IRC logs for #openttd on OFTC at 2019-05-05
⏴ go to previous day
00:00:07 <michi_cc> "I have a patch" is sometimes usefull :)
00:14:47 <DorpsGek_II> [OpenTTD/OpenTTD] michicc updated pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjngD
00:21:21 <spnda> I as always love the random bluescreens
00:25:06 <peter1138> They're not random.
00:25:17 <peter1138> They're because your hardware is faulty.
00:25:49 <spnda> Just checked Windows Logs and it's some Critical Error of Discord it seems
00:28:44 <milek7> ..discord can't cause bluescreen
00:28:46 <peter1138> No, that would not be causing a bluescreen.
00:28:57 <spnda> But I have checked RAM and GPU, both are fine
00:28:58 <peter1138> A hardware fault could cause Discord to crash though.
00:29:32 <spnda> I'll probably reseat my CPU when my thermal paste arrives
00:30:37 <peter1138> Grr, annoying pain elementals
00:53:19 <DorpsGek_II> [OpenTTD/OpenTTD] YJSoft commented on pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjn2E
01:23:43 <glx> hmm seems fontcache.cpp escaped the "use override specifier" codechanges
01:34:07 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 commented on pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjn2x
02:12:20 *** Supercheese has joined #openttd
02:33:51 <DorpsGek_II> [OpenTTD/OpenTTD] kiwitreekor commented on issue #7562: Feature Request: Add drag functionality to land purchase tool https://git.io/fjnar
02:48:01 <DorpsGek_II> [OpenTTD/OpenTTD] michicc updated pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjngD
02:48:17 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjna6
03:50:54 *** ChanServ sets mode: +v tokai
05:39:41 *** supermop_Home has joined #openttd
06:45:39 *** Supercheese has joined #openttd
07:44:43 *** gareppa has joined #openttd
08:13:34 *** andythenorth has joined #openttd
09:11:52 <andythenorth> Horse 88% complete
09:11:57 <andythenorth> it was 94% but eh
09:20:12 <andythenorth> I am about to make it 85% :P
09:20:21 <andythenorth> I made a new FIRS v4 Steeltown
09:20:25 <andythenorth> needs many new wagons :P
09:28:44 <andythenorth> in UK railway terms, it's unambiguously 'plate wagon'
09:28:51 <andythenorth> probably fails to translate
09:36:21 <DorpsGek_II> [OpenTTD/OpenTTD] Hexus-One updated pull request #7568: Change: Default setting of "Advertise" in server browser to "Yes" https://git.io/fjn0m
09:36:55 <DorpsGek_II> [OpenTTD/OpenTTD] Hexus-One commented on pull request #7568: Change: Default setting of "Advertise" in server browser to "Yes" https://git.io/fjnrl
09:40:04 *** sla_ro|master has joined #openttd
09:43:22 *** stefino has joined #openttd
09:50:10 <Eddi|zuHause> andythenorth: i'd probably file those under "flat wagon"
09:50:51 <andythenorth> germany seems to
09:51:41 <Eddi|zuHause> and wikipedia doesn't seem to know the term "plate wagon"
09:54:39 <andythenorth> 'low-side wagon' is used by model train shops
09:58:11 * andythenorth lost in UIC wagon codes
09:58:35 <nielsm> low-side wagon is pretty clear imo
09:59:11 <nielsm> danish seems like it's called an open wagon, but that may as well be a minerals car without cover
10:00:59 <andythenorth> the low-side ones seem to be classified mostly as flat, because sides are removable
10:01:14 <andythenorth> that's not relevant for newgrf purposes :D
10:07:20 <andythenorth> what to refit them to
10:07:39 <andythenorth> the sides are 1px high, not sure they can carry bulk cargos like coal ;P
10:08:23 <Eddi|zuHause> they can be used for bulk cargos like sand that you use on construction sites (the lower sides makes it easier to access)
10:09:34 <Eddi|zuHause> probably machine parts and stuff
10:12:49 <Eddi|zuHause> that's what i meant
10:13:16 <Eddi|zuHause> that's also what i meant :)
10:14:10 <andythenorth> DB wagon manual says " for transporting bulk materials, such as loose rock, aggregate, and sand, or concentrated loads, like iron and steel products, natural stone, semi-finished products, machinery and vehicles"
10:16:23 <andythenorth> not my taste, but the czech guys can definitely CGI well
10:19:49 *** Progman has joined #openttd
10:37:03 <Eddi|zuHause> but they didn't separate the trolley and tram catenary
10:38:23 <Eddi|zuHause> also, i can't imagine that traffic sign to work right
10:49:56 <Wolf01> Meh, tried to scroll an image again
10:50:37 <Eddi|zuHause> that happens all the time :p
10:51:40 <Wolf01> Also, I figured out after 2 months that I didn't cover the forest with the station catchement area in mashinky
11:49:32 *** arikover has joined #openttd
11:50:28 *** andythenorth_ has joined #openttd
11:52:57 *** arikover has joined #openttd
11:59:54 *** andythenorth_ has joined #openttd
12:23:55 <DorpsGek_II> [OpenTTD/website] auge8472 commented on pull request #85: Change: HTML5-elements and corresponding CSS-rules for the blog article template https://git.io/fjnoS
12:25:55 <DorpsGek_II> [OpenTTD/website] auge8472 commented on pull request #84: Change: semantic HTML-elements for the pages main sections and corresponding CSS-rules https://git.io/fjno9
12:34:33 <DorpsGek_II> [OpenTTD/website] auge8472 commented on pull request #84: Change: semantic HTML-elements for the pages main sections and corresponding CSS-rules https://git.io/fjnod
13:21:29 *** chomwitt has joined #openttd
13:31:45 *** andythenorth has joined #openttd
13:41:59 *** Progman has joined #openttd
13:56:26 <andythenorth> it's not symmetrical :P
14:00:26 <andythenorth> I could make it symmetrical
14:00:56 <Wolf01> Nope, that's another one :P
14:10:18 *** Smedles has joined #openttd
14:17:02 <DorpsGek_II> [OpenTTD/OpenTTD] clzls commented on issue #7567: Chinese(simplified) can not be shown correctly https://git.io/fjn6O
14:30:21 <DorpsGek_II> [OpenTTD/website] auge8472 updated pull request #85: Change: HTML5-elements and corresponding CSS-rules for the blog article template https://git.io/fjmXN
14:43:43 <andythenorth> is someone rebuilding the website then?
14:51:16 <DorpsGek_II> [OpenTTD/OpenTTD] michicc updated pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjngD
14:53:51 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7572: GDI engine for font glyph rendering as a replacement for FreeType (take 2) https://git.io/fjn6y
14:59:22 *** TinoDidriksen has joined #openttd
14:59:53 *** TinoDidriksen is now known as Guest1823
15:00:08 *** Guest1823 is now known as TinoDidriksen
15:30:46 <DorpsGek_II> [OpenTTD/website] auge8472 commented on pull request #85: Change: HTML5-elements and corresponding CSS-rules for the blog article template https://git.io/fjnin
15:50:07 *** HerzogDeXtEr has joined #openttd
15:50:30 *** frosch123 has joined #openttd
16:13:18 *** supermop_Home has joined #openttd
16:18:15 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on pull request #7213: Feature: BFS-based river generator https://git.io/fjnih
16:18:46 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN approved pull request #7568: Change: Default setting of "Advertise" in server browser to "Yes" https://git.io/fjnPe
16:18:55 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN merged pull request #7568: Change: Default setting of "Advertise" in server browser to "Yes" https://git.io/fjn0m
16:21:47 <Samu> change inflation defaults too
16:21:58 <Samu> to off, i think most ppl prefer off, i dunno
16:23:20 *** andythenorth has joined #openttd
16:24:00 *** Progman has joined #openttd
16:29:31 <supermop_Home> hi andythenorth
16:32:04 <andythenorth> yolo is such a WOT blitz thing
16:59:11 <peter1138> Hmm, well, docs, I suppose.
17:03:19 <supermop_Home> so moprv seems like it works ok with master
17:12:03 <peter1138> It's weird how everything in it is disabled by default, though.
17:13:41 <peter1138> Also the sound effect (as least for the steam powered trucks) is a wet splot sound. Very odd :p
17:14:19 *** Wormnest has joined #openttd
17:23:17 <supermop_Home> I have never played with sound so never knew
17:23:58 <supermop_Home> one issue I have now is that capacity is shown twice in buy menu
17:24:05 <supermop_Home> but is correct elsewhere
17:30:37 * andythenorth has gone Horse mad
17:38:16 <peter1138> What do you mean by "shown twice"?
17:57:43 <andythenorth> isn't that just the trailers?
18:04:21 <peter1138> Only happens when cargo filter is enabled
18:05:07 <peter1138> Might just be literally drawing it twice :p
18:05:23 <supermop_Home> anything I can change to fix it?
18:05:53 <supermop_Home> i think i'm just going to put this on bananas as is for now to get more eyes on it
18:12:08 <peter1138> No, I suspect it's an OpenTTD bug.
18:22:57 <supermop_Home> why the hell has mlss been downloaded 300k times
18:23:12 <supermop_Home> who the hell is downloading that broken trash?
18:23:27 <glx> there's a "select all" button
18:24:06 <Eddi|zuHause> supermop_Home: how are people supposed to know it
18:24:15 <Eddi|zuHause> without downloading it?
18:25:59 <supermop_Home> hmm i really need a better name than 'Mop's Generic NRT Road Vehicles'
18:26:53 *** tobias is now known as Guest1836
18:27:16 <peter1138> There isn't a select all button.
18:29:30 <glx> in content download window ?
18:31:03 <Guest1836> hi all, I got trouble installing latest ottd version 1.9.1 on ubuntu 19.04 because the "libiculx60" package is missing... however I'm not sure how to install this dependency. Did anybody encouter this issue and could help me out?
18:33:04 <glx> don't know if it's easy to solve
18:33:51 <glx> openttd depends on an old ICU version
18:34:48 <Guest1836> hmm, yes.. the libculx60 package doesnt seem to be available for ubuntu disco dingo :(
18:36:26 <Guest1836> is it possible to compile this package manually somehow?
18:41:29 <glx> peter1138: oh right "select all" is only when joining a server
18:48:08 <Guest1836> thanks for the info glx
18:49:36 <andythenorth> so intermodal cars refit all cargos?
18:49:45 <andythenorth> and express intermodal cars are limited refits?
18:54:19 *** gelignite has joined #openttd
18:57:32 *** arikover has joined #openttd
19:00:49 <peter1138> glx, well yes, but that only selects what the server has in the save, right?
19:16:25 *** Progman has joined #openttd
19:24:01 <peter1138> There was a fly in my tea :(
19:24:10 <andythenorth> how many carolies?
19:24:43 * _dp_ once found a microusb connector in a tea
19:31:22 <peter1138> I... have eaten too much again though :/
19:31:52 <andythenorth> are we supposed to be doing these bloody docs btw? :)
19:32:14 <andythenorth> if you want me to help say so, but I will just moan a lot
19:32:27 <andythenorth> TB made me do lots of boring stuff and ignored the moaning
20:31:52 *** Supercheese has joined #openttd
20:49:09 *** arikover has joined #openttd
21:09:21 <nielsm> but commenting it out does not affect the bug
21:10:46 <nielsm> actually reading some more code, it only affects string sorting
21:44:57 <andythenorth> I suspect I am destroying nmls spritecache
21:45:03 <andythenorth> nmlc info: 67120 sprites, 48 cached, 0 orphaned, 42603 duplicates, 24469 newly encoded (native)
21:45:14 <andythenorth> seems like a lot of new encodes for a change to a single wagon
21:45:22 <andythenorth> just 1 png actually changed
21:45:50 <andythenorth> the graphics regenerate every time, I wonder if that invalidates the cache
21:49:11 <peter1138> I ruined a lot of nuts today. They were mascerated.
21:49:22 <peter1138> And now they're being digested.
21:49:52 <frosch123> andythenorth: it uses the timestamp of the png files
21:50:06 <frosch123> so, if that changes, nml can't do anything
21:57:30 <andythenorth> frosch123: I could educate my graphics generation though
21:59:22 <andythenorth> reinventing make myself though...terrible idea :P
22:00:17 <frosch123> switch to cmake, just to break devzone compilation even more :)
22:00:45 <andythenorth> diffing all the pngs eh
22:01:42 <frosch123> how long does the nml encoding part take :p
22:02:44 <frosch123> i had a patch to make nml encoding multithreaded, but i did not fancy adding a -j option
22:05:27 <andythenorth> I will time it in a minute :)
22:05:54 <frosch123> there is a command line option to print the time of every nml processing step
22:07:09 <frosch123> maybe --verbosity=4
22:10:57 <frosch123> so, all your encoding takes 10 seconds
22:11:18 <frosch123> if we make it use 8 cores, you would save 8.75 seconds
22:11:24 <andythenorth> even though I blitz the cache
22:11:43 <andythenorth> is it worth not destroying the cache though?
22:11:54 <glx> using 1 big png or many small ones ?
22:12:44 <glx> maybe your deps are broken (knowing how you could fail at writing Makefile ;) )
22:13:01 <LordAro> peter1138: did i win?
22:13:23 <andythenorth> glx: the makefile is blind to the pngs
22:13:31 <andythenorth> it just runs all graphics generation every time
22:14:05 <glx> and the generation touches all the pngs then
22:14:16 <andythenorth> sprites are composited from multiple files, detecting which ones cause which output files to need rebuilt...
22:14:48 <glx> you could generate temp png and compare their checksum
22:15:06 <andythenorth> the generation is fast enough to do that, and then only replace ones that have changed
22:15:24 <andythenorth> I should stop destroying the cache :(
22:16:29 <glx> just need a smarter generation
22:16:48 * frosch123 is amazed by the statistics for iron horse
22:17:34 <frosch123> 1 va2 register, 2 concurrent spritesets, but 25k sprites :p
22:17:48 <andythenorth> I have been considering rewriting it
22:17:55 <andythenorth> to optimise for the compiler
22:17:58 <andythenorth> instead of the author
22:18:10 <andythenorth> currently it's optimised for ease of adding vehicles
22:18:54 <glx> theorically the author should not care about optimising for the compiler
22:19:01 <andythenorth> I need to count how often I compile to check cargo sprites vs how often I check vehicle sprites
22:19:07 <glx> it's the compiler job to optimise stuff
22:19:16 <andythenorth> I might have to move all the cargo sprites into graphics layers
22:19:36 <andythenorth> which are orders of magnitude slower to QA compared to spritesheets
22:19:41 <andythenorth> but don't change very often
22:20:29 <andythenorth> there is a *lot* of repetition of cargo sprites composited into specific vehicles
22:21:10 <glx> I guess compiler should detect that and factorise
22:21:37 <andythenorth> I don't think it can do that for actual graphic
22:23:21 <andythenorth> about 72 or so rows of "wagon composited with cargo"
22:23:58 <andythenorth> it's just very slow to check the results
22:24:08 <andythenorth> have to run a game, and test every cargo
22:24:14 <andythenorth> takes literally hours
22:26:13 <glx> can't use some kind of return base + cargo_offset ?
22:26:50 * glx knows nothing in grf world
22:26:59 <andythenorth> I use layers for pantographs on electric trains
22:27:06 <andythenorth> because they have up / down states
22:27:53 <andythenorth> I know the location points for the pan in the spritesheet, so I (1) composite the train + pan in a spritesheet, for fast checking (2) just use layers in game, and don't need to check them
22:27:59 <andythenorth> could do same for cargos
22:28:12 <andythenorth> draw them all in spritesheet, but use layers in game
22:29:57 <andythenorth> let's test if it's faster :)
22:32:31 <andythenorth> removing the cargos is 0.4s faster on graphics generation
22:33:19 <andythenorth> maybe about 2s faster for nml
22:35:39 <andythenorth> frosch123: 42603 duplicates <- is that significant / worth worrying about?
22:36:29 <frosch123> the grf will be big
22:36:37 <frosch123> nml does not yet skip duplicates in the output
22:37:05 <frosch123> not sure whether ottd would
22:37:46 <frosch123> container 2 format could handle duplicates and share their space. but the support is incomplete
22:39:37 <andythenorth> presumably I could stop generating so many?
22:39:46 <andythenorth> generated code tends to be noisy and redundant
22:40:00 <frosch123> do you know which sprites are duplicate?
22:40:11 <frosch123> closed wagons with loading stages or something?
22:40:21 <andythenorth> something like that
22:40:43 <andythenorth> can I get it to sort them into sets? :P
22:40:45 <frosch123> hmm, i wonder how nml counts them
22:40:56 <frosch123> does it compare the pixel data, or just the source location
22:42:06 <andythenorth> I was assuming it counts all sprites, not just graphics?
22:42:26 <frosch123> no, it's just graphics
22:43:11 <frosch123> those statistics are about the graphics cache
22:44:04 <andythenorth> there might be a lot of redundant sprites due to the poor choices I've made about spriteset templates
22:44:30 <andythenorth> for legacy reasons, I repeat the same 4 spriteset templates for *every* vehicle
22:44:43 <andythenorth> there are about 300 vehicles
22:46:40 <frosch123> @calc 67120 / 300 / 4
22:46:40 <DorpsGek> frosch123: 55.9333333333
22:47:01 <frosch123> 56 spritesets per vehicle? (without orientations)
22:47:57 <frosch123> maybe you also have wagon sprites for pantograph up/down :)
22:48:35 <frosch123> hmm, i guess wagons have a lot of sprites due to the cargos
22:48:39 <andythenorth> some wagons have 70 or 80 spritesets
22:48:45 <frosch123> so, hmm, engines and wagons are very different
22:49:02 <andythenorth> I should try and unpick this
22:55:45 *** Wormnest has joined #openttd
23:12:27 <DorpsGek_II> [OpenTTD/OpenTTD] dorobouNeko opened pull request #7573: Fix #7561: Remove assumption between power and cost https://git.io/fjnyI
23:14:50 * andythenorth removing duplicate spriteset templates
23:14:55 <andythenorth> no difference to timings so far :P
23:21:05 <andythenorth> maybe it cuts about 2s off parsing, but that could be just noise
23:21:18 <frosch123> how much file size?
23:21:36 <andythenorth> oof, I didn't keep a record
23:21:42 <andythenorth> once I have a commit, I'll check
23:29:03 <andythenorth> frosch123: previously 8,606,125 bytes (9.4 MB on disk), now 7,828,566 bytes (7.8 MB on disk)
23:29:19 <andythenorth> for the nml file
23:30:11 <frosch123> but, smaller nml is probably also good :)
23:33:13 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7573: Fix #7561: Remove assumption between power and cost https://git.io/fjnyl
23:33:44 <frosch123> everyone using std::sort does the same mistake
23:33:58 <frosch123> at work and in ottd
23:34:02 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7573: Fix #7561: Remove assumption between power and cost https://git.io/fjny8
23:34:12 <andythenorth> frosch123: resulting grf is identically sized
23:34:25 <andythenorth> no surprise, I've just consolidated spriteset templates
23:34:42 <frosch123> oh, i thought you now had less duplicates
23:34:43 <andythenorth> same size is actually a positive outcome, I didn't break anything :P
23:35:03 <andythenorth> duplicates didn't change yet
23:38:24 <andythenorth> 40 minutes, to save 2s on parse step :D
23:38:37 * andythenorth doesn't want to calculate if that was worth it
23:50:37 <andythenorth> @calc 42885 - 42603
23:50:47 <andythenorth> wow, adding one wagon created 282 duplicates
23:53:10 <DorpsGek_II> [OpenTTD/OpenTTD] dorobouNeko commented on pull request #7573: Fix #7561: Remove assumption between power and cost https://git.io/fjny6
23:53:31 *** Gustavo6046 has joined #openttd
continue to next day ⏵
