Config
Log for #openttd on 25th November 2020:
Times are UTC Toggle Colours
00:20:36  *** HerzogDeXtEr has quit IRC
00:39:47  *** Flygon has joined #openttd
00:50:32  *** gelignite has quit IRC
01:09:04  *** dP has joined #openttd
01:09:04  *** dP is now known as _dp_
02:14:11  *** Execthts has quit IRC
02:15:53  *** Exec has joined #openttd
03:05:59  *** debdog has joined #openttd
03:06:00  *** glx has quit IRC
03:06:22  *** Wormnest has quit IRC
03:09:23  *** D-HUND has quit IRC
04:03:39  *** D-HUND has joined #openttd
04:07:03  *** debdog has quit IRC
05:48:46  *** sla_ro|master has joined #openttd
06:17:29  *** tokai|noir has joined #openttd
06:17:29  *** ChanServ sets mode: +v tokai|noir
06:24:28  *** tokai has quit IRC
06:29:31  *** nielsm has joined #openttd
06:45:33  *** nielsm has quit IRC
06:47:07  *** WormnestAndroid has quit IRC
06:47:20  *** WormnestAndroid has joined #openttd
07:08:09  *** andythenorth has joined #openttd
07:34:58  <andythenorth> topic :P https://twitter.com/rumpl/status/1330811090889027584
07:36:22  *** longtomjr has joined #openttd
07:54:33  *** iSoSyS has joined #openttd
08:53:10  <Eddi|zuHause> 'alias "cd.."="cd .." <-- that's a default setting in my distro
08:54:13  <Eddi|zuHause> also, that one doesn't fit there, as it's not a typo but a backwards compatibility thing. because DOS allowed this
08:59:53  * orudge notices TrueBrain is being busy :D
09:00:18  <TrueBrain> so many passwords to reset :P
09:00:25  <TrueBrain> it is funny how they all act different on the activity
09:07:43  <TrueBrain> ugh, I am happy if we switched mailserver .. assuming the greylisting is smarter on zernebok
09:07:51  <TrueBrain> it sometimes takes FOR EVER for validation emails to arrive
09:09:10  <TrueBrain> lol @ OVH .. it sends an email to change your password .. that is also interesting
09:12:32  <TrueBrain> okay, that was by far the weirdest one yet .. not even "enter password twice"
09:20:16  <TrueBrain> lol ... - Password too long (needs to be less than 31 characters)
09:20:48  <TrueBrain> and ironicly, 31 characters does work
09:21:36  <TrueBrain> seems "less than" has a different meaning in French :)
09:32:44  <TrueBrain> okay, I think that is it :)
09:34:14  <TrueBrain> the only one I would like to reset is info@ .. let me see people who do access at all have access to 1password already :)
09:38:03  <TrueBrain> funny, fonsinchen and Rb also still have the info@ configured :) That leaves me with 1 user that accesses it, of which I do not know the IP .. hmmm :)
09:39:01  <TrueBrain> well, that must be orudge :) DEDUCTION dear Watson :)
09:39:21  <TrueBrain> best ISP name ever
09:40:07  <TrueBrain> orudge: if I reset the password on the new mailserver, does that affect your ability to imap sync it?
09:41:45  <orudge> TrueBrain: as long as I can find out what the passwords are, but if they're in 1pass, should be fine
09:42:05  <TrueBrain> okay, in that case I will reset the account on the new server only :D
09:43:07  <TrueBrain> it is under Socials
09:43:39  <orudge> OK, thanks
09:46:42  <TrueBrain> right, that brings OpenTTD to 2020 in password security too \o/
09:46:51  <orudge> Only taken about 15 years ;)
09:46:53  <TrueBrain> and the bus-factor got reduced :P
10:00:10  <orudge> TrueBrain: do we still need the openttd-cdn.org domain?
10:00:52  <TrueBrain> no, we do not. It does require me to reprovision some certificates, but that will be fine I guess
10:01:31  <orudge> It's good until next September
10:01:39  <orudge> but I can cancel the renewal if we don't need it after that
10:01:48  <TrueBrain> yeah, that is fine
10:03:32  <TrueBrain> certs are valid till May 2021, so I will fix it before that time :)
10:03:56  <TrueBrain> right, so after mail, all that remains on the old VM is supybot ... for some reason I am dreading migrating it :P
10:31:29  <TrueBrain> MySQL 5.7.21 stopped .... things you don't want to see in 2020. Well, the "stopped" part is nice
10:38:54  <LordAro> it's not *that* out of date...
10:38:58  <LordAro> even still in support
10:40:18  <TrueBrain> I am shocked I upgraded in 2018, it seems :P
10:40:55  <TrueBrain> and LordAro , in case you missed your pings: if you PM me a private email address I can forward lordaro@ to, I can give you access to the 1password .. if you like, ofc
10:41:28  <LordAro> ah yes, i'll get round to that at some point...
10:41:33  <TrueBrain> euh, well, after we switched mails, lol .. oops, small print :P
10:41:40  <TrueBrain> cool :) Just wanted to make sure you got the message :P
10:41:45  <TrueBrain> no rush ofc
10:42:23  <TrueBrain> now the question: how long am I going to wait for a reply of the 3 remaining (ex-)developers? hmmm
11:42:00  *** matt21347 has joined #openttd
12:14:34  *** WormnestAndroid has quit IRC
12:15:10  *** WormnestAndroid has joined #openttd
12:21:21  *** Flygon has quit IRC
12:23:10  *** WormnestAndroid has quit IRC
12:24:00  *** WormnestAndroid has joined #openttd
12:34:05  *** arikover has joined #openttd
13:28:11  *** iSoSyS has joined #openttd
13:49:20  <andythenorth> hmm
13:49:49  <andythenorth> every train sprite has to be provided twice, and is triggering the duplicate detector frosch gave me
13:49:57  <andythenorth> this seems less than optimal
13:51:38  <FLHerne> Why does it have to be?
13:57:49  <Eddi|zuHause> it's probably one of those deep fractal issues where you have to zoom out several layers before you can even begin fixing stuff
13:57:50  <andythenorth> to be accurate, every symmetric, non-8/8 sprite has to be provided twice
13:58:00  <andythenorth> it's a known issue
13:58:06  <andythenorth> I'm not sure it's a ticketed issue
13:58:33  <Eddi|zuHause> same sprite, different offsets, yes
13:59:06  <Eddi|zuHause> on the source code side, you can fix that with templates
13:59:15  <andythenorth> oh there is https://github.com/OpenTTD/OpenTTD/pull/6846
13:59:18  <Eddi|zuHause> in the resulting grf, they will still be duplicated
13:59:20  <andythenorth> yes it is documented
14:00:30  <andythenorth> Iron Horse has 53740 duplicates reported
14:00:35  <andythenorth> I'm not sure what % of total that is
14:01:54  <andythenorth> @calc 53740 / 105554
14:01:54  <DorpsGek> andythenorth: 0.509123292343
14:01:59  <andythenorth> yeah about half is dupes
14:02:24  <andythenorth> some of the dupes are due to poor choices by me
14:02:39  <andythenorth> the majority are handling offsets for vehicles < 8/8
14:03:16  <andythenorth> I don't know if removing them would have any measurable impact on filesize or compile time
14:03:24  <andythenorth> but 20MB seems awfully large for a train grf
14:03:32  <andythenorth> and 25s compile time is very long
14:04:29  <Eddi|zuHause> well, it's 50% of realsprites, which is certainly a lot. no mention of ratio between realsprites and pseudosprites, so difficult to tell overall effect
14:04:49  <andythenorth> this is realsprites
14:04:54  <andythenorth> based on counting paths to images
14:06:19  <andythenorth> I did a crude test by cutting the realsprite template from 8 to 4
14:06:29  <andythenorth> [this has 'interesting' effects in game]
14:06:43  <andythenorth> grf size drops from 20.4MB to 14.9MB
14:06:52  <Eddi|zuHause> well, even if you reimplement the offsets to be centered/symmetrical, you need a new backwards compatibility layer
14:07:26  <Eddi|zuHause> could be a grfv9 thing
14:08:19  <andythenorth> the effect on compile time is negligible
14:08:28  <andythenorth> 1s reduction on 24s run time
14:08:44  <andythenorth> but the grf filesize looks about 25% saving?
14:08:49  <Eddi|zuHause> that sounds about right
14:09:29  <andythenorth> I'm concerned about grf use of bandwidth
14:09:34  <Eddi|zuHause> realsprites is mostly a 1:1 read/write operation, no complicated calculations needed
14:09:54  <andythenorth> every time I ship 20MB of grf, shipping 14MB would be better
14:10:32  <andythenorth> I guess in a world where web pages are often 1MB
14:10:37  <Eddi|zuHause> there's a layer of PNG decoding, which has a don't care on sprite size, and depends only on file size
14:10:39  <andythenorth> it's not worth being concerned about
14:11:47  <andythenorth>  I guess my worry is scalability
14:11:56  <andythenorth> adding 2 more rosters gives a 50-80MB grf
14:12:08  <andythenorth> where it could be a 45-60MB grf
14:12:09  *** y2kboy23_ has joined #openttd
14:12:53  <Eddi|zuHause> my computer science brain screams "that's both linear scaling, nothing to worry about"
14:13:13  <andythenorth> an alternative is to publish each roster as a grf
14:13:18  <andythenorth> which the compile can achieve
14:13:26  <andythenorth> maintenance and so forth would be consolidated to one repo
14:13:31  <andythenorth> compile time would be improved
14:13:43  <andythenorth> but then how does Reddit find which Iron Horse grf to install?
14:14:34  <Eddi|zuHause> ironhorse-oceania.grf ironhorse-eurasia.grf
14:14:42  *** iSoSyS has quit IRC
14:15:02  <andythenorth> plausible
14:15:05  <Eddi|zuHause> ironhorse-eastasia.grf
14:15:17  <andythenorth> also the joke is good
14:15:37  <Eddi|zuHause> yes, i spent a total of 20 seconds coming up with that :p
14:16:45  <FLHerne> andythenorth: I think people are much more likely to discover the other rosters as separate grfs
14:16:51  *** y2kboy23 has quit IRC
14:16:54  <FLHerne> They'll be visible in the download list
14:17:12  <FLHerne> (witness how many people don't realize FIRS has more than one economy)
14:17:37  * andythenorth considers removing everything but Steeltown from FIRS
14:17:38  <FLHerne> Also, you're not wasting clients' and BaNaNaS' bandwidth downloading rosters they don't want
14:18:08  <andythenorth> the other objection is that I might be tempted to make different parameters for different grfs :P
14:18:17  <andythenorth> which lacks discipline about design choices
14:18:26  <Eddi|zuHause> FLHerne: when have you last looked at the fraction of bandwidth from "just download everything on bananas" people?
14:18:43  <andythenorth> hmm, there would be some shenanigans around versions
14:18:51  <andythenorth> I would really not split the repo and codebase
14:19:05  <andythenorth> so the revisions would all march in lockstep
14:19:26  <andythenorth> and the tags
14:19:49  <andythenorth> I could have a core repo, and add it to each roster as a remote?
14:19:56  <andythenorth> seems faff
14:20:05  <Eddi|zuHause> you'd have to release all grfs combined when making a new tag, which has some ... corner cases
14:20:27  <andythenorth> min. compatible version could be handled ok
14:20:37  <andythenorth> but tags hmm
14:22:01  <andythenorth> oof can grfs have dependencies, like the AI libs?
14:29:59  <Eddi|zuHause> you can probably set them in bananas
14:33:01  *** nielsm has joined #openttd
14:33:11  <arikover> Yes you can set them in Bananas (That's what I did for the NARS Passenger addon, dependency NARS 2.51): https://bananas.openttd.org/package/newgrf/ac1d0101/2016-05-28T21:04:49+00:00
14:35:56  <TrueBrain> the day andythenorth finds out that you can have GRFs depend on each other :P Funny, we can even learn andythenorth new tricks :D
14:41:39  <andythenorth> TrueBrain I know almost nothing about grfs
14:41:49  <Eddi|zuHause> TrueBrain: i'm not convinced that is actually a good thing :p
14:41:58  <andythenorth> fortunately we have irc :P
14:42:11  <TrueBrain> stackoverflow ftw .. wait .. is there a NewGRF section on stackoverflow?
14:42:40  <andythenorth> I could put all the trains in a core grf, disabled
14:42:49  <andythenorth> then release grfs that depend on core
14:42:52  <andythenorth> and turn trains on/off
14:43:28  <Eddi|zuHause> how does solve the "you need to download 50MB of trains you don't use" part?
14:43:34  <Eddi|zuHause> +that
14:44:01  <andythenorth> don't 'people' just download everything anyway?
14:44:09  <andythenorth> and no it doesn't solve it
14:44:12  <andythenorth> oof
14:45:26  <Timberwolf> Speaking of sprite duplication, is there a way to have a station layout which combines e.g. a platform sprite that has 5 load states and a roof (or other object) which only has 1 load state?
14:47:22  <Eddi|zuHause> isn't that a thing sprite composition should be able to do?
14:47:54  <Timberwolf> Interesting, I didn't consider composition.
14:50:34  <Timberwolf> Is it supported for stations? I can only find it referenced in Action2/Vehicles.
14:51:37  <Eddi|zuHause> i honestly don't know... station spec was always a bit ... weird
14:53:25  <andythenorth> #rewrite etc
15:12:40  <Eddi|zuHause> last time i brought this up, frosch ranted at me about features that stations have that objects can't do, like repeating by drag&drop
15:19:05  *** Laedek has quit IRC
15:56:45  *** y2kboy23_ has quit IRC
15:57:50  *** y2kboy23 has joined #openttd
16:00:33  <TrueBrain> should we use IPFS for BaNaNaS? :P
16:02:19  <FLHerne> No, https://github.com/philipl/pifs
16:02:25  <TrueBrain> lol .. AWS can be fun:
16:02:36  <TrueBrain> 8:56 AM PST: Kinesis is experiencing increased API errors in the US-EAST-1 Region. This is also causing issues with (long list of services). This issue has also affected our ability to post updates to the Service Health Dashboard.
16:03:58  <TrueBrain> FLHerne: why stop at Pi? :D
16:06:39  * andythenorth has train questions
16:07:54  *** Progman has joined #openttd
16:08:09  <LordAro> andythenorth: trains go choo
16:08:15  <LordAro> and occasionally choo choo
16:08:26  <andythenorth> FLHerne containers as boxes (seacans) in most cases, or show flatracks with cargo sprites, for appropriate cargos (like steel coils)?
16:08:45  <andythenorth> I already did bulktainers and so on
16:12:40  <andythenorth> https://www.europages.com/filestore/opt/product/3e/ca/FlatRackContainer_8fa21f6f.jpg
16:19:37  <FLHerne> andythenorth: I think custom containers would be neat
16:20:20  <andythenorth> I have the sprites
16:20:25  <andythenorth> I just need to write a compositor
16:20:38  <FLHerne> Never enough tooling :p
16:21:20  <andythenorth> pipe flatrack https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQb6kaMKLTWck5QrpqmqfWhe4URzEPXgGL98Q&usqp=CAU
16:23:20  <FLHerne> andythenorth: Are you going to do the little BR guard containers so that each wagon can carry 6 pax?
16:23:31  <andythenorth> tempting
16:24:20  <andythenorth> https://brickset.com/sets/4559-1/Cargo-Railway ?
16:25:10  *** Progman has quit IRC
16:26:54  <Eddi|zuHause> FLHerne: that github link is great, except for when he forgets to mention that storing the position somewhere eats into the "100% compression"
16:28:05  <FLHerne> Well, yeah
16:28:48  <FLHerne> I think it has a compression ratio considerably < 1 for almost any data
16:30:19  <Eddi|zuHause> or... almost no data
16:31:11  <Eddi|zuHause> for any compression algorithm, the fraction of data that can be significantly compressed tends towards 0
16:31:39  <FLHerne> I'm not sure what you're trying to say
16:32:01  <Eddi|zuHause> the trick to compression algorithms is to find sets of data that overlap heavily with the data that reasonably will be used as input
16:33:26  <FLHerne> Yes
16:34:41  <FLHerne> In this case, not only is most data not similar to early digits of π, the way it processes input byte-wise means it even sucks for data that *is* like π
16:35:23  <FLHerne> So it's useless for practically any conceivable input
16:35:32  <Eddi|zuHause> maybe i'm the one that is confused, what you meant by "compression ratio < 1"
16:36:29  <FLHerne> < 1 is "better off not doing it" levels
16:37:02  <Eddi|zuHause> maybe you meant <0
16:37:12  <FLHerne> https://en.wikipedia.org/wiki/Data_compression_ratio
16:38:28  <Eddi|zuHause> i was starting from the claim of "100% compression", which would imply 1-compressed/uncompressed
16:41:10  <FLHerne> Oh, that's a slightly odd way to put it
16:42:20  <Eddi|zuHause> that's like the 2nd paragraph of the page you linked
16:42:24  <FLHerne> Then yes, it achieves <0% compression for all plausible inputs if you count the metadata
16:45:19  <Eddi|zuHause> someone probably studied that properly, instead of making wild guesses :)
16:46:53  <FLHerne> ?
16:48:55  <FLHerne> "compression ratio considerably < 1" is a correct statement, just a different way of phrasing it from the weird one they used
16:55:59  <Eddi|zuHause> no, i mean that you have no proof for that statement
17:00:35  <FLHerne> I read about it when it came out, so I know that someone studied it properly ;-)
17:08:32  *** Samu has joined #openttd
17:37:33  *** Wormnest has joined #openttd
17:48:07  *** nielsm is now known as Guest6857
17:51:55  <DorpsGek_III> [OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master https://git.io/Jkyi2
17:51:55  <DorpsGek_III>   - Update: Translations from eints (by translators)
17:53:11  *** Progman has joined #openttd
17:53:33  *** Guest6857 has quit IRC
18:07:14  *** otetede has joined #openttd
18:12:53  *** frosch123 has joined #openttd
18:21:31  *** Wolf01 has joined #openttd
18:23:25  *** gelignite has joined #openttd
18:23:51  <frosch123> TrueBrain: oh no, we got hacked!
18:24:31  <andythenorth> oh it's like the theme of my day recurring
18:24:35  <andythenorth> oof
18:24:59  <frosch123> search hacked by duckduckgo - whenever i try to search the wiki i get redirected to duckduckgo.
18:25:24  *** HerzogDeXtEr has joined #openttd
18:25:26  <andythenorth> rude
18:27:00  <Wolf01> Ha! So it isn't like that by design XD
18:28:09  <TrueBrain> Do I want to know what you are smoking frosch123?
18:29:10  <frosch123> you are the dealer
18:29:14  <TrueBrain> owh, info@ ... I could have guessed :P
18:29:46  <TrueBrain> https://pasteboard.co/JC1ZY1i.png
18:29:48  <TrueBrain> to share with the group
18:30:37  <Eddi|zuHause> these evil hackers nowadays. they stop at NOTHIGN!
18:30:47  <andythenorth> that would be trivially addressed
18:30:59  <frosch123> TrueBrain: i am sure the same hacker that installed duckduckgo also did reset all passwords
18:31:02  <andythenorth> Search: [input] [Go] (uses DuckDuckGo)
18:31:05  <andythenorth> wat?
18:31:09  <andythenorth> all the passwords also got reset?
18:31:11  <Eddi|zuHause> they must be gathering at illegal darknet sites like this "Internet Relay Chat"
18:31:14  <andythenorth> definitely escalate that
18:31:17  <andythenorth> it's a sign of breach
18:32:11  <andythenorth> longtomjr o_O Horse Wagons fork? https://www.reddit.com/r/openttd/comments/k0weho/good_freight_wagon_gfx_set/
18:32:20  <andythenorth> Iron [something]?
18:32:30  <TrueBrain> I send that dude a nice reply :)
18:32:33  <TrueBrain> full of xmas cheers!
18:32:44  <andythenorth> I should get a reddit account
18:32:50  <andythenorth> predict how well that would go?
18:32:59  <TrueBrain> I really appreciate the dude emailing us btw, that is really awesome (no sarcasm; that is awesome)
18:33:09  <andythenorth> oh no, FIRS is broken :( https://www.reddit.com/r/openttd/comments/k0vlvi/firs_question_recyclable_industry_chain_suddenly/
18:33:13  <andythenorth> Suddenly also
18:33:55  <TrueBrain> seriously, how cute is that .. mailing someone the search is hacked :D
18:34:02  <TrueBrain> this made my day :)
18:35:39  <LordAro> frosch123: cute
18:43:53  <TrueBrain> I cannot believe someone emailed that :) ITS SO FLUFFY
18:45:21  *** Samu has quit IRC
18:46:01  <TrueBrain> (and for those that don't get the reference: https://www.youtube.com/watch?v=hCEVTcL1bJ0 )
18:55:50  *** tokai has joined #openttd
18:55:50  *** ChanServ sets mode: +v tokai
19:02:47  *** tokai|noir has quit IRC
19:10:06  <andythenorth> hmm
19:10:15  <andythenorth> all containers go one way in Horse
19:10:23  <andythenorth> empty trains don't show any
19:10:27  <andythenorth> I guess they scrap them :P
19:13:13  <Timberwolf> Make it like Factorio barrelled liquids. Industries accepting containers will stop production if the empty containers aren't taken away.
19:13:44  <andythenorth> o_O
19:14:09  <andythenorth> deliver them to other places
19:17:54  <frosch123> funnily, in an ottd economy empty containers are like a secondary cargo. so they would pay more to make players transport them
19:18:56  <Wolf01> Make the price negative, so you pay to transport them
19:27:43  <arikover> Empty containers are in fact transparent in ottd.
19:27:50  <andythenorth> yes!
19:28:00  <andythenorth> problem solved :D
19:30:01  <Timberwolf> I think they should stay single use, so the sink industry is "container incenerator" and, post 2030, "container vapourising plant"
19:38:58  *** longtomjr has quit IRC
20:00:27  *** otetede has quit IRC
20:03:07  <TrueBrain> I got a reply from the "you are hacked" dude :) "it is still external and no external search engine can be as good as an internal one".
20:03:40  <TrueBrain> "not invented here" syndrome is BAAAACCCCKKKK
20:05:29  <andythenorth> hang on I can fix it
20:05:54  <andythenorth> "the site owner can't control the indexing, algorithm, or SERPs presentation of an external search engine"
20:06:11  <andythenorth> on the other hand, how much time do you want to spend on ngrams?
20:06:18  <andythenorth> or non-euclidean vector spaces?
20:06:18  <TrueBrain> well, I guess I spend the next few weeks writing my own search engine
20:06:22  <TrueBrain> maybe it can be better than google?
20:06:30  <andythenorth> they're quite fun
20:06:42  <andythenorth> we wrote one once
20:06:53  <andythenorth> to search YouTube
20:06:55  <andythenorth> it found cats
20:06:59  <TrueBrain> everyone write their own search engine exactly once :)
20:07:05  <DorpsGek> what about me? who cares about me?
20:07:14  <TrueBrain> NOBODY! Literally, NOBODY
20:07:16  <andythenorth> OMG, it lives
20:07:18  <TrueBrain> I almost turned you off today
20:07:22  <TrueBrain> just .. because I didn't care about you
20:07:48  <frosch123> it spies on everyone
20:08:28  <TrueBrain> I am going to be called into HR now am I not? :s
20:08:50  <frosch123> who is our hr?
20:08:51  <andythenorth> you are HR no?
20:09:05  <andythenorth> by default it falls to Senior Leadership Team
20:09:16  <frosch123> i did not receive any paycheck the last 12 years, i want to talk to controlling
20:09:21  *** glx has joined #openttd
20:09:21  *** ChanServ sets mode: +v glx
20:09:27  <andythenorth> frosch123 make an appointment
20:09:32  <andythenorth> use the calendar system
20:09:57  <frosch123> we have none :o
20:10:16  <andythenorth> also, should I try and fix the 53k duplicate realsprites in Iron Horse or not (50% of realsprites are dups)
20:11:16  <frosch123> yes, delete the reverse vehicle feature
20:11:35  <andythenorth> you mean the Livery API?
20:11:38  * _dp_ actually worked on search engine that was somewhat better than google
20:11:47  <_dp_> hi everyone btw xD
20:11:59  <andythenorth> hi
20:12:06  <frosch123> how do you measure betterness?
20:13:41  <_dp_> it did much better language processing
20:14:00  <_dp_> even nowadays google barely does any afaict
20:14:15  <TrueBrain> it is not difficult I think to do it better on some places
20:14:21  <TrueBrain> pretty sure it is difficult to do it better on all :D
20:14:38  <TrueBrain> but yeah, I am waiting for someone to write a better google
20:14:41  <TrueBrain> it is about time it happens
20:14:53  <andythenorth> Bing!
20:14:59  <TrueBrain> well, clearly not :)
20:15:05  <TrueBrain> so _dp_ , did you make it into a SaaS? :D
20:15:20  <andythenorth> what's the other one, the language processing one
20:15:22  * andythenorth forgot
20:16:12  <frosch123> deepl?
20:17:05  <frosch123> it shares a lot of letters with dp, he contributes 40% of them
20:18:13  <andythenorth> wolfram
20:18:17  <andythenorth> I haven't tried it
20:18:40  <frosch123> it's mostly mathematica aas, isn't it?
20:19:34  <andythenorth> dunno
20:19:41  <andythenorth> it was all hype some years ago
20:19:50  <andythenorth> it seems very niche
20:20:36  <frosch123> so was watson
20:21:48  <milek7> eh, b2 is really slow
20:21:55  <milek7> it is really necessary to download from it in parallel
20:22:17  <frosch123> is it using smb? :p
20:24:31  <andythenorth> how about an Action D or something in which grf declares transform to apply to reversed sprites?
20:24:40  <andythenorth> because more complexity is obviously the solution?
20:25:14  <frosch123> i had a multi-vehicle-sprite draft spec that centered sprites in all orientations
20:25:56  <andythenorth> that would be too sensible
20:25:57  <frosch123> mostly meant for long wagons, it also works for a single vehicle
20:26:14  <frosch123> i still did not figure out gist
20:26:28  <frosch123> i think a normal repo is easier to use
20:30:24  <andythenorth> hmm, can we adjust the palette?
20:30:32  <andythenorth> I need a set of 'shiny' colours
20:30:35  <andythenorth> for aluminium cargo
20:30:47  <frosch123> you need ottd 1.2 for that
20:30:50  <milek7> or maybe not, and it's just my server clogged up
20:30:50  <andythenorth> we need RGBAR
20:31:20  <frosch123> what's that second R?
20:31:28  <frosch123> we have RGBAM though
20:32:18  <andythenorth> Radiosity
20:32:36  <andythenorth> it requires your device to have a camera or colour sensor
20:32:55  <andythenorth> then shiny pixels can reflect your face back at you
20:33:34  <andythenorth> I guess the reflections should be of the game world though?
20:33:42  <andythenorth> the sun and clouds and stuff?
20:33:51  * andythenorth considers animating fake sky into aluminium ingots
20:34:02  <frosch123> ah factorio just got that with the release two days ago
20:34:16  <andythenorth> always showing us where we should go next :)
20:34:28  <andythenorth> we shipped conveyors already?
20:35:21  <frosch123> https://cdn.factorio.com/assets/img/blog/fff-363-night-lighting-after.png <- they have glowing cargos now
20:35:36  <frosch123> aw, they even have a "cdn." domain
20:36:16  <andythenorth> wow
20:36:20  * andythenorth likes it
20:36:40  <FLHerne> andythenorth: https://docs.google.com/document/d/1hKdqwFOfqPDgon-USFzZYsKWMgbHpAyJiFnjYQ7gVdA/edit
20:36:42  <andythenorth> I mostly don't like factorio style, only V's stuff
20:36:50  <andythenorth> but that lit style is great
20:38:01  <andythenorth> FLHerne is that new? I feel like I've read it before :)
20:38:50  <andythenorth> it's well written
20:39:44  <FLHerne> I don't think I've seen it before
20:41:31  <frosch123> my old company bought another company with that problem
20:41:51  <frosch123> the devs were constantly talking: we rewrote everything to modern in 3 years
20:42:02  <frosch123> and the support people constantly said: nothing works
20:42:25  <frosch123> then the money was gone, and they were bought
21:17:11  * andythenorth had an idea
21:17:29  <andythenorth> some cargos have more than one sprite they could use
21:17:36  <andythenorth> what if we could let the player choose the type?
21:18:06  <frosch123> sounds cool, let's call them cargo subtypes
21:18:19  <andythenorth> no that conflates with the livery API
21:18:31  <andythenorth> Vehicle Sprite Choices?
21:19:02  <andythenorth> nah, it needs good bacronym
21:19:23  <frosch123> can we add a settings for individual pixels? maximum customizable
21:19:29  <andythenorth> I already have those
21:19:40  <andythenorth> you have to use external plug-ins though
21:20:17  <andythenorth> the Photoshop add-on to OpenTTD is surprisingly popular
21:20:22  <andythenorth> people are using it for other things
21:27:30  <FLHerne> andythenorth: That's obviously the use-case for cargo subtypes
21:27:37  <FLHerne> Liveries using it is kind of a hack :p
21:36:03  <orudge> Just compiled OpenTTD in what must have been less than a minute on the M1. This thing does appear to be fast!
21:36:26  <andythenorth> orudge :o
21:36:29  <andythenorth> but does it blend?
21:37:02  <michi_cc> Only once :)
21:38:22  * andythenorth wonders if all OpenTTD mac performance issues are resolved
21:38:53  <frosch123> compiling and graphics are probably very different metrics
21:39:21  <frosch123> orudge: what does the fps window say?
21:42:00  <andythenorth> i9 mac compiles OpenTTD in about 1m15s
21:42:08  <andythenorth> using 17 threads
21:42:22  <andythenorth> pegs the CPU to mx
21:44:50  <frosch123> why is it that slow?
21:45:04  <frosch123> my 10 year old i5 with 4 cores takes 1m49s
21:45:29  <andythenorth> mac intel laptops under-perform horribly
21:45:34  <andythenorth> they're really poor
21:45:49  <frosch123> in theory you should be 8 times faster than me
21:46:05  <frosch123> 4 times the threads, twice the clock
21:46:38  <dwfreed> not everything can be parallelized, and clock speed itself is a terrible metric anymore
21:46:44  <andythenorth> 4 times the sticker price
21:46:49  <frosch123> give it some seconds for the linking, so 15s should be the goal?
21:46:49  <andythenorth> 4 times the marketing hype
21:47:00  <andythenorth> 4 times the thermal throttling
21:47:11  <andythenorth> 4 times the buying cheap under-specced intel parts
21:47:14  <dwfreed> macs have terrible cooling
21:47:20  <andythenorth> yup
21:47:27  <dwfreed> intel procs have terrible thermals
21:47:35  <andythenorth> the small macs do fine for what they are
21:47:57  <andythenorth> but I bought the big one, which has a horrible CPU stuffed into a tiny enclosure
21:48:59  *** jottyfan has joined #openttd
21:50:27  *** jottyfan has quit IRC
21:54:55  *** jottyfan has joined #openttd
21:54:55  <andythenorth> to be fair, it is overall the best computer I have ever owned
21:55:01  <andythenorth> but not for performance reasons :P
21:56:09  *** frosch123 has quit IRC
21:59:04  *** sla_ro|master has quit IRC
21:59:13  *** jottyfan has quit IRC
22:03:27  *** andythenorth has quit IRC
22:05:23  *** jottyfan has joined #openttd
22:15:44  <orudge> 44 seconds to build. Haven't checked fps yet but will give it a
22:15:51  <orudge> shot tomorrow I expect
22:19:07  *** arikover has quit IRC
22:23:05  *** gelignite has quit IRC
22:24:32  <LordAro> 44 seconds to build is fast
22:35:30  *** jottyfan has quit IRC
22:40:35  <LordAro> curious, my msys build has decided to pick up xaudio
22:40:47  <LordAro> but then fails due to missing XAUDIO2_DLL_A
22:40:59  <LordAro> glx: don't suppose you've seen this before?
22:41:49  <LordAro> oh hang on, msys terminal, not mingw terminal
22:41:53  <LordAro> weird.
22:41:55  <glx> it's not supposed to find xaudio IIRC
22:42:10  <LordAro> there's an xaudio2.h file, it's just... different to what's expected
22:42:47  *** Wolf01 has quit IRC
22:43:36  <LordAro> /mingw64/x86_64-w64-mingw32/include/xaudio2.h
22:47:10  <LordAro> nope, mingw still picking this thing up
22:47:29  <LordAro> i'll amend the test script to make sure XAUDIO2_DLL_A exists
22:47:46  <LordAro> dunno why i've not seen this before, i've not updated in ages
22:49:38  <glx> indeed xaudio detection fails in my msys2/mingw
22:49:52  <glx> -- Performing Test XAUDIO2_FOUND -- Performing Test XAUDIO2_FOUND - Failed
22:50:05  <LordAro> wonder how i've managed to acquire that file
22:50:09  <LordAro> must have installed something...
22:52:02  <LordAro> /mingw64/x86_64-w64-mingw32/include/xaudio2.h is owned by mingw-w64-x86_64-headers-git 8.0.0.6001.98dad1fe-1
22:52:09  <LordAro> that's a fairly basic package...
22:57:57  *** HerzogDeXtEr has quit IRC
23:03:44  *** Laedek has joined #openttd
23:12:57  *** matt21347 has quit IRC
23:39:36  <DorpsGek_III> [OpenTTD/OpenTTD] LordAro opened pull request #8345: MSYS XAudio build fix https://git.io/JkSqF
23:56:52  *** Progman has quit IRC

Powered by YARRSTE version: svn-trunk