IRC logs for #openttd on OFTC at 2025-01-24
β΄ go to previous day
01:01:02 <_glx_> hmm weird emscripten failure in #13372
02:03:31 *** herms has quit IRC (Quit: bye)
03:14:41 *** Wormnest has joined #openttd
03:17:00 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:47:04 *** gnu_jj_ has joined #openttd
03:50:14 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
03:57:12 *** Wormnest has quit IRC (Quit: Leaving)
06:11:41 *** keikoz has quit IRC (Ping timeout: 480 seconds)
07:30:20 <peter1138> Aww boo, my bread didn't bake overnight :(
07:45:14 <peter1138> Stops me eating it all before lunch I guess.
07:46:20 <peter1138> > A new callback could be called specifically on construction/consist-rearranging/station-leaving with extra info in var10/18
07:47:27 <peter1138> Yeah, order-of-operations, especially of CB36, means that leaving persistent storage as a free-for-all probably won't work well.
07:54:09 <peter1138> Hmm, if there's a callback do I care about the result? :D
07:56:48 <andythenorth> peter1138: if it's only for setting storage?
07:56:54 <andythenorth> do we have any precedent?
08:05:29 <peter1138> Wonder what extra info would be needed/useful.
08:14:35 <peter1138> Hmm, order-of-operations make sense anyway...
08:20:59 <peter1138> So far I'm calling on-build, on service-at-depot, and on consist change with CCF_ARRANGE set (for trains that might mean it runs twice on
08:21:08 <peter1138> servicing at a depot)
08:25:46 <andythenorth> is it unsafe to run it for load/unload, railtype change, and other triggers?
08:26:51 <peter1138> As always it depends what you do with it.
08:29:24 <andythenorth> I am not qualified to answer π
08:29:39 <andythenorth> 'desync' is a magical cargo cult word I say here
08:30:38 <andythenorth> I understand the concept of state diverging, but I don't have a picture of which operations cause that (and which don't)
08:31:03 <andythenorth> probably better that grf authors *don't* understand desyncs
08:34:38 <andythenorth> ok this "Size Class" badge is probably fine
08:34:49 <andythenorth> no need for wagon sizes in the buy menu
08:36:01 <andythenorth> it's a bit visually repetitive on the text in the buy menu without them
08:36:13 <andythenorth> but at 2x UI zoom, they're often cropped anyway
08:37:44 <peter1138> THe whole window is 2x the 1x size, so they'll be equally cropped...
08:38:14 <peter1138> Size Class Class Size Size Class
08:38:29 <peter1138> Length: Extraordinarily Long.
08:46:11 <andythenorth> hmm you raise a good point
08:46:19 <andythenorth> what size class is that? π
08:46:41 <andythenorth> should I just put the length in numbers, as fractions of a tile?
08:47:00 <andythenorth> or I could use decimals that have confusing rounding
09:14:29 *** mindlesstux has joined #openttd
09:18:14 <truebrain> how on earth did they manage to do that? π
09:18:37 <truebrain> even the content of the `.git` folder got flattened π
09:18:47 <truebrain> `mv */* mynewfolder/` or something? π
09:26:55 <truebrain> I like how we are all not replying in the place GitHub wants us to reply
09:33:08 <peter1138> As they are building for docker, they probably actually want a dedicated server build.
09:33:34 <LordAro> `-DOPTION_DEDICATED=ON` probably yes
09:58:36 <LordAro> truebrain: got any ideas about that emscripten failure? doesn't feel like anything i could've caused...
09:59:21 <truebrain> I haven't looked at any CI failure; so currently, no π Will take a look a bit later
10:00:40 <LordAro> my only thought is that FindXaudio2 is resetting CMAKE_REQUIRED_FLAGS, i'm not sure about cmake's scoping rules and if that might be affecting subsequent includes
10:01:42 <truebrain> I also have `upload failed: generated/openttd-nightlies/folders.yaml to s3://***/openttd-nightlies/folders.yaml An error occurred (InternalError) when calling the PutObject operation (reached max retries: 2): We encountered an internal error. Please try again.` .. and retrying doesn't solve anything π¦
10:01:51 <truebrain> sometimes there are weird bugs when uploading to Cloudflare's R2
10:02:00 <truebrain> mostly there isn't a problem in the world π
10:05:25 <LordAro> i wish i had your optimistic worldview
10:06:07 <truebrain> and no, I have no clue why Emscripten fails. It does seem to have to do with your change, but clearly that makes it a side-effect π
10:06:33 <truebrain> but fact is that atomic detection fails after, and succeeds before your change π
10:22:09 <truebrain> visibility is important π
10:25:32 <xarick> holy wow, never seen a disconnection this large
10:27:48 <merni> that's one big waterfall
10:30:26 <xarick> trying to figure out what happened
10:31:22 <xarick> there was a river widening here that managed to demolish like 5 or 6 tiles worth of rivers
11:10:58 <peter1138> Was it the radio antenna? (I don't know the build order or rules)
11:14:14 <xarick> nop, objects require towns, so it's before
11:18:14 <peter1138> Hmm, is eints something I can set up for external translations?
11:18:41 <peter1138> Maybe I should ponder the text-file based external translations anyway.
11:24:18 <peter1138> Yay, freshly made bread.
11:34:26 <xarick> oh, pff, it was the same issue as in the PR...
11:34:40 <xarick> terraform ruining a future tile
11:37:03 <xarick> this is really difficult to debug, maybe you would be interested in the debug code?
11:41:46 <xarick> #13366 workaround "fixes" it
12:47:45 *** kuka_lie has joined #openttd
13:36:33 *** akimoto has joined #openttd
14:03:05 <LordAro> more accurately, multithreaded code/cpus are extremely difficult to reason about
14:03:17 <andythenorth> has anyone ever found multithread good?
14:03:27 <_glx_> mutlithread is a big footgun
14:03:37 <LordAro> andythenorth: well you were able to type that while having other things running
14:04:09 <andythenorth> remember co-operative multi-tasking? π
14:04:45 <ahyangyi> too unyielding to use that?
14:06:31 <andythenorth> really must be lunch time
14:07:07 <andythenorth> I used to think parallel was good (or map:reduce stuff), but even that has footguns in python
14:07:25 <andythenorth> pickle fails in quite unexpected ways
14:07:31 <andythenorth> modules have import overhead
14:07:38 <andythenorth> actual isolation of state can be hard
14:08:05 <andythenorth> results depending on results means that multi-pass scheduling is needed
14:14:04 <xarick> I have at least one more fix, but it's really invasive, large
14:14:16 <xarick> but garantees rivers are always connected
14:15:03 <xarick> not sure I go through with it
14:15:14 <peter1138> RISC OS was co-operative multi-tasking.
14:30:23 <andythenorth> co-operative multi-crashing
14:30:32 <andythenorth> so was Mac OS prior to OS X
14:31:57 <peter1138> Before Windows 95, I believe.
14:32:28 <peter1138> > Cooperative multitasking is still used on RISC OS systems.
14:41:32 <xarick> interesting, StationType::Dock
14:41:45 <xarick> these mini changes I need to be aware of
14:56:11 <xarick> hmm, AirportTypes are not enum class yet
15:16:26 <brickblock19280> Well it would break if for instance I gave all of these historical versions and then use the game one for the current flag if that then changes I would have an incorrect flag for the unit
15:30:13 *** pickpacket7 has joined #openttd
15:30:47 *** pickpacket has quit IRC (Ping timeout: 480 seconds)
15:30:47 *** pickpacket7 is now known as pickpacket
15:36:29 <truebrain> Is: "buy more memory" a valid response? π
15:46:50 <kuhnovic> It's nicer than sending them off to download more RAM π
15:49:06 <LordAro> "-t --trace Print the names of the input files as ld processes them."
15:49:32 <andythenorth> I downloaded more RAM
15:50:00 <peter1138> Can someone translate for me.
15:50:04 <andythenorth> I used this company that sends you a simple device to expand your RAM
15:50:44 <peter1138> (If you have a badge, and another badge, it will be broken?)
15:52:32 <andythenorth> I assumed brickblock19280 was just trolling π
15:52:56 <peter1138> I don't, but I don't understand the point.
15:53:37 <brickblock19280> I mean it's an edge case but I just want it clarified how and when flags will be updated
15:54:46 <brickblock19280> Since if you always want period appropriate flags you should probably just define a copy of the current one
15:56:50 <andythenorth> oh you mean flags are versioned?
15:57:03 <andythenorth> like "union_jack_final_final"
15:57:36 <andythenorth> (if nobody points out that "union jack" is only valid for boats, I won't mind)
15:57:43 <andythenorth> although that does raise the question of boat flags
15:58:13 <andythenorth> and whether flagging for e.g. Panama or Liberia is a running cost advantage
15:58:30 <andythenorth> [buyable variants for flags]
15:59:09 *** Wormnest has joined #openttd
15:59:18 <peter1138> I still don't under the premise of the question.
15:59:53 <peter1138> If you "install" period correct flags, then whatever uses those period correct flags will... use those flags.
15:59:56 <andythenorth> imagine Trump takes Greenland
16:00:02 <andythenorth> how many stars on the US flag?
16:00:07 <andythenorth> before or after 2025
16:00:31 <andythenorth> the question is about the spec yes?
16:00:42 <andythenorth> so if a flag changes by date, use varact2 to change the sprite
16:01:00 <andythenorth> if it's a wholly new flag, then it's a new name string
16:01:36 <peter1138> If I ask for a US flag, and get a US flag, is that "broken" if it's not the right period of US flag? Is that was "broken" means?
16:04:24 <andythenorth> whether that happens depends on the flag vendor
16:04:56 <andythenorth> if flags need to be historically appropriate, select a vendor that can fulfill that
16:05:07 <andythenorth> and review their contractual terms and SLA
16:06:39 <peter1138> Maybe the question is more about "what happens if two badges are loaded with different icons"?
16:07:11 <andythenorth> toast was not a good lunch
16:07:13 <peter1138> Maybe the question is more about "what happens if two NewGRFs load one badge with different icons"?
16:12:53 <brickblock19280> peter1138: Yes so those who do not want that need to be informed to mirror the current flag from OpenTTD.grf
16:16:21 <peter1138> To be honest it feels like you are inventing a problem.
16:19:52 <peter1138> Also remember that badges are there for the benefit of players. If period-correct flags (as an example) benefit players, then what's the problem with having them. (And vice-versa if they don't benefit players.)
16:22:42 <peter1138> Should we include all those designs? Would it benefit players to do so?
16:25:26 <peter1138> (I chose the US flag as an example of one that's relatively easy to find a list of all the different versions.)
16:25:42 <peter1138> Or andythenorth did :)
16:25:42 <brickblock19280> I'm not asking that
16:25:56 <peter1138> See. I don't understand the question still :)
16:26:07 <brickblock19280> Whatever it really isn't a big deal
16:30:07 <LordAro> didn't we get rid of all the flags when we got rid of the languages in the server list?
16:30:21 <LordAro> iirc it was at least partly because of this problem
16:30:59 <peter1138> The problem there, as far as I know, was using country flags to denote language is a bad idea.
16:31:19 <LordAro> but it's not like which flags are possible isn't contentious either
16:31:37 <peter1138> It's a flag that exists.
16:32:40 <LordAro> so what are you suggesting? if it exists in ISO3166 (or whatever it is) it should be included?
16:33:57 <peter1138> I'm suggesting I have a list of flags that have a name associated with them, and they are currently built into a NewGRF that provides some badges.
16:35:05 <peter1138> (What are you suggesting?)
16:36:07 <LordAro> feels like adding another 'currency' problem though
16:36:12 *** HerzogDeXtEr has joined #openttd
16:48:04 *** gelignite has joined #openttd
16:52:18 <wensimehrp> LordAro: It's fine to use the ROC flag as a historical flag
17:27:03 <andythenorth> can we have badges for different drivers?
17:46:41 *** gwyd4016 has joined #openttd
17:46:41 <gwyd4016> Properties stored as badges
17:46:58 <andythenorth> kind of what badges are π
17:47:03 <andythenorth> but static props, not variables
18:08:10 *** akimoto has quit IRC (Remote host closed the connection)
18:10:03 <andythenorth> in-game badge editor?
18:10:13 <andythenorth> with perfect quality?
18:21:43 <peter1138> Heh. I mean, clearly it's not enough, but...
18:40:26 <truebrain> I guess having it max is a good thing now π
18:48:23 *** ChanServ sets mode: +v tokai
18:55:08 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
19:11:49 *** greeter has quit IRC (Remote host closed the connection)
19:21:02 *** greeter has joined #openttd
19:51:57 <peter1138> Hmm, why is my interactive rebase no longer working...
20:00:31 <andythenorth> maybe you're using the wrong flags?
20:00:42 <andythenorth> or did the flags change?
20:49:38 <andythenorth> `"newgrf/CA\12\22"` is not a valid badge π
20:52:53 <andythenorth> I'm sure I've converted grfid formats somewhere
21:00:29 <peter1138> CA\12\22 is an NML thing.
21:01:16 <andythenorth> yeah GPT has written me a converter
21:01:57 <johnfranklin> xarick: The billions shift from side to sideβ¦
21:02:45 <andythenorth> obvs GPT has got it wrong
21:05:12 <xarick> I need merges to happen
21:05:25 <xarick> my last PR makes no sense without them
21:06:50 <andythenorth> have they passed the unit tests?
21:06:56 <xarick> the culmination of all those tiny PRs merged will allow me to do a final one
21:07:54 <andythenorth> hmm I need it to not `0x` prepend
21:07:55 <andythenorth> "newgrf/0x43411222",
21:08:48 <peter1138> Don't think 0x is right.
21:09:35 <andythenorth> not sure what to do with it, but eh, it's there in the debug π
21:10:06 <peter1138> Design an icon for Iron Horse :)
21:24:23 <andythenorth> cargo display side quest? π
21:35:36 <_jgr_> I'm not sure a tarpaulin is the right kind of container for stuff like chlorine or acids π
21:40:48 <truebrain> andythenorth: Do you supply glasses with that wagon? π
21:42:39 <talltyler> β64 cargo types isnβt enoughβ
21:59:58 <xarick> hmm, I need a way to not return false on partially built rivers
22:00:19 <xarick> if i am to add TestRiverConnection
22:02:15 <xarick> do you guys want my debugging function TestRiverConnection in master?
22:08:25 <peter1138> talltyler, I have a branch for that.
22:28:58 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:53:30 *** kuka_lie has quit IRC (Quit: Lost terminal)
23:04:11 <xarick> okay, bed, cyas good night
23:10:25 *** gelignite has quit IRC (Quit: Stay safe!)
23:29:07 *** Flygon has quit IRC (Remote host closed the connection)
23:42:04 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:55:22 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
continue to next day β΅