IRC logs for #openttd on OFTC at 2019-05-11
⏴ go to previous day
00:14:22 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 updated pull request #7270: Introduce CMake (and removing all other project-related code) https://git.io/fhbqc
00:17:57 *** Hazzard was kicked by DorpsGek (Wrong channel. Retry in #openttdcoop.)
00:23:40 <glx> but my script has a list of autokick ottdcoop commands :)
02:32:55 *** SimYouLater has joined #openttd
04:50:18 *** Wormnest has joined #openttd
04:55:33 *** kiwitree has joined #openttd
06:09:31 *** Supercheese has joined #openttd
07:17:36 *** FLHerne has joined #openttd
07:27:35 *** andythenorth has joined #openttd
08:09:53 *** Supercheese has joined #openttd
08:15:12 <peter1138> Morning. Is it lunch? Twelveses? Elevenses?
08:22:54 <andythenorth> NRT rail crossings? o_O
09:03:53 <andythenorth> volume factor for cargos? o_O
09:39:34 *** gelignite has joined #openttd
09:39:59 *** Progman has joined #openttd
09:47:58 <LordAro> maybe there should've been some sort of setting conversion for that
09:51:05 *** kiwitree has joined #openttd
09:53:40 *** sla_ro|master has joined #openttd
10:43:58 * andythenorth wonders if nml can add days to a date
10:44:04 <andythenorth> might be an expression for that
10:45:18 <LordAro> andythenorth: maybe you should add one, if not :p
10:45:50 <andythenorth> looks like vehicle prop 2A is days
10:46:10 * andythenorth has all vehicles introduced on same day of year currently
11:11:00 <Eddi|zuHause> NML can certainly do that
12:06:57 *** Wolf01 is now known as Guest2294
12:11:09 <Wolf01> Hmmm, with fritz I can't even contact destiny 2 servers... at least with the other modem I could login and get dropped after 3 minutes
12:12:01 <Eddi|zuHause> it wants to protect you from insanity
12:12:30 <Eddi|zuHause> is it a true fritzbox or one of those branded/castrated ones?
12:12:37 <Wolf01> I waited 5 months to play destiny 2 again...
12:14:38 <Wolf01> I suspect destiny 2 servers are really picky, I don't have any problem wit any other service
12:15:03 <Eddi|zuHause> do they let you choose different gateways?
12:16:24 <Wolf01> I can change everything, but I don't know what my isp supports
12:16:42 <Eddi|zuHause> i mean the destiny 2 login page
12:17:30 <Wolf01> I could try with US or Asia servers
12:17:52 <Wolf01> I can't change any option from the game
12:18:05 <Wolf01> I'm not even reaching the menu
12:19:23 <Wolf01> Opened the 2 required ports, even enabled upnp for my pc, nothing changed
12:21:52 <Eddi|zuHause> there's a guy who reported he could switch the network to his cell phone, connect, and then switch back to his regular network to play on
12:22:34 <Eddi|zuHause> otherwise, i couldn't find any details how to select specific gateways
12:23:58 <Wolf01> The only thing I know is that my isp uses port 3074 for remote router management, and that's one of the 2 required ports for destiny
12:24:24 <Wolf01> But I had that problem only with the isp's modem, not with other ones
12:29:16 <Wolf01> Did a full scan and repair of game files, seem to have worked
12:29:43 <Wolf01> At least it passed the login
12:30:16 <Eddi|zuHause> you forgot to defrag your hard disk
12:32:50 <Wolf01> Ok, the STUN is even open now, it was type 2 before
12:43:46 <peter1138> LordAro, main issue is it's hidden as an advanced setting.
12:45:27 * peter1138 pats his trusty Mikrotik routers.
13:08:40 <DorpsGek_II> [OpenTTD/OpenTTD] michicc opened pull request #7586: Fix #7463: Promote scroll mode setting to basic category. https://git.io/fjC57
13:09:17 <Eddi|zuHause> .. what a coincidence :p
13:13:23 *** HerzogDeXtEr has joined #openttd
13:22:09 <Wolf01> So, Germany started to build e-highways
13:23:57 <Eddi|zuHause> yeah, but it would probably need some significant fraction of highways converted to actually take off
13:24:56 <Eddi|zuHause> ... and not some other country coming along building a slightly incompatible system
13:27:23 <Wolf01> We have the same rail system everywhere in EU [citation needed], wouldn't we be able to come up with the same e-highway infrastructure?
13:28:09 <Eddi|zuHause> ... except we don't
13:28:49 <Eddi|zuHause> we have incompatible rail gauges (spain, russia), incompatible loading gauges (britain), in compatible electrification systems (everywhere), ...
13:29:10 <Eddi|zuHause> and we were close to having incompatible wagon coupling systems a bunch of times
13:29:22 <Eddi|zuHause> and let's not talk about signalling systems
13:29:51 <Eddi|zuHause> where even ETCS is still incompatible with ETCS everywhere
13:30:22 <Wolf01> I was sarcastic, we don't even speak the same language in the same State
13:31:00 <Eddi|zuHause> at least we solved the incompatible currency problem... oh wait
13:47:42 <peter1138> Managed to lose my keys this morning :(
13:48:59 <peter1138> Is it lunch time yet?
13:56:38 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN approved pull request #7586: Fix #7463: Promote scroll mode setting to basic category. https://git.io/fjCdw
14:05:55 <DorpsGek_II> [OpenTTD/OpenTTD] michicc merged pull request #7586: Fix #7463: Promote scroll mode setting to basic category. https://git.io/fjC57
14:20:46 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh updated pull request #7398: Fix #7371: Avoid dependency on foundations of town tile during saveload https://git.io/fjJ49
14:50:16 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN approved pull request #7398: Fix #7371: Avoid dependency on foundations of town tile during saveload https://git.io/fjCF3
14:53:10 <peter1138> Jesus Christ these Nazis are agressive.
14:53:19 <peter1138> Hell Knights are easier to kill.
14:58:04 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh merged pull request #7398: Fix #7371: Avoid dependency on foundations of town tile during saveload https://git.io/fjJ49
14:58:10 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh closed issue #7371: Kdtree is built too early in savegame loading process https://git.io/fhjxX
15:08:36 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh updated pull request #7516: Memory measurement and limits for Squirrel https://git.io/fjYfW
15:13:30 <peter1138> Plasma rifle seems to be the way forward.
15:17:59 <peter1138> Mmm, Turkish delight.
15:18:49 <peter1138> I probably ought to mow the lawn while it's not raining.
15:20:32 <Eddi|zuHause> how is it not raining?
15:30:14 *** frosch123 has joined #openttd
15:32:45 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN approved pull request #7516: Memory measurement and limits for Squirrel https://git.io/fjCFD
15:38:48 <nielsm> and that was a savegame upgrade PR so every other of those needs rebasing again
15:39:49 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7575: Feature: Add industry production graph https://git.io/fjCFd
15:42:10 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7575: Feature: Add industry production graph https://git.io/fjCFb
15:44:12 <nielsm> I usually switch volume measurements to m^3 and avoid the kilolitres measurement :P
15:45:38 <peter1138> I'm sure you do but not every done.
15:47:00 <peter1138> There's also imperial vs metric tons.
15:47:20 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh updated pull request #7353: Feature: Measure vehicle capacity utilisation efficiency https://git.io/fhho4
15:49:31 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7575: Feature: Add industry production graph https://git.io/fjCbv
16:03:34 *** andythenorth has joined #openttd
16:11:55 <peter1138> That's the lawn done. Pretty quick at the moment, it's all shortish already.
16:17:58 * peter1138 succumbs to a nice beer.
16:48:27 *** snail_UES_ has joined #openttd
16:49:10 <snail_UES_> hi guys, Andy posted an interesting question on my thread…
16:49:46 <snail_UES_> if I want to modify a wagon’s capacity, there are two ways to do it: (1) through the “Capacity” callback, or (2) through the “properties” callback, using the “capacity” property
16:49:54 <snail_UES_> are there any differences between the two methods?
16:50:10 *** gelignite has joined #openttd
16:53:12 <peter1138> Yes, capacity property is more efficient.
16:53:37 <peter1138> You should only use the callback if you want to make the capacity change based on other things.
16:59:14 <snail_UES_> I do make capacity change based on a few things… but I’m still using capacity property to do that
16:59:31 <snail_UES_> I feel it simplifies my code a lot (especially because I also have to change loading amount)
16:59:55 <peter1138> Eh? If you're using the capacity property, it can't change.
17:00:52 <peter1138> I'll rephrase that. If you're not using the capacity/properties callback, it can't change.
17:01:12 <peter1138> I misread your question :(
17:01:25 <peter1138> There's a capacity property that should be used for fix capacity.
17:01:37 <peter1138> The callbacks should only be used if capacity should change.
17:05:35 <Samu> round 18 finished, now without inflation
17:05:47 <Samu> and now, redoing round 5
17:06:35 <Eddi|zuHause> there's two callbacks
17:06:47 <peter1138> Hmm. There's a refit capacity callback.
17:06:52 <peter1138> And the CB36 callback.
17:06:59 <Eddi|zuHause> one is subject to the default capacity multiplier for goods/mail etc.
17:07:01 <Eddi|zuHause> the other is not
17:07:07 <Eddi|zuHause> i forgot which one is which
17:08:25 <peter1138> I don't see two. Only refit and CB36.
17:10:43 <Eddi|zuHause> so CB15 is not subject to the multiplier, CB36 is
17:14:10 <snail_UES_> petern1138: I’m using CB36
17:14:42 <Eddi|zuHause> snail_UES_: with CB36, capacity will double if you refit to goods
17:14:48 <snail_UES_> ah I see. When doing this, I never refit passengers or mail. Only other goods
17:14:55 *** Wormnest has joined #openttd
17:14:58 <snail_UES_> oh, an not goods either
17:15:04 <snail_UES_> only other types of cargo
17:15:09 <snail_UES_> so it will be the same?
17:15:40 <snail_UES_> that’s why, in the game, the two methods had the same final effects
17:21:39 <frosch123> the only difference between cb36 and cb15 are the capacity multipliers
17:21:52 <frosch123> cb36 is meant for unknown/general cargos. cb15 is meant for control freaks
17:22:43 <Eddi|zuHause> if snail is not a control freak, then i don't know who is :p
17:23:16 <frosch123> Eddi|zuHause: i am pretty sure cb36 capacity was added by someone not being aware of cb15
17:23:33 <Eddi|zuHause> i like the clarifying use of graphs on that page
17:23:57 <Eddi|zuHause> frosch123: i'm not convinced that is the case
17:26:50 <Eddi|zuHause> (note: i was being ironic. that page is horrible documentation)
17:27:28 <peter1138> Was it me or Lakie?
17:27:52 *** Supercheese has joined #openttd
17:27:59 <snail_UES_> CB36 capacity is useful because it can be set in parallel as loading amount… usually the two get changed together to be consistent
17:28:21 <snail_UES_> otherwise I’d have to change capacities through CB15, and then use CB36 anyway to change loading amount
17:28:44 <peter1138> Yeah, prefer CB36, I think.
17:28:45 <snail_UES_> CB36 allows to pick two birds with a stone
17:29:20 <peter1138> Not really, it's still two separate calls within the game.
17:29:44 <snail_UES_> ah… well, at least code-wise, they’re in the same statement
17:29:49 <peter1138> But efficiency-wise it is better as you are only checking for CB36 instead of both CB15 and CB36.
17:30:06 <DorpsGek_II> [OpenTTD/OpenTTD] kiwitreekor updated pull request #7575: Feature: Add industry production graph https://git.io/fjc3i
17:30:43 <frosch123> peter1138: when cb15 succeeds, cb36 is not called. so no difference in efficiency
17:31:31 <snail_UES_> cb36 would still be called for loading amount, though
17:31:40 <peter1138> Except for every time the game looks up graphics it'll go through the test for callbacks.
17:31:41 <frosch123> (both are called in depot only, so efficiency is super-no-pointo)
17:31:44 <Eddi|zuHause> the amount of callbacks you check in the callback switch does not affect efficiency...
17:32:20 <snail_UES_> well it’s more efficient code-wise, so it’s more than enough for me :p
17:32:36 <DorpsGek_II> [OpenTTD/OpenTTD] kiwitreekor commented on pull request #7575: Feature: Add industry production graph https://git.io/fjCAc
17:33:36 <Eddi|zuHause> we should add a JIT compiler so it can optimize switches during execution :)
17:34:14 <frosch123> there actually already is
17:34:47 <frosch123> switch cases are reordered and merged to enable binary search for large switches
17:35:32 <frosch123> which may imply that few large switches are better than many small switches
17:36:29 <Eddi|zuHause> but a true JIT compiler might optimize over several switches?
17:37:14 <Eddi|zuHause> like, avoid redundant calculations, and stuff
18:33:38 *** andythenorth has joined #openttd
18:38:59 <DorpsGek_II> [OpenTTD/OpenTTD] kiwitreekor updated pull request #7575: Feature: Add industry production graph https://git.io/fjc3i
18:43:55 <DorpsGek_II> [OpenTTD/OpenTTD] kiwitreekor commented on pull request #7575: Feature: Add industry production graph https://git.io/fjCxM
18:44:53 *** Progman has joined #openttd
18:46:39 <DorpsGek_II> [OpenTTD/OpenTTD] kiwitreekor updated pull request #7575: Feature: Add industry production graph https://git.io/fjc3i
20:07:45 *** Wormnest has joined #openttd
20:18:26 *** Hazzard has joined #openttd
20:31:41 *** chomwitt has joined #openttd
20:47:52 <andythenorth> vehicle variants? :)
20:53:07 *** gelignite has joined #openttd
21:01:37 <frosch123> wasn't that some concept art by V?
21:07:17 <V453000> it was my concept train set, the first one after nuts
21:07:29 <V453000> obscenely oversized machinery
21:08:01 <frosch123> izded is 5 letters, not suitable for grfid
21:08:18 <frosch123> luckily "grfid" is 5 letters as well
21:11:25 *** ChanServ sets mode: +v tokai
21:17:46 <V453000> BTW from quick tests by 1. removing the fancy drawing logic from wagons (it only applied in super extreme and not necesary case), 2. making a single graphics switcher per engine, therefore removing all the "+ count_veh_id" and removing some count_veh_ids and other switches, I'm pretty much back to the performance that earlier NUTS was getting
21:18:12 <V453000> I guess it could still be better if I did a big rewrite and functionality changes, but for now that's a nice result :)
21:18:30 <V453000> I need to do more code splitting, at the moment I finished 1 engine, but for testing that's representative already
21:18:40 <V453000> a single graphics switch for 1 vehicle is 10k lines though :D
21:20:29 <frosch123> V453000: can you get away with using "position_in_consist"?
21:20:58 <V453000> instead of the count_veh_id?
21:20:59 <frosch123> basically, variables with parameter like "count_veh_id(PARAMETER)" are way more expensive than those without
21:21:18 <V453000> yeah that's what I assumed, and now I have seen the effect of that myself :)
21:21:44 <V453000> but I might do a bigger NUTS rewrite somewhat soon with which I could remove them
21:23:08 <andythenorth> what's the compile time though?
21:23:27 <V453000> don't care, acceptable
21:24:04 <V453000> it's actually decent as the graphics are cached
21:24:14 <V453000> it'll get worse, I guess I'll need to add significant amount of lines
21:24:25 <frosch123> andythenorth: 3 seconds probably
21:26:06 <V453000> I can somehow measure it right?
21:26:48 <V453000> yeah there was some nmlc verbose command
21:26:48 <andythenorth> 30s that's very quick
21:26:56 <frosch123> nmlc --verbosity=4 or something
21:26:58 <andythenorth> yeah verbose=4 or something
21:27:35 <V453000> there's also something for not showing the errors right?
21:27:40 <V453000> white pixels and unreferenced blocks
21:28:49 <frosch123> you can also "tag" sprites which are supposed to contain white and animated colors
21:29:18 <V453000> well I'm not sure right now which ones are actually truly correct :)
21:30:09 <V453000> looks like a pain to do at this point :D
21:30:19 <V453000> hm i guess the verbosity is overridden by --quiet
21:30:45 <frosch123> well, i guess "| grep -vi warning" won't work :p
21:30:59 <V453000> generating actions 3.5
21:31:15 <V453000> collecting real sprites 0.7
21:31:40 <glx> frosch123: in mingw it should ;)
21:31:49 <V453000> so yeah about 30s total
21:37:08 *** snail_UES_ has joined #openttd
21:37:52 <V453000> isn't your compile like 1.4s andy?
21:37:54 <andythenorth> maybe it's time to refactor Horse
21:37:59 <andythenorth> no, it's like 1m10s
21:38:06 <V453000> also I'm about to add a bunch of code so you just wait :D
21:38:17 <andythenorth> I'm considering removing a lot
21:38:46 <andythenorth> frosch123 so providing lots of vehicles sprites for many cargos bloats the spritecache?
21:39:29 *** Supercheese has joined #openttd
21:40:49 <peter1138> 20:20 < frosch123> basically, variables with parameter like "count_veh_id(PARAMETER)" are way more expensive than those without
21:40:59 <peter1138> Hmm, so, persistent storage for vehicles?
21:41:16 <peter1138> Calculate once in the depot, use the result for drawing
21:44:53 <frosch123> i think those variables are mostly used for the visuals
21:45:21 <frosch123> so i think changing ottd to only resolve visible sprites is a better solution
21:45:51 <peter1138> I probably had a patch for that...
21:46:17 <peter1138> But, zoomed out...?
21:46:34 <frosch123> currently it resolves the sprites when vehicles move
21:46:43 <frosch123> instead it should only invalidte them
21:47:08 <frosch123> some issue was, that ottd stored the top-left corner of the sprite in the vehicle-position hash
21:47:27 <frosch123> i can't remember whether i changed that 2 years ago, or whether it was unfinished
21:47:39 <peter1138> How about detecting if tick_counter is used, and if not, caching the resolved spritegroup?
21:47:55 <frosch123> but position hash should use the center x/y position which does not depend on sprite size
21:48:33 <peter1138> Something else about using the sprite size to mark the viewport dirty.
21:48:54 <frosch123> we already have maximum vehicle sprite size computed to determine depot grid size
21:49:06 <frosch123> same can be used to decide whether a veihicle is possibly visible
21:57:05 <peter1138> So we need V453000's slow NUTS and his savegame to test ;p
21:59:00 <peter1138> Hmm, caching the spritegroup I guess is more hassle, loads of other variables.
22:00:08 <frosch123> the realism control freaks already complain about the cached recoloring
22:04:12 <Markk> Hello everyone, how can I force the to save the config I've made in the settings? Just by closing and opening the game again?
22:04:46 <Markk> My computer crashed without me being able to close OpenTTD properly, and the changes I'd done in settings was restored to default again.
22:04:50 <frosch123> pretty sure there is also a console command
22:06:09 <Markk> Any idea on what it could be or how I can find the command? :)
22:07:53 <Markk> It took me about 3 minutes to get the settings to what I want, so it's a real time saver. So thank you frosch123!
22:26:36 * andythenorth wonders if Horse makes OpenTTD slow
22:27:18 <V453000> just build 700 trains :P
22:27:50 <V453000> from my current experience if you're not going apeshit on count_veh_id 20 times on every single 4/8 vehicle, it's probably not too bad
22:29:06 <V453000> Elapsed time: 0h:00m:33s
22:29:11 <V453000> dam, .bat timestamp :D
22:35:13 <andythenorth> yeah I do use that though
23:05:38 <Eddi|zuHause> V453000: well, depends on what you're trying to do, but you could maybe use STORE_TEMP to slightly reduce it? :)
23:30:18 *** andythenorth has left #openttd
continue to next day ⏵