IRC logs for #openttd on OFTC at 2020-03-03
⏴ go to previous day
00:18:20 *** tokai|noir has joined #openttd
00:18:20 *** ChanServ sets mode: +v tokai|noir
01:05:56 <DorpsGek_III> [OpenTTD/OpenTTD] ya-isakov commented on issue #8029: Black screen on SDL2-enabled OpenTTD, Sway and SDL_VIDEODRIVER=wayland https://git.io/Jv23P
02:34:28 *** Flygon_ has joined #openttd
03:16:22 *** Wormnest__ has joined #openttd
05:47:58 *** WormnestAndroid has joined #openttd
08:52:33 *** gelignite has joined #openttd
09:14:03 *** ChanServ sets mode: +v tokai
11:13:53 *** Smedles has joined #openttd
13:22:31 *** Laedek_ has joined #openttd
13:24:42 *** Smedles has joined #openttd
13:30:17 *** Flygon_ has joined #openttd
13:47:19 *** gelignite has joined #openttd
15:45:33 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 commented on issue #8029: Black screen on SDL2-enabled OpenTTD, Sway and SDL_VIDEODRIVER=wayland https://git.io/Jv23P
16:44:40 *** sla_ro|master has joined #openttd
17:15:09 *** Wormnest__ has joined #openttd
18:18:13 <Samu> hmm dredging site near a coast, and an AI trying to load sand via trains
18:18:45 <Samu> it's not working, no sand goes to the station
18:19:07 <Samu> serve_neutral_industries setting
18:19:27 <Samu> working as intended, but breaking ais
18:25:56 *** Smedles has joined #openttd
18:29:47 *** Progman has joined #openttd
18:48:02 <Samu> KrakenAI, AIAI, NoCAB are failing
18:52:08 <Samu> i wonder now why are thei failing :p
18:56:17 *** tokai|noir has joined #openttd
18:56:17 *** ChanServ sets mode: +v tokai|noir
19:17:24 *** andythenorth has joined #openttd
19:19:22 <Samu> EA is spamming me with emails
19:20:35 <Samu> You've been a member of Answers HQ for six years!
19:20:46 <Samu> really? i don't even know wtf this is
19:43:44 *** frosch123 has joined #openttd
19:46:05 <Eddi|zuHause> taxis seem like the worst mode of transport in C:S
19:47:28 <andythenorth> web based train newgrf maker
19:47:43 <andythenorth> step 1: teach openttd to compile grf from remote GH repo
19:47:50 <Eddi|zuHause> i'm fairly sure you had this idea before
19:47:59 <andythenorth> it was a TERRIBLE idea last time
19:48:02 <andythenorth> this time it is good
19:48:21 <Wolf01> Port the compiler to javascript
19:49:08 <Eddi|zuHause> i've got this offer to become a PHP developer for a bank... let's say i'm... hesistant
19:49:49 <andythenorth> so many reasons to question that
19:50:17 <andythenorth> Wolf01: wasm compiler
19:50:21 <andythenorth> wasm openttd in browser
19:50:36 <andythenorth> two tabs, one edits train value, other reloads it in openttd live
19:51:02 <andythenorth> we could put the editor directly into openttd, but I think that would be a TERRIBLE IDEA
19:51:16 <Wolf01> Web based OTTD is real
19:52:19 <andythenorth> against all probability
19:52:30 <andythenorth> we could put a GitHub client into OpenTTD
19:52:35 <andythenorth> then edit vehicles in place
19:52:37 <andythenorth> then commit back
19:52:52 <Wolf01> THAT is a terrible idea
19:53:02 <andythenorth> we could put a GH client cross-compiled into WASM into a browser in OpenTTD in WASM
19:55:17 <FLHerne> andythenorth: Your ideas are steadily getting worse
19:56:47 <frosch123> he watched the METAL video too often
20:12:40 <LordAro> it's the sort of bad idea where you just laugh at it, rather than try to explain why it's so bad
20:14:44 <TrueBrain> But with enough bad ideas, there is a single brilliant one :)
20:15:53 <andythenorth> the WASM future will happen
20:16:07 <andythenorth> one day, we'll even run JS on servers :P
20:16:31 <TrueBrain> still dreading the day we have an OS running in a browser
20:16:36 <TrueBrain> wait .. that already happened 10 years ago
20:17:05 <andythenorth> when I learnt Macromedia Director, the first thing I thought was 'can I make the GUI part of the OS?'
20:17:25 <andythenorth> then I made a slideshow instead :P
20:18:17 <andythenorth> making all the window / file manager part of an OS would be at least twice as much work
20:18:31 <TrueBrain> at least, that is correct
20:20:16 <andythenorth> I miss cross-platform fully graphical runtimes :(
20:20:45 <andythenorth> I was going to learn Unity, but then it turned out to be broken
20:21:00 <andythenorth> oh there was also Silverlight
20:21:38 <andythenorth> eh wat, Unity still exists?
20:21:45 <andythenorth> I thought it died under crippling CVEs
20:21:51 <TrueBrain> and it is kicking ass
20:22:58 <andythenorth> oh it was the browser plugin that got banned
20:25:31 <TrueBrain> right, time to write parsers for the Base* types
20:30:37 <andythenorth> extra train info for Horse?
20:31:34 <TrueBrain> what-ever you feel like ;)
20:35:02 <frosch123> maybe buy more flour, sugar and rice?
20:47:31 <Eddi|zuHause> have you been to a supermarket lately? :p
20:48:44 *** Eddi|zuHause is now known as Eddi|zuHause2
20:48:47 *** Eddi|zuHause2 is now known as Eddi|zuHause3
20:48:47 <frosch123> shopping is entertaining, whenever there is an empty shelf, you check what is sold there normally, and then wonder who tf would stockpile that
20:51:11 <TrueBrain> "Exception: option metadata:blitter is missing" <- hmmmmm
20:51:21 <TrueBrain> seems not all base graphics have that entry :D
20:52:01 <frosch123> i wonder whether it is even used :p
20:52:03 <TrueBrain> musa made it mandatory :)
20:52:14 <frosch123> i thought that kind of stuff is auto-detected
20:52:38 <TrueBrain> I dunno .. just reading implementations :)
20:52:53 <TrueBrain> "Exception: option origin:default is missing" <- seems that is optional too
20:52:55 <frosch123> hmm, ah, there was some logic depending on "some sprites are 32bpp" vs "most sprites are 32bpp"
20:53:37 <TrueBrain> seems md5 of base graphics does not depend on content of .obg file
20:53:45 <TrueBrain> so possibly this is worth cleaning
20:54:38 <TrueBrain> I don't get what [origin] does tbh :P
20:54:54 <frosch123> it is displayed when the md5sums are incorrect
20:55:03 <TrueBrain> ah .. okay, that makes sense
20:55:10 <TrueBrain> and I guess it used to be that you had to tell that per file
20:55:15 <TrueBrain> and now you can also set a "default"
20:55:37 <frosch123> ottd loads the baseset with most files present, and dispalys a message when one file is missing or broken
20:56:03 <TrueBrain> so musa is a lot more strict
20:56:06 <TrueBrain> I guess that is not a bad thing
20:56:22 <frosch123> TrueBrain: the "per file" is for original graphics, "openttd.grf" is "redownload ottd", rest is "get all files from your ttd cd"
20:56:52 <TrueBrain> yeah .. but 2 base graphics we have on file doesn't define a "default", but do for each file it contains
20:56:52 <frosch123> so, probably not used by any bananas set :)
20:57:11 <TrueBrain> but those are older non-active ones
20:57:16 <TrueBrain> so I guess it is fine to just leave them or what-ever
20:58:17 <frosch123> it's probably hard to trigger those messages, users would have to modify the tars they download
21:00:39 <TrueBrain> when will OpenGFX and co be bumped to 1.0?
21:00:43 <TrueBrain> as I guess they are done?
21:01:36 <Samu> I found a bug with AITileList_IndustryProducing
21:01:50 <frosch123> when people were working on it, it was not considered done
21:02:22 <Samu> if serve_neutral_industries is disabled, this still returns tiles
21:02:31 <Samu> it should return no tile
21:02:41 <frosch123> just noone is interested in finishing the remaining draft sprites
21:03:33 <frosch123> no idea, i think rb was the only contributor
21:03:47 <TrueBrain> I think this is a nielsm to wrap up to 1.0 if possible :)
21:03:58 <frosch123> maybe he hoped someone would do a real soundset somewhen :)
21:05:00 <frosch123> last year there was some dude on the forums mailing original authors of opensfx, and asking them for a more free license
21:05:43 <frosch123> they got some responses, but it's still not debian core
21:06:13 <frosch123> anyway, why do you care about version numbers?
21:06:28 <TrueBrain> was just annoying me, to see < 1.0, for something that is not moving :)
21:06:57 <Samu> GetStationCoverageRadius should be deprecated
21:07:04 <frosch123> i guess you can name nml 1.0 when you implement stations :p
21:07:11 <Samu> but that would break many AIs that use it
21:07:13 <frosch123> but the rest has no real agenda
21:08:14 <frosch123> i still don't get why people invest so much time into music sets
21:08:35 <TrueBrain> people just fancy it, I guss
21:08:52 <frosch123> unless a game can play context-specific music matching the game state, external players are always superior
21:09:22 <TrueBrain> Exception: the short name is not 4 long
21:09:36 <frosch123> what did old bananas do?
21:09:41 <TrueBrain> "name = Holiday Island"
21:09:46 <TrueBrain> how is this parsed in OpenTTD ...
21:10:05 <TrueBrain> all Base Music sets have ".mid", but 1 has ".gm"
21:10:39 <TrueBrain> 4 base music sets have wrong "shortnames"
21:10:51 <TrueBrain> shortname = Holiday
21:11:19 <TrueBrain> and yes, we have shortname and ShortName and short_name
21:11:32 <frosch123> i can select those sets in-game
21:11:45 <TrueBrain> only first 4 chars are used
21:11:51 <TrueBrain> so the error is correct :) But it is harmless
21:12:52 <frosch123> "default = Holiday"
21:13:06 <frosch123> someone entered "holiday" into every field they did not understand :)
21:16:18 <TrueBrain> Exception: option metadata:description.da_da set but not expected
21:16:21 <TrueBrain> ah, you can also translate those
21:17:28 <TrueBrain> dunno .. random entry
21:17:52 <TrueBrain> could also validate language-tags, I guess
21:19:52 <TrueBrain> "Exception: option metadata:palette set but not expected" <- in a Base Sounds set :D
21:22:29 <nielsm> frosch123: the windows version of TTD had the music files named as .gm and .awe and .fm (iirc) but they were just standard midi files
21:22:51 <frosch123> sound palettes are an early version of sound fonts?
21:22:54 <TrueBrain> so I can rename .gm to .mid, and it would still work fine?
21:23:02 <TrueBrain> frosch123: pretty sure someone copy/pasted too much :P
21:23:25 <nielsm> well the original_win music set expects the files named as .gm
21:23:35 <nielsm> just make sure the .obm file refers to the right file names :)
21:23:50 <TrueBrain> yeah, but I am creating the validation for them :)
21:24:35 <nielsm> I'd say rather check that the files have a standard midi file header, than validate file extension
21:25:21 <TrueBrain> first is a good idea, sure
21:25:35 <TrueBrain> but all but 1 have the .mid extension ..
21:25:53 <TrueBrain> basically, I have to use the extension to know what type it is, after which I validate it really is that type :P
21:26:02 <TrueBrain> (so .grf means NewGRF, and after that it is checked it really is a NewGRF)
21:26:30 <nielsm> yeah just assume .mid or .gm are midi files
21:27:07 <nielsm> really, all music sets should just use .mid and nothing else, using .gm for a non-ttd music set is misunderstood
21:27:24 <TrueBrain> print(set(keys_in_section) - set(keys))
21:27:24 <TrueBrain> TypeError: 'type' object is not iterable
21:27:28 <TrueBrain> okay .. that error is new to me ...
21:27:47 *** ChanServ sets mode: +v tokai
21:29:01 <TrueBrain> okay, I need to fix up some Base Sets in order to import them, but it is all very minor
21:29:03 <TrueBrain> and should not impact md5
21:29:19 <TrueBrain> what is the default for blitter .. hmm
21:30:10 <TrueBrain> in the base-graphics section
21:30:33 <glx> base graphics are 8bpp usually
21:31:16 <TrueBrain> "usually" is not the answer I am looking for :D
21:31:21 <TrueBrain> can't find where this blitter entry is read ..
21:31:30 <frosch123> this->blitter = (item != nullptr && *item->value == '3') ? BLT_32BPP : BLT_8BPP;
21:31:41 <TrueBrain> all entries are read with "fetch_metadata("palette");"
21:31:55 <TrueBrain> so if the first char is "3", it is 32bpp
21:31:59 <TrueBrain> now that is lazy coding :D
21:32:22 *** HerzogDeXtEr has joined #openttd
21:32:30 <frosch123> it will work until ottd 3.0 or something
21:34:03 <TrueBrain> okay, added a shit-ton more validation on base-sets .. found a few boo-boos .. but that concludes validating upload-files for most part, I guess ..
21:34:09 <TrueBrain> owh, right, "mid/gm" and "cat" files
21:34:16 <TrueBrain> after that .. md5sums ..
21:34:55 <glx> haha just got an email from central bank of nigeria on my openttd.org address
21:35:13 <TrueBrain> YOU WON THE LOTTERY!
21:35:16 <frosch123> i think i already deleted that one from info
21:35:16 <TrueBrain> We are so happy for you :)
21:35:47 <frosch123> oh, 20:03 to info, 20:16 to me
21:36:07 <Samu> i received 8 "badges" from EA, their Answers HQ, 1 email per badge
21:36:20 <Samu> i dont even use their service
21:36:39 <TrueBrain> "Exception: option names:piano_roll.mid is missing" <- it now even detects issues like this, sweet :)
21:37:18 <TrueBrain> default = Licensed under a variant of the MIT license; see license.txt for details. Contact the author at furber.k@outlook.com.
21:37:23 <TrueBrain> frosch123: I think nobody understands [origin]
21:37:28 <andythenorth> frosch123: did I miss a memo about stockpiling? o_O
21:37:37 <glx> yesterady it was a products inquiry "The attached file contains two products list so we have combined them with an iso for easier download. Kindly find purchase inquiry as attached to this email then click on the iso file to find inquiry lists. "
21:37:48 <frosch123> TrueBrain: unless ottd requires it, you can just delete it
21:37:55 <frosch123> no sane way to trigger it
21:38:09 <TrueBrain> so why did someone add it, I wonder :D
21:38:13 <nielsm> I think the origin section is only supposed to tell you how to obtain the data files the set requires if you don't have them
21:38:15 <glx> they are inventive these days
21:38:21 <nielsm> i.e. original media to copy the files from
21:38:32 <nielsm> (copyrighted media files that can't be distributed)
21:39:03 <TrueBrain> "DEBUG(grf, 1, "No origin warning message specified for: %s", filename);"
21:39:09 <frosch123> andythenorth: weird people started panicking on saturday, and are now wondering what they shall do with flour, sugar and green kale... and are now looking for info how to cook that
21:39:41 <TrueBrain> well, I just leave it as is .. what-ever :P
21:40:01 <frosch123> TrueBrain: it was useful when people copied original files, it's pointless for bananas downloads
21:40:24 <TrueBrain> but OpenTTD "requires" it, I guess
21:40:37 <glx> frosch123: but flour and sugar can be kept forever, no hurry
21:40:50 <andythenorth> I am not going to panic until deliveroo delivery times go up past 1.5 hours
21:40:57 <frosch123> glx: their children will be happy :)
21:41:11 <TrueBrain> so how to validate mid/gm/cat .. what are their headers ..
21:41:18 <Samu> Fix: AI/GS Creating a list of tiles that will accept or produce cargo for a given industry were not considering whether the industry could only be serviced by its neutral station only
21:41:28 <frosch123> prefix those questions with "nielsm" :)
21:42:12 <TrueBrain> nielsm: k, that is mid/gm? Also cat?
21:42:28 <TrueBrain> no, cat is different ..
21:42:33 <nielsm> nobody should be uploading cat music to bananas
21:42:35 *** tokai|noir has joined #openttd
21:42:35 *** ChanServ sets mode: +v tokai|noir
21:42:41 <nielsm> there aren't any tools to create it afaik
21:42:44 <TrueBrain> nielsm: OpenSFX kinda requires you doing so? :)
21:43:04 <nielsm> ah, well cat files don't have any magic to recognise
21:43:12 <glx> Samu: Fix: [AI/GS] Consider neutral station setting when creating AITile lists
21:43:19 <nielsm> but you can try to parse the directory and see if it makes sense
21:43:29 <nielsm> sound effect cat files should always have a fixed amount of entries too
21:43:39 <glx> or tile list, as it's for both
21:43:52 <TrueBrain> nielsm: I have no clue what cat files are :D Happen to know where it is parsed in the code or something to go by?
21:44:52 <glx> cat meaning catalog but many different format in real world IIRC
21:45:22 <Samu> ScriptTileList_IndustryAccepting and ScriptTileList_IndustryProducing, they already have long names
21:45:48 <glx> it's ok for function names
21:46:00 <glx> will be hard to shorten the names
21:47:44 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick opened pull request #8030: Fix: [AI/GS] Consider neutral station setting when creating tile lists https://git.io/Jva4v
21:48:02 <TrueBrain> surprisingly, all midi files are valid midi files :P
21:48:19 <frosch123> TrueBrain: looks like cat has no real header, it's just binary data
21:48:34 <TrueBrain> so nothing to do any validation on?
21:48:38 <TrueBrain> size-bytes or anything?
21:48:46 <nielsm> yeah cat files start with an index of pointers into the file
21:49:15 <nielsm> first 32 bits (little endian) is a pointer to the data of the first data contained
21:49:30 <nielsm> everything before that should be the index
21:49:59 <nielsm> so just read the first dword and divide by 4 to get number of entries
21:50:21 <TrueBrain> code appears to be a tiny bit more complex nielsm
21:50:31 <TrueBrain> but frosch123 's link works, I guess
21:50:36 <frosch123> is the number of sounds fixed?
21:50:53 <frosch123> then the first 4 bytes only have two valid values
21:50:53 <nielsm> sound effects cat files should always have the same number of entries
21:51:41 <TrueBrain> okay, nevermind, I just process these "raw" :P
21:51:57 <nielsm> I don't think you need to validate the contained sound data
21:52:01 <nielsm> just that the header makes sense :)
21:52:04 <frosch123> TrueBrain: first dword must be 73 or 73+0x80000000
21:52:42 <frosch123> nah, there is another factor 8
21:52:52 <frosch123> 73*8 or 73*8+0x80000000
21:52:57 <nielsm> you can also verify that all entries in the file index are valid pointers into the file
21:52:59 <Samu> i made it return earlier
21:54:27 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick commented on pull request #8030: Fix: [AI/GS] Consider neutral station setting when creating tile lists https://git.io/Jva4O
21:55:45 <TrueBrain> frosch123: works, cheers :)
22:03:15 <TrueBrain> okay ... md5 of all single objects ... except for heightmaps .. (I am trying to avoid a fp.seek .. tricky for heightmaps :D)
22:06:53 <TrueBrain> there is the checksum, w00p
22:07:57 <TrueBrain> someone uploaded an AI Library as AI :)
22:08:01 <TrueBrain> well, that is never going to work :P
22:08:32 <TrueBrain> so many mistakes that BaNaNaS made :P
22:08:40 <TrueBrain> who wrote that validation?!
22:09:19 <Samu> last one that was failing is OtviAI, let's check
22:15:17 <Samu> it doesn't start building, but still have it in its plans
22:16:20 <frosch123> i love it when you debug for an hour, and then only need to add two short lines
22:18:00 <Samu> oh, WormAI was also affected, it's also fixed
22:22:48 <TrueBrain> hmm .. I wonder if common users know how to create a ob[gms] file, with the md5sums in there :P
22:23:16 <TrueBrain> but okay .. I think the validation code is ready now .. which is a lot quicker than I expected :)
22:23:29 <frosch123> the music set authors were rather common people
22:24:22 <TrueBrain> now validating every BaNaNaS upload .. to see if I did my work correctly :P
22:25:31 <frosch123> andythenorth: why does eints request that newgrf translators use the same amount of colour coded in translations, while it does not require that for gs and ottd?
22:25:59 <TrueBrain> I think the flow will be something like this: upload a grf, it looks up if this is a new version of an existing, or a new entry. It asks you to validate the finding. It asks you to fill in the blanks, like: upload readme.txt, upload license.txt, etc. If that is all done, you can hit publish
22:26:06 <TrueBrain> this works for newgrf, scenarios and heightmaps
22:26:17 <TrueBrain> for AI/GS (library), you can either upload your nut-files
22:26:21 <TrueBrain> or a zip with them in there
22:26:26 <TrueBrain> and the flow goes from there
22:26:38 <TrueBrain> for Base Sets, you can start with uploading a ob[gms], and the grfs after
22:26:57 <TrueBrain> so basically: upload all the files, you get asked some questions to make sure we understood your request
22:27:03 <TrueBrain> should be easy enough
22:27:22 <frosch123> glx: on the "commits" tab you can see the diff per commit
22:28:47 <glx> yeah didn't notice there were 2 commits ;)
22:28:59 <TrueBrain> even then you needed the "hide whitespace" :P
22:29:03 <frosch123> hmm, gh's split-view diff is substandard
22:29:25 <frosch123> other tools highlight diffs based on characters in that case
22:29:41 <TrueBrain> it does at the bottom
22:29:48 <TrueBrain> but for some reason it detected a remove too :P
22:31:37 <Samu> how do i request a backport requested?
22:32:00 <Samu> it's a serve_neutral_industries fix which is a new feature for 1.10
22:32:13 <nielsm> ask someone to add the tag
22:33:05 <Samu> well in that case, 8020 also needs it
22:33:12 <TrueBrain> well done frosch123 :)
22:33:28 <andythenorth> frosch123: eints, don't know :)
22:33:33 <TrueBrain> Samu: how was that asking?! :o
22:34:54 <Samu> i dont think 8009 is yet ready for the bakcport requested tag
22:36:34 <Samu> cool, thx, wondering if there's still time those will get into 1.10
22:38:18 <Samu> 8001/8009 is difficult to solve
22:39:00 <Samu> better not rush it into 1.10
22:42:40 <TrueBrain> 1200 lines of code to validate BaNaNaS uploads .. lol
22:42:46 <TrueBrain> and it aint done yet :P
22:47:41 <Eddi|zuHause3> if you write more than 100 lines in a day, you're probably doing it wrong
22:48:19 <DorpsGek_III> [OpenTTD/eints] frosch123 opened pull request #7: Add new string codes from OpenTTD 1.9 and 1.10 https://git.io/JvaBM
22:48:45 <frosch123> gh still sorts commits by author date, instead of by commit date
22:54:34 <TrueBrain> only 25 uploads did not validate; 80% of that is OpenGFX, and 100% are Base Sets :D
22:56:02 <Eddi|zuHause3> nobody needs those... ship it :p
22:57:41 <TrueBrain> awh, AI/GS detection is broken :P
23:03:20 *** Progman has joined #openttd
23:20:36 <DorpsGek_III> [OpenTTD/eints] frosch123 merged pull request #7: Add new string codes from OpenTTD 1.9 and 1.10 https://git.io/JvaBM
continue to next day ⏵