joined #openttd
05:32:57  *** Speeder_ has joined #openttd
05:35:52  *** dvim has quit IRC
05:36:15  *** dvim has joined #openttd
05:39:17  *** christoph[m] has quit IRC
05:39:22  *** reldred has quit IRC
05:39:32  *** Heiki[m] has quit IRC
05:39:33  *** reldred has joined #openttd
05:39:47  *** Speeder__ has quit IRC
05:46:08  *** christoph[m] has joined #openttd
05:49:04  *** Heiki[m] has joined #openttd
06:33:18  *** keoz has quit IRC
06:37:26  *** Smedles has quit IRC
06:38:26  *** Smedles has joined #openttd
07:18:45  *** andythenorth has joined #openttd
08:13:46  *** cHawk- has quit IRC
08:46:31  *** HerzogDeXtEr has joined #openttd
09:26:50  *** cHawk has joined #openttd
09:29:19  *** cHawk has quit IRC
09:57:48  <andythenorth> HELLO
09:57:53  <andythenorth> IS THE LOLZ?
10:03:46  <TrueBrain> for you, always my friend :)
10:04:22  <andythenorth> I am a very serious person
10:04:24  <andythenorth> no lolz
10:08:06  <orudge> TrueBrain: I've sent a PR to your personal OpenTTD repository to fix a macOS issue
10:08:09  <orudge> Your builds otherwise seem to work quite nicely
10:08:16  <TrueBrain> \o/ \o/
10:09:00  <orudge> Now to make it universal :D
10:09:02  <TrueBrain> and the naming of the bundles is fine etc?
10:09:07  <orudge> Seems to be
10:09:11  <orudge> x86_64 vs arm64
10:09:28  <TrueBrain> such a waste of bandwidth, those universals ... really, Mac did not think that through :P
10:09:32  <orudge> In the download page we should probably describe it as Intel vs Apple Silicon
10:09:37  <TrueBrain> lets waste 50% of the bandwidth .. well, it was 66% for a while :P
10:09:42  <TrueBrain> yeah, we will :)
10:10:07  <orudge> For Apple Silicon, we need the `openttd' binary from one of jobs, then we glue it into the other job before running cpack
10:10:37  <TrueBrain> hmmm
10:10:44  <TrueBrain> can it also be run after both openttd binaries are created?
10:10:54  <orudge> We'd need to rebuild the .dmg though
10:10:58  <orudge> but yes
10:11:05  <orudge> you could run it, build them separately, then re-run cpack
10:11:05  <TrueBrain> well, you dont need to build the dmg ofc
10:11:13  <orudge> but let's see
10:11:14  <TrueBrain> just run a normal 'make'
10:11:17  <TrueBrain> don't make a cpack
10:11:21  <TrueBrain> combine the job, run cpack
10:11:22  <orudge> well yes
10:11:25  <orudge> of course
10:11:25  <orudge> :)
10:11:29  <TrueBrain> that would be ideal from a workflow perspective
10:11:36  <orudge> OK
10:11:39  <orudge> Might play with that later
10:11:49  <TrueBrain> Upload job will fail for now btw, but I can solve that cleanly .. will apply that a bit later :)
10:32:23  <orudge> Hmm, why did that fail I wonder
10:32:25  <orudge> It works on my machine
10:39:04  *** ask6155 has joined #openttd
10:39:13  <ask6155> hello
10:40:10  <ask6155> When I see people videos about openttd, I see their main menu. it is a snowy background. How do I change my background?
10:40:32  <ask6155> For me it always shows up as tropical
10:40:49  <TrueBrain> every major release we change the main menu background; 1.9 is different from 1.10, etc
10:40:57  <TrueBrain> you cannot change that yourself :)
10:41:15  <TrueBrain> we have a competition before every new release
10:41:17  <TrueBrain> and the best background wins
10:41:23  <TrueBrain> and gets to be showcased for a whole year
10:41:41  <ask6155> oh... so it's not possible to even change the landscape?
10:42:33  <TrueBrain> the background on the main menu cannot be changed; you can play snowy landscapes if you like
10:42:39  <TrueBrain> just select the right climate before starting a new game
10:42:55  <ask6155> I thought that maybe if I press the sub tropic button It would change the background :)
10:44:25  <LordAro> that would be a neat tick
10:44:26  <LordAro> trick*
10:44:57  <LordAro> ask6155: it can be changed, but involves swapping out the opntitle.dat file (which is just a save game)
10:45:16  <TrueBrain> don't spoil it! :P
10:45:48  <ask6155> hmm maybe I'll put my save ;)
10:46:37  *** keoz has joined #openttd
10:51:49  *** ask6155 has quit IRC
10:52:03  *** ask6155 has joined #openttd
10:52:17  <ask6155> do you guys also help with newgrfs?
10:53:52  <TrueBrain> that is such a generic question, I doubt that will get an answer; we work on: ask the question, policy, and you will get an answer or not :D
10:55:03  <ask6155> I and my friends have been playing vannila tropical for quite some time. I was thinking of adding a few industries to 'spice it up'
10:55:41  <ask6155> I found ECS and it felt very big. I decided to go with improved town industries.
10:57:05  <ask6155>  It says that to transport new cargos, it require newgrf vehicles but it doesn't specify which ones
10:57:43  <ask6155> so which ones should I download?
10:58:41  <ask6155> https://github.com/2TallTyler/improved_town_industries/blob/db04743b00e421593c370181465af3ddae1ee84b/README.md
10:59:22  <TrueBrain> personally, I have no clue :) But possibly others in here will .. it is a workday, so expect slow responses ;)
11:00:08  <ask6155> I'll hang out
11:18:00  <andythenorth> there are loads
11:18:16  <andythenorth> ogfx+ trains and ogfx+ RVs should be the default recommendation
11:18:28  <andythenorth> basically default vehicles, extended and with full cargo support
11:18:40  <andythenorth> don't use anything made by andythenorth, you can't trust him
11:22:29  <orudge> TrueBrain: OK, that pull request I sent you should be good to go. :)
11:22:39  <TrueBrain> orudge: sweet, will check it out after lunch :)
11:22:44  <TrueBrain> first winning with Creeper World 4 :D
11:22:49  *** Gustavo6046 has quit IRC
11:22:53  *** Speeder_ has quit IRC
11:23:16  *** Speeder_ has joined #openttd
11:24:38  *** dvim has quit IRC
11:24:51  *** dvim has joined #openttd
11:28:24  *** grossing has quit IRC
11:28:26  *** grossing has joined #openttd
11:34:18  *** innocenat__ has joined #openttd
11:34:27  *** mindlesstux_ has joined #openttd
11:34:56  *** ask6155 has quit IRC
11:37:39  *** innocenat_ has quit IRC
11:37:39  *** innocenat__ is now known as innocenat_
11:39:19  *** mindlesstux has quit IRC
11:39:19  *** mindlesstux_ is now known as mindlesstux
12:40:15  *** WormnestAndroid has quit IRC
12:40:28  *** WormnestAndroid has joined #openttd
13:22:18  *** glx has joined #openttd
13:22:18  *** ChanServ sets mode: +v glx
13:39:50  <TrueBrain> orudge: thinking some more about it, if we want to do universals, possibly good to do all 3: Intel-only, M1-only, "Use this it will be fine"
13:40:03  <TrueBrain> so people don't have to download twice the size from what they need :D
13:40:34  <TrueBrain> it has the added benefit that people who know about their CPU, are not like: is this supported on my CPU?!
13:40:35  <TrueBrain> :)
13:41:00  <TrueBrain> I wonder if the user-agent javascript detection library we have can detect ARM vs x64
13:42:02  <glx> TrueBrain: https://github.com/glx22/OpenTTD/pull/5/files
13:42:36  <glx> probably not optimal
13:44:03  <TrueBrain> glx: nice :D
13:44:07  <TrueBrain> why not optimal?
13:44:50  <TrueBrain> please do abc on ci-build.yml before you PR this to upstream .. "doxygen" comes for sure before "zlib1g-dev" :D
13:45:08  <TrueBrain> and if you want to CI this, make it into its own step please
13:45:26  <glx> well the ci step was just a test :)
13:45:53  <TrueBrain> It is not a bad idea, honestly
13:46:14  <TrueBrain> but I like the PR; it saves a lot of custom code in the workflow :)
13:46:49  <orudge> [13:39:50] <TrueBrain> orudge: thinking some more about it, if we want to do universals, possibly good to do all 3: Intel-only, M1-only, "Use this it will be fine" <-- certainly a possibility too
13:46:52  <TrueBrain> orudge: seems the ua-parser.js, if updated, will detect Intel from M1, so we could have our auto-selection pick the right dmg :)
13:47:19  <TrueBrain> (we do the same for win32 vs win64, for example)
13:47:21  <orudge> Yes
13:47:27  <orudge> There should be an arm64 or similar in the user-agent
13:48:02  <orudge> Keeping them separate with user agent detection should be fine
13:48:12  <orudge> Eventually Apple will stop supporting Intel hardware, though that might be several years away
13:48:40  <TrueBrain> how popular the M1 reviews are, pretty sure our user-base on Intel drops quick
13:48:45  <TrueBrain> (those 10 people will quickly buy one)
13:48:46  <TrueBrain> :P
13:49:03  <TrueBrain> we will have to experiment at bit, honestly :)
13:50:02  <TrueBrain> https://github.com/TrueBrain/OpenTTD/actions/runs/415491950 <- failed because I was silly, but binaries are available in artifacts
13:50:08  *** DasPoseidon has joined #openttd
13:50:30  <TrueBrain> hmm ... no .pdb file .. hmm
13:58:39  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain requested changes for pull request #8329: Build AppImage https://git.io/JIo6G
14:06:05  *** nielsm has joined #openttd
14:06:10  <TrueBrain> orudge: owh, I think we should add a cache for the OSX dependencies btw
14:06:17  <TrueBrain> speeds everything up a bit :D
14:06:27  <TrueBrain> only no clue how :P
14:06:40  <TrueBrain> possibly glx knows, as I believe he did that for Windows once
14:07:13  <glx> well I use an existing action for that
14:07:18  <glx> check CI :)
14:07:28  <TrueBrain> can't we reuse that for Mac?
14:07:31  <glx> yes
14:07:34  <TrueBrain> or not now till orudge 's  stuff is merged?
14:07:38  <glx> it works everywhere
14:07:52  <LordAro> didn't work for apt, afaict
14:08:02  <TrueBrain> " Error: connect ETIMEDOUT 13.107.42.16:443" <- on "upload-artifact" .. ffs GitHub ...
14:08:18  <TrueBrain> LordAro: we were talking about vcpkg :P
14:08:23  <LordAro> ah
14:08:28  * LordAro disappears
14:08:29  <glx> vcpkg is available on windows, linux and mac, so the vcpkg action works everywhere
14:08:46  <TrueBrain> glx: but he has 2 patches on top of vcpkg :)
14:09:46  <glx> I don't know how it can be handled
14:10:19  <TrueBrain> okay, so we will have to wait for that, and switch vcpkg after
14:10:20  <TrueBrain> sure
14:11:51  <glx> maybe it just works if patch is applied before the vcpkg action
14:12:02  *** Flygon has quit IRC
14:12:20  <glx> IIRC the action just checks the commit hash
14:13:04  <glx> should be easy to test I guess
14:13:17  <TrueBrain> if you wouldn't mind? That would be much appreciated :)
14:13:23  <TrueBrain> I am currently testing if the action works on Mac at all
14:13:34  <TrueBrain> also testing Windows arm64 :P
14:14:05  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIoXq
14:14:48  <glx> windows arm64 will probably fail on cpu.cpp
14:15:00  <glx> because there's a #warning
14:16:11  <TrueBrain> oops .. I did a "find / | grep vcpkg" on the Mac node
14:16:14  <TrueBrain> I regret that action :D
14:17:28  <TrueBrain> will NSIS work for arm64? :D
14:17:48  <LordAro> bet it won't
14:18:39  <TrueBrain> the Mac actions is already 4 minutes production stdout .... do I want to know how many lines I printed? :p
14:18:52  <TrueBrain> I just wanted to know if vcpkg was installed and in which folder :P
14:19:15  <TrueBrain> it is installed
14:19:26  <glx> hmm I think the info is in orudge's yml
14:19:47  <TrueBrain> he installs his own vcpkg
14:19:50  <TrueBrain> in /usr/local/share/vcpkg
14:19:56  <TrueBrain> so I was wondering if that was still correct
14:20:05  <TrueBrain> but I regret my simple step to validate that :D
14:20:21  <TrueBrain> so far just 30k log lines, lot of: permission denied :D
14:20:34  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIo12
14:20:47  <TrueBrain> ah, ARM failed on host-tools, ofc
14:20:58  <TrueBrain> glx: how difficult is it to only build host tools with MSVC?
14:21:25  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIo1H
14:21:44  <glx> well from MSVC it's easy, but I think it should be doable with cmake action
14:21:49  <glx> will check the doc
14:21:56  <TrueBrain> <3
14:22:41  <TrueBrain> LordAro: found a solution to your i386 need ... AppImage! :)
14:23:05  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8329: Build AppImage https://git.io/JIoMv
14:23:13  <glx> looking at the code he doesn't install his own vcpkg, but "updates" the existing one
14:23:23  <LordAro> TrueBrain: oh no
14:23:25  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIoMf
14:23:42  <TrueBrain> LordAro: not good enough?
14:23:50  <glx> so if this step is done before vcpkg action it should just work for caching
14:24:41  <TrueBrain> it is one slow step .... lol
14:24:43  <TrueBrain> 2+ minutes
14:25:21  <orudge> [14:06:10] <TrueBrain> orudge: owh, I think we should add a cache for the OSX dependencies btw
14:25:24  <orudge> [14:06:18] <TrueBrain> speeds everything up a bit :D
14:25:27  <orudge> Yes
14:25:29  <orudge> I was going to do something like that based on the Windows one
14:25:37  <orudge> Was hoping to wait for my vcpkg PRs to be merged though
14:25:55  <orudge> also need to wait for another vcpkg PR to update cmake ideally, because the current version is borked on Silicon. Though maybe it works OK on an Intel host.
14:26:16  <orudge> ARM for MSVC I can do at a later stage as a separate PR once everything is live if you prefer, TrueBrain
14:26:19  <orudge> I have done it locally
14:26:22  <orudge> I think perhaps a patch or two was necessary
14:26:36  <orudge> The vcpkg dependencies would need to be updated there too
14:26:42  <glx> vcpkg action will manage vcpkg exe build I think
14:27:07  <orudge> [14:17:29] <TrueBrain> will NSIS work for arm64? :D <-- arm64 windows will run x86 apps. Apparently MS is working on an x86_64 emulator now too.
14:27:09  <TrueBrain> orudge: okay; I will build a normal upstream vcpkg Mac OS Intel variant, so people can see if that works
14:27:16  <TrueBrain> makes it a bit easier in the workflow if it does
14:27:50  <orudge> Yep
14:28:06  <TrueBrain> if your PRs hit upstream, we can add arm64 to it easily :)
14:28:07  <orudge> I'm happy if for the time being we just want to get everything moved over to GitHub Actions on Mac Intel, and ARM Intel can follow as soon as vcpkg etc is stable
14:28:16  <TrueBrain> exactly
14:28:20  <orudge> The PRs are approved, just waiting for merge
14:28:25  <TrueBrain> otherwise this will become such a dependency nightmare :D
14:28:44  <TrueBrain> so I leave it in as a matrix-step etc
14:28:44  <orudge> s/ARM Intel/ARM Mac/ :)
14:29:07  <TrueBrain> so ideally I also have MSVC already build host-tools in one folder, and let the other use that
14:29:12  <TrueBrain> so we are prepared there for arm64 too
14:30:00  <orudge> Nice
14:30:35  <TrueBrain> btw, glx , we need a -DOPTION_TOOLS_ONLY setting tbfh :P
14:30:39  <TrueBrain> I now use -DOPTION_DEDICATED=ON to avoid some dependencies :D
14:30:45  <orudge> Now I just need to either attempt to stick ARM64 Windows on my M1 Mac in a VM (don't know if any hackers have got that working yet) or obtain an ARM64 Windows laptop, but don't think I have room for any more hardware just now. Or money, to be fair. :P
14:32:04  <TrueBrain> ghehe
14:32:12  <TrueBrain> I didn't even know it existed, tbh :P
14:32:16  <TrueBrain> Apple hyped it a lot better :D
14:32:56  <andythenorth> Hype Machine
14:33:16  <TrueBrain> and lol glx, lukka/run-vcpkg is already at @6 :P
14:33:18  <TrueBrain> that goes quick
14:38:20  <DorpsGek> [OpenTTD/OpenTTD] azubieta updated pull request #8329: Build AppImage https://git.io/JTZha
14:39:02  <TrueBrain> that is quick
14:39:42  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8329: Build AppImage https://git.io/JIoSC
14:40:27  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIoSV
14:40:53  *** DasPoseidon has quit IRC
14:41:01  <TrueBrain> hmm .. if I skip the linux targets, anything depending on that, isn't executed either
14:41:04  <TrueBrain> that is not what I meant ..
14:42:01  *** DasPoseidon has joined #openttd
14:44:06  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8329: Build AppImage https://git.io/JIo9B
14:46:06  <orudge> TrueBrain: btw, when you're doing the vcpkg build for macOS, be sure to set the MACOSX_DEPLOYMENT_TARGET env variable to 10.9 before starting :)
14:46:37  <TrueBrain> I only replaced your vcpkg step with the actions one :)
14:46:50  <TrueBrain> the question is .. does it work: https://github.com/TrueBrain/OpenTTD/actions/runs/415580135
14:46:54  <TrueBrain> (I don't have a Mac, I cannot test :P)
14:49:36  <glx> TrueBrain: I think you can add buildWithCMakeArgs: '--target tools' to cmake action
14:49:52  <orudge> TrueBrain: I can test :)
14:50:43  <TrueBrain> glx: cool, trying it now :)
14:51:30  <glx> haha I was doing the same test for vcpkg and osx
14:52:48  <TrueBrain> meh, why is a "skipped" state  not considered successful ... this is annoying GHA!
14:55:09  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIoQA
14:55:17  <TrueBrain> glx: seems to work as advertised :)
14:55:38  <orudge> TrueBrain: Mac build seems to work fine :)
14:55:43  <TrueBrain> orudge: cool :)
14:55:54  <TrueBrain> so that makes arm really easy once vcpkg is ready for it
14:56:09  <DorpsGek> [OpenTTD/OpenTTD] azubieta updated pull request #8329: Build AppImage https://git.io/JTZha
14:56:44  <orudge> Now, a fun project might be to get the generation of the Microsoft Store appx working via GitHub Actions, to save me having to deal with it by hand
14:57:00  <TrueBrain> go for it :D
14:57:16  <orudge> but I will wait for the dust to settle on this first I think :)
14:57:47  <TrueBrain> seems the only thing I need to figure out, how to deal with Linux builds for nightlies ... the rest should now be ready :D
14:58:19  <TrueBrain> okay Mac build is 4 minutes faster this way, that is nice
14:58:21  <TrueBrain> caching rullessss
14:58:40  <TrueBrain> when did MSVC become the faster target ... silllyyyyy
14:58:58  <TrueBrain> glx: Error: ..\src\cpu.cpp(72): fatal error C1021: invalid preprocessor command 'warning'
14:58:59  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIo7K
14:59:01  <TrueBrain> as you already mentioned :)
14:59:07  <TrueBrain> but the host tools stuff works, sweet :)
14:59:21  <DorpsGek> [OpenTTD/OpenTTD] azubieta commented on pull request #8329: Build AppImage https://git.io/JIo7D
14:59:27  <orudge> TrueBrain: I do have patches for that
14:59:30  <TrueBrain> https://github.com/TrueBrain/OpenTTD/runs/1537869805?check_suite_focus=true for the curious :)
14:59:37  <orudge> to implement some of the cpu.cpp stuff on ARM
14:59:38  <TrueBrain> yeah, we can do that after this branch is merged :D
14:59:48  <TrueBrain> but it means the workflow is ready for it
14:59:50  <orudge> :)
15:00:23  <azubieta0> Hello TrueBrain, do you have a cross-compilation setup for OpenTTD somewhere ?
15:00:41  <TrueBrain> these days, "cross compiling" means using another docker container
15:00:56  <TrueBrain> so that strongly depends on if that is possible :P
15:01:30  *** azubieta0 is now known as azubieta
15:01:33  <TrueBrain> but it is an app-file per target, that is too bad :) Was hoping you could just list the archs :P
15:02:25  <azubieta> TrueBrain sadly some package names may chance from one arch to another
15:02:42  <azubieta> that's why we need different recipes
15:02:55  <azubieta> also the repositories change
15:03:17  <TrueBrain> we have found that on debian/ubuntu, it mostly only is changing the "amd64" into "i386", etc
15:03:28  <TrueBrain> a single variable, basically :D
15:03:39  <azubieta> i386 is simple, arm is a very different thing
15:04:01  <TrueBrain> well, with vcpkg it turns out, it is just as simple as changing "x86_64" into "arm64" :D
15:04:22  <azubieta> really !? I need to check out vcpkg
15:04:43  <TrueBrain> https://github.com/TrueBrain/OpenTTD/blob/release-workflow/.github/workflows/release.yml#L299
15:05:14  <TrueBrain> we unpack source, we run vcpkg with "arch", we build host-tools, we build target binary, done
15:05:32  <azubieta> that's awesome
15:06:01  <TrueBrain> but okay, vcpkg is only useful for static builds, I guess
15:06:07  <TrueBrain> so our linux builds are just switching containers
15:06:13  <TrueBrain> (and otherwise build natively)
15:06:18  <TrueBrain> doing that for arm is more difficult
15:06:23  <TrueBrain> (requires QEMU, basically)
15:06:23  <glx> static or embeded dynamic (like macos)
15:06:32  <TrueBrain> glx: tomato tomato honestly :)
15:06:47  <TrueBrain> isn't static the same as embedded dynamic? :D
15:06:54  <TrueBrain> (in the practical sense of the word)
15:07:05  <glx> but on linux cross compiling should not be too hard either as we use cmake
15:07:18  <LordAro> TrueBrain: i think linux arm usage is so small (and 95% of that is raspberry pis), that we can rely on users/package repo maintainers to compile it themselves :p
15:07:18  <TrueBrain> and Docker supports running arm containers via QEMU :P
15:07:36  <TrueBrain> LordAro: who knows; but I wasn't considering arm atm
15:07:38  <glx> it's still build native tools, then configure cmake to cross compile
15:07:55  <TrueBrain> I was more trying to see if we fix your i386 itch :P
15:08:00  <TrueBrain> +can
15:08:50  <TrueBrain> poeh, AppImage resulsts are kinda big .. hmm
15:09:49  <azubieta> TrueBrain let's see if we can make it smaller
15:10:20  <azubieta> please check the .bundle.yml file inside and see if some of the packages listed there could be dropped
15:14:02  <azubieta> notice that we have to embed everything that may not be present in the target system
15:14:07  <TrueBrain> haha, freepats is in there ofc .. ugh ..
15:14:17  <TrueBrain> fuck midi :P
15:15:11  <TrueBrain> anyway, that is something for another day :) Did not expect you to fix those things that quickly ;)
15:15:27  <azubieta> :D
15:15:36  <glx> ok updated my docs PR
15:16:29  <TrueBrain> funny, you cannot output something from a job if it might contain a secret :D
15:16:33  <TrueBrain> GHA is pretty clever :P
15:16:48  <TrueBrain> but too bad you don't have "settings" or something, secrets, but just not secrets
15:16:59  <TrueBrain> some things are not useful to put in the workflow itself, as forks take over those variables
15:17:02  <TrueBrain> hmm ... have to think about this
15:17:13  <TrueBrain> "<botocore.awsrequest.AWSRequest object at 0x7fa8fb8836d8>"
15:17:20  <TrueBrain> lovely aws-cli error
15:17:23  <TrueBrain> like ... thank you?
15:17:41  <glx> quite useful error :)
15:18:28  <TrueBrain> owh, and yes, pdbs are not exported in LordAro's workflow
15:18:31  <TrueBrain> no clue how to fix that :)
15:18:34  <TrueBrain> neither are -dbg.debs
15:18:44  <LordAro> i thought i fixed pdbs?
15:18:55  <TrueBrain> I dunno .. they are not in your artifacts
15:18:56  <LordAro> dbg.debs aren't even generated, afaik
15:19:07  <LordAro> dunno how to make them
15:19:26  <TrueBrain> owh, in your version they are
15:19:32  <TrueBrain> just with a very weird name
15:19:56  <LordAro> ah yes, i hadn't worked out how to give them a proper name yet :p
15:20:08  <TrueBrain> I can fix that :)
15:20:09  <LordAro> you did that bit first
15:20:59  *** Gustavo6046 has joined #openttd
15:21:00  <DorpsGek> [OpenTTD/OpenTTD] glx22 opened pull request #8369: Docs https://git.io/JIoNC
15:21:45  <glx> oups I forgot to set a proper name
15:22:16  <LordAro> is it worth building docs any more frequently than nightly?
15:22:37  <glx> if someone wants to check warnings
15:22:45  <LordAro> there's thousands of them :p
15:22:48  <TrueBrain> we should make warnings an error :P
15:22:59  <TrueBrain> maybe finally someone is going to fix them :D
15:23:29  <TrueBrain> setting a strategy without matrix is not really useful btw :)
15:23:34  <TrueBrain> I noticed more people did that
15:24:17  <TrueBrain> I leave it up to LordAro if he wants it in the CI; I am fine with having it in there, but I can see your argument it is not helping anything
15:24:30  <LordAro> i don't follow?
15:24:32  <orudge> So, how do I actually test rdtsc() in OpenTTD? I see a pathfinder performance timer thing, how do I see that? Is it via a debug channel?
15:24:42  <glx> TIC/TOC ?
15:24:58  <TrueBrain> LordAro: what part did you not follow?
15:25:06  <LordAro> TrueBrain: matrix/stratgy thing
15:25:20  <TrueBrain> LordAro: glx too, and I saw a PR earlier, sets "fail-fast: false" in strategy
15:25:22  <TrueBrain> without a matrix
15:25:35  <TrueBrain> not sure what the expected outcome of that command is, but that doesn't do anything as far as I know :)
15:25:35  <LordAro> ah right
15:25:43  <LordAro> yes, that's a bit pointless :)
15:25:55  <TrueBrain> somewhere that is being copy/pasted from :)
15:26:51  <orudge> glx: OK, so it's just for developers profiling then basically?
15:27:02  <glx> yes
15:28:38  <TrueBrain> LordAro: I wonder if copying the PDB should be part of CPack configuration
15:28:48  <LordAro> TrueBrain: probably, yeah
15:28:55  <LordAro> i don't know how cpack works
15:29:05  <TrueBrain> me neither! HIGH FIVE!
15:29:19  <LordAro> it might just be my internet connection, but i can't actually get the doxygen output from glx's PR anyway
15:29:49  <TrueBrain> it is just too darn much :P
15:29:59  <TrueBrain> I wonder if anyone has ever used the source documentation ....
15:30:07  <TrueBrain> I guess the AI and GS API docs do run without warning?
15:30:14  <TrueBrain> could be useful to enable error on warning there?
15:30:23  <glx> I saw warnings for api
15:30:39  <TrueBrain> that would be bad, I guess?
15:31:30  <glx>  /home/runner/work/OpenTTD/OpenTTD/src/script/api/script_event.hpp:87: warning: parameters of member AIEvent::AIEvent are not (all) documented
15:31:52  <TrueBrain> I would expect that shit like that should be fixed? (honest question :D)
15:32:41  <glx>  /home/runner/work/OpenTTD/OpenTTD/src/script/api/script_road.hpp:399: warning: argument 'start' from the argument list of AIRoad::BuildOneWayRoadFull has multiple @param documentation sections <-- this one seems fun
15:33:30  <glx> and some not documented return type
15:33:54  <TrueBrain> LordAro: openttd-20201211-release-workflow-g17d6fbf2c2-windows-x86.pdb.xz <- ALMOST correct :)
15:33:59  <TrueBrain> x86 -> win32 :D
15:34:13  <TrueBrain> but that should be an easy fix :)
15:34:23  <LordAro> :)
15:34:29  <TrueBrain> but I found the pdbs :D
15:34:46  <TrueBrain> now for the -dbg.deb .. I think CPack supports that
15:35:02  <TrueBrain> if anyone has the time, I would welcome a PR :)
15:35:08  <TrueBrain> for now, I am off; ciao :)
15:42:02  <orudge> OK, so attempting to access the performance counter on M1 crashes. Internet rumours suggest the same on rpi4, curiously. So we can just go without I guess.
15:51:19  *** iSoSyS has joined #openttd
16:01:20  *** Cursarion has quit IRC
16:30:33  *** Wormnest has joined #openttd
16:35:52  *** Progman has joined #openttd
16:41:39  *** arikover has joined #openttd
16:43:18  *** frosch123 has joined #openttd
17:11:46  *** arikover has quit IRC
17:15:25  <DorpsGek> [OpenTTD/team] frosch123 commented on issue #103: [ta_IN] Translator access request https://git.io/JIguF
17:39:24  <andythenorth> yo!
17:39:30  * andythenorth has finished working for the week
17:39:36  <andythenorth> shall we merge something?
17:40:17  <frosch123> i am somewhat in an nml mood, but there are no PRs to review, and no issues except writing docs
17:41:26  <andythenorth> I can't get excited about writing docs
17:41:31  <andythenorth> we could debug the dual-headed thing
17:41:37  <andythenorth> and find out how badly I wrote my code :P
17:41:45  <frosch123> 12
17:41:50  <andythenorth> 12/10
17:41:55  <frosch123> exactly :)
17:42:06  <andythenorth> I removed randomisation from the dual headed container train
17:42:09  <andythenorth> was easier
17:43:50  <andythenorth> stations are out of scope, right?
17:45:07  <frosch123> fixing bugs is easier, more obvious what the result is supposed to be like, so less discussions
17:45:24  *** jottyfan has joined #openttd
17:47:09  *** jottyfan has quit IRC
17:47:12  <FLHerne> Hm, I'd better make more PRs
17:48:27  <andythenorth> make install PW=16 2>&1 | grep -v "returns a constant" is good
17:48:35  <andythenorth> but ideally a suppression would be better
17:49:36  <andythenorth> so are we ever going to do partial compiles and linking? o_O
17:50:08  <FLHerne> Needs lots more refactoring before that's feasible, I think
17:50:33  <andythenorth> what about building in support for reversed vehicles?
17:50:36  <andythenorth> o_O
17:50:56  * andythenorth looks what the duplicate sprite count is in Horse :)
17:51:09  <andythenorth> oh 50% :D
18:01:36  <FLHerne> How does 'support for reversed vehicles' work?
18:02:20  <FLHerne> Vehicle flag, and it automagically creates mirrored sprites with different offsets?
18:03:22  <frosch123> it's about sprite offsets
18:03:33  <frosch123> but that is only sanely solvable by patching ottd
18:05:39  <andythenorth> I solved it insanely
18:05:45  <andythenorth> adding a reasonable amount of bloat
18:06:17  <andythenorth> iirc it's just a transform
18:06:18  *** sla_ro|master has joined #openttd
18:06:34  <andythenorth> sprites [0..3] are exchanged for sprites [4..7]
18:06:41  * andythenorth hasn't checked code though :D
18:07:12  <FLHerne> Oh, reversed like that
18:07:44  <frosch123> though you can patch nml to no longer report them as duplicates
18:08:04  <frosch123> yeah, that may be the best solution
18:09:11  <andythenorth> the dups I have come from a grep you gave me frosch123
18:09:13  <andythenorth> grep generated/graphics generated/iron-horse.nfo | sed 's/^[0-9]*//' | grep -v ' 1 1 0 0 normal' | sort | uniq -c | sort -n > dups.txt
18:11:42  * andythenorth wonders if eliminating 35k comments from the nfo would have any measurable performance impact
18:11:48  <andythenorth> all // Name: foo
18:14:31  <andythenorth> hmm yes the reversed variant just transpose the sprites https://github.com/andythenorth/iron-horse/blob/master/src/global_constants.py#L186
18:14:36  <andythenorth> or the offsets rather
18:18:50  *** tokai has joined #openttd
18:18:50  *** ChanServ sets mode: +v tokai
18:25:42  *** tokai|noir has quit IRC
18:44:23  <andythenorth> so do we have nml bugs?
18:44:27  * andythenorth doesn't find many
18:45:49  <frosch123> i fixed 4 last week
18:46:21  <frosch123> i wrote some stupid regression tests for it
18:46:40  <andythenorth> \o/
18:46:47  <frosch123> when i write regression tests, i pick random variables to switch on
18:46:52  <frosch123> so the tests are really stupid :p
18:47:38  <andythenorth> hmm do I really want to support...nvm
18:48:01  * andythenorth had a nice compile where different magic was orthogonal by vehicle type
18:48:11  <andythenorth> now I have started mixing the magic together in new vehicles
18:48:15  <andythenorth> oof unwise
18:49:09  <andythenorth> oh the magic is fine, I'm just using it wrong
18:49:22  <andythenorth> quite often I surprise myself that things work :P
19:11:25  *** DasPoseidon has quit IRC
19:29:11  *** Wolf01 has joined #openttd
19:30:19  *** tokai|noir has joined #openttd
19:30:19  *** ChanServ sets mode: +v tokai|noir
19:30:31  <Wolf01> Hmmm. I have nothing to play
19:37:13  *** tokai has quit IRC
19:37:48  *** DasPoseidon has joined #openttd
19:41:40  *** tokai|noir has quit IRC
19:43:56  *** tokai has joined #openttd
19:43:56  *** ChanServ sets mode: +v tokai
19:48:23  <andythenorth> Wolf01 2048!
19:48:37  * andythenorth uses that as the default worst option
19:48:54  <Wolf01> I was thinking about freecell or minesweeper
19:49:20  <Wolf01> Minesweeper VR maybe
19:49:45  <Wolf01> 4€ on steam
19:50:55  *** tokai|noir has joined #openttd
19:50:55  *** ChanServ sets mode: +v tokai|noir
19:51:33  <Wolf01> Shit... I need to move to play it, I don't have space
19:53:25  <andythenorth> freecell!
19:53:31  * andythenorth plays spider solitaire
19:53:33  <andythenorth> thanks for reminder :)
19:54:15  <Wolf01> Solitaire VR also on steam
19:55:22  <Wolf01> I would play spider, but that game is full of ads
19:56:01  *** DasPoseidon has quit IRC
19:56:36  <Wolf01> One thing I can't fix is to make No Man's Sky working, it starts etc, but I can't control it with anything in VR mode
19:56:58  <TrueBrain> orudge: did someone offer to pay our monthly bill? Do I understand your reply to him correct? (his initial email is not in your reply, so it is a bit guessing)
19:57:12  <Wolf01> I purchased the VR kit exactly to play NMS...
19:57:38  *** tokai has quit IRC
19:58:41  <andythenorth> hmm solitaire game uses 500% of CPU, drains battery in 4 mins
19:58:48  <andythenorth> coin mining much? o_O
19:59:43  *** tokai|noir has quit IRC
20:00:02  <frosch123> https://dpaste.org/sqTc <-we could use an industry test case, for both old and new produce syntax
20:00:27  <frosch123> but otherwise we mostly lack testing the error output
20:00:56  <frosch123> we only test nml examples that actually compile, no expected failures, so no testing of the error messages :)
20:03:03  *** DasPoseidon has joined #openttd
20:03:54  <frosch123> TrueBrain: he found a wealthy nigerian prince :)
20:07:16  *** DasPoseidon1 has joined #openttd
20:07:20  *** DasPoseidon has quit IRC
20:07:20  *** DasPoseidon1 is now known as DasPoseidon
20:48:47  *** gelignite has joined #openttd
20:50:16  *** supermop_Home has joined #openttd
20:50:20  <supermop_Home> hi
20:54:39  <Wolf01> o/
21:03:06  *** iSoSyS has quit IRC
21:07:28  *** tokai has joined #openttd
21:07:28  *** ChanServ sets mode: +v tokai
21:09:52  <Wolf01> I really ended up playing solitaire
21:10:39  <frosch123> in VR?
21:11:17  <Wolf01> No, the microsoft one, I'll wait for a discount for that :P
21:11:17  *** DasPoseidon has quit IRC
21:12:55  <andythenorth> I played spider
21:13:00  <andythenorth> forgot how annoying it is
21:13:11  <andythenorth> was that really 75 minutes to win 1 game?
21:19:22  *** Gustavo6046 has quit IRC
21:19:51  *** Gustavo6046 has joined #openttd
21:19:54  <Wolf01> I played 2 spider and 2 freecell, lost 1 spider, won all the others
21:19:59  <supermop_Home> andythenorth sounds about right
21:20:45  <supermop_Home> haven't played freecell in awhile
21:23:21  <supermop_Home> hmm I've already messed it up
21:30:49  <supermop_Home> there are ads in solitare now?
21:47:54  <orudge> TrueBrain: sort of, they sent a 0 donation :)
21:48:01  <TrueBrain> that is .. amazing honestly
21:48:26  <orudge> They offered to pay for a month or two
21:48:30  <orudge> Yes
21:48:35  <orudge> Very generous
21:48:43  <orudge> I will reply back to thank them
21:48:57  <LordAro> nice
21:49:13  <andythenorth> :o
21:54:25  <andythenorth> oof why did Wolf01 make me play solitaire :P
21:55:46  <frosch123> is it better than sudoku?
21:55:53  <andythenorth> never tried :)
21:56:39  <frosch123> really?
21:56:41  <frosch123> impressive
21:56:59  <andythenorth> well I helped child #1 do a sudoku once
21:57:05  <andythenorth> but not all of it
21:59:41  <frosch123> https://www.janko.at/Raetsel/Uebersicht.htm <- when you need more puzzles for your kids
21:59:54  <andythenorth> oof so many there
21:59:57  <frosch123> i doubt anyone can solve all of them in their lifetime
22:00:31  <frosch123> andythenorth: some puzzles are actually good. sudoku is quite boring though
22:01:02  <andythenorth> :)
22:03:53  <frosch123> never counted the puzzles on that page
22:04:21  <frosch123> but there are >200 types of puzzles, some of them with >100 instances
22:04:24  <andythenorth> is spider solitaire the worst?
22:04:30  * andythenorth can't tell if it's just luck or not
22:05:39  <frosch123> isn't spider the one where you can switch to only play with a single suit?
22:05:45  <andythenorth> yes I think so
22:05:47  <frosch123> pretty hard to loose those
22:06:51  <andythenorth> I am playing 2 suit
22:15:54  *** frosch123 has quit IRC
22:38:54  *** Wormnest has quit IRC
22:46:06  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8369: Add: [CMake] targets to generate documentation https://git.io/JIKHc
22:46:44  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8369: Add: [CMake] targets to generate documentation https://git.io/JIKHW
22:46:55  <TrueBrain> that moment you are sure you are only going to make a single comment ..... :P
22:47:17  *** WormnestAndroid has quit IRC
22:47:41  *** WormnestAndroid has joined #openttd
22:48:47  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain requested changes for pull request #8369: Add: [CMake] targets to generate documentation https://git.io/JIKHz
22:48:50  <TrueBrain> two more for the road in that case :D
22:56:47  <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8369: Add: [CMake] targets to generate documentation https://git.io/JIKQJ
22:57:09  <glx> 2 replies in one (using the review trick)
22:59:14  <TrueBrain> Smart :)
23:01:52  <glx> and I edited my replies to add some references
23:10:06  <DorpsGek> [OpenTTD/OpenTTD] glx22 updated pull request #8369: Add: [CMake] targets to generate documentation https://git.io/JIoNC
23:10:34  * andythenorth wonders
23:10:40  <andythenorth> if we'll ever ship some default grfs
23:10:52  <andythenorth> so players get trams, instead of confusion
23:15:10  *** Wormnest has joined #openttd
23:16:47  *** HerzogDeXtEr1 has joined #openttd
23:16:58  *** gelignite has quit IRC
23:19:32  *** andythenorth has quit IRC
23:22:15  *** HerzogDeXtEr has quit IRC
23:27:24  *** sla_ro|master has quit IRC
23:36:38  *** WormnestAndroid has quit IRC
23:38:38  *** keoz has quit IRC
23:40:51  *** Wolf01 has quit IRC
23:41:53  *** HerzogDeXtEr1 has quit IRC
23:43:26  *** tokai|noir has joined #openttd
23:43:26  *** ChanServ sets mode: +v tokai|noir

Powered by YARRSTE version: svn-trunk