IRC logs for #openttd on OFTC at 2024-04-05
⏴ go to previous day
00:43:55 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
00:48:50 <peter1138> Should be asleep, of course, but...
00:50:30 *** bertvvvs has quit IRC (Quit: You have a nice day.)
00:53:41 <peter1138> Hmm, George seems to have gone for "actual numbers" or something, instead of prototypical models. Do we need some flag so that only one of a kind can be built...
00:54:33 <peter1138> (The stats on these are all the same, just the introduction year increases)
00:56:07 *** emperorjake has joined #openttd
00:56:08 <emperorjake> Limiting how many of a certain vehicle can be built, seems like a "bad feature" to me
00:56:53 <peter1138> Quite. As does making 8 almost identical engines.
01:19:24 *** johnfranklin has joined #openttd
01:19:24 <johnfranklin> I heard something very terrible happened
01:56:23 *** Wormnest has quit IRC (Quit: Leaving)
02:59:45 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:03:00 *** gnu_jj_ has joined #openttd
03:06:07 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
04:43:32 <DorpsGek> - Update: Translations from eints (by translators)
04:51:12 *** D-HUND is now known as debdog
05:12:01 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:24:08 <andythenorth> peter1138: yes, then the UK sets can be limited to the number of real class 58s built
06:38:52 <emperorjake> or JP+ Engines can be like, no you only get 1 of these, it's a prototype
06:49:11 <emperorjake> would it be per map or per company?
07:13:56 <andythenorth> if you scrap it, you can't get it back
07:14:13 <andythenorth> but we'd have to add vehicles-in-vehicles, so you could low-loader it across the map
07:24:16 <peter1138> One way to solve #12100 is to use a different flag for random flipping, and only (try to) copy the manually flipped flag.
07:24:36 <peter1138> But then there would be 4 different flags related to reversing rail vehicles...
07:30:05 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
07:36:21 <andythenorth> who reported that anyway?
07:37:54 <LordAro> peter1138: re #12233 - what happens if there are 5 orders? looks suspiciously like it would overlap the lower text
07:38:14 <peter1138> It's always limited to 4.
07:39:21 <peter1138> vehicle_gui.cpp:1625
08:16:41 *** Smedles has joined #openttd
08:55:25 *** gelignite has joined #openttd
09:40:01 *** lemuria0685 has joined #openttd
09:40:01 <lemuria0685> has there been any progress on openttd 14 or are we still being held back
09:40:17 <lemuria0685> (wait, does jgr 0.58.1 already have vanilla 14 features)
09:41:13 <andythenorth> nothing is known to have changed with the upstream issues
09:42:07 <andythenorth> alternatives to waiting have been explored, but not fruitful
09:43:46 <kamnet> andythenorth: JGR does have 14 features. You can also just be playing the 14.0-RC3 release right now.
09:44:20 <andythenorth> discord reply strikes again?
09:44:20 <lemuria0685> yeah i think i'm good, i'm a jgr player so won't be affected that much by the xz drama
09:45:05 *** matusguy has joined #openttd
09:46:17 <kamnet> Sorry didn't have my glasses on.
09:46:56 *** bertvvvs has joined #openttd
09:52:32 *** reldred has joined #openttd
09:52:32 <reldred> lemuria0685: It’s probably holding up JGR bumping out any new releases as well depending what his release build chain looks like
09:53:21 *** matusguy_ has joined #openttd
09:59:37 *** matusguy has quit IRC (Ping timeout: 480 seconds)
09:59:37 *** matusguy_ is now known as matusguy
10:04:55 <_jgr_> I could work around it, but it's not worth the bother as there's no rush from my end
10:50:12 <lemuria0685> Yeah I'm not exactly starving for new JGR features, so take your time...
10:51:22 <lemuria0685> ~~I would like to cast a curse on the xz backdoor attacker/s where every mode of transport with a schedule they take will always be severely delayed. Canceled flights, 4 hour late trains, and so much more~~
11:29:21 *** kuhnovic has joined #openttd
11:29:21 <kuhnovic> xarick: I'm curious if this fixes the issues you recently saw with ships getting stuck
11:32:41 <peter1138> Will we ever fix line drawing or should I switch to rect drawing... hmm.
11:33:22 *** gelignite has quit IRC (Quit: Stay safe!)
11:36:32 <peter1138> And is this acceptable...
11:54:11 <peter1138> Failure, I have nothing to salad with.
11:54:29 <andythenorth> my salad was toast
12:00:32 <peter1138> Cheese hot cross bun, and a packet of fridge raiders.
12:06:27 <peter1138> Well, I guess it's better anyway.
12:26:29 *** talltyler has joined #openttd
12:26:29 <talltyler> Yes please, groups nested with whitespace only is really hard to read when you have a lot of groups
12:27:44 <peter1138> I came up with a decent enough algorithm to determine which levels continue with only one iteration of the completed list, so I'm happy with that 😄
12:29:47 <talltyler> Nice, I look forward to reviewing it 🙂
12:33:03 <peter1138> If I could do SetBit/ClrBit concisely, it would be shorter 🙂
12:34:04 <peter1138> Technically `x ? SetBit(mask, bit) : ClrBit(mask, bit);` works, but it's not really in the spirit of ternaries.
12:35:51 <peter1138> `if (x != HasBit(mask, bit)) ToggleBit(mask, bit);` probably too, but then it's not clear what is happening.
12:38:13 <_jgr_> Existing code tends to use SB for this case
12:39:20 <andythenorth> VariantContainers? 😛
12:41:28 <peter1138> Interesting, SB() (like the others) also returns the value along with changing it in-place, but that feature is not used anywhere.
12:42:53 <peter1138> Given my next line of code is assigning the value to something else, I could do that. But I think that may just be confusing.
12:49:21 *** nielsm has quit IRC (Ping timeout: 480 seconds)
13:34:42 <xarick> kuhnovic: there are still some lost ships
13:36:18 <kuhnovic> xarick: I saw lost ships in my saves too, but not all lost ships are stuck ships. Sometimes the AI blocks its own routes, and the ships are legitimately lost.
13:37:26 <xarick> less frequent I guess, but they still get in that limbo
13:38:22 <kuhnovic> But is it really a limbo? Do you have a save that shows it? (Preferably not 70K ships on a 4K map)
13:42:09 <xarick> i guess im gonna start a new save
13:43:48 <xarick> starting a new game and got a crash
13:43:55 <xarick> wondering if it's a me problem
13:45:59 <xarick> my debug code is not resilient enough when generating rivers
13:48:51 <xarick> building rivers, that is
14:03:52 <peter1138> As per #12043, should the tree lijnes have shadows...
14:18:07 <peter1138> Kinda think it's too prominent.
14:45:09 <talltyler> Yeah, the circle is missing its shadow
14:45:57 <peter1138> The circle is separate, that's a sprite and I'm not touching that here.
14:46:51 <LordAro> dynamic shadows when?
14:47:03 <LordAro> implement ray tracing just to display the UI
14:47:08 <peter1138> Depending on the time of day?
14:47:49 <LordAro> "-" looks fine to me, fwiw
14:59:46 *** gelignite has joined #openttd
15:14:38 *** matusguy has quit IRC (Remote host closed the connection)
15:23:41 <xarick> Can't get any ship lost 😦
15:28:26 <xarick> maybe i need more samples
15:40:03 <j_n> peter1138: what's going on here
15:40:07 <j_n> I don't remember planes falling out of the sky like that
15:48:17 <peter1138> You're just not trying hard enough.
15:48:46 <j_n> I have a 64x[max map size possible in JGRPP] map
15:48:58 <j_n> seeing if a yate haugan can reach the other end
15:49:33 <j_n> it'll probably crash long before it even has a chance of reaching the other end
15:49:51 <j_n> on an unrelated note, I should probably try playing on a thin map one day
15:50:00 <peter1138> If they had fuel, perhaps.
15:50:16 <j_n> I was pretty sure aircraft would fall out of the sky if they flew for too long
15:52:05 <j_n> I should probably try this with a road vehicle one day
15:54:22 <xarick> should crash the busses
15:54:38 <xarick> or ships or trains and so on
15:55:42 <j_n> reminds me of that pull request for OpenTTD that made hoverboats able to freely traverse land tiles and destroy trains
15:56:09 <j_n> would've made for a great April Fools update
15:58:40 *** _zephyris has joined #openttd
15:58:40 <_zephyris> peter1138: I posted the necessary sprites in the bug report 🙂
15:59:05 <peter1138> Sure but I'm not touching the sprites here.
15:59:30 <_zephyris> peter1138: More contrast is better!
16:00:13 <_zephyris> It's not too much visual noise, and it increases contrast. Those are the fundamental rules of good visual readability.
16:00:22 <_zephyris> (thanks for taking a look 🙂 )
16:01:09 <peter1138> The sprites, you can just open a PR to change then if you wish.
16:01:23 <truebrain> Lol, the xz github repo is no longer blocked.. it is removed 😛
16:01:33 <peter1138> While they're in the same area, the changes are not really related.
16:03:09 <andythenorth> j_n: it was never intended for merge 🙂
16:03:14 <andythenorth> it was a joke for a livestream
16:04:09 <peter1138> Someone poured their soul into that patch, only for the devs to REFUSE it!
16:04:54 <truebrain> Poor frosch, will he ever become a dev to avoid this problem?
16:09:18 <FLHerne> I linked all my friends to that gronk video
16:09:45 <FLHerne> every single one of them said something like "...what is this nonsense I don't get it"
16:10:07 <FLHerne> we must all have gone mad
16:12:31 <FLHerne> it's still hilarious but I don't know why
16:23:18 *** HerzogDeXtEr has joined #openttd
16:39:48 *** Wormnest has joined #openttd
17:20:49 <kamnet> Who is in charge of tt-wiki.net?
17:21:26 <peter1138> Well I can't be expected to read all of a single comment...
17:21:53 <peter1138> Probably orudge. tt-forums is down too.
17:49:43 <andythenorth> not sure I'm going to watch this 😛
17:50:03 <andythenorth> YouTube has put me into a clickbait hole
17:50:07 *** silent_tempest has joined #openttd
17:50:07 <silent_tempest> Huh okay did not know that Github would update that PR automatically.
17:56:46 <xarick> I can't trigger the ship issue in a new game 😦
17:57:00 <xarick> but it still happens in that 70k save 😦
17:57:53 <xarick> it's rarer, but still happens
18:00:01 <j_n> andythenorth: well there was a real backdoor in .xz
18:00:07 <j_n> but it didn't really wreck anything
18:00:12 <j_n> my distro version wasn't even affected
18:01:31 *** locosage has joined #openttd
18:01:52 <locosage> someone ran a static analyzer on openttd and made an article out of it
18:02:42 <xarick> can't reproduce it yet on a new game, kuhnovic
18:02:50 <locosage> well, I guess that someone is related to said analyzer development xD
18:04:48 <kuhnovic> Good, I might have just fixed it 😉
18:05:10 <LordAro> locosage: would've been nice if they actually reported the issur
18:13:06 <LordAro> 479 usages of lengthof in the source
18:13:12 <LordAro> bit of an effort to replace them
18:14:50 <LordAro> maybe just needs s/lengthof/std::size/ and seeing what breaks :D
18:15:29 <peter1138> It's not like I've not been doing them slowly...
18:15:38 <locosage> just replace and wait for the next article xD
18:15:41 <peter1138> But usually with more appropriateness than just replacing lengthof with sizeof
18:19:17 <peter1138> Often it's replacing with iterators, range-for, or a begin/end pair.
18:24:21 <peter1138> That got rid of both a lengthof and an endof...
18:32:56 <peter1138> Some of these I touched ages ago but they're either long lost in a branch or... a stash.
18:33:05 <peter1138> But easy enough to just redo them.
18:38:40 <merni> Peter "I have two patches for that now" 1138
18:42:38 <merni> Sounds like something for Discord channel #add-on-development :p
18:44:27 <xarick> technically not blocked, there's a detour to the north
18:47:39 <merni> well the "blcoked" lock is useless anyway
18:47:46 <merni> doesn't provide a useful route
18:53:59 <xarick> vehicle lists still taking lots of cpu time 😦
18:57:56 <peter1138> You'll get bored of this...
18:59:26 <talltyler> Before or after you get bored of this?
19:01:41 <peter1138> I think the fact I'm doing it again instead of just pulling out my old patches suggests I will not get bored 🙂
19:01:50 <talltyler> CodeQL says I’m a bad reviewer 😛
19:02:29 <peter1138> Fix one thing, introduce another,sigh.
19:04:55 <peter1138> We have CodeQL because we're all otherwise human. And it's hard to spot things that are slightly outside the code...
19:05:25 <peter1138> I do smirk at the suggestion that we use "lengthof" because we googled how to do it in C++...
19:06:46 *** keikoz has quit IRC (Ping timeout: 480 seconds)
19:18:41 <talltyler> And listened to the AI-generated quick answer
19:26:58 <xarick> kuhnovic: same problem apparently
19:28:40 <xarick> needs the right set of conditions to trigger it
19:29:34 <xarick> two path choices once again
19:34:31 <xarick> increase NUMBER_OR_WATER_REGIONS_LOOKAHEAD by 1 on reverse... but ehm, nevermind, not going to be easy to implement
19:34:57 <andythenorth> so....articulated ships then? 🙂
19:35:08 <andythenorth> or multiple holds?
19:35:17 <andythenorth> hmm if we had ships-in-ships....
19:35:25 <andythenorth> we could have 'hold' vehicles
19:46:04 <xarick> or maybe SHIP_REVERSAL_PENALTY needs to be a bigger number, sufficient enough not to be ambiguous in path choice
19:47:31 <xarick> was thinking about 1 million
19:51:45 <peter1138> Okay, that's enough for the moment...
19:53:15 <peter1138> So preferences for this hierarchy tree, shadow or no shadow?
19:54:07 <xarick> a big penatly solved this specific case 🙂
19:54:17 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
19:55:18 <locosage> shadow is not a bad idea in principle
19:55:24 <locosage> peter1138: but this looks meh imo
19:57:04 <locosage> maybe will help if icon has shadow too
19:57:15 <locosage> also maybe less dark shadow
19:57:15 <xarick> `constexpr int SHIP_REVERSAL_PENALTY = YAPF_TILE_LENGTH * 10000;` solved this case
19:57:22 <locosage> but then it won't be the same as text I guess
19:57:30 <peter1138> If it's just the icon, that's separately solvable, as I mentioned earlier.
19:57:39 <peter1138> Well text does have a shadow.
19:58:31 <peter1138> Oh, except it's mostly black, so doesn't 😄
19:59:10 <peter1138> My view is that it's visible without a shadow, and it is not really important enough to need to stand out.
20:00:54 <andythenorth> peter1138: no shadows
20:01:05 <andythenorth> rule of "don't add visual effects unless clearly better"
20:01:05 <xarick> question, does yapf have a safeguard for cost overflow?
20:02:33 <xarick> maybe an assert doing cost < 0 would be sufficient
20:05:27 <xarick> no more than 2147 reverses
20:09:29 <peter1138> Not having shadows saves me from having to implement a deferred line drawer (to ensure the "shadows" are drawn first)
20:11:29 <peter1138> Not exactly sorter, just creating a list of lines, then drawing them once they're all known.
20:11:48 <peter1138> If there was a z-buffer... 😄
20:13:22 <talltyler> #12430 has an annotation
20:24:47 *** Smedles has joined #openttd
20:55:20 <_zephyris> There's an argument that the vehicle purchase/refit trees shouldn't have shadows to match the text, but then they should be solid black, to match the text. While the settings tree should be orange and shadowed, to match the text.
20:57:31 <_zephyris> The orange lines on the mid gray background is very low contrast, just look at a greyscale screenshot to see the problem.
20:58:20 <xarick> peter1138: vehicle number and group name are a bit too close
20:58:29 <_zephyris> This would be more consistent/readable
21:02:44 *** gelignite is now known as Guest198
21:02:48 *** gelignite has joined #openttd
21:03:55 <peter1138> xarick: Make an issue or a PR.
21:05:22 <kuhnovic> xarick: Is there a savegame for "this case" 😉?
21:05:49 <kuhnovic> I guess it needs further investigation. It's an annoying issue for sure.
21:06:22 *** Smedles has joined #openttd
21:06:26 <xarick> in 2 years there was no ships in limbo
21:06:39 <xarick> 1 million for the reverse cost
21:06:41 *** Guest198 has quit IRC (Ping timeout: 480 seconds)
21:06:54 <xarick> but I'm risking an overflow
21:07:56 <xarick> should be sufficient enough in most cases
21:09:43 <kuhnovic> I'm afraid this doesn't fix anything. Apparently there are still cases where an equilibrium is found that causes the ship to "ping-pong" .
21:10:15 <kuhnovic> Changing the penalty apperently just result in different cases where such an equilibrium is reached
21:11:13 <kuhnovic> xarick: Well, it's better than nothing. Please make sure the viewport is looking at the location where the problem is, otherwise I'll never find it
21:11:34 <xarick> I think 1000 is a too small penalty
21:12:19 <xarick> enough tiles to trigger an alternate path in reverse to be cheaper ...
21:12:25 <kuhnovic> Which is roughly what you get if you make the ship turn around without doing any 90 degree turns
21:12:46 <kuhnovic> So in terms of actual penalty it makes sense
21:12:52 <xarick> especially also because the intermediate region on reverse also changes
21:13:15 <kuhnovic> xarick: What do you mean with that?
21:13:49 <xarick> it accounts +1/-1 region, if you seen the video
21:16:41 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:16:42 <kuhnovic> Ah right, yeah that's problematic. And maybe that's where the problem lies. The high level PF doesn't add a penalty to the reverse region. Maybe it should.
21:17:54 <kuhnovic> Ok bedtime now, but I'm going to give this some more thought. Thanks for tortuse testing Xarick, you always manage to find some edge cases!
21:18:30 <xarick> im going to test 100k penalty, then 10k penalty
21:24:14 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:31:58 <xarick> i take too much time getting openttd version...
22:01:47 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
22:06:58 <peter1138> Our text truncation seems to be a bit weird when drawing LTR text in RTL mode.
22:34:59 <peter1138> And this highlights a scaling issue...
22:37:15 *** gelignite has quit IRC (Remote host closed the connection)
22:37:24 *** gelignite has joined #openttd
22:54:31 *** gelignite has quit IRC (Remote host closed the connection)
23:13:21 <rau117> peter1138: This was done to give possibility to buy any old version in replacement of the only locomotive available which changes over time.
23:13:21 <rau117> With the “vehicles never expire” setting disabled, only 1-2 current versions should be possible to buy.
23:13:21 <rau117> Although we can add some kind of icon that displays the (im)possibility of buying a locomotive at the current date.
23:13:21 <rau117> An indirect solution would be possibility to hide all obsolete vehicles by ctrl+click on “hide” button (when vehicles never expire 1).
23:13:21 <rau117> It will help not only in `xussr` with dozen of variations, but also in `iron horse` with **hundreds **of variations.
23:29:38 <kamnet> Man I wish we had a real iOS developer now 😦
continue to next day ⏵
