IRC logs for #openttd on OFTC at 2023-09-03
⏴ go to previous day
00:31:39 *** Flygon has quit IRC (Read error: Connection reset by peer)
00:43:58 <ahyangyi> 🤔 isn't the max map size 4096x4096
02:25:14 *** Wormnest has quit IRC (Quit: Leaving)
03:03:11 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:57:02 *** D-HUND is now known as debdog
04:19:13 *** tokai|noir has joined #openttd
04:19:13 *** ChanServ sets mode: +v tokai|noir
04:26:06 *** tokai has quit IRC (Ping timeout: 480 seconds)
05:25:41 *** HerzogDeXtEr has joined #openttd
06:04:11 <ahyangyi> ahyangyi: Today I figured it out: I can at least create a thread in Discord channel #development-forum 😛
07:16:02 *** gelignite has joined #openttd
07:25:03 <peter1139> Yes, but above 1024 is mostly pointless.
08:51:29 *** johnfranklin has joined #openttd
08:51:29 <johnfranklin> peter1139: what about multiplayer
08:56:12 <johnfranklin> I would say 2^22 tiles with 2^10 on one side is the minimum size of "realistic" cooperating (sometimes 1024\*2048 is also kind of okay, but my server maps are always 2048\*2048 or above)
09:23:22 <truebrain> another user has returned! 😛
09:23:50 *** frosch123 has joined #openttd
09:23:50 <frosch123> we got new aws invoices, but it fluctuates so much per month, i still can't tell how much the migration saved
09:24:07 <truebrain> owh, the bill is in? Why didn't I get an email .. odd
09:24:27 <frosch123> well it has been 2 months now, so 2 bills :p
09:25:02 <truebrain> weird comment, given I did check the one from 2 months ago 😛
09:25:05 <frosch123> the july one is the third highest bill in 18 month, the augutst one is the second lowest one in 18 month
09:25:07 <truebrain> not sure how your billing periods work 😄
09:25:25 <frosch123> i only know the monthly mails
09:25:56 *** ketsuban has joined #openttd
09:26:19 <truebrain> weird I didn't get a CC .. I got them till last month 😛
09:26:23 <truebrain> I wonder what AWS changed this time
09:26:33 *** ketsuban is now known as Guest1622
09:26:34 *** Guest1622 is now known as ketsuban[d]
09:27:29 <truebrain> anyway frosch123 , where your math goes a bit wrong, the migration finished in July
09:27:39 <frosch123> ketsuban[d]: if you now relogin to bananas, you should see the migrated content
09:30:13 <frosch123> truebrain: various things in the bills are weird. for example 2022-02 is one of the lowest bills (short month?) while 2022-03 is one of the highest bills. so maybe i have to consider the number of weekends (4 vs 5) per month, or something :p
09:31:32 <truebrain> 2022-03 was the month where AWS was like: I am going to charge you for CloudWatch (basically, logs)
09:31:39 <truebrain> that was a really weird thing they did
09:31:54 <truebrain> never try to understand AWS bills; you will fail
09:32:01 <frosch123> ok, so 2022-03 is high for different reasons?
09:32:27 <frosch123> truebrain: yeah, looks like it :p
09:32:39 <truebrain> 780GB of traffic from AWS this month .. I wonder if I can reduce that even further ...
09:32:52 <frosch123> i tried to see whether there was a change, but so far it's just noise
09:33:23 <truebrain> it was 590GB before migration; so those things are always a bit funny
09:33:28 <truebrain> it makes you go: huh? did I fuck up?
09:33:31 <truebrain> often it isn';t the case
09:33:48 <truebrain> often it is just increase in traffic .. summer months etc
09:34:36 <truebrain> anyway, the bill of this month is really simple: 100 dollar for EC2 instance, 90 dollar for traffic, that is it.
09:34:53 <truebrain> I removed all other services and costs (well, there is 1 dollar for backups and backend storage, but that isn't important)
09:35:11 <truebrain> the bills before that had a wide range of other services, which sometimes jumped for no clear reason 😛
09:35:53 <frosch123> truebrain: before taxes?
09:35:59 <truebrain> so next job is going to be to zoom in on the traffic, and see if that can be reduced ..
09:36:06 <truebrain> yeah, AWS bills are always before taxes, till the email you get
09:36:37 <frosch123> ok, the gap looked like a tax amount 🙂
09:37:14 <truebrain> so basically the bill in the AWS dashboard says: 188.92 dollar, and we charged 228.60 dollar
09:37:21 <truebrain> it doesn't even clearly say why those two values are different
09:37:29 <truebrain> like "you know" attitude
09:37:58 <truebrain> they try, but it is really unpleasent to try to figure out where costs go
09:38:05 <truebrain> I am happy I don't have to do this on an enterprise level 😄
09:38:35 <truebrain> hmm, the traffic really is weird. Before migration it was clearly lower.
09:38:40 <truebrain> and we offload more to Cloudflare
09:38:45 <truebrain> so what is causing it to be higher? 😄
09:40:00 <truebrain> the only thing I can think of, is more BaNaNaS downloads over TCP, as that would be the most hungry of them all
09:40:30 <truebrain> 2.7 MiB/s is handled by HTTP, and 192 KiB/s over TCP .. so that is in a fair balance, I guess (6% over TCP, 94% over HTTP)
09:40:38 <truebrain> still, why that 6%? 😄
09:41:29 <truebrain> TURN is only 21 KiB/s
09:41:46 <truebrain> (and multiplayer, except TURN, is 5.4 KiB/s)
09:42:19 <truebrain> that is a fun graph 😄
09:47:03 <_zephyris> ketsuban[d]: And thanks for having a go at better bridge newgrf support 😀
09:49:31 <_zephyris> Hah, oops, dumb mixup. I blame the toddler trying to grab my phone from me!
10:11:57 <Eddi|zuHause> that's a weird evolution from "my brother typed that" :p
10:32:39 <andythenorth> search box in vehicle list? 😄
10:32:41 <andythenorth> also autoreplace?
10:32:50 <andythenorth> search box is useful
10:35:18 *** emperorjake has joined #openttd
10:35:18 <emperorjake> Very useful, as it works for additional text too
10:35:25 <Eddi|zuHause> but what should it search? group names, train names, model names?
10:36:40 <andythenorth> whatever the current implementation searches
10:36:54 <andythenorth> we can extend it later if we need to
10:37:10 <Eddi|zuHause> if i type in "passenger", what should be displayed? all vehicles loading passengers, all wagons that have "passenger" in the name, or ...?
10:37:16 <andythenorth> is it already solved in JGRPP?
10:39:27 <Eddi|zuHause> the suggestion is probably valid, but i'm not touching gui code... that's like a week of rabbit holes until you can even begin to code
10:40:32 <Eddi|zuHause> to program in a gui framework you first have to learn to think like the designer of the gui framework
10:40:52 <Eddi|zuHause> and i was always quite bad at that
11:52:08 <talltyler> We have a cargo filter already, so the text filter should not look at cargos.
11:52:38 <talltyler> I would probably just filter group names and call it a day
11:53:11 <talltyler> Possibly engine names, but that sounds fiddly
12:08:44 <Eddi|zuHause> what about destination names?
12:10:27 <emperorjake> This is just for the purchase menu, or were you thinking of applying it to the train list?
12:11:02 <Eddi|zuHause> the train list. it already exists for the purchase menu
12:29:56 <andythenorth> I was thinking for autoreplace
12:30:03 <andythenorth> it's notably missing there
14:42:44 *** nickname has joined #openttd
15:30:11 *** Wormnest has joined #openttd
16:20:41 <peter1139> Oh dear, might need to replace that tyre :/
17:07:38 <truebrain> anyone happen to have some time / thought on https://github.com/OpenTTD/OpenTTD/pull/11232 ? I kinda would like to know if we like such change, before I continue to work on automatic crashdump collection 😄 (it is kinda mandatory to have a computer-readable format for that to work 🙂 )
17:15:51 <_jgr_> Are you intending to do some analysis on the incoming crashlogs, or is it more a case of someone getting an email when a crash log is automatically submitted?
17:16:18 <truebrain> the first; opt-in ofc
17:17:02 <truebrain> so similar to survey; but in case of a crash. So we send the crashlog + dump to us, and analyze it. Make it a sentry report (which deduplicates it).
17:17:56 <truebrain> but as first step I suspect I make a website where you can upload the log+dmp, and it tells you all the good pieces
17:32:41 <truebrain> (well, no, first ofc is to get that PR approved 😄 😛 )
18:19:19 <peter1139> rint sizeof(_grf_text)
18:19:21 <peter1139> (unsigned long) $0 = 20971520
18:20:02 <peter1139> I'm wondering if that can be a vector...
18:20:44 <truebrain> it is only a list of 256?
18:22:04 <peter1139> Seems to be arbitrary, judging by newgrf_text.cpp:564, it's allocated sequentially.
18:22:18 <truebrain> ah, TAB_SIZE * 256 .. lol
18:22:42 <truebrain> you would think that even a std::map would be better, if the index is important 😄
18:22:53 <truebrain> 20MB .. even if you have no GRFs loaded 😛
18:25:48 <frosch123> oops, i did not see that when changing a 3 into that 256
18:26:55 <peter1139> I'm vectorizing it :)
18:28:17 <truebrain> when did that 3 change into a 256?
18:33:16 <truebrain> so out of the 50MB we need to run, 20MB is allocated to this .. I am shocked how little memory OpenTTD needs 😛
18:35:36 <frosch123> if you are lucky it was zero-initialised and only physically allocated by the OS when needed
18:36:08 <truebrain> this would be bss not?
18:36:40 <frosch123> there is now a std::vector in every entry, not sure whether that was always the case
18:37:37 <frosch123> but if it was POD before and all zeroes, there is another section for those (not bss)
18:38:22 <DorpsGek> - Update: Translations from eints (by translators)
19:07:26 *** Kitrana has joined #openttd
19:08:43 *** Kitrana2 has joined #openttd
19:10:37 *** Kitrana1 has quit IRC (Ping timeout: 480 seconds)
19:14:10 <peter1139> Okay, debugging on native Linux is annoying -- the mouse cursor stays hidden :/
19:15:27 <andythenorth> hmm I'm logged out of my reddit sock puppet account
19:15:30 *** Kitrana has quit IRC (Ping timeout: 480 seconds)
19:15:36 <andythenorth> cookie must have been cleared 😛
19:16:13 <andythenorth> wanted to comment on world gen, but eh
19:22:08 <andythenorth> right what were we doing peter1139?
19:22:12 <andythenorth> something about vars?
19:22:17 <andythenorth> playing shooting games?
19:23:15 <peter1139> Okay, even not hiding the cursor doesn't help. The cursor is visible but nothing else can respond to mouse input :/
19:24:00 <peter1139> This is weird, it always worked fine with native Windows or WSL2.
19:26:50 <andythenorth> hmm should I virtualise macOS on macOS?
19:26:57 <andythenorth> Apple have some native framework and images for it
19:30:52 <peter1139> FIRS goes up to 0x197. 407 is somewhat less than 524288.
19:31:29 <truebrain> but that also fits in 3*
19:32:21 <peter1139> That is definitely the wrong repository.
19:33:36 <andythenorth> also kamnet's surname is Fields
19:33:50 <andythenorth> which given that he fields lots of playerbase interaction
19:34:01 <andythenorth> is pleasing nominative determinism
19:36:19 <peter1139> Hmm, I can't get GRFTextEntry to less than 40 bytes though :(
19:36:52 <peter1139> Given it's currently badly aligned, aligning it better... doesn't improve it ;(
19:45:23 <peter1139> Reshuffling it allows it to be 12 bytes instead of 24 bytes on 32-bit builds, but I doubt we care about that :D
19:46:38 <peter1139> I should test ARM, that's probably relevant.
19:52:17 <kamnet> truebrain: Thank you, moved over, suggested solution made.
19:52:39 <truebrain> I doubt anyone is actually going to read or act on that, honestly
19:52:52 <truebrain> I think if you want it fixed, you will have to grep for it yourself 😄
19:54:14 <truebrain> And if you don't know how a tool like `grep` works, you can always search via GitHub on strings
19:54:21 <truebrain> so find the right string, and throw it in there! 😄
19:58:35 <truebrain> okay, my reply in the ticket sounds a bit more passive aggressive then the intention; let's fix that 😄
19:59:26 <peter1139> Well, how many times do you need to repeat it...
20:00:17 <peter1139> It is almost like only the first line of your response on #11257 was read.
20:00:21 <truebrain> nah, lot of people just don';t know how good GitHub's search is 🙂
20:14:35 *** virtualrandomnumber has joined #openttd
20:14:51 *** virtualrandomnumber has quit IRC ()
20:22:16 <peter1139> Hmm, VS Code global search & replace seems broken for me.
20:22:41 <peter1139> Might be a combo of that and format-on-save getting stuck.
21:05:44 *** gelignite has quit IRC (Quit: Stay safe!)
21:10:58 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:23:31 <peter1139> Hmm, regex mangled :/
21:45:39 *** esselfe has quit IRC (Remote host closed the connection)
21:49:27 *** esselfe has joined #openttd
22:07:06 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:19:34 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
22:59:36 <andythenorth> ha 🙂 `reload` in console reloads the game, but without rescanning newgrfs
22:59:55 <andythenorth> which is much faster, but caused me a moment of 'wtf?'
23:26:24 <truebrain> 0-index .. it still makes my brain hurt ..
23:27:23 <truebrain> I think I am wrong.. not sure 😛
23:42:39 <peter1139> And I think the original assert is not possible any more.
23:43:35 <peter1139> Hmm, also, that stringid is uint16_t, which is a lot less than 524288.
23:50:45 <peter1139> Yeah, looks like that * 256 was not even usable. * 32 is the maximum supported by uint16_t
continue to next day ⏵
