IRC logs for #openttd on OFTC at 2019-10-20
⏴ go to previous day
00:00:24 <glx> anyway I think I need ply and pillow according to the readme
00:00:50 <andythenorth> but there are only x hours per day
00:00:55 <andythenorth> and some are needed for playing Blitz
00:03:21 <glx> pillow link in readme is 404 ;)
00:04:51 <glx> "We don’t recommend trying to build on Windows. It is a maze of twisty passages, mostly dead ends." <-- oh that's nice
00:07:02 <DorpsGek_III> [OpenTTD/nml] andythenorth opened pull request #50: Doc: fix 404 on pillow install instructions link https://git.io/JeB6X
00:07:27 <DorpsGek_III> [OpenTTD/nml] LordAro approved pull request #50: Doc: fix 404 on pillow install instructions link https://git.io/JeB6D
00:07:36 <DorpsGek_III> [OpenTTD/nml] LordAro merged pull request #50: Doc: fix 404 on pillow install instructions link https://git.io/JeB6X
00:13:03 <glx> I'm in a state where I think python is a mess to use ;)
00:14:09 <andythenorth> glx: python is a mess to use
00:14:20 <andythenorth> you have to develop strategies :P
00:14:31 <glx> pip install pillow should work, but it fails beacause it wants to build the package
00:14:39 <glx> and needs a bunch of deps
00:15:08 <LordAro> glx: if you can, always rely on any system package if you can
00:15:28 <LordAro> otherwise, there's usually ways to force pip to not try to compile
00:16:22 <andythenorth> if you isolate all your pythons, and all your projects
00:16:25 <andythenorth> then mostly it's fine
00:16:31 <LordAro> andythenorth: that too
00:16:35 <LordAro> depends on the project
00:16:36 <andythenorth> as long as setuptools doesn't break, yet again, upstream
00:17:01 <andythenorth> nearly all of my recent python failures are due to Apple breaking the Xcode libs
00:17:36 <andythenorth> the usual suspects are things like libxml :P
00:18:45 <glx> but I guess it's not the only one
00:19:49 <DorpsGek_III> [OpenTTD/OpenTTD] Wirdal commented on issue #7592: Road vehicles don't balance between multiple loading bays https://git.io/fjlL0
00:25:19 <milek7> LordAro: seems to work
00:27:17 <glx> because I'm still fighting with pillow
00:27:41 <glx> so 3.8 is really the issue
00:36:51 <glx> after some workaround (downloading a prebuilt package) pillow is installed
00:44:20 <FLHerne> peter1138: The "Support for Not Road Types" commit breaks `nmlc --nml=out.nml example_road_vehicle.grf`
00:45:38 <FLHerne> `[expression.identifier_to_print(roadtype.value) for roadtype in self.roadtype_list]` crashes when the roadtype table contains a list
00:46:10 <FLHerne> i.e. `RED_TEST: [RED_, REDR, ROAD]`
00:46:54 <FLHerne> I tried to debug it, but don't really know how roadtypes are supposed to work yet
00:47:28 * FLHerne should RTFM, but not at midnight :P
00:48:34 <FLHerne> (I rebased a couple of local commits onto master, found it crashed now, and couldn't figure out why until I realised it was crashing already)
00:58:19 <glx> it runs, I can do "nmlc --help", but something is still wrong
00:59:16 <LordAro> glx: time.clock() was removed in 3.8
01:01:11 <LordAro> process_time instead, i think
01:04:16 <glx> I don't really know python coding :)
01:22:20 <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #53: Fix #52: time.clock() has been removed in python 3.8 https://git.io/JeBii
01:23:34 <glx> I'll let you submit your setup.py change
01:35:28 *** andythenorth has joined #openttd
01:35:45 <andythenorth> FLHerne: GL RTFM :)
01:35:54 <andythenorth> I haven't written it yet
01:36:45 <andythenorth> also it wasn't crashing for me?
01:36:57 <andythenorth> wife's away, can't sleep
01:37:11 <supermop_Home> I usually sleep more in that case
01:41:37 <andythenorth> FLHerne: confirmed, that fails
01:41:47 <andythenorth> can't see a useful line number in the error message
01:42:02 <andythenorth> OTOH I am not wearing my glasses :P
01:44:24 * andythenorth figures there must be an nml output module that needs updating
01:44:36 <andythenorth> the grf and nfo output works
01:45:25 <glx> D:\developpement\GitHub\glx22\nml\examples\road_vehicle [actions +2 ~1 -0 !]> nmlc.exe --nml=out.nml .\example_road_vehicle.grf
01:45:25 <glx> [Knmlc ERROR: Input file is not utf-8 encoded: 'utf-8' codec can't decode byte 0x82 in position 5: invalid start byte
01:45:50 <glx> a totally different error
01:47:43 <andythenorth> but nmlc can't treat a grf as input :)
01:47:49 <andythenorth> so that command flherne uses is odd :)
01:48:08 <andythenorth> but nmlc --nml=out.nml example_road_vehicle.nml should be broken for you
01:48:37 <andythenorth> should be something like
01:48:41 <andythenorth> "Error: (TypeError) "expected string or bytes-like object"."
01:49:01 <glx> Location: File "D:\developpement\GitHub\glx22\nml\nml\expression\__init__.py", line 47, in identifier_to_print
01:49:35 <andythenorth> nmlc example_road_vehicle.nml should work fine
01:49:43 <andythenorth> and --nfo=foo.nfo should work
01:50:26 <andythenorth> --verbosity=4 isn't getting me any more useful info
01:50:40 <andythenorth> anyway, way to late for this :)
01:55:42 <andythenorth> ret += ', '.join([expression.identifier_to_print(roadtype.value) for roadtype in self.roadtype_list])
01:55:49 <andythenorth> is a straight copy from railtypes
01:55:58 <andythenorth> which aren't reported as failing on lists
01:56:58 <andythenorth> oh railtypes fails :)
01:57:04 <andythenorth> just tested it with iron horse
01:57:17 <andythenorth> I wonder if nmlc writing nml has been broken for some time
01:58:54 <andythenorth> also, bedtime, srsly
02:52:38 *** Xaroth72 has joined #openttd
03:27:32 *** ChanServ sets mode: +v tokai
04:41:50 *** Extrems has joined #openttd
05:51:53 *** andythenorth has joined #openttd
06:17:33 *** ist5shreawf[m] has quit IRC
06:27:02 *** Xaroth72 has joined #openttd
06:27:02 *** Thedarkb1-X40 has joined #openttd
06:27:02 *** FLHerne has joined #openttd
06:27:02 *** johanna[m] has joined #openttd
06:27:02 *** fjodor[m] has joined #openttd
06:27:02 *** patricia[m] has joined #openttd
06:27:02 *** grossing has joined #openttd
06:27:02 *** dihedral has joined #openttd
06:27:02 *** backtu[m] has joined #openttd
06:27:02 *** nolep[m] has joined #openttd
06:27:02 *** dude[m] has joined #openttd
06:27:02 *** ircer[m] has joined #openttd
06:27:02 *** iarp[m] has joined #openttd
06:27:02 *** tops[m] has joined #openttd
06:27:02 *** ist5shreawf[m] has joined #openttd
06:27:02 *** khavik[m] has joined #openttd
06:27:02 *** ZirconiumX has joined #openttd
06:27:02 *** KenjiE20 has joined #openttd
06:27:02 *** goodger has joined #openttd
06:37:33 *** HerzogDeXtEr has joined #openttd
06:42:24 *** Antheus has joined #openttd
07:12:11 *** Sheogorath has joined #openttd
07:34:40 *** Wormnest__ has joined #openttd
08:01:40 *** arikover has joined #openttd
08:29:51 *** supermop_work has joined #openttd
08:48:14 *** Progman has joined #openttd
08:48:36 *** sla_ro|master has joined #openttd
09:24:59 *** Xaroth728 has joined #openttd
09:26:12 *** ist5shreawf[m] has quit IRC
09:27:39 *** arikover has joined #openttd
09:27:39 *** HerzogDeXtEr has joined #openttd
09:27:39 *** Thedarkb1-X40 has joined #openttd
09:27:39 *** FLHerne has joined #openttd
09:27:39 *** johanna[m] has joined #openttd
09:27:39 *** fjodor[m] has joined #openttd
09:27:39 *** patricia[m] has joined #openttd
09:27:39 *** grossing has joined #openttd
09:27:39 *** dihedral has joined #openttd
09:27:39 *** backtu[m] has joined #openttd
09:27:39 *** nolep[m] has joined #openttd
09:27:39 *** dude[m] has joined #openttd
09:27:39 *** ircer[m] has joined #openttd
09:27:39 *** iarp[m] has joined #openttd
09:27:39 *** tops[m] has joined #openttd
09:27:39 *** ist5shreawf[m] has joined #openttd
09:27:39 *** khavik[m] has joined #openttd
09:27:39 *** ZirconiumX has joined #openttd
09:27:39 *** KenjiE20 has joined #openttd
09:27:39 *** goodger has joined #openttd
09:36:24 *** Kitrana has joined #openttd
09:37:44 *** ist5shreawf[m] has quit IRC
09:38:50 *** Thedarkb1-X40 has joined #openttd
09:38:50 *** FLHerne has joined #openttd
09:38:50 *** HerzogDeXtEr1 has joined #openttd
09:38:50 *** arikover has joined #openttd
09:38:50 *** HerzogDeXtEr has joined #openttd
09:38:50 *** johanna[m] has joined #openttd
09:38:50 *** fjodor[m] has joined #openttd
09:38:50 *** patricia[m] has joined #openttd
09:38:50 *** grossing has joined #openttd
09:38:50 *** dihedral has joined #openttd
09:38:50 *** backtu[m] has joined #openttd
09:38:50 *** nolep[m] has joined #openttd
09:38:50 *** dude[m] has joined #openttd
09:38:50 *** ircer[m] has joined #openttd
09:38:50 *** iarp[m] has joined #openttd
09:38:50 *** tops[m] has joined #openttd
09:38:50 *** ist5shreawf[m] has joined #openttd
09:38:50 *** khavik[m] has joined #openttd
09:38:50 *** ZirconiumX has joined #openttd
09:38:50 *** goodger has joined #openttd
09:45:27 *** ist5shreawf[m] has quit IRC
09:47:16 *** Thedarkb1-X40 has joined #openttd
09:47:16 *** FLHerne has joined #openttd
09:47:16 *** HerzogDeXtEr1 has joined #openttd
09:47:16 *** arikover has joined #openttd
09:47:16 *** johanna[m] has joined #openttd
09:47:16 *** fjodor[m] has joined #openttd
09:47:16 *** patricia[m] has joined #openttd
09:47:16 *** grossing has joined #openttd
09:47:16 *** dihedral has joined #openttd
09:47:16 *** backtu[m] has joined #openttd
09:47:16 *** nolep[m] has joined #openttd
09:47:16 *** dude[m] has joined #openttd
09:47:16 *** ircer[m] has joined #openttd
09:47:16 *** iarp[m] has joined #openttd
09:47:16 *** tops[m] has joined #openttd
09:47:16 *** ist5shreawf[m] has joined #openttd
09:47:16 *** khavik[m] has joined #openttd
09:47:16 *** ZirconiumX has joined #openttd
09:47:16 *** goodger has joined #openttd
09:53:13 *** fjodor[m] has joined #openttd
09:54:18 *** ircer[m] has joined #openttd
09:56:18 *** HerzogDeXtEr has joined #openttd
09:57:52 *** KenjiE20 has joined #openttd
09:58:08 *** FLHerne is now known as Guest5517
10:17:18 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #7592: Road vehicles don't balance between multiple loading bays https://git.io/fjlL0
10:20:57 <DorpsGek_III> [OpenTTD/OpenTTD] Eddi-z commented on issue #7592: Road vehicles don't balance between multiple loading bays https://git.io/fjlL0
10:21:56 <DorpsGek_III> [OpenTTD/nml] nielsmh commented on pull request #53: Fix #52: time.clock() has been removed in python 3.8 https://git.io/JeBDS
10:30:41 *** andythenorth has joined #openttd
10:43:57 <andythenorth> print(format((time() - start), '.2f') + 's')
10:44:19 <andythenorth> nml isn't doing anything clever with the time, I just ran it and checked
10:45:28 <andythenorth> how precise does it need to be? :P
10:46:47 <andythenorth> in fact, time() is more accurate for my purpose, because I want to know how long I'm waiting in total
11:01:21 <nielsm> datetime.time() you mean? problem with that is system clock adjustments
11:01:29 <nielsm> it's not really suitable for stopwatch use
11:01:56 <nielsm> time.monotonic() is a reliable real time stopwatch
11:04:25 <andythenorth> based on principle of not changing behaviour, I'm inclined to improve process_time
11:05:12 <andythenorth> assuming this was originally used to benchmark process timings
11:05:13 <nielsm> yeah it doesn't really matter for this use
11:05:34 <DorpsGek_III> [OpenTTD/nml] andythenorth merged pull request #53: Fix #52: time.clock() has been removed in python 3.8 https://git.io/JeBii
11:07:27 *** Wormnest__ has joined #openttd
11:21:51 *** gelignite has joined #openttd
11:48:23 *** andythenorth_ has joined #openttd
11:50:06 *** frosch123 has joined #openttd
11:50:18 <andythenorth_> “OpenTTD directories” could move from README.md to own file in /docs?
11:58:25 <frosch123> readme section 4.2 is the most referenced part of all
11:58:36 <frosch123> everyone downloading mods manually asks about that
11:59:05 <frosch123> it may even be worth to fake the numbering so that 4.2 remains :p
11:59:29 <Eddi|zuHause> andythenorth_: very -1
12:02:00 <andythenorth_> Noted frosch123
12:02:21 <andythenorth_> Seems bizarrely upside down to me
12:02:45 <andythenorth_> but I am not the audience
12:05:15 <andythenorth_> Begs the question, why are they manually downloading mods? :D
12:06:27 *** andythenorth_ has joined #openttd
12:07:45 <Eddi|zuHause> ... because still some people refuse the idea of bananas?
12:07:58 <Eddi|zuHause> ... both on the uploader and the downloader side
12:09:12 <nielsm> we could add an "install mod" feature in the UI where you can select a file from whereever and have it auto-copied to the appropriate location
12:09:28 <planetmaker> -1 on moving dir info out of readme. That's one of the readme purposes
12:09:31 <nielsm> (that's a lot of work)
12:09:57 <andythenorth_> probably not worth it just to make readme shorter :)
12:09:57 <Eddi|zuHause> nielsm: would probably also require some sort of unpacker
12:09:58 <nielsm> could even define something like .ottdmod file extension
12:10:02 <planetmaker> what will be gained by putting it in a separate file. "easier to find" is not one of them.
12:10:06 <planetmaker> readme is visible ingame
12:10:14 <planetmaker> and google doesn't care
12:10:29 <planetmaker> anyone who looks at doc dir... can look at any file. Fewer is easier to grep
12:11:54 <andythenorth_> Hmm iOS really doesn’t work for anything other than using the phone
12:12:26 <nielsm> where do you find the readme file in-game?
12:14:26 *** andythenorth_ has joined #openttd
12:14:40 <andythenorth_> Frigging ios kicked me again
12:14:58 <planetmaker> he, indeed... it is *not*... could have sworn it were, as we show every readme, except our own
12:15:08 <andythenorth_> I only have this crap because all my kids photos are on Apple
12:15:52 <Eddi|zuHause> nielsm: there should be a readme button on the intro menu
12:16:43 *** andythenorth_ has joined #openttd
12:16:48 <planetmaker> but there actually is not. But there is a completely bogus highscore table
12:18:12 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison opened issue #7783: SDL2: Not possible to type shift-T into a edit context in multiplayer https://git.io/JeByK
12:18:15 <nielsm> yeah a "help" button might be good
12:18:39 <nielsm> getting a readme viewer, links to the wiki, and stuff
12:18:47 <andythenorth_> That would add an interesting dimension to the docs rewrite
12:19:01 <andythenorth_> One more context to consider :)
12:19:42 <andythenorth_> I keep talking myself out of writing a basic manual for players
12:19:58 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison opened issue #7784: SDL2: Not possible to use the up/down keys in the console https://git.io/JeBy6
12:20:04 <andythenorth_> Quick start guide :p
12:20:36 <andythenorth_> In .md, rendered to $whatever
12:20:49 <nielsm> also the difference between "settings" and "game options" is still weird
12:21:30 <Eddi|zuHause> nielsm: but it would also be wrong to mix them
12:21:36 <nielsm> the newgrf and ai/gs configuration windows could maybe have a button to open the appropriate folder in the OS file browser (if the OS supports it)
12:22:38 <andythenorth_> Most of options should be gone
12:22:49 <andythenorth_> Split to settings or newgame
12:23:12 <andythenorth_> Town names, drive side etc are newgame concerns
12:24:04 <nielsm> baseset choices should be labeled "configuration" rather than "options" imo, and should ideally also allow selecting driver via in-game UI
12:24:05 <Eddi|zuHause> andythenorth_: but things like resolution and language should not be hidden deep in some giant settings tree
12:24:14 <nielsm> (of course changing driver still requires restarting the game)
12:24:24 <frosch123> andythenorth_: there was some german dude who wrote a manual and published it by print-on-demand, he sold 3 books or so
12:24:35 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison opened issue #7785: SDL2: Version 2.0.5 is required, but configure script only checks for 2.0 https://git.io/JeBy1
12:24:44 <andythenorth_> Eddi I’m inclined to +1 on those
12:25:26 <frosch123> resolution is such a weird setting. you only need fullscreen/windowed
12:25:34 <frosch123> what is resolution useful for?
12:25:49 <andythenorth_> My OS has a button for fullscreen
12:26:14 <frosch123> well, i also only use maximized, never fullscreen
12:26:36 <andythenorth_> Fullscreen tends to crash macs
12:27:03 <andythenorth_> $everything tends to crash macs these days
12:27:34 <andythenorth_> Terminal decline of the platform, just awaiting a usable alternative
12:27:44 <andythenorth_> Is beOS finished yet.
12:30:47 <andythenorth_> In fact Apple has been dying for the last 20 years, wish it would get it over with :p
12:32:12 <frosch123> i thought beos/haiku was atari stuff?
12:32:15 <andythenorth_> I didn’t own one until 1998
12:32:45 <andythenorth_> At the time the mac magazines were including beOS on cover CDs
12:32:53 <frosch123> i think i decided before 1998 that all mac users are weird, and i do not want to belong into that social group
12:33:00 <andythenorth_> Because Apple was dead
12:35:38 *** andythenorth_ has joined #openttd
12:37:14 *** andythenorth_ has joined #openttd
12:37:36 <andythenorth_> I guess my main observation
12:38:36 <andythenorth_> Looking at other projects, they don’t try and cover everything in README whilst also omitting some fundamentals
12:39:01 <andythenorth_> And also strewing random docs across multiple locations
12:39:44 <andythenorth_> The readme is bad because it attempts to patch over that a docs approach was never figured out happily
12:40:02 <andythenorth_> and because stuff and things
12:43:19 <nielsm> can TT-Forums be called "official forums"?
12:43:41 <nielsm> and should it be linked from inside the game?
12:45:07 <frosch123> it is linked from the homepage
12:45:25 <frosch123> how could it be more official?
13:10:33 <Guest5517> frosch123: Could have a separate file, and change section 4.2 to "read this: <link>"?
13:10:41 *** Guest5517 is now known as FLHerne
13:10:46 *** andythenorth has joined #openttd
13:15:00 *** Wormnest__ has joined #openttd
13:17:25 *** Lejving has joined #openttd
13:17:25 <nielsm> something like that maybe?
13:19:38 <planetmaker> <nielsm> also the difference between "settings" and "game options" is still weird <--- I am of the opinion for long that it should be consolidated into one - at least on the surface
13:20:18 <Eddi|zuHause> planetmaker: i'm of the very opposite opinion
13:20:35 <Eddi|zuHause> internally it could be merged, but on the UI side, separation is better
13:20:45 <Eddi|zuHause> the settings tree gets borderline unusable
13:21:10 <Eddi|zuHause> for a new user it's just massively intimidating
13:21:13 <planetmaker> No. If you want to change stuff, it's good to have one place to go.
13:21:32 <planetmaker> Of course the UI settings have a separate tab or so, so they are prominently visible
13:21:41 <Eddi|zuHause> yes, but that implies you already know it's a setting at all
13:21:44 <nielsm> I agree that the settings tree is bad usability
13:21:48 <Eddi|zuHause> and you know how to look for it
13:21:52 <andythenorth> FWIW I was of the impression that forums were no longer official
13:22:07 <andythenorth> and I think it's better that way
13:22:34 <planetmaker> Eddi|zuHause, if you have like 'load game, settings, and help' in the main menu: where do you go and change resolution? Anyone would go the right place
13:22:48 <planetmaker> as opposed to now where you also have options, and newgrfs, and ai, and scripts...
13:23:33 <Eddi|zuHause> planetmaker: a person who is looking for resolution won't klick on AI settings
13:24:20 <Eddi|zuHause> planetmaker: a person who doesn't know AI settings are even a thing wouldn't click on settings and scroll through 1000s of unrelated settings to find the ai settings
13:24:23 <planetmaker> but a graphics setting
13:24:40 <planetmaker> you however totally miss the point :)
13:24:58 <planetmaker> of course not ai. But there's at least options, and settings. And what's the other stuff?
13:25:00 <Eddi|zuHause> no, i don't, i just think your point is the wrong point
13:25:20 <planetmaker> Most games have like settings -> audio, graphics / video, controls
13:25:30 <planetmaker> and maybe then also mods, and other game settings
13:25:49 <planetmaker> none have the detail level we offer usually :P
13:26:05 <Eddi|zuHause> yes, and someone decided that tabs for the various different setting categories are bad, and introduced the tree
13:26:28 <planetmaker> IMHO this is mixing different discussions:
13:26:29 <Eddi|zuHause> ... and i think shoving more and more into that tree is the wrong idea
13:26:39 <planetmaker> * consolidating all settings to one place from which to access them
13:26:46 <planetmaker> * how to present the settings
13:27:11 <Eddi|zuHause> those are not different points
13:27:19 <planetmaker> one place trumps IMHO that a tree is possibly not the best. But having totally different concepts is worse than one concept
13:27:22 <Eddi|zuHause> just different viewing directions on the same point
13:28:03 <Eddi|zuHause> there's settings with a gameplay effect -> settings, there's settings without a gameplay effect -> game options
13:28:04 <andythenorth> by a quick count, there are over 100 settings in settings
13:28:16 <andythenorth> adding 11 more is marginal
13:28:26 <andythenorth> whereas having 2 places is significant
13:28:34 <planetmaker> Eddi|zuHause, not quite. It's not implemented such at all currently. And options and settings... is rather synonymous
13:28:49 <planetmaker> in our case options is w/o gameplay effect. Except drive side
13:28:57 <planetmaker> so it's... weired however one puts it
13:29:01 <andythenorth> 'Game options' is broken anyway
13:29:50 <andythenorth> load/9521/game_options.png
13:30:08 <andythenorth> doesn't fit, can't be resized
13:30:09 <frosch123> andythenorth: the *nice* thing about the community sites is that they are so political :)
13:30:17 <Eddi|zuHause> ok, let's put it this way (naming aside) game options -> things that are likely locale/os-specific, settings -> things that are platform-independent
13:30:19 <planetmaker> oh, and town names :D
13:30:37 <andythenorth> I would consolidate the options, then work on the tree being more usable
13:30:44 <andythenorth> but I thought we decided that years ago :)
13:30:45 <planetmaker> Eddi|zuHause, that *would* be one way one could separate it, yes. But that clearly is not the case
13:31:06 <frosch123> about settings: imho the tree should contain *all* settings, and then there should be a second gui "basic settings" duplicating stuff like "language"
13:31:07 <planetmaker> and the distinction is not self-explanatory either
13:31:11 <Eddi|zuHause> i never claimed that
13:31:19 <andythenorth> frosch123: there's a lot less tt-forums whining about how awful OpenTTD developers are since all developers quit
13:31:28 <Eddi|zuHause> i just said that joining them and mushing all together is a bad idea
13:31:30 <frosch123> actually, not all settings, all except mapgen
13:31:49 <andythenorth> mapgen is a separate project
13:32:29 <planetmaker> UI settings; mapgen and new game settings; ingame(?) settings
13:32:34 <planetmaker> what other is there?
13:32:41 * andythenorth sticks to the line that OpenTTD does not have official forums
13:32:44 <frosch123> some would be offended if reddit was linked as general info site :p
13:32:54 <andythenorth> Github, #OpenTTD
13:32:58 <andythenorth> end of officialness
13:33:31 <andythenorth> I will say though, reddit openttd is cute
13:33:33 <frosch123> planetmaker: ai/gs/grf belong to mapgen?
13:33:35 <andythenorth> I am considering joining
13:33:43 <planetmaker> @frosch123, yes, IMHO
13:33:56 <andythenorth> meh, I am confused about README
13:34:10 <andythenorth> TB suggestion was to rewrite it for casual github visitors
13:34:30 <andythenorth> but it strikes me that a casual GH visitor is in entirely the wrong place
13:34:44 <andythenorth> unless we think GH is where the website should be (totally possible)
13:35:06 <planetmaker> @andythenorth, what is the actual aim of a re-write? what flaws need to be addressed?
13:35:33 <frosch123> anyway, i agree with nielsm. when i look at the two modern games i play regulary, the main gui has maybe 2 buttons to start/load a game, and like 6 buttons to various websites
13:35:40 <planetmaker> if we want (more) documentation on GH (yes, possible), it probably should be outside the readme w/o reducing it much
13:35:52 <frosch123> ottd intro gui is very different
13:36:18 <andythenorth> TB: some more serious thought, what would really help for OpenTTD, I think, is if the README.md in the root folder was meant for people visiting
13:36:19 <andythenorth> and in docs/ or something we can put one which is meant for bundling
13:36:31 <andythenorth> then there are lots of other TB thoughts
13:36:47 <andythenorth> my actual goal isn't rewriting README
13:37:03 <andythenorth> it's cleaning up proper domain for our different types of info
13:39:23 <andythenorth> eddi's point that the tree is confusing stands though
13:40:01 <planetmaker> I don't disagree there completely. Yet in lack of an actual improvement proposal... I'm not sure how that brings forward a discussion
13:40:54 <planetmaker> The underlaying problem is that we have about a zillion settings - more than one can reasonably present in many easy-to-overlook ways
13:41:22 <frosch123> the tree is the only viable solution for the advanced settings. there is nothing wrong with a second gui duplicating the basic settings
13:41:36 <frosch123> just that everyone will disagree what settings would be the basic ones :p
13:41:43 <planetmaker> it's so many you *have* to use some hirarchical way to present them - unless you want to completely remove them from UI which is not sensible for many either. We have the basic/adv/expert for a reason
13:42:13 <planetmaker> as usual you nailed it
13:42:26 <planetmaker> we should hire you as game designer
13:43:12 <planetmaker> (just to be sure: there was no irony here)
13:44:05 <andythenorth> if only we could have intrusive player monitoring
13:44:11 <andythenorth> then we could use statistics :P
13:44:25 <andythenorth> consider that irony :)
13:46:06 <andythenorth> hmm, why did I hate the wiki so much?
13:46:39 <LordAro> out of date, for the most part
13:47:12 <andythenorth> LordAro: oh yes that
13:47:23 <andythenorth> I think for gameplay the wiki is super
13:47:33 <andythenorth> it's only the development section that makes me twitch
13:49:27 <planetmaker> hm, looks like an interesting tool :)
13:50:00 <LordAro> oh, the development section is even more out of date than the rest of it
13:50:38 <LordAro> though it concerns me that you brought it up right after the wiki...
13:51:18 <andythenorth> I am sorta kinda hoping to move nml docs to md, preferably automated
13:51:28 <planetmaker> hehe, I didn't want to mention it, LordAro :)
13:51:34 <andythenorth> whether we should have md development docs for OpenTTD is a different question
13:51:40 <andythenorth> I have no strong opinion on it
13:51:44 <planetmaker> andythenorth, you mean... the one from wiki?
13:52:27 <planetmaker> what is the benefit? would it be more current?
13:52:42 <andythenorth> it would be helpful when submitting PRs
13:52:46 <andythenorth> they would travel with docs
13:52:59 <andythenorth> i.e. docs would be required as part of PR
13:53:13 <LordAro> non-automated docs have just as much chance of becoming outdated as anything else
13:53:20 <andythenorth> recently nml PR have fallen into a hole where docs don't exist
13:53:28 <andythenorth> so nobody can write test cases
13:53:39 <andythenorth> but because there are no test cases, the PR can't be approved
13:53:39 <nielsm> though markdown doesn't look great this way
13:53:42 <planetmaker> as do obviously the grfspec docs of OpenTTD itself
13:53:48 <planetmaker> re roadtypes or so
13:53:57 <andythenorth> and because the PR is not approved, the docs can't be written
13:54:02 <andythenorth> it's quite silly
13:54:13 <andythenorth> docs with the project is entirely conventiional
13:54:39 <andythenorth> 16-cargo industry, and NRT are the recent examples
13:54:46 <andythenorth> but NML development is just basically unpleasant
13:54:50 <andythenorth> and nobody wants to do it
13:55:02 <planetmaker> point is, currently the NML docs are somewhat in sync with the grfspec, share a lot of info... moving all that into the repo... yes, fine
13:55:21 <andythenorth> nfo docs would have to stay in wiki
13:55:30 <andythenorth> nml isn't nfo :)
13:55:52 <planetmaker> why would they stay in wiki? grfspecs of OpenTTD are documented in exactly the same manner: only if s/o feels like
13:56:03 <planetmaker> it's 100% identical problem
13:56:13 * andythenorth is now convinced
13:56:34 <planetmaker> (I don't argue you're wrong with moving specs / docs to repo) Just... asking questions :P
13:56:55 <andythenorth> should the (nfo) newgrf spec go with grfcodec, or openttd?
13:57:02 <andythenorth> strikes me it's an API in OpeNTTD
13:57:09 <planetmaker> openttd, it's its api
13:57:22 <andythenorth> this historical thing where newgrf belonged to no-one because TTDP
13:57:35 <FLHerne> andythenorth: Ah, I'd not noticed that railtypes were broken because none of the grfs I was testing with used them
13:57:45 <andythenorth> I haven't bisected FLHerne
13:57:51 <andythenorth> so I don't know if it *ever* worked
13:58:01 <FLHerne> andythenorth: I suspect it probably didn't :-/
13:58:01 <planetmaker> railtypes... broken?
13:58:37 <andythenorth> compiling railtype (or roadtype) table with multiple labels in a list, using --nml=foo.nml
13:58:48 <FLHerne> planetmaker: ...if you run `nmlc --nml=out.nml` on code that uses railtype compatibility lists
13:58:49 <andythenorth> FLHerne: filing an issue? :)
13:59:07 <FLHerne> Trying to just fix it :P
13:59:52 <andythenorth> nielsm: do you think it can url parse? :)
14:00:11 <nielsm> it'd be a lot of work to make the file viewer do markdown well
14:00:24 <andythenorth> but we can parse md to other formats...
14:00:25 <nielsm> it also has another issue: files longer than 65535 lines
14:00:30 <FLHerne> It's not blindingly obvious how to, because of the usual arbitrary-format-mutation during compilation...
14:00:40 *** Kitrana1 has joined #openttd
14:00:45 <FLHerne> I do find that aspect of nml really annoying
14:00:49 <planetmaker> nielsm, is that a real issue or a pathological edge case?
14:01:35 <nielsm> the scrollbar widget limits itself to UINT16_MAX positions
14:01:56 <nielsm> widget_type.h line 673
14:01:58 <LordAro> probably not terribly difficult to change the type
14:02:03 <frosch123> it's completely fine to open the docs in a browser
14:02:30 <FLHerne> Most of the compilation steps /overwrite/ the originally parsed AST in prop variables with odd bits of intermediate state
14:02:48 <nielsm> frosch123: then we just need to make it more browser friendly than plain text is
14:02:57 <LordAro> one of the reasons the game docs weren't added to the viewer originally was that it's quite hard to work out their location when installed
14:03:09 <LordAro> they might even be gzipped
14:03:48 <frosch123> yep, linking externally is a lot easier as long as you do not move the location afterwards :p
14:04:16 <LordAro> slightly modified page on the wiki would be fine, imo
14:04:26 <LordAro> just add "update the wiki page" to the release checklist
14:04:49 <frosch123> for example: the grfcrawler search in ottd does not link diretctly to grfcrawler. there is a redirect page on openttd.org
14:05:46 <andythenorth> I have been thinking a lot about the website content
14:06:01 <andythenorth> my inclination is to keep it really simple, and for almost everything, signpost elsewhere
14:06:19 <andythenorth> we mostly do that already
14:06:47 <FLHerne> What's a unified word for {road, rail} ?
14:06:49 <frosch123> i suggest to delete the about page, and delete everything but info@ and translator@ from contact
14:06:55 <FLHerne> Er, roadtype, railtype
14:07:12 <frosch123> FLHerne: ottd calls trains/road vehicles ground vehicles
14:07:21 <andythenorth> I think 'about' should be replaced by a 'get started'
14:07:26 <andythenorth> which should probably be directly on front page
14:08:58 <andythenorth> so I am wondering about GH root having a DEVELOPMENT.md
14:09:13 <andythenorth> with signposts to all the various kinds of docs
14:09:21 <andythenorth> or whether that can go in CONTRIBUTING without bloating it
14:35:34 <Wolf01> So, humble bundle monthly going to change (but not for current subscribers), I wanted to switch to yearly payment in january but I did it now to be sure to get the 1 free month even after the change
14:40:01 *** Kitrana has joined #openttd
14:59:55 *** Wormnest__ has joined #openttd
15:40:58 <andythenorth> hmm, found some wiki to tidy :P
15:53:13 <planetmaker> nielsm, *Changelog; and... license *agreement*? Maybe just "License"?
15:54:22 <planetmaker> and tt-forums is sufficiently official. It goes by forum.openttd.org as well
15:55:03 <planetmaker> (thus cannot get more official :) )
15:55:40 <planetmaker> But if we want to put a link, maybe we link to forum.openttd.org instead of tt-forums.net
15:56:03 <andythenorth> I do think that some of the historical conflict in tt-forums was caused by the OpenTTD takeover
15:56:12 <andythenorth> I think some people felt driven out
15:56:21 <andythenorth> water under the bridge anyway
15:56:25 <andythenorth> now it's JGRPP-forums
15:56:33 <planetmaker> tt-forum is bigger than forum.openttd.org
15:57:00 <planetmaker> so I actually see there little point to argue against it...
15:57:00 <glx> yeah forum.openttd.org only link to the subforum IIRC
15:57:17 <planetmaker> but that's fine for our purpose at hand, I think
15:57:28 <andythenorth> eh we could use forum.openttd.org as the link?
15:57:35 <andythenorth> quite interesting
15:57:37 <planetmaker> that's what I argue :)
15:57:56 <andythenorth> I'll amend my draft of the readme
15:57:59 <planetmaker> we could do so for like... forever
15:58:41 <planetmaker> probably at least since the last restructuring of forums
15:59:33 <andythenorth> one day we do something about Development forum :P
16:00:37 <planetmaker> which development forum?
16:00:58 <andythenorth> OpenTTD one. We could maybe have a better sticky pointing to Github
16:01:02 <andythenorth> there is one, but it's not obvious
16:01:25 <planetmaker> was not sure you meant newgrf or openttd :)
16:01:32 <andythenorth> A single sticky like "OpenTTD development takes place on GitHub" might be better
16:02:16 <planetmaker> Well. The "wanted contributions / patches" links exactly there: to GitHub
16:02:43 <planetmaker> and the coding style... is still relevant, too. But maybe can be put in the other
16:02:43 <andythenorth> yes, just think the title could be bettter
16:02:59 <planetmaker> ok, what should it be? "Contributing" ?
16:03:56 <andythenorth> I'd be really explicit and just mention GitHub
16:04:04 <andythenorth> "Want to contribute? Visit Github"
16:04:12 <andythenorth> or "Guide to developing for OpenTTD"
16:06:45 <andythenorth> oof "Pandoc has some issues that affect almost every article that it tries to parse:"
16:09:07 <planetmaker> and I will unsticky the coding style one
16:10:47 <andythenorth> pretty good imho
16:22:02 *** Wormnest has joined #openttd
16:25:31 *** WormnestAndroid has quit IRC
16:25:42 *** WormnestAndroid has joined #openttd
16:26:01 <andythenorth> nielsm: does this change the goal for the readme?
16:26:17 <andythenorth> TB's suggestion was to aim it at GH visitors, but in-game is quite different :)
16:26:58 * andythenorth checks out the PR
16:26:59 <nielsm> but maybe yes, since being able to view it in-game means you already have the game installed and running
16:27:00 <planetmaker> I don't think that readme should aim at like "github visitors", "ingame audience" or whomever. It's more generic than one platform IMHO
16:29:02 <planetmaker> However I don't think it's really that much diverging or conflicting goals...
16:29:51 <andythenorth> I think the fact that it still needs to work for people offline is quite a constraint
16:30:14 <andythenorth> it's not really possible to write it like some GitHub projects do, which is a list of links with short descriptions
16:30:24 <andythenorth> it's more like a traditional readme.txt
16:30:41 <andythenorth> but readme.txt is possibly a very bad introduction to our GH project
16:30:58 <planetmaker> So... README.md for github :)
16:31:01 * andythenorth still thinking about the problem
16:31:26 <planetmaker> btw, I don't think that "viewing ingame" is really like "offline"
16:31:48 <andythenorth> what about CD-ROM distribution?
16:31:50 <planetmaker> I really doubt that the majority of audience has no internet connetion and browser
16:32:03 <planetmaker> you mean floppy disc 5 1/4"?
16:32:20 <planetmaker> they get the XP-version :P
16:33:02 <andythenorth> nielsm: that is one of those features
16:33:13 <andythenorth> where when you see it
16:33:21 <andythenorth> you think "why wasn't this done before?"
16:34:02 <andythenorth> we could do with an abstraction for urls :P
16:34:19 <andythenorth> hmm actually maybe not, probably more hassle than benefit
16:41:53 *** WormnestAndroid has quit IRC
16:42:06 *** WormnestAndroid has joined #openttd
16:59:58 <nielsm> should perhaps add a ToC pane to the textfile viewer to support really long files?
17:00:28 <nielsm> and use a very basic markdown-like parser to guess at the headlines
17:00:33 <andythenorth> I would sooner delegate to websites :)
17:01:23 <Eddi|zuHause> nielsm: just try parsing as markdown, and if it doesn't work, give up and draw plain text?
17:01:37 <nielsm> Eddi|zuHause: I don't mean rendering markdown
17:01:53 <nielsm> I just mean extracting headlines and showing them in a jumplist
17:02:27 <Eddi|zuHause> you can still draw things as monospace font, even after parsing markdown stuff
17:04:51 <FLHerne> (doesn't actually fix the problem yet, but at least means it only needs fixing once...)
17:08:46 *** Wormnest has joined #openttd
17:11:33 <Eddi|zuHause> deduplicating code is always good... (haven't looked at any details, or what the actual problem is)
17:13:55 <andythenorth> idea looks sound
17:14:03 <andythenorth> I haven't read every line
17:17:11 <FLHerne> Are `feat8_prop` and `not_defined_cond` the right names?
17:17:37 <FLHerne> Hopefully I read the docs right, but they replace unlabeled hex constants so I'm not 100% sure :P
17:18:47 <andythenorth> does it work? :)
17:20:27 <andythenorth> FLHerne: the chances of getting a detailed nml review from someone who actually understands it are low :)
17:20:36 <andythenorth> if the tests pass, and all relevant examples build
17:20:42 <andythenorth> and if you've been really careful
17:20:49 <andythenorth> it's 'probably fine'
17:21:03 <Eddi|zuHause> FLHerne: what's a "not_defined_cond"?
17:21:18 <andythenorth> the only way nml quality is going to improve is if some people start poking it, with the risk of breaking things
17:21:27 <andythenorth> but if it breaks in prod, we can just revert and release again
17:25:09 <Eddi|zuHause> FLHerne: then maybe "cond_tracktype_not_defined"?
17:26:00 <FLHerne> Right, that seems more readable
17:27:08 <Eddi|zuHause> FLHerne: is there precedence for the other condition type?
17:28:15 <FLHerne> Eddi|zuHause: They're there already, in the duplicated versions
17:28:23 <FLHerne> Eddi|zuHause: (added by NRT)
17:28:44 <andythenorth> the NRT support was copy-paste cargo cult
17:28:46 <FLHerne> I only have to give it a name because previously each version has its own magic number
17:29:09 <FLHerne> You didn't need to tell me that :P
18:05:04 <andythenorth> setup.py works for me on mac also
18:05:55 <LordAro> oh no, now i'm an NML contributor
18:06:41 <andythenorth> is there a final set of instructions I can paste in?
18:07:17 <andythenorth> then you can get it to publish to bundles :P
18:07:56 <glx> probably needs some secrets to be set in the project settings
18:08:17 <andythenorth> obvs. can't test for myself :P
18:08:23 <Wolf01> andythenorth: got the liebherr. Curses and swearing various gods to apply the discount with the new system. :P
18:08:38 <andythenorth> let me know how you find it
18:08:48 <andythenorth> I think my Lego days might be over
18:09:28 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on issue #42: Document how to release nmlc, including building a windows binary https://git.io/Je8vQ
18:09:55 <glx> it's even simpler now, it's just "python setup.py bdist"
18:10:10 <glx> and you get a .zip in dist
18:10:25 <LordAro> glx: should make a requirements.txt file, will simplify the dependency install
18:11:01 <glx> requirements depend on platform
18:11:05 <Wolf01> <andythenorth> let me know how you find it <- I don't think I'll build it for the next 5 years, I still have to build the bucket wheel and the rough terrain crane XD
18:11:14 <glx> cx-freeze is needed only for windows
18:11:24 <andythenorth> I built the B model for the bucket wheel, not buying any more big lego
18:11:31 <andythenorth> only bought that because the kids really wanted it
18:12:11 <LordAro> we wanted to avoid squashing #7781, right?
18:13:42 <andythenorth> happy to, if it's worth it
18:14:17 <andythenorth> just has a lot of weird semi-colons and stuff, so I didn't
18:14:40 <andythenorth> my assumption was that's the file format
18:15:01 <LordAro> it's an example file, not a description
18:15:22 <LordAro> obs_example.ini might be a better name
18:15:52 <andythenorth> give them a dir? or keep docs flat?
18:16:03 <andythenorth> I would like to move some wiki docs into GH /docs
18:16:07 <andythenorth> not all, but a few
18:16:37 <Eddi|zuHause> LordAro: i don't think it should be called ".ini" when the resulting file doesn't end in ".ini" either
18:16:48 <andythenorth> it's quite shouty
18:16:48 <LordAro> Eddi|zuHause: ...that would make sense, yes
18:17:45 <Eddi|zuHause> andythenorth: -1
18:18:00 <Eddi|zuHause> filename should be descriptive
18:18:27 <andythenorth> and if I consolidate other legacy lang file info into there?
18:18:35 * andythenorth wonders if there's any point to that
18:20:08 <frosch123> it has the most up-to-date docs about lang files
18:20:12 <Eddi|zuHause> andythenorth: so you want to move out compile stuff from the readme, to then merge other stuff into the compile stuff?
18:20:19 <Eddi|zuHause> that doesn't make much sense
18:20:32 <andythenorth> it's about the balance between how many docs are in /docs
18:20:35 <andythenorth> and how long each doc is
18:20:40 <andythenorth> and how general or specific
18:21:02 <andythenorth> all I want is it to be more canonical and have a structure
18:21:44 <andythenorth> I think the correct thing is just to drop the HOWTO from the name
18:22:29 *** m1cr0man has joined #openttd
18:28:22 <DorpsGek_III> [OpenTTD/OpenTTD] andythenorth opened pull request #7787: Doc: rename HOWTO_compile_lang_files.md https://git.io/JeBdy
18:28:38 <andythenorth> git web UI is quite nice for docs edits
18:28:46 <andythenorth> suits my taste more than wiki
18:41:55 <LordAro> andythenorth: trouble with GH web UI is that you always end up creating a branch in the repo, not in your fork
18:42:01 <LordAro> which is a bit irritating
18:42:26 <glx> GH web ui works like that
18:43:25 <glx> but we could enable auto delete branch on merge
18:44:20 <glx> but auto delete could be annoying too (for people wanting to keep the branch in their repo)
18:44:55 <glx> even if I don't find a use case for that
18:47:19 <milek7> this needs update probably?
18:51:15 *** m1cr0man has joined #openttd
18:57:30 <andythenorth> LordAro: how about a long-running docs branch? o_O
18:57:34 <andythenorth> dunno if that works
19:04:41 <andythenorth> if that's not lost knowledge :)
19:10:34 <Eddi|zuHause> concerning the settings tree: what about if the "basic" settings are a different more newbie-friendly overwiew, and the tree-view gets enabled for the more advanced settings for easier searching?
19:11:02 <Eddi|zuHause> i believe VLC does something like that
19:12:41 <milek7> why even there is this weird 'options' 'settings' split
19:15:21 <Eddi|zuHause> milek7: we already merged the historical difficulty and "patch settings"
19:15:57 <Eddi|zuHause> milek7: but afair the game options were of a slightly different quality, which didn't fit neatly into the other settings
19:17:46 <glx> it's just a regression workflow but it's a good start
19:18:49 <andythenorth> newgrf builds in 2020 :)
19:19:09 <andythenorth> I have FIRS building on Azure, and it works fine, but Azure is like being punched in the face with a cold sponge
19:22:46 <frosch123> glx: nice, it even built the extension
19:23:23 <andythenorth> nielsm: I am about to branch for converting known_bugs.txt to .md, does that mess up your PR? :P
19:23:31 <frosch123> oh, wait, i was looking at ubuntu
19:24:09 <nielsm> andythenorth: not much, if you format it well it'll mean less work :)
19:24:23 <nielsm> since currently my ToC parser for known-bugs.txt is broken
19:24:27 <frosch123> not quite sure, but i think win does not build the extension module
19:24:58 <andythenorth> the current known_bugs.txt is an unusual indented format
19:25:17 <andythenorth> I think a heading per bug, followed by plain body copy :P
19:25:44 <andythenorth> unusual / unfamiliar /s :P
19:29:50 <andythenorth> nielsm: maybe we should write a chapter based manual :D
19:33:12 <glx> hmm I remember seeing it build the extension locally
19:34:52 <DorpsGek_III> [OpenTTD/OpenTTD] James103 commented on issue #7613: Generating too many news messages at once crashes the game. https://git.io/fj0w2
19:35:30 <glx> oh my bad, I should use build not build_exe
19:35:47 <andythenorth> nielsm: does your viewer care about 80 char line limit?
19:36:06 * andythenorth never got the memo about 80 char, it's a long time since VT220
19:37:17 <nielsm> the textfile viewer can wrap but my jumplist doesn't handle that properly
19:44:13 <andythenorth> do we want me to link the #nnnn items?
19:45:14 <andythenorth> probably should eh
19:45:20 <nielsm> yeah since it's possible
19:45:47 <DorpsGek_III> - Update: Translations from eints (by translators)
20:23:25 <nielsm> hmm... for use with a viewer that supports markdown links properly that would probably be the best formatting yes, but if the viewer (as in my PR currently) doesn't format links then those end up as very long lines
20:23:43 <nielsm> though I'd really prefer to fix the viewer :)
20:25:44 <planetmaker> is there a C++ library for displaying markup? I'd very much presume 'yes' :)
20:26:46 <nielsm> I'm not sure if we can use a straight up display library, at least not without some heavy integration work with the current UI framework
20:27:03 <nielsm> but something to parse markdown to some semi-structured stuff would work
20:28:52 <nielsm> hm, <regex> is C++11, I presume that means it's okay to use?
20:30:45 <frosch123> i would still prefer just opening something in a web browser
20:31:22 <nielsm> then let's add something that translates markdown to html to the build process
20:32:57 <frosch123> why does it need to be a local file?
20:33:51 <nielsm> otherwise the version needs to know how to produce the correct url to its manuals
20:34:09 <andythenorth> why not just have make run a text parser in $scriptlanguage
20:34:22 <nielsm> (if you ignore the question of whether offline machines should be allowed documentation access)
20:34:29 <andythenorth> in newgrf docs, I parse md to multiple formats
20:34:58 <frosch123> andythenorth: yet another build-dep. python and more
20:35:00 <planetmaker> hm, there's a difference between "should" and "is" :)
20:35:14 <planetmaker> and yes... probably right: why not use a browser. Everyone has one
20:35:33 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #7788: Doc: Convert known_bugs.txt to markdown format https://git.io/JeBbN
20:35:35 <frosch123> just yesterday someone ranted at cppcon to not include unnecessariy doc generation in builds
20:35:45 <andythenorth> what if your connection is down?
20:36:07 <planetmaker> if you're outside Germany, that's unlikely :P
20:36:09 * andythenorth isn't very concerned, playing OpenTTD is not exactly critical
20:36:58 <frosch123> if you live offline, you can probably also read unformatted txt
20:37:03 <andythenorth> LordAro: GH links back out again then? :P
20:37:18 <andythenorth> it's really unclear what the best thing to do is for docs
20:37:24 <frosch123> otherwise you would have to consider bundling the wiki tutorial into the downloads
20:37:28 <andythenorth> I lost the markdown argument sometime a while ago :P
20:38:53 <andythenorth> is .md intended to be human readable, or is it source?
20:38:58 <andythenorth> I am a bit unclear
20:39:04 <andythenorth> I always use html for docs if I can :P
20:39:07 <andythenorth> but that was vetoed
20:39:29 <andythenorth> I distrust pseudo-human-readable formats, because they never work
20:44:02 <frosch123> it's rendered as html, what do you want more?
20:44:48 <nielsm> yes when viewed on github
20:45:08 <andythenorth> I am on the fence
20:45:12 <frosch123> well, who do you target? a server admin without gui? a player on a desktop pc?
20:45:21 <andythenorth> this is my question too
20:45:30 <andythenorth> I don't know who the audience is supposed to be :)
20:45:34 <andythenorth> is it Richard Stallman?
20:45:50 <frosch123> with .md you can catch both. the server guy can read readme.md with less or vi, and the desktop player can be linked to online docs
20:46:26 <andythenorth> the server guy has probably scripted emacs to render it all, and reformat it to a specific line length depending on day of week
20:46:47 <frosch123> readme section 4.2 (the most important part of readme) is important for people who want to copy newgrf onto their server, and for people who get alpha releases from forums or bundles.o.o
20:47:08 <frosch123> in both cases you want to link the docs from the forums or similar
20:47:57 <planetmaker> 4.2 is the most-often asked-for section :) By a large margin
20:48:11 <planetmaker> just today I saw like "moving to new computer" in tt-f as background
20:48:36 <frosch123> planetmaker: i told andy earlier, whatever changes in the readme, the numbering for 4.2 must not change :p
20:49:46 <planetmaker> keeping valid all the answers along the line of "rtfm, section 4.2" :P
20:50:55 <andythenorth> that will present an...interesting challenge :P
20:51:00 <andythenorth> given the way the .md renderer works
20:51:16 <planetmaker> Then soonish the docs in §4.2 will need to contain a sentence like "This section needs to stay §4.2 for hysterical raisons as everything links to §4.2"
20:51:20 <andythenorth> I will have to arbitrarily number preceeding sections
20:51:23 <frosch123> planetmaker: btw. i meant to ask: what to do with coop membership tickets? should we disable eints? and offer project people to convert their repo to git?
20:52:28 <planetmaker> frosch123, for hosting the repositories, a github repo probably is best... I started to convert some repos... but never really finished it :D
20:52:43 <DorpsGek_III> [OpenTTD/OpenTTD] BlackPlague666 commented on issue #7767: Large numbers of trains and signals cause low FPS https://git.io/Je8mF
20:52:54 <planetmaker> As to compiling NewGRFs... I wish to actually get that somewhat more automated so less human admin factor is needed
20:52:59 <planetmaker> but... well... yes
20:53:16 <frosch123> planetmaker: i am not asking about the long term plan, i wonder what to do now
20:53:44 <frosch123> there are like 3 open translator tickets, but it's kind of wrong to give them access
20:54:00 <frosch123> people only want to become translators for firs, and firs translations are currently not used
20:54:08 <planetmaker> should it be wrong for translations? Ah
20:54:43 <frosch123> firs is on gh, the hg repo keeps on getting some eints commits, but would require manual merge otherwise
20:54:55 <andythenorth> I think I _might_ have deleted the hg repo
20:55:12 <frosch123> all other projects on devzone are inactive as well, so there is kind of little point in devzone eints anymore
20:55:16 <andythenorth> possibly only from redmine, rather than the actual repo
20:55:23 <andythenorth> but afaict devzone eints is dead?
20:55:23 <planetmaker> right... so can we talk eints into committing to (various) GH repos?
20:55:47 <andythenorth> yes, it's likely quite more complex
20:55:53 <planetmaker> yes, you're right... mostly little point. I saw a few translation changes on a few repos... but very very few
20:56:27 <frosch123> planetmaker: you just need some gh bot user, like dorpsgek
20:56:51 <frosch123> and then people need to configure their gh to give push access to that one
20:57:17 <andythenorth> I doubt that I've written anything that can actually harm us
20:57:50 <andythenorth> I didn't include any malware links, not knowingly anyway
20:57:56 <andythenorth> and there's no javascript coin miner
20:59:36 <planetmaker> frosch123, so... that user needs eints access and repo access... configured on a per-project basis
21:00:04 <nielsm> that's a GH "application"
21:00:39 <nielsm> you'd give the eints instance permission to act on your behalf on the repository
21:00:46 <frosch123> planetmaker: essentially, we can setup it for firs, and assume that noone else will need it
21:01:07 <andythenorth> then I can cargo cult it :P
21:01:22 <andythenorth> presumably, whatever was done for ottd master GH is reusable?
21:01:31 <frosch123> the main thing that does not work is automatic eints setup via redmine roles
21:01:33 <andythenorth> we have translation commits?
21:01:46 <andythenorth> how do trunk translators get auth?
21:01:54 <andythenorth> oh yeah, they mail that random address :)
21:02:26 <frosch123> andythenorth: think smaller :)
21:02:39 <frosch123> we don't need a solution that works for 300 projects
21:02:53 <andythenorth> no, but Iron Horse gets translations
21:03:58 <planetmaker> frosch123, it actually would be nice to have a solution that works for more than just firs
21:04:09 <frosch123> planetmaker: so, suggestion would be, disable all current eints projects, and only readd firs and iron horse
21:04:10 <andythenorth> I just wondered, given that we have 1, why not make 1 n
21:04:26 <planetmaker> frosch123, but ... it would actually be nice, when the openttd translators could opt-in to also translate NewGRFS (and ai, basesets,...)
21:04:39 <andythenorth> reducing the number of things we have
21:04:40 <frosch123> sure, when there are some
21:05:02 <frosch123> i don't see the demand
21:05:15 <andythenorth> I did wonder about just saying 'send PRs' for newgrfs
21:05:34 <planetmaker> frosch123, so what solution would solve it for 2 but not for $configured others?
21:05:43 *** supermop_Home has joined #openttd
21:06:01 <planetmaker> I'd find that really sad
21:06:07 <frosch123> planetmaker: manually creating checkouts on the eints vm and making andy give access for some bot
21:06:10 <andythenorth> frosch123 I could just give specific people PR rights on a branch?
21:06:11 <planetmaker> I did have this year updates for some of my newgrfs
21:06:47 <frosch123> andythenorth: making eints create prs is a lot more work, than just making eints push directly
21:06:56 <andythenorth> oh I was just going to have people do it
21:07:04 <andythenorth> as an alternative low-fi solution
21:09:58 *** tokai|noir has joined #openttd
21:09:58 *** ChanServ sets mode: +v tokai|noir
21:10:39 <planetmaker> frosch123, I'd prefer to have an app which authors can integrate in their repo. Then they can apply for eints and it will be added.
21:10:51 <planetmaker> I know... maybe it's wishful thinking :)
21:11:20 <andythenorth> I have some kind of Azure connection on my account
21:11:25 <andythenorth> and I authorise it per repo
21:11:37 <andythenorth> they have some name, like GH service or something
21:12:06 <andythenorth> Installed GitHub Apps
21:12:20 <frosch123> which of those projects would you consider to point out to translators, who actually want someone to get their translations from bananas somewhen
21:12:41 <andythenorth> ^ nice, but it's engineering
21:12:52 <andythenorth> we have quite a lot of engineering already, and it only just stays alive :P
21:12:55 <frosch123> there is silicon valley on that list, i think i made the last release 5 years ago
21:13:13 <frosch123> the translations since then are only on bundles, so basically noone gets them
21:14:14 <frosch123> i think it's unfair to translators to direct their efforts to unmaintened projects
21:15:36 <planetmaker> well, ok. Let's shut it down
21:16:11 <frosch123> so: clear the list, readd gh stuff manually?
21:17:12 <planetmaker> well, but can eints handle the git repos? I guess it can... openttd uses it after all
21:17:25 <planetmaker> so, yes, changing the repo links to github manually would work
21:17:30 <frosch123> it can already do that on devzone
21:18:16 <frosch123> so, what shall we call the user?
21:18:26 <frosch123> or should we hijack dorpsgek to also work for coop?
21:18:38 <planetmaker> what about... webster?
21:19:16 <andythenorth> some ideas....around bananas 2 :P
21:19:26 <andythenorth> where releases are automatic from translations :P
21:19:39 <andythenorth> I discovered some serious latency in releasing some translations for my grfs
21:20:06 <andythenorth> which I felt bad about :P
21:20:18 <frosch123> andythenorth: you are the only one releasing at all :p
21:20:39 <frosch123> what do you think when was the last release of opensfx and openmsx?
21:20:44 <andythenorth> my stuff gets kicked off the front page of bananas quite fast
21:21:37 <andythenorth> my (no data) impression is that newgrf dev is as healthy as ever
21:21:38 <frosch123> with the exception of nutracks, none of that stuff is on devzone. and the nutrack guy is waiting for a new ssh-token to be installed on devzone
21:21:54 <frosch123> and i would rather offer them migration to gh than update the ssh key
21:22:00 * andythenorth also had a hopeful idea that bundles could die, and be handled in bananas 2 :P
21:22:07 <andythenorth> bundles should not need to exist at all, it's daft
21:22:44 <andythenorth> I mean, it's lovely, and was really useful :)
21:22:50 <planetmaker> ok, andythenorth. I shall stop the migration of stuff to a new server and spare me all that and shut down end-of-year. Got me convinced
21:22:55 <planetmaker> was that the aim?
21:23:09 <frosch123> andythenorth: around ottd 0.6 newgrf were made by artists, from 1.0 to 1.2 stuff moved to coders using vcs, now stuff is back to artists with no clue about vcs
21:23:20 <andythenorth> a few have github
21:23:44 <frosch123> planetmaker: focus on stuff that you would use yourself
21:24:07 <andythenorth> coop was great, when there were more people helping
21:24:16 <andythenorth> now it's just an unfair maintenance burden on 2 people
21:24:59 <frosch123> planetmaker: my probably flawed impression is that 90% of industry sets are forks for firs, but noone uses a vcs
21:25:20 <planetmaker> dunno about industry sets newer than FIRS :P
21:25:48 *** WormnestAndroid has quit IRC
21:25:55 <planetmaker> well, if people nowadays need convincing to use VCS... they are mostly a lost cause
21:26:07 *** WormnestAndroid has joined #openttd
21:26:21 <planetmaker> I'm reasonably shocked time and again by some of my collegues...
21:27:59 <planetmaker> looking through the list of topics on the first page: I concur
21:29:29 <planetmaker> right... ok, so what shall be done: Migrate all repositories to github.
21:29:38 <planetmaker> make devzone read-only
21:30:17 <planetmaker> And create anew from scratch whatever suits me :P
21:30:23 <planetmaker> as my pet toys :P
21:30:37 <planetmaker> we keep eints running for those who like... how do we manage accounts?
21:31:21 <frosch123> 1. make devzone readonly, 2. backup all repositories to gh as closed repositories, 3. offer people to transfer ownership of their repositories to their user, if they want to, 4. setup eints for <5 projects
21:31:40 <planetmaker> what you mean with 'closed repositories'?
21:31:49 <andythenorth> I have no particular goal for coop, but what I really like about it....was that it was truly cooperative
21:32:16 <frosch123> planetmaker: like openttd-deprecated, it has a "archived" thingie next to it
21:32:22 <andythenorth> I seem to recall that there were as many as 10-12 people all working in similar direction on coop projects
21:32:45 <andythenorth> it was a nice time
21:33:02 <planetmaker> yes... there were a lot of people doing stuff
21:33:07 <planetmaker> quite active times
21:33:50 <planetmaker> like 3 people caring for the server who were constantly around. And many more writing NewGRFs and making use of the server
21:34:19 <planetmaker> similar actually on the OpenTTD side... it kinda all shrank by 2/3
21:36:10 <frosch123> planetmaker: you already moved projects to gh/openttdcoop. how did you select those projects?
21:36:19 <frosch123> just so i can figure out which ones still need moving
21:36:49 <planetmaker> some randomly, then started alphabetically from top
21:36:53 <frosch123> oh, also, i moved all projects that i considered "mine" to my account, not to openttdcoop
21:37:03 <andythenorth> mine are all on GH
21:38:25 <frosch123> ok, i'll see whether i can get that done next week
21:39:16 <andythenorth> are issues moving?
21:39:37 <planetmaker> just for curiosity, what are your pet projects, frosch123 ? :)
21:40:27 <frosch123> planetmaker: what do you mean?
21:40:54 <planetmaker> those which you moved... I guess I can just check GH :P
21:41:07 <frosch123> just those were i made 99% of commits
21:42:27 <frosch123> i have no projects where i contributed 50% or something, either "most" or "barely anything"
21:42:30 <planetmaker> backdraw of GH: you don't find openttd-related repos any longer, if you don't know what exactly
21:43:38 <frosch123> i would go to bananas, get the name, then google github+name
21:43:40 <planetmaker> and thus... actually I cannot look-up on GH :P
21:44:05 <frosch123> i do the same for mods of other games
21:45:00 <frosch123> planetmaker: ok, it's hard for "silicon valley", but the rest should work
21:45:51 <frosch123> hmm, i see, google cannot handle CamelCase that well
21:46:07 <planetmaker> really? I don't see that working for OpenTTD
21:46:08 <frosch123> "github debugvehicles" works, "github debug vehicles" does not
21:46:33 <planetmaker> yes, that's what I mean. You only find stuff where you know very exactly what to look for
21:47:32 <frosch123> ah, well, bananas2 will provide source links
21:48:54 <andythenorth> extend action 14 :P
21:49:00 <frosch123> it gets better if you also add "openttd" to the search term
21:49:07 <andythenorth> in the openttd commits, it's interesting to de-overly some people :P
21:49:17 <andythenorth> the huge early spike is significantly Rubi
21:49:22 <andythenorth> frosch is fairly constant
21:49:33 <andythenorth> peter goes through a bunch of intense commits then a break
21:50:20 <frosch123> andythenorth: i figured that contributors are randomly missing from the stats
21:50:30 <andythenorth> did rubi commit a lot of other people's patches also?
21:50:44 <andythenorth> I seem to remember something in early logs and forums :P
21:50:57 <frosch123> yes, rubi did almost everything
21:51:12 <frosch123> took, patches, refactored them and later maintained them
21:51:19 <andythenorth> maintainer burnout :P
21:51:33 <frosch123> he is the only reason why there was a low-mark of 14 open bug reports around 1.0
21:51:53 <andythenorth> anyway, I'm not sure how to express my feelings on coop accurately :P
21:52:14 <andythenorth> coop was great, then things started to break that I can't fix, so it was less great
21:52:32 <andythenorth> but moving everything to non-shared, non-repeatable stuff seems less good
21:52:41 <andythenorth> but it's JFDI and I don't have to wait on other people
21:53:14 <andythenorth> of all the things that kill me in open source, it's the async nature of getting work done :P
21:53:43 <andythenorth> but I miss coop happy times
21:53:57 <frosch123> well, blame nml :p grf development moved back to little-tech-safy people, so vcs , coop and gh workflows are too advanced
21:54:03 *** sla_ro|master has joined #openttd
21:54:18 <andythenorth> I never wanted to adopt nml, I nearly rage quit when FIRS was converted to it
21:54:28 <andythenorth> in retrospect...that was silly :)
21:54:32 <frosch123> it was very different when yexo, ammler and planetmaker were coders for 90% of grfs, herding dozens of artists
21:55:10 <andythenorth> #openttdcoop.devzone used to be busier than #openttd
21:55:20 <glx> next step will be a GUI to generate nml ;)
21:55:27 <andythenorth> then yexo, foobar, terkhen all got jobs etc
21:55:28 <frosch123> oh my... i just remembered ammler writing nfo grfs via shell scripts
21:55:42 <andythenorth> you don't remember CPP variadic macros?
21:55:50 <andythenorth> a thing that should never have been
21:56:05 <andythenorth> glx: a GUI would be full circle
21:56:19 <andythenorth> I built a web-based 'make me a ships grf' tool
21:58:14 <glx> github actions can do that
21:58:16 <frosch123> planetmaker: for what use case? people know how to build grfs locally
21:58:20 <planetmaker> yes... yexo and rubi left gaps behind
21:58:38 <planetmaker> for those who don't :P
21:59:31 <glx> create a workflow file in a newgrf repo, then any one can copy it to their personnal repo
21:59:35 <frosch123> devzone consists of repo, eints, and bundles. repo should go to gh, eints is doable on small scale and sufficient, i just have no idea about bundles
21:59:59 <frosch123> though i guess bundles also works on small scale
22:00:45 <frosch123> glx: wasn't the problem with hosting the artefacts?
22:01:39 <glx> the problem is to pass artefacts between workflow runs
22:02:09 <glx> but in this case they just need to stay in the workflow
22:04:07 <glx> hmm I still don't understand why build_ext fails for python 3.5
22:04:22 <glx> I tried ugrading setup tools, still fail
22:05:03 <andythenorth> it's only a basic proof of concept, it does nothing properly
22:05:12 <LordAro> frosch123: andythenorth: it would be really neat to rework eints into a GH app
22:05:14 <andythenorth> but passing artefacts is fine
22:05:22 <andythenorth> the problem is creds
22:05:49 <andythenorth> for Azure I give it a GH service (revokable) and AWS service point (revokable)
22:06:23 <andythenorth> storing, e.g. bundles ssh (sftp, scp) creds directly in people's Azure accounts is a big non-non-non
22:06:25 <LordAro> glx: that feels familiar to me. could be related to how python3.5 that they're using was compiled
22:06:36 <andythenorth> and that's even if I could find the keystore tool, which exists somewhere
22:06:38 <frosch123> LordAro: what does that actually mean? gh authentication for translators?
22:06:50 <frosch123> or is there more gh hosted stuff i do not know about
22:07:11 <andythenorth> frosch123: it would be similar to the way I've given Azure access to GH
22:07:24 <frosch123> so, just oauth2 stuff
22:07:40 <glx> LordAro: but VC++ 14+ is installed
22:07:42 <andythenorth> there's a key, and so on
22:07:53 <frosch123> i have that 50% done, but did not figure out a test strategy
22:07:58 <andythenorth> and it's managed in GH as a provider, not in the consuming app
22:08:07 <glx> unless they want exactly 14.0
22:08:15 <LordAro> frosch123: i *think* you could enable eints on a repo, and then translators would then use some eints frontend (coop or otherwise) to submit translations, and eints could commit them to the repo
22:08:31 <andythenorth> how about a totally upside down idea
22:08:50 <andythenorth> (multiple apps behind the scenes, but call it bananas 2 from the user point of view)
22:08:52 <LordAro> glx: trouble is, GH actions is too new, can't find other people's solutions :p
22:08:56 <andythenorth> bananas 2 adds a project as a remote
22:09:05 <frosch123> glx: there is some --compiler=... parameter to setup.py
22:09:11 <andythenorth> bananas 2 has a dedicated translations repo, with a branch for each project
22:09:21 <frosch123> i have no idea how and whether setuptools autodetects compilers
22:09:28 <andythenorth> bananas 2 never commits to the project for translations
22:09:49 <andythenorth> when there's a push to the project, bananas 2 pulls the remote and rebases the local translations branch
22:10:05 <LordAro> seems to me like more output is required first anyway
22:10:12 <andythenorth> then, if there are changes > threshold, it publishes a trivial version bump on the content service
22:10:18 <LordAro> (and preferrably it should fail when the extension build fails)
22:10:33 <LordAro> (that's probably the "optional" in setup.py)
22:11:16 <frosch123> i can only find examples with --compiler=mingw32
22:12:15 <glx> maybe it relies on environment for python 3.5
22:12:28 <andythenorth> oof, if the project included docs, bananas 2 could then publish those, using GitHub Pages, and it's weird headless docs branch, with a path like firs/version/docs, ecs/version/docs
22:12:54 <andythenorth> one uber repo for content
22:13:02 * andythenorth wonders about the malware exploit funtimes
22:14:29 <LordAro> are you sure it's installed for the 3.5 image?
22:17:07 <glx> C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Tools\MSVC\14.23.28105\bin\HostX86\x64\link.exe
22:17:25 <glx> but maybe I should try x86 instead
22:25:36 <LordAro> probably won't exist, VS compiler paths are weird
22:30:41 <glx> yup for 3.6 and 3.7 no issue with both x86 and x64, 3.5 fails for both
22:30:57 <glx> and I "broke" some stuff ;)
22:34:05 <frosch123> he, so for 3.7 it works on windows, but not on ubuntu x86
22:34:42 <glx> x86 is windows only it seems
22:36:32 <frosch123> ah, now the 3.5 windows build contains an error about msvc 14 missing
22:36:49 <frosch123> but well, if 3.7 works :)
22:39:53 *** WormnestAndroid has quit IRC
22:40:30 *** WormnestAndroid has joined #openttd
22:49:58 <andythenorth> website release?
22:52:59 <andythenorth> LordAro: you have some unreviewed PRs from April
22:53:34 <LordAro> andythenorth: on the website? they're incomplete, or need splitting up, or something
22:53:50 <LordAro> unfortunately the other person interested in doing website restructuring got bored
22:54:42 <DorpsGek_III> [OpenTTD/website] LordAro approved pull request #105: Drop supported operating systems list from 'About' https://git.io/JeBx4
22:54:48 <DorpsGek_III> [OpenTTD/website] LordAro merged pull request #105: Drop supported operating systems list from 'About' https://git.io/Je4M4
22:55:19 *** Eddi|zuHause2 has joined #openttd
22:56:49 <andythenorth> I deliberately didn't review that PR soz
22:57:08 <andythenorth> html style is a thing that is prone to opinion-stated-as-fact
22:57:19 <andythenorth> or even misunderstanding-stated-as-fact :P
22:57:28 <andythenorth> and it's too much like my day job to trade it back and forth
22:57:39 <andythenorth> the diff looked basically fine
22:57:55 <andythenorth> but nitpicking markup is a waste of life, either the page is valid and renders, or it isn't and doesn't
23:09:09 <LordAro> andythenorth: too late now
23:09:12 *** Etua is now known as Guest5582
23:09:28 <LordAro> i'm sure i'll get into another websitey mood at some point, the PRs aren't going anywhere
23:10:47 <andythenorth> if the GH docs stuff gets sorted, a new website _might_ just appear in a PR :P
23:10:55 <andythenorth> one thing at a time :P
23:29:04 <andythenorth> most of the dev stuff could stay in wiki, it's guidance
23:29:10 <andythenorth> but code style is canonical?
23:32:34 *** andythenorth has joined #openttd
continue to next day ⏵
