IRC logs for #openttd on OFTC at 2019-03-22
⏴ go to previous day
00:20:09 <frosch123> changing the default would break ottd compilation
00:20:32 <frosch123> last time obg generation did a plain md5sum
00:45:27 <peter1138> Presets work, but there's no default :p
00:49:53 <peter1138> colour_presets = -1957633922,-1336168961,-1790426469,-1784986981,1451692562,0,0,2077322240,0,0,0,0,0,0,0,0
00:50:33 <peter1138> That's just how the settings system exports uint32s.
00:50:51 <Eddi|zuHause> also, why are those 32bit numbers and not 24bit?
00:51:21 <Eddi|zuHause> @base 10 16 -1957633922
00:51:21 <DorpsGek> Eddi|zuHause: -74AF1F82
00:51:21 <peter1138> Because there's an 8 bit contrast component.
00:51:53 <Eddi|zuHause> that doesn't seem to know about 2-complement :p
00:52:37 <Eddi|zuHause> peter1138: i'd say exporting uint as signed is a bug.
00:53:04 <peter1138> Whatever, it's how the settings system works, not how the rgb company colours work.
01:05:37 *** circ-user-3HYVs has quit IRC
01:55:11 *** circ-user-3HYVs has joined #openttd
02:37:46 <Eddi|zuHause> ... i guess it just has never come up before
02:43:17 <Eddi|zuHause> "/* Forward declare these; can't do 'struct X' in functions as older GCCs barf on that */" <-- and then it uses "struct X" in function headers anyway?
02:47:19 *** Thedarkb-X40 has joined #openttd
03:00:53 <Eddi|zuHause> "proc(ini, _settings, "patches", &_settings_newgame);" <-- that doesn't sound right
03:03:15 <Eddi|zuHause> "/* XXX - wtf is this?? (group override?) */" <-- now we're getting in deep :p
03:07:22 <Eddi|zuHause> case SDT_INTLIST: <-- i think this is the place
03:09:13 <Eddi|zuHause> "case SDT_NUMX:" has a call to "IsSignedVarMemType"
03:09:21 <Eddi|zuHause> SDT_INTLIST is missing that
03:11:18 <Eddi|zuHause> it's not checking for signedness at the seprintf line
03:11:38 <Eddi|zuHause> unconditionally using %d there
03:11:54 <Eddi|zuHause> not %d/%u like with SDT_NUMX
03:12:12 <Eddi|zuHause> that doesn't matter
03:19:55 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z opened pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjfhC
03:20:00 <Eddi|zuHause> i think that should work
03:30:34 <Eddi|zuHause> well, i have no test case :)
03:39:47 *** Supercheese has joined #openttd
03:40:36 <peter1138> colour_presets = 2337333374,2958798335,2504540827,2509980315,1451692562,707801136,3640411317,2077322240,3636918595,3640253251,3633053951,2885685385,0,0,0,0
03:40:45 <peter1138> I mean, it works, but it's not exactly... better? :pp
03:40:57 <Eddi|zuHause> well, but it's... correct
03:41:04 <peter1138> Also, loading doesn't working
03:41:21 <peter1138> Gets loaded as int.
03:41:25 <Eddi|zuHause> i was trying to decipher loading
03:42:00 <Eddi|zuHause> loading numbers has some special case for U32
03:42:12 <Eddi|zuHause> same kind of special case needed for list, i guess
03:42:17 <peter1138> if (sizeof(int) < sizeof(long)) v = ClampToI32(v);
03:44:33 <Eddi|zuHause> that function needs passing of the VarType
03:56:41 <Eddi|zuHause> so in StringToVal it parses stuff with strtoul
04:01:19 <Eddi|zuHause> i don't even see how that handles signed correctly, unless it's by accident
04:02:38 <Eddi|zuHause> also, that function seems to be locale-dependent?
04:09:55 <Eddi|zuHause> do we want to use std::strtoull?
04:11:45 <Eddi|zuHause> or we just scrap the Clamp and hope it'll be alright?
04:16:43 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z updated pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjfhC
04:16:51 <Eddi|zuHause> i'm completely unsure about that...
04:38:01 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z updated pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjfhC
04:38:12 <Eddi|zuHause> now i'm getting completely crazy
04:40:58 <Eddi|zuHause> i think that one's incorrect
04:42:48 <Eddi|zuHause> also, it's missing a file
04:43:03 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z updated pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjfhC
04:54:42 *** tokai|noir has joined #openttd
04:54:42 *** ChanServ sets mode: +v tokai|noir
05:30:14 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z updated pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjfhC
05:34:37 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjfjA
07:59:09 *** andythenorth has joined #openttd
08:24:46 * LordAro wonders when Eddi|zuHause sleeps
10:38:33 *** andythenorth has joined #openttd
10:43:38 *** Gustavo6046 has joined #openttd
11:09:39 *** techmagus has joined #openttd
12:24:03 *** gareppa has joined #openttd
12:29:52 *** circ-user-3HYVs has quit IRC
12:32:41 <Eddi|zuHause> (sorry for sniping like this)
12:34:34 <planetmaker> I actually disagree that baseset and tools should be under NewGRF. I made the choice deliberately
12:34:49 <planetmaker> should we start an edit-war?
12:34:57 <peter1138> OpenGFX isn't a NewGRF.
12:35:18 <planetmaker> with tools it can be argued that most (all?) are newgrf-related
12:35:20 <peter1138> Is OpenGFX built with nml?
12:35:22 <Eddi|zuHause> anyway. i was advised to start each sentence on a new line, so diffs would be easier to read
12:35:24 <planetmaker> but basesets are different
12:35:27 <andythenorth> baseset is baseset
12:35:31 <planetmaker> yes, OpenGFX is written in NML
12:35:36 <andythenorth> it's an essential dependency
12:35:52 <andythenorth> it's incidental that it's implemented as newgrf
12:36:19 <planetmaker> it's a collection of grf and NewGRF.
12:36:23 <Eddi|zuHause> planetmaker: ok, but then scrap the "NewGRF" entry, and call "Tools" "NewGRF Tools"?
12:36:27 <andythenorth> the newgrf header should contain: spec + toolchain entries
12:36:40 <planetmaker> there's catcodec which is a sound tool
12:36:41 <andythenorth> ultimately, it's just a blog post
12:36:53 <andythenorth> so ultimately it's blah blah and classic bikeshedding rabbit hole
12:36:56 <Eddi|zuHause> planetmaker: now you're splitting hairs
12:37:24 <planetmaker> nml is a tool for basesets and newgrfs. nforenum, grfcodec as well. and musa a tool for all kind of content
12:37:34 <andythenorth> if I wrote the post, I'd put 'Content APIs' and dump everything under that, which would confuse everyone
12:37:38 <andythenorth> but I would be correct
12:37:45 <planetmaker> Eddi|zuHause, not quite. But this distinction drove my choice for separate topic
12:38:34 <Eddi|zuHause> planetmaker: then it's "## mod support" and "### GRF Tools"
12:38:48 <andythenorth> everything about no coherent vision, GS vs NewGRF etc is because of historical evolution of content APIs
12:38:51 <Eddi|zuHause> (wording needs more work)
12:38:56 <planetmaker> or just one heading: NewGRFs, baseset & Tools
12:39:26 <andythenorth> my rationale is that basesets are content
12:39:29 <andythenorth> not content APIs
12:39:40 <andythenorth> but eh, peak bikeshedding achieved
12:39:53 <planetmaker> yellow with violet dots please
12:39:57 <andythenorth> introduce joke: "how many people does it take to write a dev blog post"
12:40:06 <andythenorth> dunno what the punchline is
12:40:17 <andythenorth> probably "we don't know, it's not finished yet"
12:40:48 <Eddi|zuHause> planetmaker: imho, the structure with both things being subgroups of one category is better. just need to find a correct name for the category
12:42:13 <planetmaker> I can well imagine to see a GS and AI section appearing some time as well
12:55:41 <DorpsGek_II> [OpenTTD/website] planetmaker commented on pull request #65: Add: Monthly Dev Post of April 2019 https://git.io/fjJT9
13:06:27 <Eddi|zuHause> planetmaker: imho discussing whether to call it the "NewGRF" section is on a similar level as convincing people to not say "Holland" or "America"
13:11:58 <Eddi|zuHause> planetmaker: i mean, we could call it the Bananas section, but people would confuse that with the hosting service
13:13:41 <andythenorth> mostly because our own thinking is distorted by historical facts
13:14:09 <Eddi|zuHause> yo, check your privilege?
13:14:45 <andythenorth> is that a missed reference? :P
13:15:35 <andythenorth> I did actually check my privilege once
13:15:45 <andythenorth> I am ahead, but not the most ahead
13:16:01 <Eddi|zuHause> there's always someone further ahead
13:21:21 <andythenorth> someone must be the most ahead
13:21:30 <andythenorth> probably Berlusconi?
13:21:39 <Eddi|zuHause> no, because it's non-transitive
13:22:18 <Eddi|zuHause> it's like nobody is ever at the top of the food chain
13:22:28 <Eddi|zuHause> because it's not a chain
13:22:34 <Eddi|zuHause> that's a simplification
13:24:14 <planetmaker> I would never have called a script "NewGRF". And just because some or many people use words incorrectly, it doesn't mean we have to use bad wording, too
13:26:20 <planetmaker> I'd have no problem with using that term
13:29:03 <Eddi|zuHause> yes, i already suggested going in that direction.
13:30:15 <Eddi|zuHause> however, we should come up with a generic-enough term to put it into the template for every future blog post
13:30:32 <planetmaker> are basesets mods? :P
13:30:59 <planetmaker> they are rather the baseline...
13:31:07 <Eddi|zuHause> they're still mods
13:31:23 <planetmaker> no, they don't modify anything
13:31:36 <Eddi|zuHause> you can modify them
13:31:41 <planetmaker> you mean they make an uncomplete game complete?
13:31:52 <peter1138> They don't... modify the gameplay though.
13:32:04 <Eddi|zuHause> no, they modify how the game looks
13:32:08 <Eddi|zuHause> that is still modify
13:32:21 <planetmaker> without baseset the game is not a game. Just broken code
13:32:26 <peter1138> Should I enter the Euromillions before I'm not allowed to? :p
13:32:33 <Eddi|zuHause> yes, but there are multiple basesets
13:33:02 <peter1138> Mostly... not very pretty.
13:33:04 <andythenorth> eh ok, so I was drafting new website words this week
13:33:33 <andythenorth> "mods" "add-ons" "packs" "content" "dlc" "customisations" ??
13:33:43 <peter1138> Funny how you give tools to make it super gorgeous, and they come up with... aBase
13:33:53 <planetmaker> graphics, sounds, scripts & tools
13:34:11 <Eddi|zuHause> planetmaker: doesn't sound very generic
13:34:14 <andythenorth> most games have mods or packs or DLC
13:34:17 <peter1138> Still waiting for pikkabase
13:34:21 <andythenorth> 'most' => the ones I see my kids playing :P
13:34:30 <andythenorth> not scientific most
13:34:50 <peter1138> But I guess that was the one that was the attempted kickstarter.
13:34:52 <andythenorth> mods tend to change behaviour
13:34:59 <andythenorth> packs tend to be graphics skins etc
13:35:06 <andythenorth> DLC tends to be ££$$€€
13:35:13 <andythenorth> but there are no fixed rules
13:35:15 <peter1138> So basesets are packs, and NewGRFs are mods.
13:35:27 <andythenorth> ["fuck knows" emoji]
13:35:40 <andythenorth> my draft website title for all this crap is "Extend"
13:35:53 <planetmaker> Eddi|zuHause, anything missing in the list of "graphics, sounds, scripts & tools"?
13:35:59 <planetmaker> that it is not generic enough?
13:36:10 <peter1138> Extensions implies stuff that would need source code changes.
13:36:12 <Eddi|zuHause> planetmaker: the problem is that it's a list.
13:36:28 <andythenorth> I wrote "Extend the base game with downloadable content, including new vehicles, industries, stations, houses, AIs, heightmaps, scenarios, gamescripts and more"
13:36:30 <planetmaker> it's not a problem if it's comprehensive
13:36:35 <andythenorth> but it's all horrible anyway
13:36:44 <Eddi|zuHause> planetmaker: it should be more concise
13:36:51 * andythenorth is solving a different related problem, if that's not obvious
13:36:57 <Eddi|zuHause> planetmaker: for a generic section headline
13:37:00 <andythenorth> the website is getting an update :P
13:37:21 <planetmaker> it's quite concise. Much more concise and descriptive than "NewGRFs" or "Mods"
13:37:57 <andythenorth> for the core game, I wrote ""Addictive retro gaming. Free to play always. Highly-rated open source game. Single player. Online multiplayer. Sandbox or competitive."
13:38:25 <andythenorth> for Develop I wrote "Contribute to OpenTTD. Join us on irc and GitHub. Ideal for experienced who want a hobby project, and for students in computer science and similar studies who want to improve skills on a successful real-world project"
13:39:10 <planetmaker> I'd not exclude people who just want to fix their personal scratch :)
13:39:15 <andythenorth> EXPERIENCED PETER'S
13:39:19 <andythenorth> that's what I meant
13:39:57 <Eddi|zuHause> andythenorth: that's not words i would put on a website
13:40:34 <andythenorth> Eddi|zuHause: it's a free opportunity for you to do better :)
13:40:42 <Eddi|zuHause> andythenorth: you're projecting too much of your own mindset onto the random reader of the website
13:41:10 <andythenorth> yes, I am a human writing words
13:41:30 <andythenorth> and acceleration in Earth's gravitational field is about 9.8m/s
13:41:36 <andythenorth> some things are just what they are
13:41:41 <andythenorth> propose better? :P
13:41:49 <Eddi|zuHause> andythenorth: for some unforseen reason it's much easier to find someone else's words bad than to come up with better words yourself
13:43:27 <andythenorth> just write the most lolz thing
13:45:00 <planetmaker> Contribute to OpenTTD. The project is run by volunteers from all kind of backgrounds. Join us on irc and Github and have a look how you can contribute
13:45:19 <planetmaker> and a link to updated wiki page(s)?
13:48:01 <Eddi|zuHause> is that an oxymoron? :p
13:48:52 <planetmaker> well. we can play catch-22. Or start somewhere
13:49:29 <peter1138> Is it lunch time yet?
13:49:40 <andythenorth> I already had crumpets
13:49:45 <andythenorth> and something I've forgotten
13:49:47 <planetmaker> same procedure as *every* time: it's tea time :P
13:49:49 <peter1138> I've got a... er... salad :p
13:49:51 <andythenorth> now it's mac and cheese
13:50:27 <planetmaker> crunchy when served with a crust of cheese
13:50:59 <andythenorth> "Contribute to OpenTTD. Find us on irc, github, and by the bikesheds"
13:53:34 <planetmaker> bring extra paint and blueprints
13:53:56 <andythenorth> "Contribute: nitpicking brings the quality"
13:54:12 <andythenorth> "Contribute: github and much lolz"
14:00:26 <planetmaker> rucola... not my preferred choice. But looks good :)
14:02:01 <Eddi|zuHause> ... "wer hat's erfunden?" :p
14:07:32 <andythenorth> peter1138: such jealous
14:07:40 <andythenorth> we should start some kind of social network
14:07:44 <andythenorth> for lunch photos
14:07:55 <Pikka> any time is tiffin time, planetmaker
14:07:57 <planetmaker> yummi cats. Fried with calamari
14:09:44 <LordAro> Eddi|zuHause: i like it
14:10:49 <Eddi|zuHause> LordAro: i hope you don't mean the green :p
14:14:56 <peter1138> What's wrong with green?
14:23:17 <Eddi|zuHause> green is the colour of poison. everyone knows that
14:25:33 <peter1138> It's the colour of low-effort healthy food.
14:26:25 <peter1138> Ooh that reminds me, I have haddock tonight again :D
14:27:14 <andythenorth> square plates: big game
14:27:30 <andythenorth> Eddi|zuHause: most of us overcome our fear of green by the age of about 15
14:27:43 <andythenorth> my kids thoroughly distrust it, they are younger
14:27:57 <Eddi|zuHause> andythenorth: too many traumatic experiences in that time
14:28:01 <andythenorth> I didn't trust green until I was about 32
14:29:01 <peter1138> So what veg do you eat if you shun green?
14:29:25 <peter1138> Bratwursts and frikadellen don't count as veg.
14:30:25 <peter1138> Oh that frivilous thing passed 3 million.
14:34:21 <peter1138> I rounded it off with a piece of simnel cake... more than doubled the carolies I bet.
14:35:06 *** WWacko1976-work has joined #openttd
14:38:44 <planetmaker> bratwurst and frikallen are secondary vegies
14:39:10 <planetmaker> vegitarians eat my food's food :P
14:44:28 <LordAro> Eddi|zuHause: your blog post update :)
14:45:12 <Eddi|zuHause> LordAro: i've had that thought for like 2 months now
15:27:14 <peter1138> Are you aware of k-d tree crashes?
15:28:19 <nielsm> yes... I ought to look at some code some time soon
15:28:42 <nielsm> been burying myself in Satisfactory for the past several days
15:28:54 *** Wormnest has joined #openttd
16:51:17 <Samu> hey pikka, i'd like your ai to be able to build more than 500 rvs
16:51:48 <Pikka> is that the current maximum?
16:53:23 <Pikka> well if you change the max_value in the info.nut, that should probably do it
16:54:58 <Pikka> well I can increase it for you in the next version, but in the meantime... :)
16:55:18 <Samu> 5000 for all vehicle types
16:56:30 <Samu> i rarely see it building ships, maybe i'm trying too distant routes
16:57:16 <Pikka> it does rarely build ships... the next version will have reworked vehicle logic which may make it build more
16:58:01 <Pikka> but in any case the routefinding between docks is very simple, so it's unlikely to build routes that aren't straight across a lake or straight along an edge of the map :)
17:02:29 <Samu> min town population of 500
17:03:04 <Pikka> that's already the minimum value, isn't it?
17:03:15 *** Progman has joined #openttd
17:03:22 <Samu> yes, but I'm looking at it
17:03:33 <Samu> can you come into this game?
17:04:20 <Samu> he builds bus stops in 2 towns, but doesn't build the buses ?
17:04:51 <Pikka> are you the server? can you see the debug log?
17:05:06 <Pikka> civi's is reasonably human-readable :)
17:05:32 <Pikka> but I can see from what it's done that it hasn't found space to build 3 stops, which is the minimum it needs to build a bus route
17:05:42 <Pikka> next cycle round it will remove the stops
17:06:02 <Samu> yeah, those little details
17:07:26 <Samu> it keeps connecting towns by roads, but without any source of income, it's not that efficient, is it?
17:07:37 <Samu> I was expecting buses by now
17:09:02 <Pikka> well, that's partly because you've set the minimum town size below the default. The towns it's started in are too small to place enough bus stops.
17:09:35 <Samu> oh, I see, so that's how it works, it needs to really place all 3 stops
17:10:05 <Pikka> ideally it places 5, but 3 is the minimum it wants
17:10:50 <Pikka> I expect he will build a train soon though, once he gets to a town with a suitable industry
17:12:17 <Samu> well, it still has 412k money to burn
17:13:44 <Samu> connecting mendhead, heh, my ai is also there pathfinding
17:14:10 <Samu> my ai pathfinds too slow, 3 years per route
17:18:08 <Pikka> okay, so, without being able to see the log, I predict the next thing he's going to do is connect charfingbourne, then try and build a railway to the gold mine at tronfingfield, which will probably fail because it's too wiggly around the water. Unless your ai connects mendhead to its network, in which case all the connected towns will be added to civi's network too and it might try to build somewhere else :)
17:19:10 <Pikka> it picked a really bad place to start, tiny towns, no food processing...
17:24:58 *** tycoondemon has joined #openttd
17:32:17 <Pikka> or it could just keep retrying building bus stops in those towns :)
17:37:26 *** Alberth has joined #openttd
17:37:26 *** ChanServ sets mode: +o Alberth
17:42:23 <Samu> says it's connecting narfingford, where is it
17:43:02 <Samu> oh i see it, that's a bit far
17:47:23 <Pikka> I guess it wants the power station
17:48:08 <Samu> my ai will probably fail pathfinding
17:48:14 <Samu> the station is on a island
17:48:49 <Samu> wants to connect gundwood to mendhead
17:51:03 <Samu> it has ~4 more months to try
17:51:30 <Pikka> I should think so, from that island
17:53:58 <Samu> updating list of connected towns
17:54:12 <Samu> a big list of towns is connected.
17:55:20 <Samu> planting trees around dindingworth
17:55:31 <Samu> can't afford to build a railway :o
17:55:53 <Samu> was trying a coal transport by rail, but gave up
17:56:25 <Samu> says it couldn't find a suitable locomotive :o
17:56:41 <Samu> maybe it's true, let me check
17:58:05 <Samu> what's the release date of wills-2-8-0 ?
17:58:17 *** gareppa has joined #openttd
17:58:39 <Samu> ewww 1945 according to wiki
18:07:00 *** Eddi|zuHause has joined #openttd
18:07:46 <peter1138> Ok I need to go and get dog food.
18:07:58 <peter1138> But should I cycle or drive?
18:08:19 <peter1138> 17:08... gonna be busy
18:10:53 <Eddi|zuHause> if you're in any vincinity to a city at this time of day, cycling will pretty much always be faster :p
18:12:15 <peter1138> Nah, I live in a town.
18:12:32 <peter1138> Hmm, I should check if Aldi have any interesting offers :p
18:13:38 <peter1138> Cast iron tagine? Do I need one?
18:14:24 <peter1138> "Online exclusive" well then...
18:16:04 <Samu> CivilAI has £233k to burn, less than half
18:16:59 <Samu> LuDiAI AfterFix finally made the 2nd route
18:22:43 <Pikka> yes, it's not going to build a ship to any of those docks... :)
18:22:59 <Pikka> but they're nice decoration :P next version it's limited to 2 docks per town btw
18:32:31 <Pikka> hmm, maybe I should set the minimum town size to 1000, it wastes a lot of time trying to build in those too-small towns ;)
18:37:19 *** sla_ro|master has joined #openttd
18:38:14 <Eddi|zuHause> ... in what world does aldi ever have interesting offers?
18:55:43 <Alberth> in the world of aldi marketing?
19:13:16 *** HerzogDeXtEr has joined #openttd
19:19:01 *** gelignite has joined #openttd
19:22:41 *** m3henry has joined #openttd
19:41:51 *** andythenorth has joined #openttd
19:55:00 *** Wormnest has joined #openttd
20:15:29 <peter1138> What do I want to do?
20:25:45 <andythenorth> vehicle variants?
20:29:43 <andythenorth> you didn't already? :o
20:35:56 *** frosch123 has joined #openttd
20:36:50 *** Supercheese has joined #openttd
20:38:51 <LordAro> andythenorth: no, 5CC
20:44:21 <frosch123> ttdp did not have enough sprites for that
20:44:41 *** HerzogDeXtEr1 has joined #openttd
20:47:28 <peter1138> I have a patch for 32 company colours.
20:47:39 <peter1138> Dynamic generation is quite simple.
20:54:01 <peter1138> But probably RGB colours is more interesting ;)
20:55:53 <frosch123> andy can also do 5cc via layers
21:05:52 *** circ-user-3HYVs has joined #openttd
21:09:29 <andythenorth> IT IS NOT THE SAME
21:09:52 <andythenorth> peter1138 hasn't made a livery-per-layer chooser yet
21:11:07 <andythenorth> best feature ever
21:15:26 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJsx
21:15:45 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJsp
21:16:02 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJsh
21:17:35 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJGv
21:19:49 <peter1138> Little Brudingstone (City)
21:21:10 <peter1138> Hmm. Loading indicators... are they handled by k-d tree?
21:21:46 <peter1138> Oh wait, I'm in the vector PR.
21:25:30 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJGt
21:25:55 <peter1138> Crap, that beer ran out :-(
21:26:03 <peter1138> I bet LordAro's gone off to the pub.
21:27:26 <LordAro> peter1138: correct, actuallly
21:27:41 <LordAro> but for cycle planning purposes!
21:28:30 <peter1138> I could check what the plan is tomorrow.
21:29:00 <peter1138> Oh, some event that needs a ticket that I don't have.
21:29:03 <frosch123> do you have beer tanker trailers for your bike?
21:29:34 <peter1138> Ah, it's a Rapha ride. Bloody expensive, them Rapha peeps.
21:30:07 <peter1138> Hmm, 7.30. Do I want to leave at 7.30 again?
21:30:12 <LordAro> peter1138: anything exciting? i'm probably going out tomorrow to test my saddlebag
21:30:21 <LordAro> frosch123: sounds heavy
21:30:38 <peter1138> "Along with exploring our local countryside, each ride will provide the chance to explore local bakeries and cafés as we (one by one) sample the delights on offer."
21:30:46 <peter1138> Well, I mean, cafes are pretty exciting.
21:30:56 <peter1138> But I tend to not bother on the fast rides.
21:32:04 <peter1138> I could just sign up and then not bother. Need to register but it's free.
21:37:34 <peter1138> Normal ride is at 9am anyway
21:44:05 <andythenorth> lolz panama canal
21:51:10 <peter1138> I ate a cookie as well :/
21:57:34 <andythenorth> those 16 cargo nml docs
21:59:16 *** DrSegfault has joined #openttd
22:01:11 <DrSegfault> I have some stations which are shorter, and in rare cases long trains stop there (due to some redirections/bypasses), causing very long traffic jams. Is there a way to let trains not stop at stations where they do not fit in?
22:01:40 <peter1138> Yeah, use non-stop orders.
22:02:09 <peter1138> Assuming it's an implicit orders
22:02:17 *** circ-user-3HYVs has quit IRC
22:02:55 <DrSegfault> peter1138: I could, but I have ~500 trains which don't use non-stop orders... :p
22:03:06 <DrSegfault> So much work to change them all
22:03:09 <peter1138> And... you didn't use shared orders?
22:03:54 <DrSegfault> Well, I did. But even then, some trains have like 100 stations, and their timetable looks clearer if you don't specify every single one of them
22:04:57 <peter1138> you'd need to put in an order before and after the offending station
22:05:05 <peter1138> have you considered just extending them? :p
22:06:06 <DrSegfault> I don't want to turn some rural 3-stations into 7-stations, that would look weird
22:06:07 <m3henry> Lordaro: size() == 0 was something I was hoping to do in another PR
22:07:27 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJG5
22:13:13 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJGh
22:46:07 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7379: Codechange: Use C++11 functions for threading https://git.io/fjJZM
22:46:25 <TrueBrain> I am sure LordAro was about to approve it :P
22:47:29 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7396: Fix: Saving SDT_INTLIST handle unsigned values properly https://git.io/fjJZH
22:47:39 <LordAro> i've not looked at it in any great detail yet
22:49:43 <TrueBrain> I am afraid you are one of the few knowledgeable enough to say anything useful about it :)
22:49:50 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJZQ
22:50:05 <TrueBrain> 11 backport-requested PRs .. guess I can do that tomorrow
22:50:09 <TrueBrain> I understood you wanted a RC2?
22:50:26 <LordAro> given it's a week until 04/01, seems desirable
22:50:38 <LordAro> but i fear i know very little about std::thread other than "if it works, it works"
22:50:52 <michi_cc> I'd think with 11 (or more backports) a RC is prudent.
22:51:11 <TrueBrain> so approve it and go go go? :D
22:51:15 <LordAro> things could be backported without making an RC, of course ;)
22:52:03 <TrueBrain> I am fine with an RC
22:52:07 <TrueBrain> I was just double checking
22:52:52 <LordAro> TrueBrain: given it *could* be helpful with debugging...
22:53:01 <LordAro> emphasis on the *could*
22:53:02 <TrueBrain> yes; many features could be useful :)
22:53:05 <TrueBrain> but we have to be careful :D
22:53:06 <michi_cc> Well, anything that is not merged in master should not be backported :)
22:53:32 <TrueBrain> but I like the preselection ofc :)
22:53:33 <peter1138> Yeah I don't think that is a backport candidate.
22:54:21 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJZb
22:54:34 <peter1138> #7365 doesn't fix anything so no point in backporting it.
22:54:36 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7366: Add: List recently executed commands in crashlog output. https://git.io/fjJZN
22:55:15 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJZx
22:56:53 <TrueBrain> I don't understand the PR .. why is a client adding AIs?
22:57:56 <TrueBrain> smells like another bug, where this is fixing the sympthons
22:58:34 <peter1138> I believe the server runs the AI but the client needs to know about the new company slot? Not sure.
22:58:46 <peter1138> Might well be the wrong way about it.
22:59:02 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJne
22:59:03 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnv
22:59:10 <TrueBrain> how I read it, the server creates an AI, and when the command is executed, it revalidates again if it really should, or something
22:59:40 <TrueBrain> for this, it broadcasts to all the clients the max ai setting every time ..
23:00:54 <TrueBrain> so all clients + server can receive a command they all dismiss
23:01:33 <TrueBrain> "ah, the server told us to do something, but given what I know now, I am going to completely ignore him"
23:02:01 <glx> yeah server should just discard before sending it to clients
23:02:50 <peter1138> I think it's because test can succeed even on the server due to the command being queued anyway.
23:03:13 <peter1138> So I guess it's wrong if test succeeded and exec fails.
23:03:40 <glx> of course test and exec should have the same result
23:04:17 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7376: Fix 001257d: Enforce the max_no_competitors and network.max_companies tests before creating an AI company in multiplayer. https://git.io/fjJnm
23:04:48 <TrueBrain> the more I look at it, the more I am sure this is one huge hack
23:04:57 <TrueBrain> "it does the job" kind of solution
23:05:41 <glx> BTW it seems original code is wrong too, there a check returning error after the early non exec return
23:05:57 <peter1138> Better solution: send the number of AIs you want to start in one command.
23:06:34 <TrueBrain> I dunno either; I just know I don't like this solution at all :P
23:06:45 <peter1138> cmake fails with kdevelop, due to a bug in cmake/kdevelop
23:06:45 <glx> probably never triggered check but it feels wrong
23:07:07 <TrueBrain> he references a commit fix which does not exist; excellent
23:07:22 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7081: Change: [Linkgraph] Pause the game when linkgraph jobs lag (#6470) https://git.io/fjJnZ
23:07:25 <LordAro> there was also something mentioned about how 5 AIs are started instead of 4 genereally, i think
23:07:53 <peter1138> Yeah, something about max number of queued commands?
23:08:06 <LordAro> (not just in multiplayer, i mean)
23:08:10 <peter1138> It's not really a huge issue.
23:08:11 <LordAro> Samu: care to elaborate?
23:08:17 <peter1138> Most people are not using AIs in multiplayer
23:08:25 <peter1138> So I think it's safe to ignore as a backport item
23:08:44 <TrueBrain> which indeed has a loop, which causes problems
23:09:06 <TrueBrain> easy solutions range from: only start AIs 1 per day
23:10:28 <TrueBrain> backports with more than one commit are a bit annoying .. it is not easy to automate if that was rebased or squashed into master :P
23:11:02 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 requested changes for pull request #7376: Fix 001257d: Enforce the max_no_competitors and network.max_companies tests before creating an AI company in multiplayer. https://git.io/fjJn8
23:11:38 <LordAro> Samu: please precisely explain the issue that #7376 fixes
23:11:45 <LordAro> and whether it only affects multiplayer
23:11:54 <Samu> it only affects multiplayer
23:12:02 <peter1138> LordAro, "No description provided" ;)
23:12:05 <Samu> it is due to commands per frame setting
23:12:21 <Samu> it can delay execution of commands by another tick
23:13:09 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnB
23:15:01 <TrueBrain> no clue if it is correct content-wise
23:15:05 <LordAro> Samu: thanks, can you put that in the PR description please?
23:15:54 <TrueBrain> glx: "calls add_custom_command(). Any add_custom_target()" :D
23:16:20 <TrueBrain> "by add_custom_command_timestamp(), then calls add_custom_target()." <- guess something went wrong with search/replace somewhere :P
23:16:22 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnR
23:16:32 <TrueBrain> glx: you keep mixing 'command' and 'target' in comments :D
23:17:00 <glx> no it's a warning to not forget to replace the add_custom_target call too
23:17:30 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnE
23:17:33 <TrueBrain> okay, that was not clear to me :)
23:18:25 <TrueBrain> I have no suggestions to fix :P
23:18:46 <glx> and the other comment is just so only treated files are touched
23:19:04 <TrueBrain> I think itreally helps if you add \n\n before that sentence, now I come to think of it
23:19:11 <TrueBrain> and something like: "Note:" in front of it
23:19:16 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnu
23:19:22 <glx> because add_custom_target can have files not timestamped, like the copy command for some baseset files
23:19:46 <TrueBrain> but okay, these kind of "hacks" to fix shortcomings in CMake will always be difficult to read :)
23:20:04 <TrueBrain> so .. look at it, see if you can make it more uinderstandable for others .. and put it in the cmake branch :)
23:20:34 <TrueBrain> right, time to watch some telly
23:20:43 *** circ-user-3HYVs has joined #openttd
23:23:53 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJna
23:30:30 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 updated pull request #7270: Introduce CMake (and removing all other project-related code) https://git.io/fhbqc
23:34:11 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnS
23:35:42 <Samu> NetworkDistributeCommands();
23:37:20 <Samu> DistributeQueue at network_command.cpp line 264
23:37:49 <Samu> this is the function that splits the queue into multiple "frames"
23:38:30 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJn5
23:40:33 <DorpsGek_II> [OpenTTD/OpenTTD] M3Henry commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnF
23:41:45 <Samu> line 996 NetworkExecuteLocalCommandQueue() comes after the DistributeQueue
23:42:03 <Samu> at network.cpp, this is when the AI companies are created
23:42:08 *** Thedarkb-T60 has joined #openttd
23:42:21 <Samu> but only if it wasn't delayed
23:43:13 <Samu> line 999 starts another tick
23:43:53 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnb
23:44:19 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7165: [core] Implement SmallVector using std::vector https://git.io/fjJnN
23:48:59 <Samu> oh, the wrong reference, gonna fix it
23:51:37 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick updated pull request #7376: Fix 011257d: Enforce the max_no_competitors and network.max_companies tests before creating an AI company in multiplayer. https://git.io/fjvY8
23:57:02 <Samu> TrueBrain, how to solve :(
23:59:31 <Samu> problem is caused by Company::GetNumItems()
continue to next day ⏵
