IRC logs for #openttd on OFTC at 2020-09-25
⏴ go to previous day
00:00:07 <frosch123> "soon" sounds like "no hurry" :p
00:02:11 <TrueBrain> Saved enough money :)
00:02:34 <TrueBrain> How else am I going to finish this massive AWS migration? :D
00:03:51 <frosch123> when i quit my last job, i avoided telling people that i did not yet apply for a new one. not everyone can handle that kind of insecurity
00:04:32 <frosch123> but waiting for the right job was definitely worth it :)
00:04:36 <TrueBrain> Or rather, I can confirm people cant always handle that :p
00:05:00 <frosch123> you learned it the hard way? :p
00:05:17 <frosch123> did they offer you food?
00:05:38 <andythenorth> I haven't had a job for years
00:05:45 <andythenorth> jobs are so weird
00:08:06 <andythenorth> I had a few years ago
00:09:04 <frosch123> you should mow the lawn more often
00:10:22 <TrueBrain> I really should .. :D
02:08:48 *** tokai|noir has joined #openttd
02:08:48 *** ChanServ sets mode: +v tokai|noir
02:17:32 *** ChanServ sets mode: +v tokai
04:38:04 <azubieta> Hello folks! I took the liberty of including my OpenTTD AppImage on appimagehub.org, I can transfer it to you if you want.
05:02:56 *** tokai|noir has joined #openttd
05:02:56 *** ChanServ sets mode: +v tokai|noir
05:37:50 *** snail_UES_ has joined #openttd
05:39:54 <DorpsGek_III> [OpenTTD/OpenTTD] no1wafer opened issue #8316: Sorting of produced cargo for industries with multiple products are inaccurate https://git.io/JUaUm
07:45:37 *** andythenorth has joined #openttd
07:48:55 *** snail_UES_ has joined #openttd
08:02:42 *** sla_ro|master has joined #openttd
08:06:27 *** andythenorth has joined #openttd
09:05:29 *** WormnestAndroid has quit IRC
09:05:42 *** WormnestAndroid has joined #openttd
09:52:50 <DorpsGek_III> [OpenTTD/OpenTTD] Galigator opened pull request #8317: Change: Service at depot resets breakdown chance https://git.io/JUamd
09:53:18 <DorpsGek_III> [OpenTTD/OpenTTD] Galigator commented on pull request #8286: Change: Service depot also reset breakdown chance. https://git.io/JUamF
09:53:19 <DorpsGek_III> [OpenTTD/OpenTTD] Galigator closed pull request #8286: Change: Service depot also reset breakdown chance. https://git.io/JJgj9
10:02:06 *** andythenorth has joined #openttd
11:23:48 <FLHerne> azubieta: Consider opening a PR adding the appimage build files to the main source tree
11:24:16 <FLHerne> (and maybe even an action to build a nightly appimage on push and do...something with it)
11:24:33 <FLHerne> TrueBrain might have some idea what 'something' should be, I don't :p
11:33:51 <DorpsGek_III> [OpenTTD/OpenTTD] michicc commented on pull request #8286: Change: Service depot also reset breakdown chance. https://git.io/JUaG2
11:39:07 <michi_cc> What's the most-used app package on Linux anyway? Is it AppImage, or Flatpack or Snap or something else? It would definitely be an appropriate replacement for the old universal linux binaries.
11:41:10 <LordAro> i still think actual generic linux binaries would be better
11:41:33 <LordAro> just needs someone to spend time setting up a centos6 image
12:18:28 <FLHerne> michi_cc: AppImage is the only one that works out-of-the-box everywhere
12:19:46 <FLHerne> Snap is pretty much Ubuntu only (it's a NIH version of Flatpak in the same way Mir was of Wayland)
12:22:19 <FLHerne> Flatpak is fairly common, but unlike AppImage the client needs to have the tools/support installed separately
12:22:25 <FLHerne> Which isn't universal yet
12:24:21 <FLHerne> I'm already mucking about with CentOS 6 images having got stuck with making the next KDevelop release, if I learn anything I might look at building OTTD in it
12:30:55 *** Gustavo6046 has joined #openttd
12:58:06 <DorpsGek_III> [OpenTTD/OpenTTD] stormcone commented on issue #8316: Sorting of produced cargo for industries with multiple products are inaccurate https://git.io/JUaUm
13:28:49 *** andythenorth has joined #openttd
13:29:19 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro opened pull request #8318: Fix: Compile warnings when asserts are disabled https://git.io/JUal2
13:30:30 <DorpsGek_III> [OpenTTD/OpenTTD] no1wafer commented on issue #8316: Sorting of produced cargo for industries with multiple products are inaccurate https://git.io/JUaUm
13:41:32 <LordAro> cmake poll: "endif()" or "endif ()" ?
13:52:00 <FLHerne> LordAro: Whatever you use for `if`
13:52:06 <FLHerne> Here it's no space for either
13:52:59 <LordAro> hmm, most of OTTD ifs use `if (foo)`
13:53:32 <LordAro> all the "standard" statements have no spaces, wheras the "control" statements do
13:54:00 <FLHerne> In theory, I think `if()` in CMake is actually a command, it's not syntactically special
13:54:30 <FLHerne> But it's a bit weird, which is why I don't think it matters so long as it's consistent
13:57:42 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro opened pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa4Z
14:02:50 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 commented on pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa4r
14:03:32 <DorpsGek_III> [OpenTTD/OpenTTD] FLHerne commented on pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa41
14:04:39 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 commented on pull request #8315: Codechange: Set CMAKE_BUILD_TYPE to default to debug if not otherwise set https://git.io/JUa4y
14:49:10 *** Feuersalamander has quit IRC
14:57:07 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro updated pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa4Z
14:58:25 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro approved pull request #8317: Change: Service at depot resets breakdown chance https://git.io/JUa0O
14:58:50 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa0s
15:10:13 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 commented on pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa0H
15:11:32 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUaEC
15:21:14 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro updated pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa4Z
15:21:52 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUauv
15:28:43 *** WormnestAndroid has quit IRC
15:28:50 *** WormnestAndroid has joined #openttd
15:34:16 <DorpsGek_III> [OpenTTD/OpenTTD] FLHerne commented on issue #8310: when I logged in to 1M city mania the error window pops up... https://git.io/JU0EE
15:39:57 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 approved pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUazH
15:42:58 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on issue #8310: when I logged in to 1M city mania the error window pops up... https://git.io/JU0EE
15:43:16 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro merged pull request #8319: Codechange: Make codestyle for CMake files consistent for 'control' statements https://git.io/JUa4Z
15:50:54 *** Feuersalamander has joined #openttd
15:51:53 <DorpsGek_III> [OpenTTD/OpenTTD] stormcone updated pull request #8308: Feature: Add cargo filter support to vehicle list. https://git.io/JU4lE
15:56:49 *** grossing has joined #openttd
16:04:41 <andythenorth> I have like 1 hour of freedom, no work, no kids, no chores
16:04:45 <andythenorth> what shall I waste it on?
16:21:48 *** WormnestAndroid has quit IRC
16:22:31 *** WormnestAndroid has joined #openttd
16:26:33 <DorpsGek_III> [OpenTTD/OpenTTD] stormcone updated pull request #8308: Feature: Add cargo filter support to vehicle list. https://git.io/JU4lE
16:47:48 *** Feuersalamander has quit IRC
16:51:57 *** Feuersalamander has joined #openttd
17:09:55 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 approved pull request #8315: Codechange: Set CMAKE_BUILD_TYPE to default to debug if not otherwise set https://git.io/JUawW
17:24:54 *** Progman has joined #openttd
17:26:16 *** matt21347 has joined #openttd
17:57:10 <LordAro> TrueBrain: for some reason i'm looking at converting OTTD to GH actions. Don't really know what i'm doing. Currently it uses an "entrypoint" to execute cmake/make/test, but wouldn't that be better as individual run steps?
17:57:17 <LordAro> s/it/the docker images/
18:09:46 <dwfreed> each RUN statement in a Dockerfile is another layer, and is also cached
18:14:05 <n00b> i ahve a question about NML
18:14:10 <DorpsGek_III> [OpenTTD/CompileFarm] LordAro commented on issue #14: Commit checker should verify generated files are updated correctly https://git.io/JUaKK
18:14:10 <DorpsGek_III> [OpenTTD/CompileFarm] LordAro closed issue #14: Commit checker should verify generated files are updated correctly https://git.io/JUaKK
18:14:27 <n00b> does NML support 32 bit integer arithmetic ?
18:15:22 <LordAro> whereas i'm suggesting baking that into the GH Action steps where it's more obvious, rather than hidden away in a sh file in another repo
18:21:54 <FLHerne> n00b: You can do whatever you want in constant expressions, they're computed by NML before generating the grf
18:22:11 <n00b> i mean expressions in switch statements
18:23:30 <n00b> example: STORE_TEMP(LOAD_TEMP(8) * 100, 9)
18:23:58 *** snail_UES_ has joined #openttd
18:23:59 <FLHerne> Sorry, I think I was misreading the spec
18:24:17 <n00b> if storage 8 has like 1000
18:24:24 <n00b> it should or should not overflow
18:24:55 <n00b> NewGRFs are allowed to operate and store 32bit integers
18:24:56 <FLHerne> I was wrong before, that should be fine
18:25:25 <n00b> can you tell me what spec did you read tho
18:25:44 <n00b> cuz i was unable to find confirmation on my own
18:26:19 <n00b> oh you are reading openttd spec, i see
18:26:22 <FLHerne> NML specs kind of suck, but mostly it just maps onto the closest NFO values
18:26:49 <FLHerne> There's some magic in NML, but only for fairly unusual stuff
18:26:52 <n00b> they dont even mention what size integers they operate on
18:37:20 <andythenorth> does it matter how many bits in most cases?
18:40:59 <FLHerne> Well, if it was 16 bits, 100k would be too big
18:42:10 <n00b> but it seems it has 32bits
18:42:31 <n00b> so should be fine, anyways i will test it now
18:44:51 *** frosch123 has joined #openttd
18:53:55 *** HerzogDeXtEr has joined #openttd
19:07:02 <andythenorth> hmm, GS editing time :)
19:07:14 * andythenorth has to learn how that works
19:12:25 <andythenorth> hmm, can't figure it out :)
19:21:33 *** Lejving has joined #openttd
19:23:55 *** b_jonas has joined #openttd
19:34:18 *** WormnestAndroid has quit IRC
19:34:46 *** WormnestAndroid has joined #openttd
20:03:13 *** WormnestAndroid has quit IRC
20:04:01 *** WormnestAndroid has joined #openttd
20:05:20 <frosch123> it also didn't find grfcodec
20:27:47 <LordAro> frosch123: that's not unexpected
20:34:17 *** Wormnest has joined #openttd
21:07:11 <frosch123> everyone knows steam engines were invented by dwarfs
21:09:42 <TrueBrain> LordAro: correct, you can use the Docker to execute commands in the workdir on the host, by volume mounting it in the docker, en executing one command at the time
21:09:50 <TrueBrain> means you can remove the entrypoint.sh completely
21:12:44 <LordAro> TrueBrain: and you'd say that's a good thing?
21:12:56 <LordAro> TrueBrain: and any ideas about it not being able to find opengfx?
21:13:11 <TrueBrain> it would mean the containers would only need to contain the compiler, so that would make them easier
21:13:36 <TrueBrain> I am not sure you understood what "container" does in GitHub Actions
21:13:44 <TrueBrain> but I suspect you are compiling OpenTTD on a GitHub runner
21:13:52 <TrueBrain> which .. doesn't have OpenGFX installed :P
21:14:14 <LordAro> there is that possibility
21:14:27 <TrueBrain> I didn't even know you could start containers, I have to say :D
21:14:31 <TrueBrain> let me read up on how they work
21:14:39 <LordAro> but i'm not sure what else that "global" container definition would work...
21:14:49 <LordAro> s/what else/how else/
21:14:53 <TrueBrain> docs say for things like MySQL etc
21:15:01 <TrueBrain> so you can connect to them
21:15:52 <TrueBrain> no, sorry, I am confusing it with services
21:16:01 <TrueBrain> let me read some more .. docs are a bit unclear :D
21:17:13 <TrueBrain> funny, it starts the container with a "tail -f /dev/null" operation :D
21:17:19 <TrueBrain> (to keep it alive, I guess)
21:17:49 <TrueBrain> the container statement does what you expect
21:17:54 <TrueBrain> I did not know that existed
21:18:03 <TrueBrain> the volume mount you have, however, "deinstalls" OpenGFX :)
21:18:11 <TrueBrain> basically, you mount an empty folder from the host to the container
21:18:14 <TrueBrain> making that folder completely empty
21:19:10 <TrueBrain> okay, this is very cool use of GitHub Actions :)
21:19:19 <TrueBrain> it even allows you to skip the CompileFarm containers
21:19:24 <TrueBrain> and just use ubuntu:16.04 for example
21:19:31 <TrueBrain> and do the "apt install" actions
21:19:38 <TrueBrain> which would be even sweeter
21:20:09 <TrueBrain> that it does not detect the version is correct, and it even tells you why :)
21:20:17 <TrueBrain> The repository will be downloaded using the GitHub REST API
21:20:17 <TrueBrain> To create a local Git repository instead, add Git 2.18 or higher to the PATH
21:20:26 <TrueBrain> it downloads an archive, which does not contain a .git folder
21:20:40 <TrueBrain> but seriously, very cool find, I did not know this existed
21:28:39 <TrueBrain> LordAro: I think you can even use a matrix to run for all the different OSes in a single yml file :)
21:29:36 <TrueBrain> no clue how you are going to get Git 2.18 on 16.04 ..
21:29:49 <TrueBrain> do the checkout on the host, not on the container, I would guess
21:30:09 <LordAro> not particularly concerned by the OS for standard CI
21:31:03 <LordAro> re: removing compilefarm containers - would make it slower, due to having to redownload the dependencies everytime?
21:31:17 <LordAro> or perhaps not worth worrying about?
21:31:25 <TrueBrain> but also always the latest
21:31:28 <TrueBrain> so mostly useful for releases
21:31:36 <TrueBrain> for CI, honestly, it might be best to skip containers
21:31:41 <TrueBrain> and just run it on the runners
21:32:17 <TrueBrain> the apt install for our dependencies might take some time
21:32:25 <TrueBrain> but I doubt it is slower than fetching the container
21:33:03 <TrueBrain> you can pick which Ubuntu version you want even
21:33:24 <TrueBrain> GNU C++ 7, 8 and 9 are installed, and CLang 10, 9 and 8
21:33:34 <TrueBrain> so that gives you a wide variety
21:35:15 <LordAro> not exactly "gcc 6 & clang 3.9" :p
21:35:37 <TrueBrain> 2014 called, they want their compilers back
21:35:54 <TrueBrain> it is the CI; you would think newer versions catch more problems :D
21:36:02 <LordAro> but these days, as long as we're actually setting -std=c++11, shouldn't have any compatibility issues
21:36:20 <TrueBrain> otherwise the nightly would catch those, honestly
21:36:41 <TrueBrain> not even that whole list
21:36:59 <LordAro> why is git 2.18 necessary?
21:37:08 <TrueBrain> something GitHub says
21:37:14 <TrueBrain> I think it is the first version with v2 protocol support?
21:37:50 <LordAro> oh, and those are enforced, presumably?
21:37:54 <TrueBrain> * Transfer protocol v2 learned to support the partial clone.
21:38:00 <TrueBrain> that is what comes to mind
21:38:08 <TrueBrain> I do not know the reason they say this :D
21:39:45 <TrueBrain> so not helpful, that issue :P
21:40:46 <TrueBrain> I was right, v2 protocol
21:41:01 <TrueBrain> it makes the checkouts, especially on OpenTTD, so much quicker :)
21:41:26 <TrueBrain> so yeah, do the checkout on the host and volume mount them in the docker if you like :)
21:41:49 <TrueBrain> it still makes me giggle you volume mounted OpenGFX away :P
21:43:02 <LordAro> hang on though, i only added that after failing in the same way the first time
21:45:51 <TrueBrain> Now that is more interesting. Add things like 'ls' to give an idea what is really on the file system etc
21:56:35 <TrueBrain> if you have a run without the volume mount, feel free to link it; I can look with you
22:09:23 *** gelignite has joined #openttd
22:31:03 <LordAro> (it'll fail in a sec)
22:32:16 <LordAro> i'd suggest that there's something weird going on with the regression run, but it's basicaly the same as the existing CI
22:32:46 <LordAro> (although the existing CI does install ogfx into /usr/share & and then redownloads it into /usr/local/share
22:39:02 <DorpsGek_III> [OpenTTD/OpenTTD] michicc merged pull request #8317: Change: Service at depot resets breakdown chance https://git.io/JUamd
22:50:09 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro merged pull request #8315: Codechange: Set CMAKE_BUILD_TYPE to default to debug if not otherwise set https://git.io/JU2X4
23:07:22 <LordAro> TrueBrain: i think it's actually not looking in /usr/share/games, only /usr/local/share/games
23:07:40 <LordAro> not sure where it's getting any of the search paths from yet...
23:12:50 <LordAro> /usr/local being the default
23:15:06 *** Progman has joined #openttd
continue to next day ⏵