IRC logs for #openttd on OFTC at 2023-02-25
โด go to previous day
00:01:19 <glx[d]> I was thinking of a flag telling vehicle ignored signal but that's not the only way to enter an already reserved section
00:02:50 <dP> yeah, but just do flag that it couldn't reserve
00:02:54 *** sla_ro|master has quit IRC ()
00:02:58 <dP> so it doesn't need to clean
00:03:22 <dP> also, this is weird, electric doesn't even want to go on free path
00:05:11 <dP> yeah, but it has no orders
00:05:23 <dP> if regular rail is removed it will go on electric part and turn back
00:06:03 <dP> is that some weird noob check?
00:06:07 <glx[d]> you can make it end on the non electrified rail
00:07:06 <glx[d]> with realistic braking
00:14:09 <dP> I don't have jgrpp for that ;p
00:14:33 <dP> is that just elrail check or does that work with any incompatible railtype?
00:14:50 <dP> I guess not since mono has no issues
00:15:29 <glx[d]> I think it's an HasPowerOnRail check
00:15:38 <dP> I just wondered if dual track train can reserve into another train reservation like this
00:15:58 <dP> but idk what newgrf to look for such trains xD
00:16:27 <glx[d]> just try the same train with an extra non elec engine
00:17:29 <dP> well, that's just like having two diesels
00:21:31 <glx[d]> `t->force_proceed = t->force_proceed == TFP_SIGNAL ? TFP_NONE : HasBit(t->flags, VRF_TRAIN_STUCK) || t->IsChainInDepot() ? TFP_STUCK : TFP_SIGNAL;` <-- oh that's why 2 clicks are needed
00:28:54 <JGR> glx[d]: The compatibility behaviour is different in my branch
00:30:50 <glx[d]> I can't remember where/when we added the check, but I remember it was to prevent "stuck" trains to stop on a tile with no power
00:38:13 <dP> I know there is a whole special flag to unstuck electric trains when changing elrail settings
00:41:08 <glx[d]> maybe it was just my imagination, I can't find anything about it
01:09:51 *** tokai|noir has joined #openttd
01:09:51 *** ChanServ sets mode: +v tokai|noir
01:16:42 *** tokai has quit IRC (Ping timeout: 480 seconds)
01:22:28 *** APTX has quit IRC (Quit: Farewell)
02:03:46 *** TROILUS has quit IRC (Quit: Ping timeout (120 seconds))
02:39:30 *** _aD has quit IRC (Quit: leaving)
02:49:21 *** Wormnest has quit IRC (Quit: Leaving)
03:40:34 *** debdog has quit IRC (Ping timeout: 480 seconds)
05:07:17 *** supermop_toil has quit IRC (Read error: Connection reset by peer)
05:07:30 *** supermop_toil has joined #openttd
05:16:50 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
08:43:38 <TrueBrain> lol, classifying heightmaps was really quick .. till I added histogram support ..
08:43:41 <TrueBrain> shocking, I know right
08:46:20 <dP> there is probably some numpy magic that can do it fast
08:46:32 <dP> still need to read the map chunk though
08:56:58 <andythenorth> we only need to classify once?
08:58:48 <TrueBrain> hmm .. how to handle I-mode images .. OpenTTD has no explicit code for it ..
08:59:33 <TrueBrain> and there are a lot of those .. hmmmmm
09:00:49 *** sla_ro|master has joined #openttd
09:01:47 <Rubidium> why did I try to look into 10499?... Now I want to make a video of OpenTTD to file a bug report, but I don't really have a clue how to do that yet :(
09:02:03 <TrueBrain> never found a better way
09:04:13 <TrueBrain> okay, documentation is wrong .. I is 16bit, not 32bit
09:04:42 <TrueBrain> and this accidentally seems fine in OpenTTD, as it just picks the first byte, which is similar to >> 8
09:04:56 <TrueBrain> (and it is a grayscale value, so that is also what we would like to have)
09:05:02 <TrueBrain> funny, how things can go right by accident
09:08:44 <TrueBrain> what is the biggest map size JGRPP supports? 16k? 32k?
09:14:09 <TrueBrain> yeah ... maybe we shouldn't support heightmaps of that size
09:14:26 <TrueBrain> or even scenarios for that matter ..
09:16:27 <TrueBrain> hmm .. some PNGs are just weird
09:16:41 <Rubidium> ugh... stupid video compression
09:17:15 <dP> andythenorth: that's not allowed, iirc max is equal to 16k*16k in amount of tiles
09:17:43 <dP> you can only do 1mil vs smth small
09:18:27 <TrueBrain> I think some heightmaps cause reads-from-stack .. fun ๐
09:18:31 <reldred> andythenorth: you can't actually generate that though
09:18:38 <reldred> all red combinations are too big
09:18:41 <TrueBrain> as in, uninitialized memory from stack ๐
09:19:09 <reldred> 16k is still quite silly, but yeah, I think you can do like 32k x 8k and other weird combos
09:19:48 <andythenorth> I made a few 16k maps to test GS
09:21:11 <andythenorth> took about 5 mins to generate
09:21:11 <dP> on a silly scale I think 1mil*256 takes the cake xD
09:21:51 <TrueBrain> `(8193, 8193)` .. someone scaled their heightmap JUST wrong
09:22:24 <Rubidium> ... but MaxX and MaxY are 8192! ;)
09:22:28 <dP> you sure it's wrong? there is sowething weird about heightmaps and border tiles
09:22:43 <dP> with 8192 tiles you get 8193 borders ;)
09:24:15 <TrueBrain> so there are PNGs with a palette of size, say, 56, and the image uses index 59
09:24:26 <TrueBrain> that is just a bit weird .. what is the expected result of such pixels I wonder ..
09:24:47 <andythenorth> started map gen at 09.21 uk time
09:28:17 <Rubidium> *sigh* my bug doesn't show up in (openttd) screenshots, and video compression makes it less apparant what the issue is
09:28:44 <andythenorth> you'll have to twitch stream it
09:29:00 <dP> there are gif recorders that just capture a small region
09:29:05 <dP> peek is a good one for linux
09:29:34 <andythenorth> trees are offensive, we should throw them out ๐
09:29:38 <Rubidium> don't care much about Windows ;)
09:30:08 <andythenorth> think this will take some time ๐
09:30:08 <dP> though recent ubuntu seems to have one built in
09:30:12 <TrueBrain> lol, so the palette says: there are 56 colours .. but when I ask: list me all the colours, it returns 60 .. I must be missing something ๐
09:31:01 <dP> hm... how do I record the screen recorder? xD
09:32:56 <TrueBrain> what's happening!! ๐ฎ
09:33:03 <andythenorth> maybe I should start a box set and come back ๐
09:33:11 <TrueBrain> Rubidium: stop murdering the game!
09:33:33 <dP> Rubidium: lol, wth is going on there? is line drawing algo unstable?
09:34:24 <dP> may be just the issue with console redrawing squares in the wrong place when stuff moves underneath
09:34:25 <Rubidium> exactly... when line width != 1, something fishy is going on. Okay, this is with pr10499 applied, but I doubt that causes this
09:38:41 <dP> dP: or line drawer is affected by the change of rects
09:39:04 <TrueBrain> okay, so Pillow has a function `colors` on their palette, which returns UNIQUE colours, not indexed colours ... that was ... unclear ... from the documentation .. wauw ...
09:42:06 <andythenorth> pillow docs are...adequate...but not amazing
09:42:36 <TrueBrain> and as it turns out, wrong too!
09:44:37 <TrueBrain> lol, CoPilot just wrote a better way to detect "flat" than I would have ๐
09:44:46 <TrueBrain> as "flat" can be at a height of 200, ofc
09:44:53 <TrueBrain> and CoPilot took that into account .. I wouldn't have ๐
09:48:03 <Rubidium> dP: it also happens with the FrametimeGraphWindow, if you make the lines 2 wide. So it's definitely more generic
09:51:26 <dP> Rubidium: well, if line drawing is affected by redraw rects it will do that everywhere
09:51:49 <dP> I had similar issue with map window but I don't remember exactly why it was happening
09:52:02 <Rubidium> true, but there are not that many places that draw lines with width > 1
09:54:25 <andythenorth> still going ๐
09:54:29 <andythenorth> started at 09.21
09:54:56 <andythenorth> ^ just doing this to amuse myself, not poking fun at JGRPP
09:56:07 <dP> jgr didn't have to ban 1mil*1mil map, no one would ever know it doesn't work ๐
09:56:35 <andythenorth> curious when it will finish
09:56:57 <andythenorth> about 25 minutes to do 10%
09:59:39 <TrueBrain> stop running the game on a raspberry pi ffs
10:00:25 <andythenorth> this is on my Sony PSP
10:00:45 <TrueBrain> hmm .. not sure if I really like the "calculate elevation" .. it is really expensive, CPU-wise ๐
10:00:54 <TrueBrain> people uploaded crazy heightmaps ๐
10:01:12 <andythenorth> can't be trusted
10:01:34 <TrueBrain> also, our interpretation of a heightmap is not super easy to deal with
10:01:38 <TrueBrain> we have these weird exceptions
10:03:59 *** HerzogDeXtEr has joined #openttd
10:08:57 <dP> you mean can't jump more than 1 height level like png?
10:13:37 <TrueBrain> takes 15 minutes to process 250 heightmaps .. lol
10:13:56 <TrueBrain> 4s on average .. owh well .. NewGRFs can be worse
10:15:28 <TrueBrain> lol, I manage to classify only 2 heightmaps as "mountainous" ๐
10:15:34 <TrueBrain> most are pretty flat, or somewhat hilly
10:15:48 <TrueBrain> makes you wonder what value this information adds, hmm ..
10:19:02 <dP> for height maps useful category would be "real world based"
10:19:10 <dP> but no way you can do that automatically
10:20:10 <michi_cc[d]> TrueBrain: What is mountainous for you? Wouldn't it also depend on what you set as "highest peak" when load it in the game?
10:20:30 <TrueBrain> it is a quality of the heightmap, not of the map generated with the heightmap
10:20:53 <TrueBrain> so I currently base it on the histogram, and see if there is a lot of elevation difference
10:21:42 <TrueBrain> ofc if you make the game generate the heightmap under strict restrictions, it can be fully flat .. but yeah, that is not a quality of the heightmap itself ๐
10:22:01 <michi_cc[d]> Okay, that makes sense, too. Something can be really high but just be a big fat plateau, and I'd not call that mountainous.
10:22:20 <TrueBrain> yup; I mentioned that earlier, that CoPilot figured that out for me ๐
10:22:23 <TrueBrain> which made me giggle ๐
10:23:11 <TrueBrain> the current algorithm is even a bit more naรฏve: it picks the 32 (out of 256) most used heights, sums those up. If that makes up more than 90% of the surface, you are flat
10:23:22 <TrueBrain> strictly seen it can mean that height=0 and height=255 are the most used
10:23:35 <TrueBrain> but the game would make that practically flat, as we cannot change that quickly in height in-game
10:25:47 <TrueBrain> now how to validate this quickest, to see if the validation is somewhat okay ... generating maps constantly is not really a sane approach ๐
10:26:03 <dP> that sounds more of a variety check than terrain type
10:26:29 <reldred> A good deal of CptK's maps for instance should definitely be flagged as mountaineous.
10:26:44 <TrueBrain> I tuned the classifier; some now are ๐
10:27:14 <TrueBrain> `Cpt.K's Riverside Mountains] resolution=high,shape=rectangle,elevation=mountainous`
10:27:23 <TrueBrain> seems to match the description ๐
10:27:37 <reldred> Did any of Cliffy Coastline series make it?
10:27:48 <reldred> or any of the Reldred's Choice ones?
10:27:49 <TrueBrain> `Cpt.K's Cliffside Hills] resolution=high,shape=square,elevation=mountainous`
10:28:03 <TrueBrain> reldred: all 3 marked as `hilly` currently
10:28:21 <TrueBrain> still tuning, but I need an easier way to visually validate what is actually going on ๐
10:28:35 <reldred> I'd probably agree with that, they're not my favorite of his maps
10:28:45 <TrueBrain> but you picked them!
10:28:53 <andythenorth> TrueBrain: can't an AI just tell us? ๐
10:29:05 <reldred> Nah he made them according to an description i made of maps of his I modified
10:29:25 <reldred> Stardust Coast should absolutely be mountainous.
10:29:36 <reldred> Thing is it really does vary depending what generation settings you use
10:29:55 <TrueBrain> again, quality of the heightmap we are classifying, not how to user uses it to generate a map from ๐
10:30:20 <TrueBrain> a movie can be 4k, but if I play it on my 1080p, it is not
10:30:25 <TrueBrain> doesn't mean the movie is not 4k ๐
10:31:11 <dP> except heightmaps don't really have that quality at all
10:32:12 <dP> if anything I'd say spiky heightmap with valleys is mountaneous and smoother is flat
10:32:17 <TrueBrain> andythenorth: haha, it even produces Python code to tell me exactly how it would classify it ๐
10:32:40 <andythenorth> seems like we could argue it all day
10:32:45 <andythenorth> but an AI would just get it right
10:32:51 <andythenorth> /me all in for AI
10:32:57 <TrueBrain> `Based on the histogram, I would summarize the elevation of this map as "varied".`
10:33:35 <TrueBrain> it actually understood the histogram, which is scary .. it even reasoned over it
10:34:57 <andythenorth> I mean....it's scary that people can do it
10:35:03 <andythenorth> we're just a walking mess
10:35:06 <TrueBrain> `Based on the histogram values, it appears that the elevation of this map is bimodal, with two distinct peaks in the distribution. The first peak is around the value of 0, indicating low-lying or flat terrain`
10:35:31 <TrueBrain> ChatGPT is still fun to toy with
10:35:40 <andythenorth> I haven't tried it, I do it all via other people
10:35:45 <andythenorth> this is the future of people
10:35:54 <andythenorth> typing into AI prompts on my behalf
10:36:23 <andythenorth> I did join the midjourney discord, but I would have got lost in there for months
10:36:54 <TrueBrain> at a certain point they kicked everyone out of that Discord
10:37:01 <TrueBrain> as in, one day I was in there
10:37:05 <TrueBrain> the next I was not!
10:37:12 <andythenorth> I know people still in it
10:37:17 <andythenorth> we use it for work stuff
10:37:26 <andythenorth> oh, JGRPP is still going ๐
10:37:55 <TrueBrain> be happy you see a status update ๐ It didn't use to do that!
10:38:52 <TrueBrain> lol, a histogram of `[2083227, 13283, 642, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]`. ChatGPT says "relatively low"
10:38:58 <TrueBrain> relatively?! That is fucking flat dude ๐
10:39:21 <andythenorth> it's not allowed to swear though
10:39:26 <andythenorth> also that's proof you're not an AI
10:40:46 <andythenorth> will I get pencil and paper and design a GS loop
10:40:54 <andythenorth> or will I sit here watching youtube
10:41:07 <TrueBrain> I am sure you will tell us in an hour which you picked
10:41:15 <andythenorth> I was hoping you would all recommend
10:41:18 <andythenorth> I would like a choices AI
10:41:26 <andythenorth> that recommends me actions
10:41:35 <andythenorth> more than just a dice roll also
10:42:10 <TrueBrain> tldr: you suck if you continue to watch youtube
10:42:29 <andythenorth> it's not aware of the need to distract the conscious mind so the subconscious can problem solve
10:42:52 <andythenorth> we need an AI with ADHD-tendencies
10:43:02 <andythenorth> it starts giving advice, then gets totally distracted
10:43:22 <andythenorth> or one like in my house "You never told me that"
10:43:26 <andythenorth> "You didn't listen"
10:44:03 <andythenorth> ^ literally every conversation, between all 4 people
10:44:14 <TrueBrain> There is a map with a histogram (in percentages) that look like this: `0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 100`
10:44:22 <TrueBrain> someone wanted to build REALLY high-up
10:44:50 <andythenorth> hmm wonder how well Dall-E makes maps
10:44:56 <andythenorth> did scrubbles try yet?
10:45:07 <andythenorth> people have wanted a map-gen tool for years
10:49:23 <TrueBrain> you know you made a boo-boo if ... `Very Flat Heightmap] resolution=high,shape=rectangle,elevation=mountainous` ๐
10:53:40 <andythenorth> big map has eaten all my RAM ๐
10:53:57 <TrueBrain> if that is all your RAM, I should worry ๐
10:54:17 <andythenorth> well the AV tool uses 1.5GB
10:54:54 <andythenorth> and Discord in Chrome is half a GB
10:55:33 <andythenorth> lol Discord in Chrome uses almost more than Photoshop
10:55:40 <andythenorth> how the world has moved on in the last 20 years
11:03:14 <Samu> Feedback on rolling back a Visual Studio update
11:21:10 <EmperorJake> Removing the tags seems like a bad idea
11:21:33 <EmperorJake> Is there a reason the search function can't just work for both old and new tags?
11:29:02 <andythenorth> because the old tags are mostly junk ?
11:29:07 <reldred> EmperorJake: agreed, simple things like typing japan and now not getting *any* of the JP+ content is a bit counter productive
11:29:08 <andythenorth> so it's not a useful search space?
11:30:21 <reldred> and given it doesn't/can't search inside descriptions means there's no real way around it either
11:30:47 <andythenorth> just ask chatGPT which are good japan grfs?
11:30:58 <Yozora> More of that, there is no way to classify geography of grf with auto classification
11:31:07 <reldred> for instance jp+ objects has JapanSet in the description but won't appear in a keyword search for 'japan'
11:32:07 <reldred> so now we basically have the worst of both worlds
11:32:37 <andythenorth> put `(Japan)` in the set name
11:32:55 <andythenorth> I added suffixes to my grfs last year or so
11:32:56 <reldred> nice so we have set names that are five miles long
11:33:00 <Yozora> andythenorth: What about old grf like tias?
11:33:17 <Yozora> New players would never be able to find it
11:33:42 <Yozora> And that's a descent grf with unique sounds
11:34:45 <Yozora> One of the Japanese themed grf's obviously
11:34:59 <andythenorth> can't find it in online content
11:35:23 <Yozora> Because full name will be toei interconnection set
11:35:33 <reldred> yeah because the search tool is now busted lmao
11:35:52 *** KeepinItRail has joined #openttd
11:35:52 <KeepinItRail> The tags help with showing what TIAS for someone who didnโt know
11:35:52 <reldred> all of it was tagged correctly beforehand
11:36:17 <Yozora> It's wery unfriendly to new players, who wants to play the game, and not looking through forums/reddit to find what they want
11:37:02 <Yozora> More of that, what if there would be 1 million train grf's? Would you look through all of them?
11:37:05 <andythenorth> I think the opposite
11:37:12 <andythenorth> the old system shows a lot of bad habits
11:37:29 <andythenorth> very much tagging for people who are familiar with tt-forums or other obscure forums
11:37:36 <andythenorth> and just junk tagging
11:37:41 <reldred> like, before we had a problem that some grf authors didn't tag their stuff correctly, so instead of fixing up tags like other major content platforms do (with say, vetted maintainers able to suggest tags, etc.) we just... throw them out entirely?
11:37:50 <andythenorth> reddit and steam shows that players can't find anything
11:37:52 <andythenorth> or understand it
11:38:06 <reldred> ahh yes, so let's just break it for *everyone*
11:39:49 <andythenorth> always the best way to effect change
11:40:18 <andythenorth> hmm 'train' is a tag, 'trains' is not
11:40:34 <Brickblock1> I think removing it is dumb the old system was not perfect and people misused it but it worked and I could find what I was looking for
11:40:53 <andythenorth> but you know what is there
11:40:59 <andythenorth> so you know how to find it anyway
11:41:20 <reldred> lmao I can't find *anything* now, not new stuff, not old stuff, not stuff that someone casually mentions
11:41:36 <reldred> unless I happen to know *exactly* what the name of it is?
11:41:43 <andythenorth> seems fine? ๐
11:41:52 <reldred> so why even have search? I can just scroll through the list alphabetically.
11:42:11 <Brickblock1> finding rukts is way harder now
11:42:16 <Yozora> Andy, did you just find tias not a while ago without help?
11:42:34 <reldred> no you had to be told exactly what it was called.
11:42:35 <Yozora> And there was a tag tias, that players could use to find it
11:42:50 <reldred> and it also just popped up in a search for 'japan'
11:43:03 <Samu> I've been thinking about the reservation issue
11:43:14 <Yozora> As well as tags: japan, tokyo, toei, kekyu and several more
11:43:24 <Samu> maybe, just maybe, there is an easier solution
11:43:43 *** sla_ro|master has quit IRC ()
11:43:46 <Samu> when a train collides with a reservation, cancel it
11:44:06 <Samu> easier said than done, but let's see if this is doable
11:44:34 <Azusa> Can we just bring back tags the way they were before the change?
11:45:21 <reldred> I think the auto classifications are great, being able to see at a glance if something is extrazoom or not for example,
11:45:37 <Brickblock1> there is also the issue that if I wan't to find a set that covers more than one country it now won't show up if you just search for one of them unless the name is stupid
11:45:37 <reldred> but removing the old tag system is just a straight up 100% regression.
11:47:17 <Tinny> reldred: Who thought that removing tags was a good idea? What was wrong with tags before? It is far more ugly to start naming sets "JP+ Japan Japanese Trains Multiple Units JNR JR EMU DMU Choochoo"...just so that you can search for something you want.... ๐
11:47:49 <reldred> people name sets bad things, let's just show the GRFID only
11:47:50 <dP> ez, fix, remove the name, put tags there ๐คฃ
11:48:14 <Tinny> dP: Sure, let's ditch set names and call things tags ๐คฃ
11:48:18 <Brickblock1> every thing is just trains
11:48:23 <Yozora> dP: There is a 32 char limit. I tried it ):
11:48:39 <Azusa> UK trains, UK trains and UK trains?
11:48:40 <reldred> if you search for trains nothing comes up
11:48:43 <Tinny> "Industry Big Set Coal 64"
11:48:55 <Brickblock1> you wouldn't even know it was from the uk
11:49:17 <Brickblock1> can't be detected by bananas
11:49:19 <Tinny> KeepinItRail: But what if I search for "british"
11:49:22 <EmperorJake> If the old tags were junk, why don't we curate them?
11:49:27 <Tinny> I wont ever find your set!
11:50:09 <reldred> The auto classification is great, I'm glad we have it, I just don't know why it needed to be at the expense of the old tag system.
11:50:18 <Brickblock1> but the old system was better than this
11:50:28 <Azusa> Why remove the old system?
11:50:29 <EmperorJake> I would go through a bunch of content and add tags for it
11:50:30 <LordAro> this is the most animated the channel has been in years
11:50:48 <LordAro> and it's just a load of people asking the same questions without waiting for answers
11:51:36 <Tinny> That aren't renaming NewGRFs stupidly
11:51:41 <Yozora> We need only one answer. When tags will be brought back
11:51:53 <Tinny> By this evening preferably
11:51:59 <Azusa> Like use the new auto classification thing along with the old manual system at the same time
11:52:16 <Brickblock1> it was like that before
11:52:23 <Azusa> This didn't have to replace the old system
11:52:26 <LordAro> ok, you've said all that
11:52:35 <LordAro> what do you think you're going to achieve by saying it again?
11:52:57 <andythenorth> lol I don't see the problem yet
11:53:41 <LordAro> i'm not even entirely sure what's happened
11:53:47 <LordAro> i scrolled up a bit, didn't find anything
11:54:47 <Brickblock1> should we all go back to using grfcrawler?
11:55:01 <reldred> nah let's go back to the old yahoo webrings
11:59:59 <andythenorth> the bit that puzzles me
12:00:07 <andythenorth> why would you use grfs you don't know the name of?
12:00:19 <reldred> ... because I like to try new things?
12:00:36 <andythenorth> but the content service is a shit way to do that?
12:00:45 <Brickblock1> the first way I found newgrfs was bananas
12:01:20 <Yozora> Imagine searching in the internet, and finding nothing because you wrote scottish train instead of british
12:01:20 <Brickblock1> and I still use it from time to time
12:01:38 <andythenorth> but why would you search random grfs in the game?
12:01:43 <andythenorth> a lot of them are just shit
12:01:45 <andythenorth> not most, but a lot
12:01:54 <andythenorth> that's a big effort
12:01:55 <Azusa> When I want to start a Japanese themed game, I would search "Japan" and download things that come up
12:02:04 <reldred> that's how I found all the korean stuff?
12:02:10 <reldred> and then found their githubs?
12:02:17 <andythenorth> I don't mean 'why?' as in 'you shouldn't'
12:02:21 <andythenorth> just genuine open 'why?'
12:02:30 <reldred> and learnt how they solved fence issues etc.
12:02:39 <reldred> OpenTTD isn't one single community
12:03:13 <Brickblock1> that is what I did when I downloaded grfs the first time I searched for sweden and found what I wanted
12:03:14 <reldred> if I want to find content from outside of the usual tt-forums/discord community I do it by... searching in the content downloader!
12:03:22 <andythenorth> I never search the content service for anything I don't already know the name of
12:03:36 <Azusa> It's kind of disheartening to see things that would have came up before the change no longer showing
12:03:40 <reldred> so you're absolutely 100% representative of the ENTIRE community?
12:03:56 <andythenorth> no, just finding the parameters of the scale
12:04:14 <andythenorth> bearing in mind that this change was made by devs, and most devs never play
12:04:24 <andythenorth> and mostly won't be ever using the content service
12:04:27 <Yozora> We have 1 person ok with new tag system and 6+ people against it
12:04:33 <Azusa> This probably affects newer players to the game trying to download newgrfs for their first time
12:04:39 <dP> I used search at some point to find new grfs as well
12:04:49 <andythenorth> the existing system is 'lol'
12:04:54 <dP> though I was searching trees so only found disappointment ๐
12:04:54 <Brickblock1> I am not against the new system but it shouldn't replaced the old one
12:04:56 <andythenorth> as in the one that just got replaced
12:05:09 <Yozora> But the previous one at least worked
12:05:10 <LordAro> Yozora: do not confuse the sample size of "people complaining about something" as an actual representative sample
12:05:32 <Yozora> Should I write an ultimate problems with the new tag system>
12:05:42 <andythenorth> on the other hand, we shouldn't confuse the sample size of 'people confused on reddit' either
12:05:51 <KeepinItRail> I donโt always know the name of restaurants I havenโt been to before, but I hope Google would should me them based on the tags the restaurant has, like โKoreanโ
12:05:52 <andythenorth> like all those people who report that FIRS has installed new trains
12:06:45 <Tinny> What has happened is that by removing tags, you can't use the search to find NewGRFs that were previously tagged with meaningful connotations. You can't search for "British" and get all the NewGRFs associated with all British trains, houses etc. You can't search for "Japan" and get all the NewGRFs that have Japanese content. This is because people don't use the NewGRF names as a tag list, because that is just silly. A trainset is called
12:06:45 <Tinny> other weird brand - sure it is weird but it provides distinction. And to make sure people were directed to the relevant trains or objects, you did a search for "british", and behold it would appear because the NewGRFs would be tagged with whatever they contain. Why would you want to search for random newgrfs? Its a case of trying to find new things; I want to boot up the game and open content manager and hunt for things that suit my need
12:06:45 <Tinny> a "Russian" train - I go on and type "Russian" and a set called "xUSSR" appears and I download that to try it out to see if it fits my needs. The tags said that "xUSSR" was a Russian trainset...but no tags, how will the search direct me to it?
12:07:28 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
12:07:59 *** WormnestAndroid has joined #openttd
12:08:17 <andythenorth> how would you find that Horse is mostly British and Irish
12:08:25 <andythenorth> or alternatively mostly UK
12:08:26 <dP> hm, does content server keep a history of what people are searching?..
12:08:40 <Yozora> In simple form, there is no way to ad geotag to the content automatically. And that tag helped a lot of people to find the stuff they wanted
12:08:59 <andythenorth> seems country is getting mentioned a lot
12:09:05 <andythenorth> is geotag desirable?
12:09:11 <andythenorth> it's very politically problematic
12:09:27 <andythenorth> e.g. tagging Taiwan might get the game banned in China
12:09:35 <Tinny> andythenorth: I never went on and searched for Iron Horse, cause I didn't know about it by name. I'd be search by country and/or "fictional".
12:09:46 <andythenorth> but the countries aren't in the tags, never were
12:10:01 <Yozora> Don't seach for excuses, we have a real problem here
12:10:03 <andythenorth> all the tags are still on bananas in global.yaml for each grf
12:10:11 <reldred> they were for the various Japan Set/JP+ content
12:10:15 <andythenorth> no we don't have a real problem, we have a cartoon train set
12:10:20 <andythenorth> keep some perspective eh ๐
12:10:37 <Yozora> Some people might play with real trains
12:11:39 <Yozora> With some special theme, and they would not be able to find it, because they have to list 255+ grf by the tag "train"
12:12:30 <Rubidium> dP: I doubt it does, primary because I do not think the content server provides search capability and I doubt a phone-home mechanism for searches would get approved
12:12:34 <Brickblock1> andythenorth: I would be more concerned by โข infringement from most train grfs
12:13:32 <andythenorth> nobody has mentioned the loss of author names so far
12:13:53 <andythenorth> but sampling tags, it's pretty common that grfs had 5 or so tags, one of which is author name
12:14:06 <andythenorth> I am just randomly clicking on global.yaml in bananas
12:14:19 <andythenorth> did anyone parse them all and plot them?
12:14:20 <Brickblock1> that isn't good either but doesn't affect most people as much
12:14:47 <andythenorth> I can't be bothered to learn how to stem all the words
12:14:58 <andythenorth> e.g. "britain" and "british" are the same semantically
12:15:09 <andythenorth> so we would want to count those all the same
12:15:27 <andythenorth> anyway all the previous tags still exist if anyone wants to analyse
12:15:41 <andythenorth> why we can't just get chatGPT to do it, I don't know ๐
12:15:57 <andythenorth> especially for the realism train sets
12:16:06 <andythenorth> just read the vehicle names, and then look up which country
12:16:34 <LordAro> searching in a manner that people have come to expect from google and similar is actually really hard
12:16:46 <LordAro> (i.e. trains vs train)
12:16:50 <reldred> we had a system that mostly worked
12:16:51 <andythenorth> we've always refused to allow manual curation because politics
12:16:57 <reldred> now we have a system that doesn't
12:17:12 <LordAro> we have a system that is different
12:17:21 <Yozora> We are not asking google, we asking for a system that would be able to provide a list of themed grfs by the name of the country
12:17:26 <LordAro> it works better for some workflows, and worse for others
12:17:26 <andythenorth> we have a system now that is experimental
12:17:29 <LordAro> that is the nature of change
12:17:34 <andythenorth> this was yolo shipped to see what happens
12:17:39 <andythenorth> and what happens is this ^
12:18:00 <andythenorth> this is design by the equivalent of guessing in the dark
12:18:16 <Yozora> What's the point of change if it makes it worse for everyone?
12:18:24 <andythenorth> to learn something of course
12:18:28 <reldred> see this is what i love about this project, new developers submit pr's and they languish for years, an established dev decides to yolo something and 'oh well, too bad, it's not worse it's just different'
12:18:30 <LordAro> does it make it worse for everyone?
12:18:34 <andythenorth> the only point of anything (other than basic needs) is to learn things
12:18:39 <kamnet> WHy didn't anybody say anything to TrueBrain earlier in the week when he was talking about it before he made the changes?
12:18:39 <andythenorth> find new ways to fuck up
12:18:41 <Brickblock1> I think I mentioned that replacing the old system entirely would be bad
12:19:04 <Yozora> Andy would not be able to find a cool grf by the original tag it was seached by everyone before
12:19:19 <reldred> LordAro: it's pretty much in terms of usability a straight up regression. The new auto classification is good, I like it, but removing the old tag system from the search is a straight up regression.
12:19:24 <Yozora> I dunno, maybe not finding something adds some detective element to the game
12:19:29 <Tinny> kamnet: Honestly, I had no idea that it was happening - like many users, I dont look at everything that is in development. Sorry not sorry
12:19:32 <andythenorth> Yozora: don't use me as a case, I never use search, it's full of junk
12:19:43 <andythenorth> this wasn't made for or by me
12:19:56 <Chris> kamnet: At no point was it clear to me that he was carpetbombing the system. I thought he was just _adding_ the 32bpp, extra zoom and newgrf type tags
12:20:00 <reldred> kamnet: maybe because not everyone has the chance to play during the week?
12:20:03 <Azusa> All I read was "we're adding a new category system to bananas", I didn't really think it would replace the current system
12:20:08 <kamnet> Tinny: Well, no, not you, but a good chunk of us in the discussion were here when it happened. ๐
12:20:11 <reldred> it's only saturday night here in Australia
12:20:24 <Azusa> I merely thought it would be a small new addition on top of what we already had
12:20:26 <andythenorth> if everyone would like to talk themselves off the ledge
12:20:33 <andythenorth> it's just a bananas content server change
12:20:36 <reldred> And I was watching the convo but I didn't notice that 'shitcan the old system entirely' was on the cards.
12:21:36 <andythenorth> this is probably the most amount of drama since train flip-in-depot was banned
12:21:45 <andythenorth> brings back nostalgia
12:21:47 <Chris> Because this is genuinely bad
12:21:57 <dP> did you forget signals? xDD
12:21:59 <andythenorth> lol no, being sick is genuinely bad
12:22:02 <kamnet> Agreed, not the desired results.
12:22:03 <reldred> Because starting drama is the only way anything gets looked at in this community?
12:22:17 <andythenorth> 960ms to set growth rate for 32 towns in GS is genuinely bad
12:22:33 <reldred> Because otherwise everyone just sticks their heads in the sand and pretends nothing is wrong?
12:22:58 <Chris> Before you could use the content downloader as a tool to explore what newgrfs were around. Now you can't even find TIAS to look for TIAS or RIMS to look for RIMS because of a change that was improperly communicated, if communicated at all
12:23:14 <Yozora> You look at it as it is drama, but it is a real problem, and at least 7 people now are mad about the change
12:23:16 <LordAro> i'm not even suggesting you're wrong, but the whole "this is literally the worst thing that has ever happened" shtick is really undermining your argument
12:23:29 <andythenorth> this has made my day ๐
12:23:41 <reldred> Oh yes, it could be way worse, so why are you complaining ๐คก
12:23:41 <andythenorth> it's just grfs, I'm not even sure why anybody cares
12:23:45 <Tinny> andythenorth: Oh I'm laughing too dont you worry
12:24:02 <LordAro> reldred: that sort of strawman bullshit is also not helping
12:24:08 <Yozora> If you don't want to restore the old system, please explain to us how to make searching by country/region in the new system
12:24:12 <reldred> I love how the default attitude in this community seems to be arrogance
12:24:39 <Tinny> LordAro: Its how you make a point, we are all different people expressing our thoughts, feelings and feedback
12:24:52 <andythenorth> Yozora: Is this the most important / primary search field?
12:24:55 <Yozora> The answer is "you can't" and this is no less a bad feature, because you could do it before
12:25:02 <andythenorth> Bananas is probably extensible
12:26:05 <andythenorth> but there's no point filling it up with internet outrage
12:26:12 <dP> andythenorth: as if that ever helped ๐
12:26:14 <andythenorth> what we want to discover are the affordances people miss
12:26:25 <andythenorth> dP: no discussions help quite often
12:26:31 <LordAro> discussions are at least recorded
12:27:10 <andythenorth> so country / region seems to be one affordance missed
12:27:26 <reldred> tags is one that's been missed
12:27:30 <reldred> you know those tags we had
12:27:37 <andythenorth> ok so that can be added
12:30:02 <andythenorth> I mean discount some of the Zorg-ness
12:30:06 <andythenorth> but there are ideas in it
12:33:41 <andythenorth> hmm wonder if we could assign star ratings randomly
12:33:44 <andythenorth> just a dice roll
12:33:53 <andythenorth> 'sorry your grf was unlucky, 1 *'
12:45:36 <glx[d]> Searching in descriptions is supposed to happen in future
12:46:15 <reldred> That would honestly solve just about all issues.
12:46:49 <reldred> Tags would still be useful; eg; 'britain, british, uk, gb' etc.
12:47:36 <reldred> Otherwise people will just make a tag/hashtag list in the bottom of their descriptions like instagram all over again lol
12:49:22 <Brickblock1> that is how it already looks but for backwards compatibility it is preferred to keep the old system anyway
12:49:40 <glx[d]> Can't admin port receive all chat ?
12:49:59 <reldred> As it currently stands, it feels like a complete regression.
12:50:19 <reldred> the new auto tags are useful, but don't really assist with searching.
12:50:45 <reldred> more just a way if you're looking at something in bananas to see what's going on with a set without having to download it
12:51:30 <Azusa> Currently intra company chat does not show to the server console
12:54:18 <kamnet> andythenorth: So other than this, what else in tags are we actually missing?
12:54:40 <LordAro> reldred: So I've been thinking about your "arrogance" comment. It's true that OSS devs across the whole software world are often accused of arrogance by users. However I don't think it's there unnecessarily - while I can't speak for anyone else, I think it's a coping mechanism of sorts. You have to understand that the /vast/ majority of feedback that any of us get is in the form of issues or complaints.
12:54:46 <LordAro> No one makes a bug report to say how good something is (nor should they). But it's very rare in the OSS world for anyone to work on something because they're getting paid, they do it because they want to. If devs took on all the unconstructive, insulting, unhelpful, "This is bad and must be undone immediately" type feedback as written, without the "arrogance" they'd all just give up and leave. I imagine
12:54:52 <LordAro> users (who are also not paying for anything) do on the whole actually want the project to continue rather than just die. I'm not quite sure of the point I'm making here, but in general remembering that there's another person on the other end who has different viewpoints from your own to consider would be a good start.
12:57:21 <Chris> kamnet: A bunch of NewGRFs are listed as their full name and not their abbreviation, so RIMS and TIAS can't be found in the in-game content downloader
12:57:59 <Brickblock1> and as Andy mentioned authors
12:58:08 <EmperorJake> Also stuff like 2cc
12:58:45 <EmperorJake> Or company colour/real liveries
12:58:47 <glx[d]> Maybe 2CC can be autodetected
12:59:26 <EmperorJake> Even if it's hidden behind parameters or refits?
13:01:36 <kamnet> Chris: glx mentioned that descriptions are ging to be searchable. In that case it seems to me like this would be a good place for such things to be placed, and we need to encourage authors to utilize it.
13:01:46 <reldred> LordAro: Okay, let's clear a few things up. I'm a senior engineer in my own right supervising the creation of a new team, a F/OSS contributor, and an active content creator for your project who has been involved with the TT community for 19 years. I have watched this project since Ludde uploaded his barely functioning C conversion of an ASM disassembly way back when. And I still stick to my statement that this project has a problem wi
13:01:46 <reldred> devs try and argue to me that having my content stolen and uploaded to Bananas was actually a good thing. I've sat and watched PR after PR after PR rot and die because nobody wants to review, and nobody wants to appoint new devs. I've had people repeatedly ask me to backport my (admittedly very simple) JGR patches to OpenTTD when you and I both know they'd just sit in the PR pile never to be touched.
13:02:32 <reldred> Talk about a love hate relationship with this project and with this game
13:03:47 <reldred> And any time anyone raises an issue, before it even gets to teeth gnashing and name calling, they get brushed off.
13:04:26 <Brickblock1> kamnet: That doesn't work now tho and would not necessarily work with old grfs
13:05:45 <reldred> This entire conversation started off as a pretty calm level headed convo with screenshots illustrating the issue, and got brushed off
13:06:13 <TrueBrain> Azusa: your discussion thread, I assume? What a politeness ๐
13:07:32 <Azusa> I could have just asked here
13:07:39 <TrueBrain> no, discussion is fine ๐
13:07:43 <TrueBrain> means others can read the reply too
13:07:49 <TrueBrain> just have to delve in actual code to find the answer ๐
13:09:25 <LordAro> reldred: I think you're applying things that haven't been true for years to the present day. I'm not aware of any content having been stolen (and reported) that hasn't otherwise been addressed properly. Certainly nothing that would suggest "this is a good thing". OTTD project as a whole has issues with issue triage and acting on PRs, and active devs (see previous for why that's hard to address) as do
13:09:31 <LordAro> many other projects. And while there's currently 74 open PRs (some of which are very old), there are 2417 closed PRs, and you have zero - so you've never even tried. Don't assume something that may have been true in the OTTD-community of old is true now
13:09:52 <Azusa> As I said, I don't believe I am seeing game chat that is sent within a company on the server console, but I also may question the privacy issues therewith
13:10:21 <TrueBrain> but I think that is a weak argument to be made, privacy, as you could also just sniff it from the network
13:15:09 <TrueBrain> put what is in my mind in the ticket ๐
13:16:18 <TrueBrain> now for the other conversation in this chat .. takes a while to read through the whole forest what the actual question is ๐ Tinny , do I translate you correctly that mostly what you miss is a region/country search capability? More than tags? Or am I missing the point you were making?
13:20:38 <Tinny> Okay, I really feel like we set off on the wrong foot - cause it's a bit of a cluser of multiple people going "oh no this isn't doing what I expect anymore"! And I'm sorry...I get that in my job too and it's always fun having to actually understand why and resolve it. Let's make a discussion thread (where can we do that?) and go back and forth with all the stakeholders ๐ค but in general, my issue is the regional/country capability - yep
13:20:38 <Tinny> different scenarios with the new system, as others have done in the chat above, and give you something to work with ๐๐
13:21:07 <TrueBrain> I absolutely love the idea of adding "what region is this set about" ๐
13:21:30 <TrueBrain> and tnx btw, for what you are saying there. much appreciated ๐
13:22:29 <TrueBrain> can you search by author currently, in-game .. let's try
13:22:37 <Brickblock1> Would it be possible to keep the old tags but only for things that have already been uploaded that way they will slowly get phased out
13:23:11 <Brickblock1> While not making it harder to find them than before
13:23:19 <TrueBrain> Brickblock1: personally, I would prefer not to. But this has everything to do with how many people used tags properly vs the rest ๐ So I am more curious what you are missing
13:23:54 <Brickblock1> I just want to have the old ones still work
13:23:54 <TrueBrain> as that is very interesting to me, to know what some people actually used tags for that was useful, and isn't covered by the classification ๐
13:23:58 <LordAro> author was definitely mentioned
13:24:02 <TrueBrain> but what value did it give you? ๐
13:24:42 <Yozora> I would add, I have used different names for underground trains (metro, subway, underground) as tags, with the name being jp+metro
13:24:45 <TrueBrain> awh, in-game doesn't search the author field, silly in-game .. well, that is fixable ๐
13:24:49 <Brickblock1> One curtain grfs it worked well but not on all of them
13:25:11 <Tinny> Looks like I've been beaten to it!
13:25:54 <andythenorth> kind of considering checking out the bananas repo and trying to parse all the global.yaml
13:26:10 <LordAro> andythenorth: sounds like an excellent plan
13:26:14 <andythenorth> we could then find eyeball all the country names
13:26:27 <TrueBrain> I have a list, don't worry ๐
13:26:43 <TrueBrain> I really like the idea of asking explictly for the region
13:26:44 <andythenorth> my suggestion would be use the ISO codes for countries
13:26:49 <andythenorth> there are also ISO for some regions
13:26:57 <andythenorth> then extend with synonyms
13:27:03 <TrueBrain> brings all the GRFs on the same level too
13:27:10 <TrueBrain> as some do, and some don't .. which is just annoying for the end-user
13:27:40 <LordAro> didn't we just get rid of the list of countries from the server list? :p
13:27:44 <andythenorth> we'd want to ideally recognise that "UK", "Britain", "British" and "United Kingdom" all resolve
13:28:04 <TrueBrain> LordAro: no, we removed the silly flags ๐ Which is a different issue ๐ ๐
13:30:36 <andythenorth> "Thereโs also a โfuzzyโ search to help people discover โproperโ countries for names that might only actually be subdivisions"
13:30:46 <andythenorth> and historical resolution
13:31:23 <LordAro> make it locale specific, so "falklands" will reso- actually i'll stop there
13:31:23 *** Lemuria has joined #openttd
13:31:23 <Lemuria> andythenorth: Oh, and good luck recognizing other languages, have fun with "Estados Unidos", "Tsina" ("China" in Tagalog), and Deutschland
13:31:53 <andythenorth> filter rather than text search?
13:31:59 <andythenorth> or type-ahead filter combobox?
13:32:43 <LordAro> filters are generally a lot more user friendly, especially when you already have all the data
13:33:51 <andythenorth> what about Isle of Sodor? ๐
13:34:56 <LordAro> does the doom soundtrack get a country code?
13:35:27 <Chris> What about multi-country sets?
13:35:29 <Rubidium> maybe just an universe map (without contentious borders) where the developer marks an area as their set covering that area. Then the user can select a location/area on that map for filtering
13:36:02 <LordAro> Chris: well that one's easy :p
13:36:15 <Rubidium> only problem is that people wanting to play a Korean-based game need to know where Korea is
13:36:21 <andythenorth> hmm yeah GIS region(s) polygons
13:36:26 <andythenorth> then intersect those with lat-long
13:36:36 <LordAro> Rubidium: about time people knew geography a bit better
13:36:39 <andythenorth> fortunately we don't have subways, so we don't need a height co-ordinate
13:36:49 <Chris> I don't envy the person that has to make the 2cc trainset country list
13:36:53 <Lemuria> Oh, and don't even get me started on territorial disputes
13:37:31 <andythenorth> agreed, don't get started ๐
13:37:59 <andythenorth> Chris: what is 2CC tagged as currently, OOI?
13:39:14 <Rubidium> although your search for Belgian trains might end up with toyland inspired sets ;)
13:39:23 <Chris> andythenorth: "train, 8bpp"
13:39:31 <andythenorth> lol no countries?
13:40:04 <TrueBrain> Chris: did you know only a very small amount of sets are multicountry ๐ But yeah, list ๐
13:40:18 <Chris> I'm looking out for my own babies okay :(
13:41:37 <andythenorth> oh my JGRPP map is going to be done soon
13:41:53 <andythenorth> 4 hrs 20 minutes so far
13:42:03 <andythenorth> so yes, might complete within 5 hours
13:42:19 <andythenorth> but there will be at least one FIRS port in a tiny lake, so I'll have to start over ๐
13:42:54 <andythenorth> with stupid number of towns and industries
13:42:59 <JGR> What do you expect to learn by doing this?
13:43:01 <andythenorth> I'm trolling myself
13:43:21 <andythenorth> I started, and then I got committed to it
13:43:30 <Samu> hey, any solution for the path reservation issue yet?
13:43:34 <andythenorth> I fully expected it to be 5 mins or so, it was last time I tried
13:43:45 <andythenorth> different map settings I assume ๐
13:44:30 <Samu> I had this idea: when a train collides with a reservation, cancel it
13:44:41 <glx[d]> the collision is not the issue
13:44:44 <Samu> but then i am too noob to check where i would do this
13:45:25 <glx[d]> you can have broken reservation before the collision even happen
13:46:03 <glx[d]> but it only happens if you touch the track/signal or force to ignore
13:47:37 <FLHerne> reldred: as a typical random outsider, I've had my PRs reviewed and accepted reasonably quickly recently
13:48:10 <FLHerne> there was definitely a time when it was pretty hopeless, but it seems better now at least for smaller things
13:48:37 <FLHerne> some big grf-API things like the bridges still seem stuck indefinitely
13:49:14 <TrueBrain> frosch: could use some ideas / suggestions on #332 here .. got a bit stuck on terrain-type-like classifier ๐
13:52:33 <Samu> the bottom train wanted to go to the right, but a train coming from the north force proceed and ate the reservation
13:53:12 <TrueBrain> LordAro: smart, I will check ๐ I was mostly just annoyed with E731 ๐
13:55:06 <glx[d]> Samu: an example where reservation is broken way before the train crash
13:55:17 <reldred> FLHerne: Yeah, I dunno, my turnaround time to having things committed elsewhere is drastically shorter than what seems to be the average here. While things lately have been better, watching well and truly battle tested features cherrypicked from JGRPP sit as PR's with no progress, or just back and forth bickering for what seems like months on end makes me not even want to bother.
13:56:28 <Samu> glx[d], my idea was that when the train that force proceed reached the reservation of the other train, cancel the entire reservation of the other train
13:56:57 <andythenorth> one the great benefits of JGRPP is that it provides a place where progress is rapid
13:57:01 <glx[d]> and post crash cleanup can't remove reservation to the depot because it's not a continuous reservation
13:57:02 <JGR> Delays on upstreaming some of my things are in part because it's not a high priority for me
13:57:12 <LordAro> reldred: i'd like to see some examples
13:57:23 <andythenorth> the advantage of one person who can make decisions ๐
13:57:32 <andythenorth> OpenTTD vanilla just isn't that
13:57:40 <LordAro> there are a few open PRs that mention "JGR", but they're all either very large, or genuinely in need of more work
13:58:09 <glx[d]> large PRs are hard to review ๐
13:58:31 <JGR> Which is why I don't usually bother submitting big features
13:58:43 <LordAro> along with the whole "JGR can move fast and break things, we don't have that luxury"
13:58:53 <LordAro> as proved by the earlier discussion
13:58:54 <glx[d]> and often we find issues month after large PR merge
14:00:03 <glx[d]> right after a release because nightlies/beta/RC are not really used
14:00:23 <andythenorth> that's the only downside of JGRPP
14:00:35 <andythenorth> it has drawn in the players who used to play nightlies
14:00:38 <JGR> My solution to this is to just release more often
14:01:05 <andythenorth> I don't use nightlies, because I play self-compiled because I have patches
14:01:12 <andythenorth> because vanilla doesn't move very quickly
14:01:52 <LordAro> trying for more frequent releases so often becomes "we can't release until this one issue is fixed"
14:02:01 <LordAro> or "just one more thing, it's nearly done"
14:02:08 <andythenorth> is that because releasing is quite expensive?
14:02:25 <LordAro> or just a lack of effort required to actually put the release together, which i think was mostly the cause of the delay of 13?
14:02:34 <LordAro> speaking of, 13.1 has been needed for a while...
14:06:51 <TallTyler> I wonder if simplifying (automating more of) the release process would ease the mental barriers to releasing often
14:07:13 <LordAro> what more can be automated?
14:07:22 <LordAro> changelog is always going to need manual finessing
14:07:43 <LordAro> beyond that it's just a few buttons (+ steam and friends)
14:08:30 <glx[d]> releasing is just a button yes, preparation for release on the other hand ๐
14:09:52 <Samu> op,s it's train 1, not 2
14:10:15 <LordAro> Samu: what probably are you trying to solve?
14:10:16 <TrueBrain> so that is what he is basically saying .. the release itself is automated, now let's automate the preparation for it ๐
14:10:32 <TrueBrain> so many small things you have to keep in mind ... like website, social posts, etc etc
14:10:51 <glx[d]> lordaro: stale reservations
14:11:35 <Samu> but at least the reservation is already cancelled
14:13:02 <TrueBrain> Chris: you perfectly describe why the tags never converged indeed ๐ Like I said, bit naive on our part ๐
14:13:02 <andythenorth> oh this bot ๐
14:13:13 <andythenorth> who is Samu now?
14:13:27 <TrueBrain> fixed in my client at least ๐
14:13:40 <TrueBrain> they had many complaints ๐
14:13:48 <andythenorth> I am enjoying the lolz
14:13:53 <andythenorth> people talking to themselves
14:14:30 <andythenorth> yeah browser-refresh fixed it for me now
14:14:56 <andythenorth> TallTyler: there are instructions for release somewhere?
14:15:04 <andythenorth> also the framing
14:15:15 <andythenorth> pyschologically, we seem to think 1 main release per year is the cadence?
14:16:20 <TallTyler> Pinned in this channel
14:16:42 <TallTyler> And quite a few after it
14:17:42 <andythenorth> well that's good
14:17:56 <andythenorth> should I move it to a .md in github?
14:18:00 <andythenorth> that's where I was looking
14:18:45 <TallTyler> Iโve only done one release but besides writing things (unavoidable), waiting for actions to finish before manually triggering the next step means it takes up a large chunk of oneโs day. Itโs not a quick โtake an hour to write stuff and press startโ, it feels like an actual job
14:19:06 <TallTyler> Just an observation, Iโm certainly not qualified to improve the process ๐
14:20:53 <TallTyler> If there were momentum for a twice-yearly release that actually brought new features (from JGRPP or otherwise) I would be happy to put in the time to complete the current process
14:22:44 <LordAro> well yeah, the other thing is whether there's actually enough stuff to warrant a release
14:22:52 <LordAro> most of the stuff in 13 wasn't there until a few months before
14:23:20 <glx[d]> and there's also how well tested is the added stuff
14:23:24 <andythenorth> early, often etc
14:23:35 <andythenorth> is anybody testing the added stuff at all?
14:23:38 <TallTyler> I also think the current explanation needs more backporting instructions, as if I were asked to do 13.1 I would not know how this works. Thereโs a link to a script but I could use a high-level explanation of what it does and why
14:23:42 <andythenorth> no devs really play
14:23:50 <andythenorth> almost all of discord uses JGRPP
14:23:57 <andythenorth> I don't use nightlies
14:24:17 <TallTyler> We still get plenty of bug reports about things that we released which werenโt quite right
14:24:19 <andythenorth> could we get a steam preview track somehow?
14:24:25 * andythenorth doesn't understand steam
14:24:37 <glx[d]> nightlies are in steam
14:24:52 <TallTyler> I got fixes to wide level crossings and multi-track rivers from the beta and RC builds
14:25:15 <andythenorth> andythenorth: so 'yes' then
14:26:02 <TrueBrain> TallTyler: know the backporting script is really new. I think I created it for 12.0? Maybe a release earlier .. but yeah, more automation == good ๐
14:32:29 *** D-HUND is now known as debdog
14:33:01 <Xarick> i wanna do something like this
14:36:34 <iSoSyS> Hello! Did anyone here ever attempted to cross compile from linux to windows using both mingw and vcpkg?
14:37:58 <iSoSyS> I was able to compile without vcpkg. But with vcpkg I'm having trouble figuring the correct flags.
14:38:12 <glx[d]> we "cross-compile" for windows arm on windows x64
14:38:20 <iSoSyS> Just wondering if anyone ever did it, so I could copy their work ๐
14:38:41 <glx[d]> for vcpkg you need to use the right triplet
14:40:54 <glx[d]> x64-mingw-static I think
14:43:29 <iSoSyS> yeah, that what I tried. But I guess only indicating the triplet is not enough. It attempts to search outside cross compilation sysroot.
14:43:29 <iSoSyS> I guess I will need two toolchain files, one to setup the cross-compilation, and then the vcpkg one.
14:43:46 <glx[d]> first the tools (without vcpkg) then the actual build using the tools
14:44:40 <glx[d]> might also need to set some cmake stuff for crosscompilation
14:45:54 <iSoSyS> I think I got it working, thanks
14:46:38 <iSoSyS> this was the command I used
14:46:38 <iSoSyS> cmake -S . -B build_vcpkg_13/ -DCMAKE_CROSSCOMPILING_EMULATOR=/usr/bin/x86_64-w64-mingw32-wine -DCMAKE_TOOLCHAIN_FILE="/opt/vcpkg/scripts/buildsystems/vcpkg.cmake" -DVCPKG_TARGET_TRIPLET="x64-mingw-static" -DVCPKG_CHAINLOAD_TOOLCHAIN_FILE="/usr/share/mingw/toolchain-x86_64-w64-mingw32.cmake" -DBUILD_SHARED_LIBS:BOOL=OFF -DCMAKE_BUILD_TYPE=Release
14:57:24 *** Wormnest has joined #openttd
14:58:56 <Samu> i doubt this is sufficient
15:04:47 <TrueBrain> I can work with most of that ๐ Nice ๐
15:14:22 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
15:15:58 <Samu> hmm block signals work different than pbs signals
15:16:16 <Samu> when reserving up to a block signal, it include the tile with the block signal
15:16:38 <Samu> that causes a false positive that the reservation is being crossed
15:17:41 <Samu> they are contiguous, but there is a block signal acting as separator
15:21:53 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
15:23:35 *** WormnestAndroid has joined #openttd
15:27:40 <Samu> how do i build block signals now? need to debug
15:29:00 <Samu> they're gone from the gui?
15:35:12 <andythenorth> now on to GS ๐
15:37:25 *** sla_ro|master has joined #openttd
15:55:07 *** Wormnest has joined #openttd
15:59:14 <FLHerne> (diesel railcar hauling seven 4-wheeled horseboxes)
16:07:23 <andythenorth> my GS is quite busy during map gen
16:36:58 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
16:37:05 *** WormnestAndroid has joined #openttd
16:47:23 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
16:49:13 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
16:50:11 *** WormnestAndroid has joined #openttd
16:55:48 <TallTyler> andythenorth: Tomorrow's event game is using PIRS and Horse and will last about 15 years. What's the correct start year for maximum Horse enjoyment? 1980?
16:55:50 *** tokai|noir has quit IRC (Quit: c('~' )o)
16:56:06 <andythenorth> anything from 1960 onwards
16:56:25 <andythenorth> 1990 or so will give faster freight
16:56:27 <TallTyler> I'm thinking 1980 because if we play until 1995, we'd get some of the 1990-gen engines ๐
16:56:37 <andythenorth> you need AV 9.8 with PIRS btw
16:56:55 <andythenorth> is it with Pikka's goal script also?
16:57:28 <andythenorth> make sure 'multiple airports per town' is allowed
16:57:33 <TallTyler> Sam, maybe, or would that be too many ships with SHARK?
16:57:55 <andythenorth> Deliverance meta appears to be Concorde ๐
16:58:06 <andythenorth> for winning ๐
16:58:13 <andythenorth> we shall see ๐
16:58:38 <TallTyler> Aircraft speed factor should be the default 1/4?
16:58:57 <TallTyler> It always seems slow to me, but I usually start earlier when airships are absolutely uselessly slow
17:01:34 <andythenorth> 'probably fine either way'
17:01:46 <andythenorth> the planes are mostly about not having to build infra
17:09:54 <Samu> unreserving is bad idea if it's done wrong :(
17:15:49 <andythenorth> I've been watching TV
17:25:18 *** Wormnest has joined #openttd
17:34:47 <andythenorth> the map finished!
17:35:07 <andythenorth> 8 hours 10 minutes
17:35:33 <andythenorth> but my checkout of JGRPP is old, and doesn't work with variants
17:35:45 <andythenorth> so I'd have to generate the map again ๐
17:38:41 <andythenorth> oh this might be a debug build
17:40:56 <andythenorth> the fps on the game appears to be 0
17:41:07 <andythenorth> the UI is completel unresponsive
17:42:58 <FLHerne> go to Mesa-style releases?
17:43:15 <FLHerne> year.quarter, release regardless of whether anything's particularly featureful
17:43:33 <andythenorth> I have worked with scheduled release trains
17:43:40 <andythenorth> whatever is on the train, goes
17:44:13 <FLHerne> it seems to me to save a lot of angst
17:44:48 <FLHerne> "do we hold up the release for this feature" "is this feature worth a release" "I need to get my feature done for the release or it'll wait a year"
17:44:56 <andythenorth> it's either done
17:45:00 <andythenorth> or there's another train later
17:46:09 <FLHerne> TallTyler: what's an "event game" ?
17:46:29 <TallTyler> Scheduled trains is thematically relevant to OpenTTD ๐
17:46:46 <andythenorth> yeah but vanilla timetabling is broken
17:47:59 <TallTyler> Oh, you canโt see that on IRC probably
17:48:33 <TallTyler> Itโs a public server game Iโm running tomorrow for a few hours, using PIRS and the Deliverance GS plus some other grfs
17:48:46 <FLHerne> which then asks me to log into discord, but eh
17:49:02 <TallTyler> 9 AM Eastern time, not sure where you are to convert that but is midafternoon in Europe
17:49:58 <TallTyler> I can share the server name and password here when it starts if any IRC people are interested, but weโll be using voice chat on Discord in addition to game chat
18:07:37 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
18:08:51 *** WormnestAndroid has joined #openttd
18:12:09 <glx[d]> samu: I feel it will be expensive, as the typical reservation use will make this check every tile
18:17:59 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
18:18:02 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
18:18:42 <Samu> wentbourne save is good for testing?
18:19:52 <Samu> I'm also worried it might break in other specific cases, such as mixed signal types
18:20:32 <Samu> or that when unreserving, it might open a path that would cause a disaster
18:20:59 <petern> Oh right, stupid Discord changing names.
18:23:31 <petern> Walk in the park for you these days ๐
18:28:11 <Samu> looks like wentbourne is only using block signals
18:29:52 *** Wormnest has joined #openttd
18:30:56 <Samu> anybody got a big train savegame other than wentbourne that uses a mix of pbs + block? I could try running AI's but...
18:31:36 <petern> All the big aavegames were openttdcoop and they shun path signals
18:32:36 <LordAro> petern: not in february
18:33:17 <LordAro> Samu: one of the coop saves? dunno if they mixed signals though
18:38:45 <Samu> doesn't seem to slow down
18:38:58 <Samu> getting the same 33 ms with and without
18:39:08 <Samu> on wentbourne, but that one uses block signals a lot
18:45:22 <DorpsGek> - Update: Translations from eints (by translators)
18:50:11 <Samu> nasty ads on openttdcoop
19:02:38 <andythenorth> Soโฆgoes it throw out GS story page buttons? ๐
19:04:13 <andythenorth> A throttled event queue is possibly the worst way to handle UI clicks ๐
19:17:22 <TrueBrain> TallTyler: I am so tempted to take you up on this offer ๐ I was thinking what we could do to simplify the system .. and I have some ideas. Time is an issue, as I have 4 projects now I want to work on, but what-ever ๐
19:18:14 <TrueBrain> mostly what holds me back is the "too many places to do things" .. but it should be relatively easy to make a piss-simple webpage that asks in a single form about everything, and when you hit the button it makes things like news posts, images, changelogs, etc
19:18:55 <LordAro> something that autogenerates the "initial" changelog would be nice
19:19:06 <TrueBrain> yeah, and fixes all the common mistakes
19:19:11 <TrueBrain> like dots, PR numbers, order, etc
19:19:33 <TrueBrain> we could make it more a point to leave good PR titles when creating PRs / merging them, that would also really be helpful
19:19:55 <andythenorth> Thought that was scripted alread? ๐ฎ
19:20:06 <TrueBrain> `git log --oneline` is not really a script
19:20:15 <LordAro> changelog form does not necessarily make a good pull request title
19:20:15 <andythenorth> Thought that was the whole point of the prefixes
19:20:27 <andythenorth> Maybe I confuse an old coop script
19:20:32 <TrueBrain> LordAro: true; but it could be more helpful ๐
19:20:50 <andythenorth> Make release in coop makefile made the changelog
19:20:55 <TrueBrain> I did consider one other thing briefly .. what if we just automatically make a release every week of what-ever is in master
19:21:14 <TrueBrain> but I soon after realised, OpenTTD kinda works as every once in a while you a reminded of a new version .. we end up on Hackernews etc
19:21:20 <TrueBrain> you don't have that with a continues deployment
19:21:39 <andythenorth> Also server-client versions
19:21:41 <LordAro> that works for web things, not so much OTTD
19:21:52 <TrueBrain> LordAro: well, with Steam that kinda changed, honestly ๐
19:22:06 <TrueBrain> but yeah, it would require a non-Steam autoupdater too ๐
19:22:08 <TrueBrain> so bench that idea ๐
19:22:25 <andythenorth> Unless we autoupdate we have a very heterogeneous fleet
19:22:36 <andythenorth> Hard to get into an mp game
19:23:15 <TrueBrain> the backporting is now as easy as we can make it I think .. the other thing that might help is to give a brief overview of PRs -not- being backported
19:23:25 <TrueBrain> although GitHub made that better now you can sort on "merged" date
19:24:08 <TrueBrain> so backport -> changelog -> next is website stuff .. it might sound silly, but we can actually use ChatGPT here to generate a first draft of a fun news message
19:24:21 <TrueBrain> we can automate the making of the website PR etc etc .. so we can smoothen out this experience by a lot
19:24:43 <TrueBrain> we already solved versioning etc .. just every branch is a bit of work; but that is just what it is (for now) ๐
19:24:54 <TrueBrain> anything else that would be useful to automate a bit more / better?
19:25:10 <andythenorth> ChatGPt release announcement
19:25:46 <andythenorth> Bananas version list?
19:26:00 <dP> andythenorth: yeah, I'd say anything less than monthly is way too frequenquent for updates
19:26:07 <dP> half of the servers are still on 12.2
19:26:23 <andythenorth> But monthly is a nice cadence
19:27:10 <TrueBrain> owh, news image ... I am sure we can find a way to autogenerate that too ..
19:27:18 <andythenorth> Has to be frequent enough to be worth automating
19:27:44 <andythenorth> Does server-client require exact same rev?
19:28:32 <andythenorth> Ok so we canโt major-minor-minor?
19:28:33 <dP> it may be possible to have some releases compatible
19:28:38 <LordAro> dP: as far as anyone else is concerned, yes :p
19:28:42 <dP> but not sure if worth the effort
19:31:24 <dP> iirc there was some release that was nearly compatible to the previous
19:31:42 <TrueBrain> is that something like "almost pregnant"?
19:32:09 <TrueBrain> or "a bit pregnant", what is the english verb
19:32:38 <dP> you could play some games without desyncs with it
19:32:49 <dP> not sure how that compares to pregnancy :p
19:33:03 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
19:35:14 *** Wormnest has joined #openttd
19:37:23 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
19:37:38 *** WormnestAndroid has joined #openttd
19:40:07 <Samu> uh oh, glx[d] road vehicle ticks over 260 ms
19:40:55 <Xarick> it's not over, I wonder what is it now
19:58:05 <Samu> moguls is causing the 260 ms...
20:01:18 <andythenorth> Probably looling for depots ๐
20:40:50 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
20:41:28 *** Eddi|zuHause has joined #openttd
20:46:58 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
20:50:06 *** WormnestAndroid has joined #openttd
20:58:06 <Samu> go to nearest road vehicle depot when there is no road vehicle depot nearby
20:58:15 <Samu> on a giant road network again
20:58:31 <TrueBrain> it is like, someone told you already ๐
21:20:26 <andythenorth> because it's been found before I think
21:20:49 <andythenorth> and because RVs don't find depots reliably
21:20:58 <andythenorth> about which I never filed a bug, just complained in irc ๐
21:21:20 <andythenorth> they just drive around looking for depots, which presumably causes a lot of pathfinding
21:21:28 <andythenorth> turn off breakdowns, problem goes away
21:22:31 <andythenorth> my information might be 15 years out of date
21:25:06 <Samu> this one is a bit of a chicken and egg situation :(
21:25:44 <Samu> go to nearest road vehicle depot needs to call the pathfinder, but this time the pathfinder fails
21:26:18 <Samu> and the nearest road vehicle depot is not set to anything, so it repeats again
21:29:28 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
21:43:03 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
21:44:54 *** WormnestAndroid has joined #openttd
21:52:51 *** Wormnest has joined #openttd
22:09:53 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
22:14:29 *** WormnestAndroid has joined #openttd
22:14:43 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
22:14:44 *** WormnestAndroid has joined #openttd
22:15:45 <glx[d]> samu: it's not the first high cpu pathfinding with RVs
22:17:05 <Samu> looks like it's unrelated to being stuck
22:17:50 <Samu> mogul tricked me into believing me it was another stuck, repeat pathfinder on vehicle controller
22:18:06 <Samu> this one is on processorders
22:19:56 <Samu> on the game this happened, there was a depot, but it was probably out of reach from the pathfinder, and wasn't returned
22:20:16 <Samu> it was far far away, past a big road network
22:21:08 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:21:38 <Samu> funny thing, when i place a depot near the vehicle, the cpu spike disappears :)
22:21:56 <glx[d]> yes because it can find one
22:24:46 <glx[d]> the issue here is order search for closest depot, can't find it, skip to next order, ...
22:25:46 *** Terkhen has quit IRC (Quit: ...)
22:25:46 *** Ammler has quit IRC (Quit: gone...)
22:25:46 *** V453000 has quit IRC (Quit: You must construct additional PYLONS to get me back!)
22:25:46 *** XeryusTC has quit IRC (Quit: So long, and thank's for all the fish)
22:25:46 *** planetmaker has quit IRC (Quit: ...und tschรผร!)
22:25:52 *** Webster has quit IRC (Read error: Connection reset by peer)
22:28:51 <glx[d]> but having only "go to nearest depot" it's an infinite loop
22:29:31 <Samu> go to nearest ship depot on a ship doesn't seem to cause an issue
22:30:15 <Samu> maybe copy behaviour from there?
22:30:56 <glx[d]> but the code is the same for order handling
22:33:16 *** Hazzard has joined #openttd
22:34:23 *** planetmaker has joined #openttd
22:34:23 *** ChanServ sets mode: +o planetmaker
22:35:12 *** Hirundo has joined #openttd
22:35:51 <Samu> ah, nevermind, nearest ship depot is just iterating over all existing depots, which is none, no wonder there's no stalls, doesn't even use pathfinder
22:36:18 *** ^Spike^ has joined #openttd
22:36:49 *** Terkhen has joined #openttd
22:36:49 *** ChanServ sets mode: +o Terkhen
22:36:50 *** V453000 has joined #openttd
22:36:56 *** ^Spike^ is now known as Guest5864
22:37:20 *** XeryusTC has joined #openttd
22:38:16 <TrueBrain> Will run that idea tomorrow; see what it yields ๐
22:38:24 *** fonsinchen has joined #openttd
22:39:20 <TrueBrain> frosch: What did you think about resolution? Not really sure about the values ..
22:40:21 <frosch> my first though was you messed up your git history, because resolution was present and removed from the newgrf classifier :p
22:40:36 <TrueBrain> Yup .. had to fix the merge conflict ๐
22:41:03 <frosch> i think the name "resolution" is fine though
22:41:05 <TrueBrain> I actually forgot we removed it :p
22:41:19 <TrueBrain> Low / medium / high as values?
22:41:24 <frosch> possibly we could reject heightmaps smaller < 256x256, they make no sense
22:41:42 <frosch> then normal/high-resolution would make sense for the rest
22:42:42 <frosch> ok, i'll add it in the comments
22:42:46 <TrueBrain> Will clean it up and see what that brings ๐
22:43:29 <TrueBrain> Will also see how to add region/country input .. can't be classified ofc, but we can ask the user
22:43:45 <TrueBrain> Sounds like a nice addition ๐
22:44:24 <frosch> i am still reading the discussion, so much text :p
22:44:48 <frosch> it's funny that the content list does not filter on description
22:44:50 <TrueBrain> Yeah, the pitchforks came out. But in the end there was a civil conversation ๐
22:45:03 <frosch> i knew that the newgrf settings do filter on description
22:45:22 <frosch> and you complained a lot about content-query sending so much data, and then it is not even used? :p
22:45:42 <TrueBrain> I hate it even more now ๐
22:47:03 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
22:47:34 <TrueBrain> (And yes .. still have to actually measure, but it is more than 1MB for the listing :p)
22:50:51 <frosch> hmm, i found some pattern: we have "legacy newgrf" which do houses, road and rail sprites by replacing the original sprites. and we have "modern newgrf" which do the same using "new houses" and "road types"
22:50:59 *** Webster has joined #openttd
22:51:23 <frosch> but labeling them like that is weird :p
22:52:18 <TrueBrain> Yeah, users won't understand the difference, nor actually care
22:52:52 <TallTyler> I bet legacy grfs are still quite popular, we have some ancient mods that have never been replaced
22:54:24 <TallTyler> Samu: It probably only fixes a corner case but I wonder if you could bail out from searching for the nearest depot if the company doesnโt have any depots
22:55:17 <TallTyler> Granted, I suspect not being able to reach a road depot at all is already somewhat unusual โ the user would need to remove road or do strange things with one-way roads to get themselves into that situation
22:55:28 <TrueBrain> Btw, did some stats on tag .. 14 ways to mention roadveh.. made me giggle ๐
22:57:30 <glx[d]> TallTyler: in this case there's a depot, but unreachable
22:59:50 <glx[d]> and in most cases it's the result of AI doing silly choices
23:00:31 <Samu> well, the AI wanted to send the vehicle to the nearest depot and stop there, probably was going to sell unprofitable vehicle
23:01:02 <glx[d]> but the AI should have a reachable depot as it build the vehicles
23:01:21 <Samu> it had a depot at the destination
23:01:28 <Samu> unreachable by the pathfinder
23:01:43 <Samu> the road between where the vehicle is and the depot is just laaarge
23:02:10 <glx[d]> there is the 10k pf node limit
23:02:42 <Samu> i could post the savegame, but better have no ais installed
23:02:54 <Samu> it has massive amounts of vehicles
23:03:53 <glx[d]> I guess it was also a huge map
23:08:28 <frosch> technically we can also dinstinguish road-vehicle and tram sets. but i think most sets add both
23:08:51 <frosch> we could also read the railtypetable from vehicle grfs, and check whether they supply narrow gauge
23:09:09 <frosch> except I don't think there is any common agreement on what "narrow gauge" means :p
23:09:28 <TrueBrain> The latter is a bit of an issue, yes ๐
23:10:09 <andythenorth> 'standardised railtype schema'
23:10:12 <andythenorth> make it official ๐
23:10:29 <frosch> I don't think the Innsbruck 2022 conventions helps us here
23:10:47 <TrueBrain> For region etc, I can see ISO codes work, as that allows for translations easier .. but a free field is easier ..
23:10:52 <andythenorth> I like Innsbruck, has a nice bridge
23:11:10 <andythenorth> TrueBrain: auto-complete ๐
23:11:39 <frosch> TrueBrain: iso codes only work for train sets, because train companies are so national
23:12:00 <frosch> for road vehicles i think it's more about continents
23:12:16 <TrueBrain> Those also have iso codes ๐
23:12:22 <andythenorth> there are only 3 RV sets anyway?
23:12:27 <TrueBrain> Just a burden to maintain
23:12:39 <frosch> do you have a list of the old tags somewhere? we could filter it for obvious junk, and see what's left over
23:13:00 <TrueBrain> But on mobile, so cant share today
23:13:17 <TrueBrain> Over 150 tags are just random garbage :p
23:13:30 <frosch> good, i need to shift my sleeping schedule anyway ๐
23:14:17 <frosch> nah, some craftpeople want to show up on monday morning
23:14:37 <frosch> and they start working like hours ealier than I normally do :p
23:15:13 <frosch> though my current work-houses are even late for me :p
23:15:51 <TrueBrain> Btw, another issue with region, is that some say British, others UK, etc ..
23:16:01 <TrueBrain> So many aliases of the same thing
23:16:19 <frosch> i assume you know the classic cgpgrey video on that?
23:16:19 *** Samu has quit IRC (Quit: Leaving)
23:16:31 <Rubidium> frosch: you mean the Dutch railways in the UK and Germany? ;)
23:16:33 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:16:46 <TrueBrain> Most likely, but can't remember, so: share!
23:18:28 <frosch> Rubidium: aren't most trainsets about steam engines though?
23:18:35 <TrueBrain> Also others like Dutch and Netherlands, etc etc .. guess picking a certain way to represent that wouldn't hurt ๐
23:19:26 <frosch> if you add all iso codes, it becomes search-only again, no dropdowns ๐
23:19:46 <frosch> continent-only could be a dropdown
23:19:46 <TrueBrain> Well, only a handful are used atm
23:20:10 <TrueBrain> The list is not as long as you might think ๐
23:20:37 <TrueBrain> But I will make a list and we can see ๐
23:20:43 <TrueBrain> Author aliases is also a think
23:20:52 *** WormnestAndroid has joined #openttd
23:20:53 <Rubidium> glx[d]: at least those run in the Netherlands. I'm talking about it running trains from London to towns like Nottingham
23:20:55 <TrueBrain> Timberwolf for example ๐
23:21:30 <TrueBrain> His Github name is totally different ..
23:21:56 <andythenorth> took me a while to work out michi on github ๐
23:22:22 <TrueBrain> For now: sleep well ๐
23:23:02 *** Wormnest has joined #openttd
23:26:12 <Timberwolf> Oh yes, I used GitHub for my day job first and never made a new account so it has my real name.
23:27:11 *** herms has quit IRC (Quit: bye)
23:28:34 *** keikoz has quit IRC (Ping timeout: 480 seconds)
23:41:08 *** ChanServ sets mode: +v tokai
23:47:04 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
23:50:56 *** Eddi|zuHause has joined #openttd
continue to next day โต