IRC logs for #openttd on OFTC at 2023-06-17
⏴ go to previous day
00:10:50 *** Eddi|zuHause has quit IRC (Read error: Connection reset by peer)
00:14:23 *** Eddi|zuHause has joined #openttd
01:08:21 *** DemonBigj782 has joined #openttd
01:15:02 *** DemonBigj781 has quit IRC (Ping timeout: 480 seconds)
02:47:56 *** DemonBigj782 has quit IRC (Quit: Leaving)
02:48:20 *** debdog has quit IRC (Ping timeout: 480 seconds)
02:50:27 *** D-HUND is now known as debdog
03:00:04 *** herms has quit IRC (Quit: bye)
03:57:29 *** demonbigj781 has joined #openttd
05:42:30 *** HerzogDeXtEr has joined #openttd
06:46:34 *** Extrems has quit IRC (Ping timeout: 480 seconds)
07:10:29 *** Extrems has joined #openttd
08:09:44 <Rubidium_> LordAro: would you care to take another look at #10994? You approved it, but I had made some conflicts in it due to the StringParameters* -> StringParameters& change, so I needed to rebase it
08:35:10 <TrueBrain> 91GB missed, 52GB hit ... the cache is getting warmer
08:36:48 <TrueBrain> I still see 1.7.1 clients and 1.8.0 and 1.10.3 making use of BaNaNaS .. lol
09:13:09 *** Flygon has quit IRC (Remote host closed the connection)
09:25:26 <Rubidium_> thanks LordAro! I really should make fewer conflicting PRs
09:29:12 *** nielsm has quit IRC (Ping timeout: 480 seconds)
10:27:54 <pickpacket> TrueBrain: omg... How old are those versions? 😂
10:30:39 <Rubidium_> 6, 5 and 3-ish years old?
10:31:31 <Rubidium_> there are still 1.3.0 servers; those are a mere 10-ish years old
10:51:28 <Eddi|zuHause> versions 1.x roughly translate into 201x
11:01:25 <petern> OH: I'm old enough to have heard of OpenTTD.
11:38:02 *** gelignite has joined #openttd
11:38:50 *** gelignite has quit IRC (Read error: Connection reset by peer)
11:42:37 *** gelignite has joined #openttd
11:50:03 <TrueBrain> right, had to patch the Nomad cluster in order to get CPU information out of it .. so now I have to roll over the cluster again .. let's see how well it works 🙂
12:02:25 *** gelignite has quit IRC (Quit: Stay safe!)
12:29:30 <TrueBrain> there we go, now I can finally see what services are using CPU and which are not ... and as it turns out: except for expected spikes, nothing is actually happening CPU wise .. it is crazy how efficient serving HTTP is these days 🙂
17:56:53 *** stormcone has joined #openttd
18:11:00 <stormcone> something looks strange with bananas, on the website you can see new newgrfs uploaded today, but the commits are missing from the repo, and neither show up ingame
18:13:56 <TrueBrain> Means they didn't actually upload, sadly .. let's see why ..
18:15:06 <andythenorth> got the most bizarre python bug, can't for the life of me see what causes it 😛
18:15:58 <TrueBrain> "failed to push some refs" .. hmm .. why oh why ..
18:16:44 <TrueBrain> owh, I think I know why ...
18:17:03 <TrueBrain> fucking branch protections ..
18:17:15 <TrueBrain> I migrated 4 services, and 4 times I forgot about them
18:17:26 <TrueBrain> now I need someone to upload anything new to BaNaNaS 😛
18:17:50 <TrueBrain> so andythenorth , got something for me you could upload? 🙂
18:18:49 <andythenorth> we need test.grf
18:19:15 <TrueBrain> normally I have a preview version of BaNaNaS for that 😛
18:19:20 <TrueBrain> but in this case, I just need to flush the queue 🙂
18:20:05 <andythenorth> unfortunately I have nothing 😦
18:23:59 <TrueBrain> hihi, I am excellent at cheating .. but still it fails, let's see ..
18:25:50 <TrueBrain> lol ... you can allow bots to bypass the required for pull-requests, but they still need to checks to succeed .. which is impossible without a pull-request
18:28:03 <TrueBrain> that is an interesting choice of GitHub ... especially as they made an explicit checkbox to mark to allow bots to bypass PRs .. but not one for them to bypass CI checks
18:30:28 <TrueBrain> stormcone: tnx for the poke
18:31:46 <stormcone> thanks for the fix too :)
18:32:13 <TrueBrain> rather happy the stuff was still in the queue 🙂
18:33:17 <TrueBrain> and now the CI is happy again too; how joyful
18:35:34 <stormcone> btw the irc topic still says 13.1 :P
18:36:40 <TrueBrain> now I am curious if any other uploads were done that are never processed between yesterday and 6 hours ago ... uploads to BaNaNaS are rare, but at least 4 uploads today .. days without, and all of a sudden .. 😄
18:40:38 <DorpsGek> - Update: Translations from eints (by translators)
18:44:16 <petern> I didn't block it today
18:47:05 <andythenorth> my Horse code contains a lot of classes for wagons, the classes assign a list to a name property ```randomised_candidate_groups```
18:47:34 <andythenorth> the property is not assigned anywhere else in the code ever, that I can see
18:47:56 <andythenorth> after assignment, if I print the property, the list contains the expected elements
18:48:30 <andythenorth> but after all the classes are initialised, some of the list contents are changed
18:48:40 <andythenorth> it's as though lists.extend() has been used, but it hasn't
18:49:31 <andythenorth> I don't think I've got any of them as class properties instead of instance properties
18:49:41 <andythenorth> I don't see anywhere I should be using .copy() or .deepcopy()
18:49:55 <TrueBrain> okay, the upload of a single user was lost; but he uploaded 5 new versions of that same GRF today .. so I guess the 4 versions that got lost in the void yesterday aren't all that important 😄
18:56:57 <Ahyangyi> I think you called `.extend` on stuff in `super_group`, which comes from `all_randomised_candidate_groups`, which comes from individual `randomised_candidate_groups` members
18:57:50 <TrueBrain> yeah, okay, only those 4 uploads were done, and after that bananas-api restarted, meaning their metadata got lost. But the user uploaded 5 more versions of that NewGRF after that. So I guess I am lucky here, and no actual uploads were lost 😄 \o/ Stupid branch protections ...... 😛
19:00:21 <andythenorth> Ahyangyi: how did you even begin to debug that 🙂
19:00:34 <andythenorth> that might be a place I need to use copy or deepcopy
19:11:38 <andythenorth> Ahyangyi: solved 🙂 thankyou so much, I spent more than an hour staring at code and printing values for this
19:26:13 <andythenorth> Horse with search string
19:26:22 <andythenorth> I don't set the name prop, only the callback
19:35:53 <glx[d]> maybe you need some basic info in the prop
19:39:45 <petern> Was that turned into a github issue?
19:42:19 <glx[d]> oh I have an issue to decode
19:43:00 <andythenorth> I'm going to set the static prop
20:04:02 <andythenorth> flattening the text stack into a property, in python 😦
20:04:06 <andythenorth> can't see a trivial way to do that
20:07:07 <discord_user_03329cf> Is search something in development?
20:07:40 <discord_user_03329cf> I’ll take that as a yes
20:08:22 <discord_user_03329cf> Does it just look at the vehicle name or does it look at stats or other things? Would it work with tags?
20:08:35 <glx[d]> it looks at many things
20:08:49 <glx[d]> but I don't have the details in head
20:09:06 <discord_user_03329cf> Fair enough
20:10:46 <discord_user_03329cf> Would be cool if you could have tags so if for example you searched for “br class 158” it would also find you the carriages needed to make it longer in rukts for example
20:12:05 <andythenorth> it would also find 'class'
20:12:11 <andythenorth> and everything containing 'br'
20:13:16 <discord_user_03329cf> So each thing with a space is seen as a search of its own?
20:15:02 <andythenorth> hmm this is making me grumpy
20:15:17 <andythenorth> in favour of simple text stack
20:15:24 <andythenorth> but now I have to put it all back
20:15:27 <andythenorth> AND have the callback
20:17:31 <glx[d]> hmm #10868 test save hangs at load, but still waiting for a crash
20:20:41 <andythenorth> did we ship the text search yet?
20:22:25 <andythenorth> can't see 10519 in the changelog
20:22:53 <andythenorth> but changelog hasn't been updated since Nov 22
20:26:24 <glx[d]> master changelog will be updated for 14.0-beta1
20:27:35 *** nielsm has quit IRC (Ping timeout: 480 seconds)
20:29:42 <andythenorth> dunno what to do about this name search thing
20:30:20 <andythenorth> for variants, it's really necessary to use the name callback
20:30:33 <andythenorth> and for name callback it's much preferable to use the text stack
20:35:43 <JGR> It would make sense to have the text filter execute the name callback
20:36:02 <JGR> It's nearly a one-line change
20:37:22 <JGR> This is already done for sorting by name
20:38:59 <petern> andythenorth apparently wants to work around it in python, instead of raising an issue in github.
20:39:03 <andythenorth> would we give it a unique context?
20:39:26 <andythenorth> petern: pure laziness
20:39:48 <andythenorth> so that we can keyword stuff the result 😛
20:40:06 <JGR> The text filter should match whatever text is one the screen that is being filtered
20:40:18 <andythenorth> well it already searches description 🙂
20:41:03 <petern> Not sure if the hierarchy is known at that point. Maybe it is.
20:49:05 <glx[d]> ok #10868 is an uncatched Script_FatalError, let's try to catch it
21:00:45 <andythenorth> how do I raise an issue now?
21:02:10 <glx[d]> do you want the issue to be handled ?
21:02:17 <andythenorth> it's not a bug or a crash
21:02:21 <petern> Write it on your blog, and then when nobody sees it for 2 years, complain that we ignore everything.
21:02:37 <andythenorth> github says forum suggestions is correct place
21:02:46 <petern> It sounds like a bug to me, to be honest.
21:05:31 <glx[d]> 7 minutes and it's still loading the 64x64 test game
21:05:47 <glx[d]> I hate when crash are not instant
21:06:47 <glx[d]> "Maximum memory allocation exceeded by 854826 bytes when allocating 536870912 bytes"
21:07:16 <glx[d]> how can the save be created but not loaded ?
21:08:59 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
21:10:09 <Rubidium_> some tree-like structure?
21:12:13 <glx[d]> ```function TestController::Save()
21:12:13 <glx[d]> return { x = array(6e7) };
21:13:48 <glx[d]> wondering if I should kill the AI or just remove the data from stack and let it continue
21:15:50 <JGR> The crash seems to be in some game log management
21:16:31 *** Eddi|zuHause has joined #openttd
21:17:41 *** stormcone has quit IRC (Quit: Bye!)
21:17:56 <glx[d]> yes because the Script_FatalError is not catched
21:18:58 <glx[d]> which cause it to be catched during AfterLoad()
21:19:10 <JGR> Making a save and loading it with that same AI in 13.1 does not crash
21:19:54 <glx[d]> memory errors throw in master
21:19:56 <JGR> Doesn't crash in my branch either
21:26:28 <glx[d]> depends on max allowed memory
21:28:21 <glx[d]> I use script_max_memory_megabytes = 1024
21:30:33 <glx[d]> when saving the array is allocated to the right size, but on loading it's done by doubling size when needed
21:32:34 <JGR> How does James find all these?
21:36:43 * Rubidium_ just assumes James is a good tester wanting to improve OpenTTD
21:42:42 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:54:40 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:01:55 *** Eddi|zuHause has quit IRC (Read error: Connection reset by peer)
22:03:07 *** Eddi|zuHause has joined #openttd
22:11:18 <petern> andythenorth: fixed it for you
22:16:52 <andythenorth> added a test grf
23:51:23 *** ChanServ sets mode: +v tokai
23:58:02 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵