IRC logs for #openttd on OFTC at 2020-12-13
⏴ go to previous day
00:03:59 *** rptr_ has quit IRC (Quit: Leaving)
00:08:41 <TrueBrain> Nice solution glx! I will test it tomorrow, but it looks lovely clean :)
00:09:44 *** sla_ro|master has quit IRC ()
00:22:18 *** Progman has quit IRC (Remote host closed the connection)
00:49:27 *** Delzur has quit IRC (Remote host closed the connection)
02:18:58 *** k-man has quit IRC (Ping timeout: 480 seconds)
02:19:03 *** WormnestAndroid has joined #openttd
02:51:56 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
02:52:39 *** WormnestAndroid has joined #openttd
02:55:42 *** tokai|noir has joined #openttd
02:55:42 *** ChanServ sets mode: +v tokai|noir
03:02:38 *** tokai has quit IRC (Ping timeout: 480 seconds)
03:47:23 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:59:50 *** Laedek has quit IRC (Read error: Connection reset by peer)
07:53:22 *** HerzogDeXtEr has joined #openttd
08:22:15 *** WormnestAndroid has quit IRC (Remote host closed the connection)
08:22:27 *** WormnestAndroid has joined #openttd
08:23:06 *** andythenorth has joined #openttd
08:23:23 * andythenorth stayed up way too late drawing trains :|
08:23:32 <andythenorth> I mean wtf, what kind of behaviour is that even?
08:34:12 *** Progman has joined #openttd
09:03:02 *** jottyfan has joined #openttd
09:52:00 <TrueBrain> andythenorth: you are a bit crazy yes, but aren't we all ? :)
09:52:25 <andythenorth> crazy talk TrueBrain
10:05:32 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain requested changes for pull request #8372: Add: [CMake] Option to only build tools https://git.io/JIQDS
10:09:21 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8372: Add: [CMake] Option to only build tools https://git.io/JIQHY
10:10:47 *** jottyfan has quit IRC (Quit: jottyfan)
10:53:27 <TrueBrain> so many emails from GitHub about cancelled workflows .. no clue what that is about :P
10:56:56 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
10:59:31 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
11:00:52 <TrueBrain> right, I can build master, pull requests, tags, both manual, automated, and on event
11:00:56 <TrueBrain> that all checks out and works as expected :D
11:07:07 <FLHerne> andythenorth: Three-part wagon where the middle bit is the cargo and goes between the angles of the outer ones?
11:07:14 <TrueBrain> our server has been deleted ... the operation was irreversible .. bye OVH server ... you were good to us ...
11:07:34 <FLHerne> Like CETS vehicles, but with the invisible spacers not being invisible
11:11:11 *** jottyfan has joined #openttd
11:11:59 <andythenorth> FLHerne 3/8 + 4/8 + 3/8 for a twin bolster, with the cargo in the middle
11:12:11 * andythenorth wonders if it's TMWFTLB
11:12:16 <andythenorth> or a good party trick
11:14:26 *** gelignite has joined #openttd
11:16:32 *** jottyfan has quit IRC (Quit: jottyfan)
12:45:54 <Timberwolf> andythenorth: delve into the horror of the curvature variables?
12:46:45 <Timberwolf> (You still end up with relative alignment between the vehicles drifting, not sure if it's possible to get a nice pivot without exhaustively providing sprites for every single x/y relative offset.
12:47:24 <andythenorth> I did wonder about curve var :P
12:50:04 <FLHerne> Needs a way to specify just offsets by varaction2?
12:54:34 <andythenorth> just branch to different groups
12:54:56 <andythenorth> offsets are in realsprites, so you're limited options there
12:56:42 <andythenorth> hmm But Ground Types?
13:11:29 <Eddi|zuHause> andythenorth: that image link gives me 403, but i imagine it's like a wood car?
13:12:38 <Eddi|zuHause> yes. pretty much what i expected
13:13:23 <Eddi|zuHause> i wanted to do something like that in CETS, but never got anywhere close
13:13:34 <andythenorth> I know I could do it
13:13:38 <andythenorth> I'm not sure I should :P
13:13:47 <andythenorth> there are enough party tricks already
13:13:49 <Eddi|zuHause> probably easiest is to treat it as one wagon when loaded, and two wagons when empty
13:14:11 <andythenorth> the easiest would just be to bend the cargo when going round corners I think
13:14:18 <andythenorth> I already bend some vehicles
13:14:24 <Eddi|zuHause> well, the easiest non-horrible thing :p
13:15:36 <Wolf01> Hmmm, would it be possible to tell in the newobjects window the reason why an object is not selectable for construction?
13:19:00 <Eddi|zuHause> sounds like a large project
13:21:00 *** frosch123 has joined #openttd
13:28:01 <andythenorth> discord suggested just being able to paint ground tiles
13:29:17 <Eddi|zuHause> sure. for empty tiles that might work. but what about rails, roads, houses, etc.?
13:31:43 <andythenorth> it would simplify a lot of station sets eh
13:43:33 <TrueBrain> so on a scale from 1 to 10 .. how mad you think people will be if I start reviewing my own PRs? :D
13:52:47 <TrueBrain> frosch123: https://github.com/OpenTTD/OpenTTD/pull/8197 <- what was your plan with this? I kinda like having CPack also bundle a source, as we are now doing it in the workflow. But your second point is kinda a problem I guess .. how do you see that? Is something like this PR worth considering, or is it a nice idea but not something we should merge?
14:03:53 <Timberwolf> Also, bridge heads are slightly steeper (by 1-2%) than 'native' slopes on hills. Thanks Simon.
14:04:30 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8189: Update: generate Mac bundle on install rather than packaging https://git.io/JIFHb
14:04:33 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed pull request #8189: Update: generate Mac bundle on install rather than packaging https://git.io/Jf12c
14:06:27 <frosch123> TrueBrain: not much. i tried it, was disappointed in the result, and hoped someone else could fix it :)
14:07:07 <TrueBrain> well, I was wondering the same :) As I rather have these things in CMake than in the GitHub Actions workflow :P
14:07:25 <TrueBrain> but I also don't really see a way to do it as smooth as we can do it in the workflow currently
14:07:50 <TrueBrain> we run "find . -iname .hg -or -iname .git -or -iname .svn | xargs rm -rf"
14:07:56 <TrueBrain> (from a fresh checkout)
14:08:24 <TrueBrain> guess the way to avoid #2, is to do a git check for a clean repo
14:08:33 <TrueBrain> and ignore the same list as in gitignore
14:09:11 <frosch123> yes, that works for a farm
14:09:15 <TrueBrain> owh,that is what glx did
14:09:29 <TrueBrain> he ignores everything from git ignore .. did not know git had a command for that :)
14:09:34 <frosch123> my concern was that i cannot tell people to run in in their checkouts
14:09:56 <frosch123> but maybe noone would do that anyway
14:10:08 <TrueBrain> well, with glx his changes, there should be no problem I think?
14:10:18 <TrueBrain> it only puts files in the tar that are committed
14:13:15 <TrueBrain> glx is our new peter, he has a patch for everything :P
14:13:56 <TrueBrain> very tempted to just approve it honestly :D
14:13:57 <frosch123> hmm, looks like a missed that point from glx' patch
14:14:04 <frosch123> so, yeah, looks like glx solved both issues
14:14:13 <TrueBrain> frosch123: he didn't mention it :P I had to read the code and test out that 'git' command to find out :P
14:14:27 <TrueBrain> he is sneaky like that :D
14:15:06 <frosch123> TrueBrain: you are lucky you missed the irc drama around 8144
14:15:43 <TrueBrain> ... who was against it? :P
14:15:47 <andythenorth> oof that looks too close to my day job
14:15:48 <frosch123> it belongs to those things everyone has an opinion about
14:16:02 <andythenorth> title case vs sentence case is the new vi vs emacs
14:16:19 <TrueBrain> in the PR reads most people reached a consensus about it :)
14:16:25 <TrueBrain> I mostly like: he just applied set rules
14:16:39 <TrueBrain> as that goes, I rarely care about what we do, as long as it is consistent :P
14:17:26 <frosch123> which is a very dull page to read :)
14:17:46 <frosch123> it was copied from some templates, and a lot of things just do not apply to ottd
14:17:57 <TrueBrain> as these things go :)
14:18:06 <TrueBrain> but how many people ragequit if I approve that PR? :P
14:18:57 <frosch123> i don't think anyone notices
14:19:12 <frosch123> ragequitting is more about things that are unfixable
14:19:29 <frosch123> or add tons of technical debt
14:19:44 <frosch123> like when you break the savegames in a bad way, or the apis
14:19:57 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8144: Change: Improved English texts - capitalization. https://git.io/JIbkV
14:20:43 <TrueBrain> well, currently I am a bit worried about only have 1 person that wants to review non-refactor-non-cmake PRs :P And he doesn't have a lot of time for it :)
14:23:34 <TrueBrain> the amount of open PRs is not -that- bad .. most just need an opinion, honestly
14:25:38 <TrueBrain> seems LordAro gave his, so yeah, some testing, and go go go :)
14:26:50 <TrueBrain> I guess I will run through a few next week, see how far my lack of C++ is going to shoot me in the foot :D
14:26:57 <LordAro> mm, that one should be fine
14:27:05 <andythenorth> how many are there?
14:27:06 <LordAro> (which should be easy)
14:27:08 <TrueBrain> there are a few more that look fine
14:27:20 <TrueBrain> and others that nobody just dared to close for what-ever reason :P
14:27:44 <TrueBrain> either it is the correct fix or it is not
14:28:09 <andythenorth> be nice if they fitted on 3 pages
14:28:11 <frosch123> TrueBrain: well, i would have closed that pr. but i have explanation-fatique, so i ignore things that i disagree with :)
14:28:43 *** sla_ro|master has joined #openttd
14:28:55 <TrueBrain> well, yes, as I said: these PRs need an opinion :)
14:29:04 <TrueBrain> and pretty sure nobody is going to agree on a single opinion
14:29:08 <TrueBrain> so ... yeah, who ever is first :P
14:29:33 <TrueBrain> for me it belongs in the category: it won't change how I play, and it helps how others play ..
14:29:37 <TrueBrain> if it doesn't hurt anyone, lets go!
14:29:45 <frosch123> abuot 8322. i just have no idea
14:30:00 <TrueBrain> someone has to check if it is a bug or a mis-feature
14:30:01 <_dp_> 8008 can use hotkey if it doesn't already
14:30:02 <frosch123> last time i used windows it had insane dos-compatibility issues with all-uppercase filenmaes
14:30:05 <_dp_> double click is nonsense
14:30:46 <TrueBrain> the PR is either way weird .. some things become lowercase, others uppercase .. it feels like it is not the proper fix :P
14:31:03 <TrueBrain> but okay .. many more of these PRs, pretty quick to process .. I will see what I can do next week :)
14:31:14 <TrueBrain> _dp_: your opinion; it is easy to dismiss change, but it doesn't make the feature wrong :)
14:31:21 <frosch123> let's count the hours until you burn out :)
14:31:39 <TrueBrain> frosch123: 16 years and counting, I know what I get myself into, no worries :)
14:32:35 <TrueBrain> frosch123: but yeah, it does require a high: I-don't-fucking-care-about-your-opinion mood :)
14:33:03 <TrueBrain> I like the Dark Mode .. reading GitHub is so much easier (for me) now :)
14:34:51 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8008: Feature: Double-click to rename vehicle https://git.io/JIb8d
14:36:12 <TrueBrain> what i fun to read in 8008 to see what kind of editors people are used to :P :D
14:37:46 <_dp_> is double click even used for anything currently?
14:37:49 <frosch123> where does 8008 mention editors?
14:38:01 <TrueBrain> frosch123: nowhere :) But I can see the different UX interactions :D
14:38:38 <TrueBrain> double-click-to-edit is/was pretty common for a while with a certain group of applications
14:38:54 <TrueBrain> haven't seen it much lately, as it is a dark-pattern
14:39:05 <frosch123> hmm, ah, versionone does that
14:39:36 <_dp_> can make it ctrl-click, it's already hopeless xD
14:39:40 <TrueBrain> and for example that you mention maximize/minimize, that made me giggle, as you are clearly "that kind of user" :D (not meant in any bad way btw)
14:39:53 <TrueBrain> personally for example I never double click on my window titles :P
14:39:56 <TrueBrain> I have icons for that :D
14:39:58 <frosch123> _dp_: some lists use double click on items to trigger the default button
14:40:18 <frosch123> ottd has support for 3/4/n clicks, but i don't think we use them
14:40:38 <frosch123> TrueBrain: i use it for maximize, a lot :)
14:40:46 <TrueBrain> in general, I think it is save to say that OpenTTD has an UX that is from 1994, and even back then it was an unusual one :P
14:41:46 <frosch123> oh, we use double click on edit boxes to open the on-screen-keyboard
14:41:53 <TrueBrain> the thing with all these kind of interactions, they are opinions. There is no "correct" way. There is only consistency :P
14:41:59 <glx> I use it sometimes, but the most annoying for me is the sensibility of "shake to minimise all other windows"
14:42:21 <frosch123> oh, yeah, i am so happy that i can disable all mouse gestures in kde
14:42:27 <frosch123> windows drives me mad with them
14:42:29 <TrueBrain> so 1000 people 1001 opinions on the matter; it is not really important honestly :)
14:42:50 <TrueBrain> what OS was it .. I think the Mac I had did it .. when you move your mouse around real quick, it becomes REALLY big
14:44:56 <TrueBrain> but anyway ... what to do with my own PRs .. I don't want to pressure 1 person to make the time for it, but I also don't want to wait weeks for anyone to take a look at it .. I can reduce the review to: I want to know about A B C and D if they are a good idea, to speed things along?
14:45:50 <frosch123> 7497 is by a factorio player :)
14:46:02 * andythenorth moves the cursor around really fast
14:46:22 <andythenorth> accessibility FTW
14:46:42 <TrueBrain> andythenorth: it scared the shit out of me the first time :P
14:46:54 <TrueBrain> (I tend to do these kind of things when I am bored / thinking)
14:47:53 <_dp_> frosch123, nvm, I typed 7479 xD
14:49:14 <andythenorth> merge the separation patch :P
14:49:28 <andythenorth> can we have a bleeding-edge branch?
14:49:36 <andythenorth> oh wait, that's JGR
14:54:21 <frosch123> andythenorth: not helpful. you need blood for bleeding
14:54:52 <_dp_> btw, onscreen kb is kinda broken with ui scaling
14:55:00 <TrueBrain> that is the solution: I should target my PRs to the JGR repo :D
14:56:19 <_dp_> does jgr ever reject patches? xD
14:56:29 <frosch123> TrueBrain: wasn't there a setting to not save config files
14:56:37 <TrueBrain> frosch123: there is; -x I believe
14:57:14 <TrueBrain> weirdly enough, that is not handled in the place I expected :D
14:57:29 <frosch123> that's exactly what i am currently baffled about :)
14:57:40 <TrueBrain> what a weird way to add such switch, lolz
14:58:01 <Eddi|zuHause> first time i hear about -x
14:58:12 <TrueBrain> read the manual more often? :)
14:58:31 <Eddi|zuHause> must be in a weird part of the manual :p
14:58:43 <andythenorth> _dp_ JGR rejects patch requests
14:59:20 <frosch123> does AfterNewGRFScan really keep a reference to another local variable?
14:59:32 <Eddi|zuHause> _dp_: plenty of times he said "XY patch is not being considered"
14:59:49 <TrueBrain> frosch123: yes ... I am baffled by that just now too
14:59:57 <TrueBrain> not sure I enjoy reading this part of the code :P
15:00:24 <frosch123> i have no idea what it does :)
15:00:33 <TrueBrain> which is also puzzling
15:00:39 <frosch123> there are two variables, and some pointers to them, which are sometimes compied, sometimes not
15:01:06 <TrueBrain> I am very tempted to remove '-x'
15:01:36 <frosch123> i though it was important for servers
15:01:38 <Eddi|zuHause> a NoX branch? :p
15:01:49 <_dp_> yeah, I'm using -x on servers
15:02:03 <TrueBrain> it seems it is likely it doesn't do what you expect :P
15:02:16 <_dp_> though mb it's not doing anything useful in my case anymore
15:03:01 <TrueBrain> frosch123: it seems to be a very complicated way to do .... well ... not much ..
15:03:19 <TrueBrain> but if I read it right: the local "save_config" is set to true by default, and false when -x, in a delayed way
15:03:33 <TrueBrain> so if scanning NewGRFs fails, it does not write the config
15:04:12 <TrueBrain> (so it is only set to true after OnNewGRFsScanned is called)
15:05:21 <TrueBrain> glx: addition request: can you make 'make docs' work too with OPTION_TOOLS_ONLY?
15:05:45 <TrueBrain> (no is an answer, I don't know if that is possible :D)
15:06:48 <frosch123> TrueBrain: no idea what the delay is used for though
15:07:18 <glx> hmm docs also needs parts of script_api
15:07:41 <frosch123> there is no case where OnNewGRFsScanned would not be called
15:07:43 <glx> and doesn't care about tools
15:07:47 <TrueBrain> "(svn r23224) -Codechange: first load the config file partially so we can push scanning AIs to later in the process (when the GUI is showing the progress bar)"
15:07:49 <frosch123> so, no idea what those pointers are aboutr
15:08:02 <glx> maybe OPTION_DOCS_ONLY :)
15:08:22 <TrueBrain> glx: otherwise forget about it ;)
15:09:10 <TrueBrain> so initially only a part of the config is loaded, and if that would be written back, that would be ba
15:09:29 <TrueBrain> anyway, for the PR, good point, I forgot there was a -x. Let me fix that :D
15:10:53 <glx> ok updated version seems to work
15:11:05 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JINTP
15:12:23 <glx> if you want to check on your side there's only one commit to cherry pick
15:13:09 <glx> I didn't touch the original commit
15:13:12 <frosch123> oh wait, there is the videodriver mainloop right in the middle of the function
15:14:22 <TrueBrain> frosch123: stop finding the ugly stuff in the code :P
15:14:58 <TrueBrain> glx: will do in a bit :)
15:15:11 <frosch123> well, i read over it, and got confused by the "shutdowngame" at the end :p
15:15:38 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JIW2Y
15:15:42 <TrueBrain> frosch123: solved the ugliness :)
15:16:20 <TrueBrain> well, with -x, the console command "save_config" still stores it to disk
15:16:26 <TrueBrain> but ... if you are that explicit about your request :D
15:18:00 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JIN3f
15:19:16 <DorpsGek> [OpenTTD/OpenTTD] michicc approved pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JINsX
15:19:19 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JINs1
15:19:31 <TrueBrain> frosch123: not sure that is the same behaviour
15:19:39 <TrueBrain> pretty sure the game loop already starts before the callback is ccalled
15:19:45 <frosch123> i have no idea why the code is written the way it is
15:20:23 <TrueBrain> well, exactly for the reason I just pointed out :D :)
15:20:30 <TrueBrain> the NewGRF scan thingy shows a GUI
15:20:38 <frosch123> ah, right, that's it. the scanning is in a thread, and the mainloop is actually the blocking thing
15:21:26 <TrueBrain> just that he used a local variable instead of a global
15:21:30 <TrueBrain> that was just ... holy crap :P
15:22:09 <DorpsGek> [OpenTTD/OpenTTD] frosch123 approved pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JINcs
15:22:27 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #7000: Some NewGRF variables concerning railtypes https://git.io/JINca
15:22:44 <TrueBrain> you found a very nice bug honestly :D
15:22:50 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #7000: Some NewGRF variables concerning railtypes https://git.io/JINCY
15:23:34 <glx> TrueBrain: well docs failed (not surprising)
15:24:10 <TrueBrain> glx: ah, yes, I assumed it could be used too :D I can fix that!
15:24:11 <DorpsGek> [OpenTTD/OpenTTD] andythenorth commented on pull request #7000: Some NewGRF variables concerning railtypes https://git.io/JINlJ
15:25:41 <TrueBrain> glx: but it shows that a "docs-only" target would be good too :)
15:25:56 <TrueBrain> docs should ignore any flag
15:25:59 <TrueBrain> and run on all source
15:26:02 <TrueBrain> not sure that is the case :P
15:26:25 <TrueBrain> not sure that list of PREDEFINED is maintained :D
15:27:06 <glx> docs just needs script_window, and the list is unmaintained I'd say, SDL2 is not in it
15:27:34 <frosch123> how do i add a comment to a commit message?
15:27:34 <glx> oh and find_version of course
15:27:44 <frosch123> tb's essays make that necessary :p
15:27:53 <glx> frosch123: second line ane more
15:28:04 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JINRa
15:28:10 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8358: Add: save openttd.cfg immediately on changing a setting https://git.io/JIW2Y
15:28:25 <TrueBrain> frosch123: oh-oh :D
15:28:39 <TrueBrain> hope it wasn't for that-what-I-just-merged :(
15:29:10 <frosch123> TrueBrain: 8367 needs a "-c" in the example in the commit message :)
15:30:08 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain dismissed a review for pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JINsX
15:30:11 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JIE7t
15:31:18 <TrueBrain> with that patch -c is so much more useful all of a sudden :D It is funny, as it annoyed me 10 years ago already :P
15:31:25 <TrueBrain> just never could grasp what the fuck was going on :D
15:32:02 <glx> hey #8358 added warnings ;)
15:32:28 <DorpsGek> [OpenTTD/OpenTTD] frosch123 approved pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JINgE
15:33:25 <TrueBrain> how does this compile at all?!
15:33:41 <TrueBrain> _save_config is only defined "extern" :P
15:35:13 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8373: Fix 8bef06a2: don't use "extern" when you want to implement a global https://git.io/JINVV
15:35:31 <TrueBrain> and now I wonder why the CI didn't fail on a warning .. :P
15:35:36 <TrueBrain> I am so used in projects that it does ..
15:36:09 <glx> macos wouldn't build at all if it fails on warnings
15:36:23 <TrueBrain> so we enable it for all targets but Mac OS?
15:37:08 <TrueBrain> which we don't do on the CI :)
15:37:27 <TrueBrain> glx: did you want to add the docs stuff too in your PR? Otherwise if you remove the "Test" commit, I will approve :D
15:38:00 <LordAro> TrueBrain: clang has warnings too :p
15:38:03 <glx> I can try to add an option for docs while at it
15:38:55 <TrueBrain> LordAro: not on the CI
15:39:01 <TrueBrain> I am not saying: enable it in CMake
15:39:08 <TrueBrain> I am saying: enable it on the CI for targets that are now clean :)
15:40:09 <TrueBrain> can't believe I fucked up "extern" :P Cannot believe it linked fine locally ...
15:41:00 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JIE7t
15:41:18 <LordAro> there *was* a warning for clang
15:41:25 <LordAro> maybe cmake disables the warning somehow?
15:42:09 <TrueBrain> sorry michi_cc , I now caused conflicts in your C++-string-loveing-PR :(
15:42:44 <TrueBrain> LordAro: the only target on the CI complaining atm, is the Mac OS target; which is pretty awesome btw :D
15:43:18 <LordAro> there's also some release-specific warnings
15:43:32 <LordAro> but i don't think any of them are particularly hard to fix
15:44:20 <TrueBrain> so much quicker to see what was going on :)
15:44:41 <TrueBrain> LordAro: release for 16.04 also has 3 warnings
15:44:47 <TrueBrain> pretty sure they are fake :P
15:45:30 <DorpsGek> [OpenTTD/OpenTTD] andythenorth commented on pull request #8268: Change: Sort cargo payment rates legend in order of (initial) cost instead of alphabetically https://git.io/JINyI
15:50:18 <LordAro> yeah, those are the ones i saw
15:50:29 <LordAro> probanly false positive indeed, though would be good to check :)
15:50:57 <LordAro> be nice if those cache warnings weren't there either
15:51:10 <LordAro> i suppose a vcpkg action upgrade didn't fix them?
15:56:24 <TrueBrain> in Dutch we call that "handjeklap"
15:56:25 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8373: Fix 8bef06a2: don't use "extern" when you want to implement a global https://git.io/JINxG
15:56:37 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8373: Fix 8bef06a2: don't use "extern" when you want to implement a global https://git.io/JINVV
15:57:10 <TrueBrain> I still do not understand why that linked btw
16:01:56 <TrueBrain> 57 mails from GitHub today
16:02:02 <TrueBrain> most informing me things failed or were cancelled :P
16:09:01 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8360: Add: allow using SDL2 in "relative mouse mode" https://git.io/JIAte
16:09:28 <frosch123> TrueBrain: c++17 simplifies the "extern" stuff
16:09:42 <frosch123> but not in a "not funny" way :p
16:10:12 <frosch123> *but in a "not funny" way
16:10:37 <TrueBrain> LordAro: it is in the PR description ... :P
16:10:40 <TrueBrain> frosch123: do I want to know? :)
16:10:56 <frosch123> exactly what i am worried about
16:12:12 <frosch123> but if your feet at strongly grounded: you no longer need the definition/initlaisation in the cpp. with c++17 you can put everything in the header and declare it "inline"
16:13:50 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8360: Add: allow using SDL2 in "relative mouse mode" https://git.io/JIAOd
16:14:00 <TrueBrain> frosch123: ... owh boy
16:14:15 <TrueBrain> (it is the reason I don't really like C++ .. it has even more magic :P)
16:14:50 <frosch123> it's like newgrf and mediawiki. keep adding stuff while keeping compatibility with old crap
16:15:05 <frosch123> but c++20 will save us all :)
16:15:47 <frosch123> it removed the need for #include, so we can fork a saner but link-compatible language (like kotlin for java)
16:16:12 <TrueBrain> removed the need for #include? How do you mean?
16:16:18 <TrueBrain> another way to define headers?
16:16:54 <frosch123> it's replaced by "import", which does not require to pass macros in both directions, so the imported thing can be written in another language
16:18:20 <frosch123> essentiaily: no more pre-processor for importing things
16:19:40 <TrueBrain> frosch123: that somewhat sounds like a sane direction for C++ :P
16:21:01 <frosch123> all modern/maintained languages (c++, rust, javascript, python) are adding the same features. they just have to deal with different amounts of old baggage
16:21:45 <frosch123> any they happily copy syntax from each other, which sometimes looks very out of place :p
16:23:32 <TrueBrain> I read the other day Rust has this lovely dbg! command, that is really cool :D
16:23:37 <TrueBrain> people are trying to port that to other languages
16:24:21 <TrueBrain> basically, just slap dbg!() around what you want to know the value of, no matter what part of what statement it is, and it will tell you
16:24:32 <TrueBrain> no more endless "printf" statements :D
16:24:59 <TrueBrain> and the 'printf("A");', 'printf("B");', .. statements :P
16:25:48 <TrueBrain> LordAro: I am still not sure what you meant with your questions on relative mouse mode :) Let me know if you meant to ask something else :D
16:27:46 <TrueBrain> (as in, that I did not understand the question :P)
16:31:25 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:33:53 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8360: Add: allow using SDL2 in "relative mouse mode" https://git.io/JIAav
16:34:24 <TrueBrain> frosch123: it could, but last_position is only used by that one function :P
16:34:38 <TrueBrain> UpdateCursorPosition(), that is
16:34:45 <TrueBrain> so I was over-optimizing tbh
16:34:53 <TrueBrain> (local stack variable vs class variable)
16:37:12 <frosch123> TrueBrain: ok, it's only related to that queued_warp
16:37:23 <frosch123> which is the road you did not want to get down :)
16:38:00 <TrueBrain> queued_wrap is no longer relevant for relative mode
16:38:17 <glx> hmm I broke stuff it seems
16:38:25 <frosch123> yes, it was a work-around to get relative mode to work, when it did not work :)
16:38:32 <TrueBrain> frosch123: really? Lol ...
16:38:39 <TrueBrain> I considered it a way to unify all video drivers
16:38:56 <TrueBrain> glx: honestly, if you can do the API CMakeLists.txt nicer, that wouldn't be a bad thing
16:39:07 <TrueBrain> personally, I am not so much after minimizing what DOCS_ONLY does
16:39:12 <TrueBrain> but more that you don't get warnings about missing SDL etc
16:39:24 <TrueBrain> (and I couldn't care less if TOOLS_ONLY and DOCS_ONLY are merged btw)
16:40:01 <frosch123> TrueBrain: queued_warp is about multiple mouse-move events getting delivered to ottd, before ottd can reset the fixed mouse posistion. so when things lagged ottd multiplied the relative movement
16:40:19 <frosch123> that behavior was very different between os, and possibly betweens wms
16:40:24 <TrueBrain> frosch123: that part I got from the comments :D (which were really welcome :P)
16:40:35 <frosch123> "relative movement" moves that problem to sdl to solve :p
16:40:48 <TrueBrain> so it is better, basically
16:41:18 <TrueBrain> the thing I left in UpdateCursorPositionRelative(), is that it mimics the current behaviour when next to a border
16:41:34 <TrueBrain> Relative Mode still gives you relative movement .. I negate that and clamp you to the border
16:41:38 <TrueBrain> (as the old method does that)
16:41:41 <TrueBrain> this is what last_position is needed
16:41:55 <TrueBrain> it is not as much an improvement (or not doing it isn't as much as a regression)
16:41:58 <TrueBrain> it is just different :P
16:42:11 <TrueBrain> now they are 1-on-1 the same with UpdateCursorPosition()
16:42:18 <TrueBrain> I am just not sure that is the right call here
16:42:55 *** m1cr0man has quit IRC (Quit: G'luck)
16:42:58 <TrueBrain> (but again, I wanted to minimize the impact this would have)
16:43:19 *** m1cr0man has joined #openttd
16:44:49 <TrueBrain> so as alternative, and this might be in line with what LordAro meant, is that we drop the wrapping mode from SDL2, and use the RelativeMode, and don't do the last_position stuff
16:44:57 <TrueBrain> SDL2 is new in 1.11, right? Or did it make it into 1.10?
16:45:37 <frosch123> sdl2 is the only one delivering relative movement for now, so we have to keep queue_warp for the other video drivers
16:46:03 <TrueBrain> hmm, now I come to think of it ... I was only considering Full Screen
16:46:14 <TrueBrain> relative mode has a difference in behaviour in windowed mode, if I am not mistaken
16:46:27 <TrueBrain> currently, if you move your mouse out of the window, it is out of the window
16:46:33 <TrueBrain> with relative mode, I am not sure you can :P
16:46:50 <TrueBrain> remember the remark you made about needing to press escape in the Preview stuff with emscripten?
16:46:57 <TrueBrain> that would be the default with relative mode, I guess
16:47:18 <TrueBrain> not sure how SDL does that :)
16:47:23 <TrueBrain> I cannot test it :P
16:49:01 <frosch123> i fetched your branch, i'll test it on native linux :)
16:49:18 <TrueBrain> Windows X11 server is ... special
16:49:52 <frosch123> oh, it's a driver parameter
16:49:56 <TrueBrain> (I have 3 cursors: Windows, X11 and OpenTTD :P)
16:49:59 <LordAro> TrueBrain: sdl2 definitely in 1.10
16:50:00 <TrueBrain> yes, -v sdl:relative_mode
16:50:18 <LordAro> possibly earlier, tbh
16:50:35 <TrueBrain> ah, yes, relative_mode makes you not being able to escape your window when it is windowed :D
16:50:57 <TrueBrain> LordAro: it was introduce Sep 2019
16:51:04 <TrueBrain> pretty sure that earlier is not an option :D
16:51:41 <TrueBrain> right, dinner time!
16:52:07 <frosch123> yeah, i cannot even leave it with alt+tab :)
16:52:27 <frosch123> do you have to enable it on start? is it possible to only enable relative mode when scrolling?
16:53:03 <frosch123> otherwise the parameter should be renamed from "relative_mode" to "capture_mouse_hard" or so
16:53:45 <frosch123> i cannot even kill ottd that way :)
16:54:26 <TrueBrain> Lets name it different from what SDL does, that is not confusing at all :p :p
16:54:41 <TrueBrain> Lol, for me I can leave fine with alt tab
16:55:59 <frosch123> it works if i get another window in front of it
16:56:10 <frosch123> i guess it's due to my "focus follows mouse-cursor"
16:56:33 <TrueBrain> So relative mode for sure is a bad idea in the generic approach
16:56:55 <TrueBrain> I can ifdef it for Emscripten only otherwise?
16:57:06 <frosch123> yes, if it works like this, it is unuable for anything but browser :)
16:57:50 <glx> hmm the more I look at my changes, the more I think I'll just merge the 2 options as OPTION_DOCS_TOOLS
16:58:19 <frosch123> unless you can capture "_cursor.fix_at = .." and use it to enable/disable it
16:59:35 <TrueBrain> Not sure there is a benefit in that case
16:59:44 <TrueBrain> Just adds code, I think
17:00:58 <frosch123> i still think calling it "relative_mode" is misleading. the observable behavior is "confine mouse to window".
17:02:04 <frosch123> but i woudn't know why you would want that either. so i guess no setting and just #ifdef EMSCRIPTEN ?
17:07:12 <TrueBrain> Emscripten needs it for fix_at to work .. so yeah, capture_mouse better describes it
17:08:03 <TrueBrain> I expected it to work better given SDL had it as option
17:09:16 <TrueBrain> I just hate ifdefing so much code
17:10:21 <frosch123> you can leave it in gfx.cpp
17:10:46 <frosch123> ifdefs are fine inside sdl2_v.cpp
17:11:08 <frosch123> os code is supposed to contain ifdefs, otherwise it looks suspicious
17:16:18 <michi_cc> TrueBrain: Hmm, your -c patch broke running the regression tests from MSVC when using CMake to build project files (shows how good I test)
17:18:34 <TrueBrain> michi_cc: and how good our CI is :p
17:19:00 <michi_cc> Command line is "regression/stationlist.exe -x -c regression/regression.cfg", started from the build dir where the lang folder is in.
17:19:18 <michi_cc> So loading the lang files depend on the working dir to be build.
17:19:26 <TrueBrain> Hahaha, owh boy .. it depends on weird behaviour ...
17:20:27 <TrueBrain> when we talked it over, we considered it such a weird usecase :P
17:20:35 <TrueBrain> but I guess it does mean we need yet-another-search-path
17:20:53 <michi_cc> Oh, and it's not just the lang files. Also applies to baseset files, maybe ai compat too.
17:21:05 <TrueBrain> workdir is no longer on search-path with -c, no
17:21:17 <TrueBrain> but why does the CI pass in that case? That is odd
17:21:30 <michi_cc> Or the way the CMake project output is organized is sub-optimal.
17:21:58 <michi_cc> Is the CI using project files or Ninja?
17:22:28 <michi_cc> Also, there might as well be differences between running MSBuild from the command line and what the GUI does.
17:23:55 <michi_cc> Well, Ninja is not doing exactly the same thing as the project files, so there's probably the reason.
17:25:36 <TrueBrain> hmm, possibly it is a bit more subtle
17:25:52 <TrueBrain> that is not the right english word :P
17:26:07 <TrueBrain> is it? It looks weird
17:26:38 <TrueBrain> can you copy an OpenGFX in regression/baseset for me, and see if it works?
17:26:51 <TrueBrain> as lang/ is expected next to the openttd binary, even with -c
17:27:24 <TrueBrain> it might also be the reason the CI works
17:27:35 <TrueBrain> as I think in your local install, you have OpenGFX in "content_download" folder
17:27:41 <TrueBrain> (in your global space)
17:28:07 <TrueBrain> so the -c fix, shows a problem that was hiding behind the surface :)
17:29:39 *** Tirili has quit IRC (Quit: Leaving)
17:29:43 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8360: Add: allow using SDL2 in "relative mouse mode" https://git.io/JIxlK
17:29:46 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed pull request #8360: Add: allow using SDL2 in "relative mouse mode" https://git.io/JIWai
17:34:55 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8355: Add: support for emscripten (play-OpenTTD-in-the-browser) https://git.io/JIn4l
17:34:59 <michi_cc> TrueBrain: The first message I get is literally "No available language packs (invalid versions?)
17:35:00 <michi_cc> ". When I copy build/lang into build/regression, next message is about missing basset.
17:35:18 <TrueBrain> that is very odd ... on Linux, the lang folder is not inside the regression folder
17:35:35 <TrueBrain> time to debug this shitty search path stuff :P
17:37:25 <TrueBrain> SP_BINARY_DIR, ///< Search in the directory where the binary resides
17:37:26 <glx> I guess project file doesn't output exe in build
17:37:29 <TrueBrain> is on the searchpath
17:37:44 <TrueBrain> hmmm .. good point glx
17:37:47 <TrueBrain> if it isn't, that makes sense
17:37:54 <TrueBrain> before the -c fix, the workdir was also on the searchpath
17:38:23 <TrueBrain> and indeed, MSVC was one of those targets that made subfolders in the build to store the executable
17:38:31 <TrueBrain> Debug/ and Release/ if I am remembering correctly
17:38:36 <michi_cc> regression/stationlist.exe -x -c regression/regression.cfg
17:39:09 <TrueBrain> regression/stationlist.exe?! :D
17:39:12 <glx> hmm yeah, regression target copies exe anyway
17:39:28 <michi_cc> The CMake project files are set up to copy (or create?) two exes in the regression folder, probably because of Win/Console exe variants.
17:39:37 <TrueBrain> so that is what breaks
17:39:46 <TrueBrain> set(OPENTTD_EXECUTABLE "regression/${REGRESSION_TEST}.exe")
17:39:56 <TrueBrain> set(OPENTTD_EXECUTABLE "regression.exe")
17:40:01 <TrueBrain> or what-ever, should fix it I guess
17:40:10 <TrueBrain> owh, ${REGRESSION_TEST} as most likely used to allow -j runs
17:40:12 <michi_cc> Which file is that in?
17:40:18 <TrueBrain> cmake/scripts/Regression.cmake
17:40:40 <TrueBrain> the line above also needs changing
17:40:53 <TrueBrain> regression/${REGRESION_TEST} -> regression_${REGRESION_TEST}
17:41:33 <TrueBrain> I think I would consider the case of MSVC a bug, as expecting "lang" to be next to the executable is a fair expectation I would say (for development builds)?
17:42:18 <TrueBrain> that still leaves basesets btw .. SP_AUTODOWNLOAD_DIR is weirdly implemented
17:42:46 <michi_cc> Changing it to use _ makes the error dialogs on startup go away, but I still get "Regression did not output anything; did the compilation fail?", which is a pain to debug as isn't any good way to see the actual regression output.
17:42:59 <TrueBrain> did you copy OpenGFX in baseset?
17:43:39 <TrueBrain> guess you should be able to execute the command from a powershell, not? But yeah, the regression was a nice idea, but in general, it always has sucked when it breaks :P
17:43:56 <michi_cc> I have baseset in SP_SHARED_DIR
17:44:06 <TrueBrain> not in a content_download folder?
17:44:08 <TrueBrain> (just to be sure :D)
17:44:37 <michi_cc> Why should I be using OpenGFX if the superior TTD base set exists? :P
17:45:09 <michi_cc> Okay, there are of course also OpenGFX versions somewhere on my disk :)
17:45:13 <TrueBrain> do I have a working MSVC ...
17:46:05 <glx> hmm execute_process() may need a WORKING_DIRECTORY
17:46:11 <TrueBrain> haha, signing in to GitHub fails via MSVC .. oops
17:46:21 <TrueBrain> glx: the working directory no longer matters if you use -c now
17:46:46 <frosch123> in what directory is the .obg file?
17:46:50 <FLHerne> there is only one search-path for 'content_download'.
17:46:51 <FLHerne> In other words, when using '-c', it will not look in '~/.openttd'
17:46:53 <FLHerne> inside the 'content_download' folder.
17:47:02 <frosch123> original baseset is unusual since the .obg is not with the .grf files
17:47:23 <TrueBrain> FLHerne: because when someone implemented content_download in 2007 or 2009, was .. being a bit weird :P
17:47:39 *** jottyfan has joined #openttd
17:48:08 <TrueBrain> well, that was already mentioned in the commit message, so pretty sure you didn't want to know that why :D
17:48:20 <TrueBrain> FLHerne: more a question is, is that a problem :)
17:50:23 <TrueBrain> (honest question btw)
17:51:08 <TrueBrain> better english would have been: "it will not look in "content_download" folder of "~/.openttd" :P
17:53:23 <FLHerne> It's a bit of an inconvenience when I create a new config file to test different settings, or for a patchpack, and my downloaded content isn't there anymore
17:53:43 <TrueBrain> FLHerne: well, functionality-wise, my patch didn't change that
17:53:49 <TrueBrain> but yes, I was annoyed by that behaviour too
17:53:53 <FLHerne> In other terms, it'll mean zBase lovers will download it even more :p
17:53:57 <TrueBrain> that is why I noted it in the commit message :P
17:53:59 <FLHerne> Yeah, it's not a regression
17:54:09 <TrueBrain> as it took me completely by surprise
17:54:23 <FLHerne> I just don't understand why you excluded that dir while improving the other things
17:54:25 <TrueBrain> Visual Studio wants to restart my computer after an update .. wtf? :(
17:54:36 <TrueBrain> FLHerne: I did not exclude it explicitly
17:54:39 <TrueBrain> it was always like this :)
17:54:57 <TrueBrain> OpenTTD uses search-paths to find files
17:55:07 <FLHerne> (if I read correctly, ~/.openttd is now included for non-content_download?)
17:55:09 <TrueBrain> it has a single search-path called SP_AUTODOWNLOAD_DIR
17:55:13 <TrueBrain> which is "content_download" basically
17:55:50 <FLHerne> Really? I remember having to copy my local grfs around because they weren't there
17:55:50 <michi_cc> TrueBrain: Okay, all clear on the further regression error, the path to editbin.exe changed with latest MSVC update and somehow CMake didn't pick that up until I completely cleared the cache and recreated the project files. So the "no output" error was due to failed conversion to console exe.
17:55:58 <glx> <TrueBrain> Visual Studio wants to restart my computer after an update .. wtf? :( <-- and a restart is not equivalent to shutdown and start next morning
17:56:04 <FLHerne> Perhaps I was doing things Wrong :p
17:56:09 <michi_cc> TrueBrain: Replacing "regression/${REGRESSION_TEST}.exe" with "regression_${REGRESSION_TEST}.exe" is till necessary, though.
17:56:13 <TrueBrain> FLHerne: if you put them in content_download, yes :)
17:56:22 <TrueBrain> FLHerne: let me draw something for you in a bit, maybe that makes it clear
17:56:24 <TrueBrain> and it needs change btw ;)
17:56:38 <FLHerne> I made that mistake years ago, but I don't think I've done it recently...
17:56:44 <TrueBrain> michi_cc: can you PR that change? Sounds like the correct solution to me honestly :)
17:57:07 <TrueBrain> glx: it is these days; at least for Updates
17:57:11 <TrueBrain> you have Update and Shutdown, these days
17:57:24 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIW6A
17:58:33 <michi_cc> That is actually doing a restart with a shutdown at the end.
18:01:10 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIxSH
18:01:25 <TrueBrain> michi_cc: lolzzz .. owh, windows ...
18:01:45 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIx9m
18:04:42 <TrueBrain> michi_cc: btw, commit message is too long :P :P
18:05:45 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
18:05:58 *** WormnestAndroid has joined #openttd
18:06:16 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8372: Add: [CMake] Option to only build tools https://git.io/JIxQ0
18:06:26 <TrueBrain> glx: except for the name, looks perfect to me :)
18:06:50 <TrueBrain> OPTION_NO_OPENTTD :P
18:07:03 <TrueBrain> I liked OPTION_TOOLS_ONLY, hard to think up another one :)
18:08:27 <michi_cc> TrueBrain: Regression is now of course failing on the CI :)
18:08:56 <michi_cc> The build output is completely unhelpful.
18:11:56 <TrueBrain> FLHerne: updated gist to make it more clear (examples)
18:12:57 <TrueBrain> michi_cc: owh, and now it fails ... lolz
18:14:38 <TrueBrain> and if you execute "ctest -C debug", it works?
18:15:13 <glx> hmm just based on timing displayed, maybe it's the output
18:15:34 <glx> if it fails to load the save
18:16:36 <michi_cc> I let CMake generate MSVC project files that run the regression for, I have no clue where ctest slots in in that.
18:17:35 <TrueBrain> and I cannot sign in to MSVC anymore, as it errors on the GitHub signin page .. lovely :D
18:18:52 <michi_cc> Okay, if I manually execute ctest in the build dir that was created by building the MSVC project files, the tests pass for me.
18:19:34 <glx> D:\developpement\GitHub\glx22\OpenTTD\out\build\x64-Debug\0 : + error : Registering two scripts with the same name and version'
18:19:34 <glx> D:\developpement\GitHub\glx22\OpenTTD\out\build\x64-Debug\1 : + error : 1: ai\regression\main.nut'
18:19:35 <glx> D:\developpement\GitHub\glx22\OpenTTD\out\build\x64-Debug\2 : + error : 2: D:\developpement\GitHub\glx22\OpenTTD\out\build\x64-Debug\ai\regression\main.nut'
18:19:55 <TrueBrain> I have seen that error before ...
18:20:54 <TrueBrain> it happens when it picks up both regressions if I am not mistaken
18:21:20 <TrueBrain> no, they are named different
18:21:49 <TrueBrain> it opens the same one twice? Lol
18:23:09 <TrueBrain> glx: can you try setting the working_directory to "regression", just for shit and giggles?
18:25:21 <TrueBrain> lolz ... yeah, that should solve it
18:25:23 <TrueBrain> holy crap, this code :)
18:28:30 <TrueBrain> and this is why it worked for the CI, I guess .. as the working-directory was different
18:28:41 <TrueBrain> seems you are right after all glx :D Let me know if that was really the case ;)
18:29:06 <glx> if I set working directory it's worse
18:31:39 <TrueBrain> that is also a bit weird
18:32:16 <TrueBrain> if you run openttd with '-dmisc=4', can you tell me the "added as search path" you get?
18:34:58 <andythenorth> do industry instances have any UID on the map (accessible to newgrf?)
18:37:07 *** iSoSyS has quit IRC (Remote host closed the connection)
18:38:44 <TrueBrain> glx: and just to confirm, that was a run on latest master, right?
18:41:35 <frosch123> andythenorth: yes, the northern-most tileindex
18:42:29 <andythenorth> Satisfaction Guaranteed GS
18:42:56 <frosch123> so you passed the 4-letter-abbreviation test
18:43:02 <andythenorth> yup, that's the main thing
18:43:22 <andythenorth> you're allocated between 0 and 100% of towns, and you have to keep them happy
18:43:36 <andythenorth> by fulfilling arbitrary goals
18:43:52 <andythenorth> there's a high score, and some kind of speed run is possible
18:44:25 <andythenorth> goals (would ideally be modularised) would be things like transport x passengers per month
18:44:32 <andythenorth> deliver x tonnes of alcohol
18:44:44 <andythenorth> deliver coal to the power station
18:44:45 *** arikover has joined #openttd
18:44:47 <andythenorth> transport away the rubbish
18:45:03 <frosch123> can you turn rubbish into alcohol?
18:45:11 <andythenorth> or vice versa also
18:45:23 <andythenorth> instead of having to know about all the industries and industry grfs
18:45:25 <TrueBrain> frosch123: I btw integrated the "relative mouse" into emscripten .. it indeed isn't too bad, as it doesn't need all the if()s :D
18:45:35 <andythenorth> I wondered about communicating via town registers
18:45:48 <andythenorth> how many are there? 255?
18:47:46 <andythenorth> so I could use known locations, and store information about, let's say, 10 industries
18:48:07 <andythenorth> storing the industry ID in one register, and whether it is satisfied in another register
18:50:50 <TrueBrain> hmm, MSVC doesn't want to compile OpenTTD for me :(
18:50:55 <TrueBrain> game_includes something
18:51:51 <TrueBrain> file(STRINGS "${FILE}" LINES REGEX "^void SQ${APIUC}.*_Register\\(Squirrel \\*engine\\)$") fails
18:51:56 <frosch123> andythenorth: that makes little sense. unless you want to loop over all 256 register somewhere, it sounds even worse
18:52:14 <andythenorth> dunno :) my idea for implementation is probably not important
18:52:34 <TrueBrain> lol: /src/script/api/||" cannot be read.
18:52:38 <TrueBrain> glx: any idea by any chance?
18:52:47 <andythenorth> there was a proposed industry<->GS callback, but it fell into drama
18:53:36 *** arikover` has joined #openttd
18:53:42 <glx> TrueBrain: not really, if I run manully in command line it seems correct
18:53:43 <andythenorth> mostly I'm looking for reasons to deliver stuff to towns, without imposing runaway growth
18:53:48 <andythenorth> town growth is sooooo boring
18:53:51 <TrueBrain> glx: no, the error I get :)
18:54:01 <TrueBrain> sounds like a separator or something?
18:55:08 <TrueBrain> at the end of SCRIPT_API_BINARY_FILES is ||(setFAIL_LINE=3&goto:ABORT)
18:55:12 <TrueBrain> doesn't sound right :P
18:55:51 <TrueBrain> pretty sure that was not suppose to be given to the application :D
18:56:05 <michi_cc> We are obviously looking at a gene that was highly conserved during evolution as it is essential to the development of all living organisms. Even small mutations result in un-viable offsprings :p
18:57:20 *** arikover has quit IRC (Ping timeout: 480 seconds)
18:57:36 <glx> TrueBrain: and SCRIPT_API_FILES looks good ?
18:59:19 <TrueBrain> where would I need to find out where it looks good?
19:00:03 <TrueBrain> it seems the add_command adds a ||, as in, if you fail, do this
19:00:15 <TrueBrain> but for some reason that is sent to the command executed
19:00:50 <michi_cc> Which CMake version are you using?
19:01:23 <TrueBrain> I just freshly installed Visual Studio 2019
19:01:47 <TrueBrain> 3.14 I have in my global
19:01:56 <TrueBrain> guess that is a bit old :P I assumed Visual Studio comes with its own?
19:02:05 <michi_cc> Okay, so you use the "open directory" from Visual Studio?
19:03:24 <michi_cc> The one that is available for me in the Visual Studio Tools Command prompt is 3.18 and for building the project files I use CMake+GUI 3.17.3. Maybe your PATH is set up in a way that it somehoe picks the 3.14 exe up?
19:04:10 <TrueBrain> so SquirrelIncludes.cmake gets a bunch of arguments
19:04:20 <TrueBrain> which works, except that there is this ||(setFAIL_LINE=3&goto:ABORT)
19:04:24 <TrueBrain> which is also given as argument
19:05:01 <TrueBrain> which gives 0 hits when searching for it
19:07:00 <TrueBrain> my thinking is always: if I have it on a fresh install, many more people must have this :P
19:10:28 <TrueBrain> no clue what causes that "setfail" to be added :D
19:11:44 * andythenorth looks what blocks FIRS 4 release
19:11:51 <andythenorth> oof it's a car crash :P
19:12:15 <andythenorth> I made a mistake and listed at least 2 algorithms as being so broken it's unshippable
19:12:28 * andythenorth wonders if anyone ever notices
19:13:40 <TrueBrain> ugh, I hate being stuck at such shitty problems to debug a completely different issue
19:14:11 * andythenorth sends TrueBrain a cookie of support
19:15:33 *** ChanServ sets mode: +v tokai
19:16:27 <TrueBrain> and the editor keeps trying to "help"
19:16:34 <TrueBrain> when you type ${, it comes up with the weirdest things :P
19:19:39 <TrueBrain> if(ARG STREQUAL "||") set(ARG_READ NO) endif()
19:19:50 <TrueBrain> still no clue what it was doing on the arguments, but what-ever
19:21:15 <andythenorth> oh yeah FIRS 4 got stuck on industry closure, I see
19:21:35 <andythenorth> I have to provide it (why?), it doesn't work (why?), nobody figured out how to make it work (why?)
19:22:06 <andythenorth> the FIRS 3 code doesn't work with > 3 cargos, and has to be rewritten completely, so eh
19:22:36 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
19:24:16 <TrueBrain> so, it works both from MSVC and from PowerShell for me, on master :D
19:27:41 <TrueBrain> I have a suspicion ... hmmmm ... how are GitHub runners configured ...
19:28:04 <TrueBrain> no, that wouldn't be possible .. hmm ... eeeeuuuuuhhhhh
19:28:06 <TrueBrain> how to debug this :P
19:29:45 <andythenorth> FIRS 4 for christmas? :P
19:29:54 <andythenorth> to play in browser OpenTTD? :P
19:32:29 <TrueBrain> ctest --verbose is your friend :)
19:34:59 *** arikover` has quit IRC (Remote host closed the connection)
19:35:12 *** arikover` has joined #openttd
19:39:52 <TrueBrain> glx / michi_cc : the problem is that Windows seems to pick up on ".\regression\" wrong in the search path
19:40:00 <TrueBrain> and considers files in the current folder too
19:40:07 <TrueBrain> so the current folder is twice on the search-space
19:40:13 <TrueBrain> (once by WORKING_DIR and once by BINARY_DIR)
19:40:40 <TrueBrain> so it seems it needs to be the full path :P
19:42:14 <TrueBrain> this is why it did work on the CI before michi_cc's patch :P
19:44:56 *** m1cr0man has quit IRC (Quit: G'luck)
19:47:22 <TrueBrain> fixing that is annoying, as this code is retarded
19:47:33 *** m1cr0man has joined #openttd
19:52:35 <TrueBrain> why did I want to fix this again? Can I just run away? :P
20:01:46 <frosch123> so you made it to 5 hours 20 minutes :)
20:02:02 <TrueBrain> no no, I have been calling this personal directory stuff bullshit after 2 minutes
20:02:10 <TrueBrain> this is utter crap code
20:19:32 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain dismissed a review for pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIx9m
20:19:35 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIxSH
20:19:36 <TrueBrain> sorry michi_cc , abusing your branch, as we need both patches for it to work :P
20:20:09 <TrueBrain> the code did become slightly cleaner :)
20:20:42 <TrueBrain> now lets hope it works ....
20:22:07 <TrueBrain> glx: still can't think of a better name :(
20:22:12 <TrueBrain> OPTION_TOOLS_ONLY doesn't cover it
20:22:27 <TrueBrain> but OPTION_TOOLS_DOCS isn't really explanatory ..
20:22:37 <TrueBrain> OPTION_TOOLS_AND_DOCS_ONLY? :D
20:23:15 <glx> long option name, but means you don't use it by accident
20:23:52 <TrueBrain> well, maybe we can do OPTION_DOCS_ONLY and OPTION_TOOLS_ONLY, which both do the same
20:24:01 <TrueBrain> I mean, nobody is going to complain you can do "make docs" with OPTION_TOOLS_ONLY
20:24:15 <TrueBrain> (well, the first person we greet with a big FUCK YOU BABY) :D
20:25:21 <glx> ah yes I can make something
20:28:44 <michi_cc> TrueBrain: Patch works locally with MSVC project files as well.
20:29:59 <michi_cc> Can't approve though, as it's still listed as my PR :)
20:30:31 <TrueBrain> well, if you agree, I can approve :P
20:30:48 <TrueBrain> but maybe we could ask frosch123 realllllyyyyyyyy nice to look at a heap of garbage :D
20:30:56 <TrueBrain> oops, wasn't selling it :P
20:31:58 <frosch123> what? my storage room?
20:33:03 <frosch123> why do you know what my storage room looks like?
20:33:27 <michi_cc> Doesn't every storage room look like that?
20:34:35 <frosch123> wtf. i thought the mess was only inside fileio. but there is the same stuff in win32?
20:35:02 <TrueBrain> I did call it a heap of garbage for a reason :)
20:35:09 <TrueBrain> and nobody made nice functions to do common operations
20:35:15 <TrueBrain> so instead ... it is a shitpile of shit
20:35:19 <glx> frosch123: yeah windows unicode handling
20:35:28 <TrueBrain> did I mention it is really poorly written? :P
20:35:51 <TrueBrain> I love that unix is in fileio.cpp
20:35:53 <TrueBrain> and windows in win32.cpp
20:35:59 <TrueBrain> that is just to fuck with you a TINY bit more :D
20:36:01 <michi_cc> fileio is supposed to be "platform independent" (expect of course for the XDG intrusions) while win32 contains platform specific stuff.
20:36:05 <frosch123> c++17 has filesystem library :)
20:36:31 <TrueBrain> michi_cc: we failed at that like 15 years ago I am afraid :)
20:36:45 <TrueBrain> pretty sure my name is also on a few of those lines, so I can call it names :P
20:37:06 <frosch123> at least you can blame it on truelight
20:37:15 <michi_cc> Some of that should probably have been in unix.cpp
20:37:46 <michi_cc> But then there are stray __HAIKU__ and WITH_COCOA fragments mixed in.
20:38:03 <TrueBrain> an abstraction layer here would have been better :)
20:38:43 <glx> all that stuff is pre-c++
20:40:56 <frosch123> ah, it's a better-hide-ws-changes diff
20:42:17 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIpjU
20:42:27 <TrueBrain> let's see if I didn't fuck up ^^ :D
20:42:55 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIpjU
20:43:27 <TrueBrain> all the new bells and whistles \o/ :)
20:52:34 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIhf1
20:53:25 <glx> stupid MSVC autocompletion on cmake files
20:53:43 <TrueBrain> glx: it is REALLY annoying :P
20:54:05 <glx> it tries to help, but really doesn't
20:55:23 <TrueBrain> nice call frosch123 :)
20:58:20 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIxSH
21:10:09 <frosch123> is _config_file always an absolute path?
21:11:05 <TrueBrain> that is the issue I was trying to solve :)
21:11:48 <frosch123> _searchpaths[SP_WORKING_DIR] = config_dir; <- isn't that an relative path then?
21:11:58 <TrueBrain> and that is what is failing on Windows
21:12:11 <frosch123> but unix is fine with it?
21:12:17 <TrueBrain> from my testing, yes
21:12:31 <TrueBrain> I would rather make it an absolute path
21:12:39 <TrueBrain> but we had no abstraction function that does that, what I could find
21:12:56 <TrueBrain> so I was balancing "do I fuck this up when I try to write this" vs "this always has worked, so do I care" :P
21:12:56 <frosch123> it makes no sense to add one now, when people want to switch to c++17
21:13:00 <frosch123> it's so much easier then
21:13:13 <TrueBrain> haha :) Well, that settles that balance :)
21:13:44 <frosch123> so, i guess ,whatever makes the regression run now?
21:14:53 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIhmc
21:15:30 <DorpsGek> [OpenTTD/OpenTTD] frosch123 approved pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIhmz
21:15:42 <frosch123> what was the blocker for c++17 again?
21:16:02 <TrueBrain> I leave merging up to you michi_cc :)
21:16:26 <LordAro> frosch123: i'd like to wait until the next debian stable
21:16:52 <LordAro> oh no, we did look at compiler compatibilities, didn't we?
21:17:05 <frosch123> why that? it alreaedy supports it
21:17:32 <LordAro> oldstable doesn't though?
21:17:40 <LordAro> and i think that's still supported?
21:17:48 <frosch123> you have to go to centos/enterprise-scamware to get a compiler that does not support c++17
21:18:17 <LordAro> oh no, stretch was eol in july
21:18:47 <LordAro> in which case, eh ¯\_(ツ)_/¯
21:20:26 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIhOe
21:20:56 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIhO3
21:20:58 <TrueBrain> brrr, LordAro scared me ...
21:21:07 <LordAro> TrueBrain: did you enable discussions?
21:21:30 <TrueBrain> LordAro: you were very verbose you didn't want it
21:21:58 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIhOw
21:22:04 <LordAro> TrueBrain: it was enabled on the OTTD repo
21:22:38 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIhON
21:22:53 <TrueBrain> LordAro: that was not me, as again, you were REALLY verbose about not wanting it :)
21:23:02 <TrueBrain> no clue why btw :P It sounds really useful for stuff like C++17 etc
21:23:30 <LordAro> i wouldn't say i was REALLY verbose about it
21:23:38 <LordAro> just mild disapproval
21:24:12 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIh3W
21:24:43 <TrueBrain> these kind of functions, we can just try them out, honestly
21:24:48 <TrueBrain> but they did a nice job
21:24:51 <TrueBrain> you can set your own categories
21:25:03 <TrueBrain> we can always try it out to talk over bigger points in life
21:25:06 <LordAro> it overlaps too much on the forum, imo
21:25:17 <TrueBrain> like ... we ever talk on the forum?
21:25:23 <TrueBrain> when was your last post? :P
21:25:28 <DorpsGek> [OpenTTD/OpenTTD] frosch123 opened pull request #8376: Codechange: Switch to C++17 on all platforms. https://git.io/JIh3F
21:25:34 <frosch123> let's see what the farm says
21:25:35 <LordAro> true, but it would just be another place to be ignored
21:25:43 <TrueBrain> with that attitude, sure
21:25:46 <TrueBrain> but if you never try things
21:25:53 <TrueBrain> and we could as well close up shop :)
21:26:26 <TrueBrain> I am running "make -j" over OpenTTD now ... I want to see my computer suffer :P
21:26:43 <TrueBrain> OpenTTD is not big enough to make that bark :)
21:27:02 <TrueBrain> just over 300 processes alive ... :P
21:27:08 <frosch123> i bind-mounted the build dir to a tmpfs
21:27:16 <TrueBrain> sorry, you gave me flashbacks of someone who run that on a 10k files project, on a production server
21:27:42 <frosch123> i did not profile it again after i got an ssd
21:27:42 <TrueBrain> I learnt there that there is no longer a 32k process limit :)
21:27:52 <LordAro> i did effectively a -j1200 once
21:27:54 <frosch123> but on the hdd i could reduce the -j by one, since it no longer waited on it
21:27:56 <LordAro> didn't work out so well
21:28:07 <frosch123> LordAro: everyone who compiled andy stuff did that once
21:28:17 <frosch123> firs was a fork bomb at some point
21:28:24 <glx> frosch123: CI shouldn't have issue, it's running on ubuntu 20.04
21:28:49 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8376: Codechange: Switch to C++17 on all platforms. https://git.io/JIhGJ
21:29:06 <TrueBrain> anyway, back to the GitHub Discussion, mostly what I think it can help for LordAro , is that we are a very closed-off small group talking about important things, like C++17. Doing it more in public could attract people who want to help out
21:29:24 <TrueBrain> of course we have to slap that one person who misbehaves once in a while
21:29:26 <TrueBrain> but that is nothing new :D
21:29:48 <TrueBrain> as those people commenting on the CMake PR, that was not really .. contributing :P
21:29:57 <TrueBrain> so there is a risk and a balance :)
21:30:15 <TrueBrain> I dislike tt-forums for development work, always have, as the layout it just shitty and I lose track
21:30:19 <TrueBrain> it is not made for development discussion
21:30:34 <dobby> how do you make a server to play online with friends ???
21:30:48 <TrueBrain> LordAro: let's do it when we have something to discuss :) It seems C++17 is no longer a discussion now ;)
21:31:31 <LordAro> no reason not to go to C++14, at the very least
21:31:37 <LordAro> i'm surprised it wasn't picked already
21:32:31 <frosch123> i don't understand the error. do they define stupid macros again?
21:32:48 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8375: Fix: [Actions] cleanup ci-build workflow to be up-to-date https://git.io/JIpjU
21:32:59 <dobby> i followed that but no one can get in :/ dont know why though
21:33:34 <frosch123> opt::optional <- what is that?
21:33:53 <frosch123> oh, we have a custom std::optional implementation, iguess that's to be deleted in that pr
21:33:55 <LordAro> something that isn't std::optional?
21:33:58 <DorpsGek> TrueBrain: OpenTTD uses TCP and UDP port 3979 for server <-> client communication, UDP port 3978 for masterserver (advertise) communication (outbound), and TCP port 3978 for content service, a.k.a. BaNaNaS (outbound)
21:34:28 <DorpsGek> [OpenTTD/OpenTTD] diddledan commented on issue #8370: In game music does not play (Ubuntu / Snap version) https://git.io/JIiLo
21:34:57 <frosch123> well, something for another day
21:35:02 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
21:35:27 <TrueBrain> it looks good to me, but I wanted to be sure :P
21:36:01 <glx> I quickly checked locally (just with the target view in MSVC)
21:36:55 <TrueBrain> Mac and Windows work
21:38:33 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
21:39:00 <glx> -DOPTION_DOCS_ONLY=yes \ <-- options are ON/OFF ;)
21:39:09 *** WormnestAndroid has joined #openttd
21:39:15 <TrueBrain> clearly an old copy paste :P :P
21:39:16 <glx> but yes works too of course
21:39:34 <TrueBrain> seems this job is going to work :D
21:39:39 <TrueBrain> means I can also compare the binaries with the latest nightly
21:39:43 <TrueBrain> see if they are similar enough :)
21:40:51 <TrueBrain> so the Mac build wants to say we have to target 10.14 for C++17 ... :P
21:42:31 *** dobby has quit IRC (Quit: Page closed)
21:45:54 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #8374: Fix: [CMake, MSVC] Don't copy regression exe's into a subfolder as th… https://git.io/JIxSH
21:46:59 <TrueBrain> found a nice bug :D Changelog is always 1 commit long :P
21:47:05 <TrueBrain> the changelog script uses "git log"
21:47:09 <TrueBrain> but we fetch at depth = 1
21:47:49 <glx> and there's not manifest step yet
21:48:19 <TrueBrain> manifest.yaml is calculated by OpenTTD/workflows
21:48:25 <TrueBrain> the one currently done by Azure is not used :D
21:50:45 <michi_cc> Random snippet regarding C++17, OSX and file handling: "std::filesystem - just don't go near it, I think even with 10.14"
21:50:50 <TrueBrain> okay, nightlies appear identical \o/ :D :D
21:51:03 <TrueBrain> michi_cc: that bad?
21:51:08 <LordAro> michi_cc: but that's the main feature
21:51:18 <LordAro> main feature worth using*
21:52:01 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
21:52:54 <michi_cc> LordAro: The real main feature would the to make out save/load system not unconditionally undefined behaviour without stupid lambda tricks.
21:53:33 <TrueBrain> tell me about it :(
21:53:37 <TrueBrain> can't we make mind-readers?
21:53:59 <TrueBrain> LordAro: if you find the time, at the bottom of the 8371 PR are some questions I don't have answers for. Would love your input on it :)
21:58:24 <DorpsGek> [OpenTTD/website] TrueBrain opened pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIhBu
22:00:16 <TrueBrain> oh-oh, website no longer wants to build .. lets retrigger and hope it was just a boo-boo :P
22:00:30 <DorpsGek> [OpenTTD/workflows] TrueBrain opened pull request #10: Add: improve workflows for upcoming OpenTTD release workflow https://git.io/JIhRs
22:01:34 <TrueBrain> ruby fails ... for some reason
22:12:17 *** sla_ro|master has quit IRC ()
22:21:54 *** Wormnest has joined #openttd
22:22:18 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:24:45 <DorpsGek> [OpenTTD/website] TrueBrain opened pull request #178: Fix: also install ruby-etc in order for "gem update" to work https://git.io/JIhwr
22:24:48 <TrueBrain> this not not why you dockerize stuff ^^
22:25:19 <TrueBrain> I would expect that info@ would be emailed about the failures, but it is not .. hmm
22:26:54 <glx> it started 3 days ago it seems
22:27:02 <TrueBrain> we should now get email at info@ for all failed workflows triggered by DorpsGek :)
22:27:03 <TrueBrain> should never happen, so it should be fine :)
22:27:53 <TrueBrain> but searching for "etc" is ... well, no :P
22:27:56 <DorpsGek> [OpenTTD/website] LordAro approved pull request #178: Fix: also install ruby-etc in order for "gem update" to work https://git.io/JIhof
22:28:58 <DorpsGek> [OpenTTD/website] TrueBrain merged pull request #178: Fix: also install ruby-etc in order for "gem update" to work https://git.io/JIhwr
22:29:15 <DorpsGek> [OpenTTD/website] TrueBrain updated pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIhBu
22:29:55 <TrueBrain> the git interface of alpine is not responding, so I cannot look if it was a change they made :)
22:30:27 <TrueBrain> anyhoooooo ... deployed the schedule on my fork, to see if it triggers correctly tomorrow :)
22:31:24 <TrueBrain> and I think I know a way to make sure it triggers on 20:00 CEST
22:31:41 <TrueBrain> just trigger it at 1800 and 1900 UTC, and check if it is 2000 CE(S)T :D
22:32:38 <TrueBrain> fixing changelog.txt will be a bit annoying ... maybe we should just do a full checkout
22:32:49 <DorpsGek> [OpenTTD/website] orudge commented on pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIhK9
22:33:08 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIhKA
22:34:12 <DorpsGek> [OpenTTD/OpenTTD] orudge commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIh64
22:34:36 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:34:59 <LordAro> TrueBrain: given we're not particularly worried about the time it takes to build nightlies/releases, don't see an issue with doing a full(er) checkout
22:35:06 <DorpsGek> [OpenTTD/website] TrueBrain commented on pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIh65
22:35:25 <TrueBrain> LordAro: that was also the balance I made, so we are in agreement :)
22:36:05 <TrueBrain> LordAro: but downloads from GitHub releases is free! :P
22:36:11 <TrueBrain> anyway, any opinion is fine by me :)
22:37:57 <DorpsGek> [OpenTTD/website] orudge commented on pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIhPE
22:38:55 <TrueBrain> wait ... CPack produces .zip files?
22:39:47 <DorpsGek> [OpenTTD/website] orudge commented on pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIhXR
22:40:37 <orudge> TrueBrain: I assume it does just now :)
22:40:51 <orudge> Will look in the next few days
22:40:57 <TrueBrain> 1.10 was before CMake :D
22:41:03 <orudge> That might be it then :)
22:41:36 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIh18
22:41:55 <DorpsGek> [OpenTTD/website] TrueBrain commented on pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIh1V
22:43:02 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIhMe
22:44:17 <TrueBrain> okay, I have no clue how to skip Linux for nightlies ... last time I tried, it didn't execute the last step :P
22:44:21 <TrueBrain> I will figure something out :)
22:44:35 <TrueBrain> LordAro: I tried the -dbg.deb the CPack method, but it turns out that is not implemented for applications atm
22:44:50 <TrueBrain> there is a stalled patch fixing that .... and that is about it :P
22:45:29 <TrueBrain> I will most likely make an issue for the .deb dependency issue and the lack of -dbg.deb, as it might be a while before we figure out how to do all that :)
22:45:39 <glx> IIRC last time I enabled zip for macos it broke
22:45:55 <TrueBrain> what debian versions to build for, hmmm
22:46:45 <LordAro> might as well add bullseye as well
22:46:54 <TrueBrain> not released yet, is it?
22:47:02 <TrueBrain> so for a release, that sounds a bit odd?
22:47:11 <andythenorth> ouch it bedtime?
22:47:17 <DorpsGek> [OpenTTD/OpenTTD] orudge opened pull request #8377: Feature: Switch to bzip2 compression for DMG files (supported by OS X 10.4+) https://git.io/JIhyJ
22:47:39 <orudge> Right, I'd better get back to wrapping presents :)
22:47:51 <LordAro> TrueBrain: not "released", but it's fairly solidly testing
22:47:58 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8377: Feature: Switch to bzip2 compression for DMG files (supported by OS X 10.4+) https://git.io/JIhy4
22:48:13 <TrueBrain> LordAro: I would suggest we do not add targets that are not released yet for our releases
22:48:15 <orudge> TrueBrain: yes, I have :)
22:48:29 <orudge> Should save a bit of money on the macOS bandwidth bill ;)
22:48:31 <TrueBrain> LordAro: say we release 1.11, and when bullseye releases, the deb doesn't work for it?
22:49:03 <LordAro> TrueBrain: i'm reasonably certain that can't happen
22:49:13 <TrueBrain> (honest question out of interest)
22:49:37 <LordAro> well, apart from the fact that it should be freezing in march, i don't think debian upgrades its packages in an incompatible way
22:50:08 <TrueBrain> on sid, this does happen from time to time, not?
22:50:12 <TrueBrain> as that is a bit the idea :P
22:50:27 <TrueBrain> fair .. but for an unrelease target, I can imagine it happens
22:50:39 <LordAro> the only real concern is icu, which is a 'versioned' package
22:50:47 <LordAro> i.e. icu57, icu63, etc
22:51:25 <LordAro> and.. hmm, there is icu68 in experimental at the moment (opposed to icu67 in testing/sid)
22:51:34 <LordAro> TrueBrain: alright, we'll hold off on bullseye for now
22:51:41 <TrueBrain> ... I JUST ADDED IT!
22:51:54 <TrueBrain> honestly, adding it / removing it .. it is simple and quick
22:51:56 <LordAro> would want to add it before 1.11.0 release though
22:51:59 <TrueBrain> so lets do it when they go in freeze
22:52:08 <LordAro> as 1.11 will definitely miss the main repos
22:52:14 <TrueBrain> I will now run it to see if it builds, just to gather the info
22:52:16 <michi_cc> Anybody replaced icu-layout for Linux yet?
22:52:25 <TrueBrain> michi_cc: you want a list of people doing that?
22:52:27 <LordAro> turns out its a lot of work
22:52:42 <LordAro> and pango pulls in a lot of ugly looking libraries
22:53:02 <LordAro> (which is the last i heard of it, about a year or more ago)
22:53:41 <michi_cc> There's supposed to be Harfbuzz, but I'm not clear which parts it's doing exactly.
22:53:58 <TrueBrain> the last few people that looked into it, never returned
22:53:58 <LordAro> i feel like one or more of them have been deprecated
22:54:01 <TrueBrain> so I suggest you don't look into it :)
22:54:05 <TrueBrain> I like having you around :D
22:54:37 <michi_cc> Well, doing it for Windows and OSX was quite enough for me, thank you very much.
22:55:18 <TrueBrain> that is the correct attitude :)
22:55:27 <TrueBrain> harfbuzz, last I worked with it, felt really incomplete
22:55:31 <TrueBrain> but this was .. 4 years ago?
22:55:50 <TrueBrain> it was really like: ICU doesn't want to do it anymore, but they do not provide any sane alternative ... :P
22:58:41 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #8377: Feature: Switch to bzip2 compression for DMG files (supported by OS X 10.4+) https://git.io/JIhyJ
22:58:57 <TrueBrain> look at us, all being productive and shit :P
23:00:22 <TrueBrain> stretch: 6 minutes, buster: 7 minutes 20 seconds, bullseye: 7 minutes 35 seconds
23:00:26 <TrueBrain> I hate modern compilers
23:00:28 <TrueBrain> they are so slow :P
23:01:01 <TrueBrain> okay, all three "just work"
23:01:45 *** andythenorth has quit IRC (Quit: andythenorth)
23:03:37 <TrueBrain> we can produce a full release in 9 minutes and 43 seconds :D
23:03:44 <TrueBrain> that is possibly faster than we ever could
23:03:47 <TrueBrain> it used to be 30+ minutes :P
23:04:01 <TrueBrain> but okay, going to add some seconds for a full checkout
23:04:13 <TrueBrain> a checkout used to take 8 seconds .. .and now ...
23:04:39 <TrueBrain> btw, you can just trigger a release of anything manual
23:04:46 <TrueBrain> from a Pull Request, from a tag, from a branch
23:05:18 <TrueBrain> (click any of the Linux jobs)
23:09:20 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
23:11:22 *** azulcosta has joined #openttd
23:11:53 <TrueBrain> ugh, MacOS runners are out for a piss again
23:11:59 <TrueBrain> owh, right, I was going to rename Mac OS to MacOS
23:12:05 <TrueBrain> too late now .. not going to do another change to the CI
23:13:33 <glx> TrueBrain: if with needs.<job id>.result checks could work, but probably a pain to write :)
23:14:03 <TrueBrain> glx: the problem is, to read that variable, you need to have the job as a "needs"
23:14:08 <TrueBrain> and if you do, your job is skipped
23:14:16 <TrueBrain> not even evaluated, if a "need" job skips
23:14:59 <TrueBrain> the problem is in this sentence: "Identifies any jobs that must complete successfully before this job will run."
23:15:10 <TrueBrain> they consider skipped non-successful :(
23:16:04 <TrueBrain> hmm .... I have another thing I want to try ..... lets see if I can be sneaky
23:16:44 <glx> "If a job fails, all jobs that need it are skipped unless the jobs use a conditional statement that causes the job to continue." <-- not very clear
23:16:58 <TrueBrain> okay, that is interesting
23:17:01 <TrueBrain> I will try that after this try :D
23:21:22 *** iSoSyS has quit IRC (Remote host closed the connection)
23:23:08 <glx> but I think there's a default jobs.<job_id>.if when jobs.<job_id>.needs is present, but it can be overwritten
23:23:24 <TrueBrain> I am testing it out as we speak :)
23:23:37 <TrueBrain> if: needs.source.result == 'success' && needs.docs.result == 'success' && (needs.linux.result == 'success' || needs.linux.result == 'skipped') && needs.macos.result == 'success' && needs.windows.result == 'success'
23:23:56 *** gelignite has quit IRC (Quit: Stay safe! Stay at home! Stop the chain reaction!)
23:23:57 <TrueBrain> it is not like it mutates often
23:24:30 <TrueBrain> MacOS runners are really stuck again :P
23:24:44 <TrueBrain> hopefully after next week they behave better :)
23:24:47 <TrueBrain> maybe they are all M1s!
23:24:59 <TrueBrain> the documentation was a lie ...
23:25:17 <TrueBrain> let me try with 'true' or something
23:26:48 <azulcosta> hello to all. I've uploaded my very first grf pack to bananas 24h ago... it's listed on bananas newgrf list.... but won't show up for download ingame for users.... Am I missing something ?
23:29:54 <azulcosta> well, i was myself late to uploading the pack to github... but don't know if that was the problem..
23:32:33 <TrueBrain> glx: one thing at the time :)
23:32:34 <glx> (but bananas doesn't care)
23:32:49 <TrueBrain> azulcosta: by the looks the item should be listed on BaNaNaS indeed; not sure why not. I will put it on my list to check out tomorrow
23:32:57 <TrueBrain> it is a bit late here now, won't be able to figure it out :P
23:33:07 <TrueBrain> I think it might be related to your minimal version
23:33:21 <azulcosta> well, i've only sent pack to github last hour....
23:33:26 <TrueBrain> it is a valid one, but we might have made a boo-boo with releases, causes that to fail
23:33:36 <TrueBrain> but I will get back to you on that :)
23:34:11 <azulcosta> i've set minimal version compatible, or grf version (used code=08) ?
23:34:38 <TrueBrain> now related to what glx is saying, you marked your GRF as GPLv2. That license defines you should also give your source code (how you got to build the GRF)
23:34:42 <TrueBrain> that is what should be on your GitHub
23:34:45 <TrueBrain> not the compiled GRF
23:34:51 <TrueBrain> but for example your NML source
23:34:59 <TrueBrain> or what-ever you used to get to the GRF result
23:36:10 <TrueBrain> it needs smaller experiments to figure out what will work :)
23:36:15 <TrueBrain> if you are bored, feel free :P
23:36:41 <glx> azure dependencies were easier to manage :)
23:37:05 <TrueBrain> okay, changelog.txt works now too
23:39:11 <TrueBrain> owh, I have one more idea to try ...
23:42:25 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8378: Remove undefined behaviour from save/load code https://git.io/JIjs7
23:42:44 <michi_cc> Okay, this PR will probably fail in interesting ways on the CI, but that is something for tomorrow.
23:46:17 <TrueBrain> but it doesn't show anything on skipped jobs
23:46:58 <TrueBrain> owh, I did the step debugger
23:47:01 <TrueBrain> lets try the runner debugger
23:47:11 <TrueBrain> please don't leak secrets, please don't leak secrets ... :P
23:47:54 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #8378: Remove undefined behaviour from save/load code https://git.io/JIjs7
23:48:02 <michi_cc> Okay, I lied, but now I'm really off :)
23:48:10 <TrueBrain> sleep well michi_cc
23:51:47 <glx> azubieta: I see your newgrf in master (so in nightly)
23:52:29 <TrueBrain> I think the problem is we never bumped the 1.10.2 client
23:52:37 <TrueBrain> so that version simply is not going to work
23:52:49 <TrueBrain> but I have to validate what the client sends :)
23:52:54 <TrueBrain> something for tomorrow
23:53:24 <TrueBrain> the job is being skipped without telling what is evaluated
23:53:44 <TrueBrain> so .. they give a lot of logs
23:53:46 <TrueBrain> but not for this part :D
23:53:56 <glx> yeah logging doesn't help in our case
23:54:58 <TrueBrain> so ugly solution it is for now
23:54:59 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
23:55:50 <TrueBrain> there is also "if: always()"
23:55:55 <TrueBrain> but I need a smaller test setup to test this :)
23:57:05 <TrueBrain> "The exception is, if you set ‘if: always()’ in job2, job2 will be executed even job1 failed." I find on the forums
23:57:07 <TrueBrain> so there is hope there
23:58:40 <TrueBrain> if: "always()&&(needs.jobA.outputs.job_status=='success')" is suggested to work
continue to next day ⏵
